Research Paper
pkunz206
ContentServer1.pdf
Eternal Sunshine of the Spotless Mind? Mortality Salience, Death-Thought Accessibility, and Self-Forgiveness
John M. McConnell
Wheaton College
ABSTRACT Terror management theory claims the quintessential indicator of cultural adherence is human self-esteem, and self-esteem is vital to suppressing death-thoughts into the uncon- scious to buffer against existential fear. Guilt—an emotional response to cultural-based rule violations—should therefore be an important motivation for self-forgiveness. In four studies, mortality salience, death-thought accessibility, and self-forgiveness were negligibly related, and offense severity, conciliatory behaviors, perceived forgiveness, and effort provided negligible moderation. In Study 1, results did not support the initial interactional-moder- ated-mediation model. In three follow-up studies, experimental-causal-chain analyses had unsupportive findings. I provide implications for theory and research.
How happy is the blameless vestal’s lot! The world forgetting, by the world forgot. Eternal sunshine of the spotless mind! Each pray’r accepted, and each wish resign’d — Alexander Pope (1717/1796)
The depth and breadth of self-forgiveness literature has lagged behind interpersonal forgiveness scholarship since their commencement, but self-forgiveness has increased in popularity among researchers more recently (Hall & Fincham, 2005; McConnell, 2015). The number of self- forgiveness publications has exponentially grown from the early 1990s to present day (McConnell, 2015). Why, how, and should people forgive themselves have now been the subjects of partial and comprehensive reviews, including a major edited volume (Davis, Worthington, Hook, & Hill, 2013; Davis et al., 2015; Fisher & Exline, 2010; Hall & Fincham, 2005; Lavelock, Griffin, & Worthington, 2013; McConnell, 2015; Mullet, Neto, & Rivi�ere, 2005; Shan & Xu, 2008; Tangney, Boone, & Dearing, 2005; Webb, Hirsch, & Toussaint, 2011; Wohl & McLaughlin, 2014; Woodyatt, Worthington, Wenzel, & Griffin, 2017; Worthington, Witvliet, Pietrini, & Miller, 2007). However, historical work in self-forgiveness and its modern wave of popular- ity have mostly produced conceptual foundations and atheoretical empirical findings. To encourage greater theoretical specification as well as to warn against mono-methodological, mono-measurement-type, and
mono-analytical biases, McConnell (2015) reviewed the extant literature and proposed a conceptual-theoretical- empirical (C-T-E) framework for self-forgiveness. McConnell’s C-T-E framework organizes multiple phe- nomena in a broad conceptual structure, provides rela- tion specificity through theoretical foundations, and packages scientific support in a coherent manner (Fawcett, 1993). McConnell overviewed a multitude of midrange theories to refine conceptual propositions and empirical findings related to why, how, and should peo- ple forgive themselves. In the current series of experi- ments, I empirically tested if terror management theory (TMT; Greenberg, Pyszczynski, & Solomon, 1986; Pyszczynski, Greenberg, & Solomon, 1999) is an explan- ation for why people forgive themselves as specified in McConnell’s C-T-E framework. First, I broadly overview conceptual and empirical scholarship in the origins of self-forgiveness. Next, I specifically discuss self-forgive- ness embedded within a TMT framework. Finally, I describe four TMT self-forgiveness experiments along with their implications for further theoretically driven research.
Psychological origins of self-forgiveness
A substantial body of conceptual and empirical work has highlighted the harmful personal effects of self- unforgiveness (e.g., Davis et al., 2015; McConnell,
CONTACT John M. McConnell [email protected] School of Psychology, Counseling, and Family Therapy, 501 College Avenue, Wheaton, IL 60187-5501, USA. � 2019 Taylor & Francis Group, LLC
BASIC AND APPLIED SOCIAL PSYCHOLOGY 2018, VOL. 40, NO. 6, 341–373 https://doi.org/10.1080/01973533.2018.1513361
2015). For this reason, scholars view self-forgiveness, in part, as a utilitarian or adaptive process that allows people to move beyond the harmful biological, psy- chological, social, and existential/spiritual effects of acute and chronic self-conscious emotions. Consistent with Lazarus and Folkman’s (1984) stress and coping model, self-forgiveness may be a fundamental well- being safeguard (McConnell, 2015; Worthington et al., 2007). Worthington and colleagues (2007) indicated that self-forgiveness is both a problem-focused and an emotion-focused coping strategy that allows individuals to buffer against the unhealthy outcomes of self-conscious emotions. In support, a large body of correlational research has found that self-forgiveness is positively related to physical, psychological, social, and existential/spiritual well-being (see Davis et al., 2015; McConnell, 2015). In Davis et al.’s (2015) meta- analytic review of correlational research, they found self-forgiveness related to both physical and psycho- logical well-being. Of particular importance to the current studies, Davis et al. found self-forgiveness was related to life-satisfaction/meaning-in-life. This finding also is consistent with qualitative research. For example, people have sought self-forgiveness (Ferrell, Otis-Green, Baird, & Garcia, 2014) and forgiveness (Heflick, 2005) near the end of their lives.
Although it is reasonable to suggest self-forgiveness is related to well-being in multiple domains of human functioning based on the aforementioned findings, atheoretical and nonexperimental methods do not support that self-forgiveness causes increased well- being, nor do they support that self-forgiveness has existed to regulate well-being (Kline, 2015; Spencer, Zanna, & Fong, 2005; Trafimow, 2015). In an attempt to provide theoretical specification and increase experimental rigor in this subfield, McConnell (2015) indicated that TMT may be a cogent explanation for the psychological origins of self-forgiveness and its connection to (existential) well-being in correl- ational research.
Terror management theory
According to TMT, avoiding thoughts of death moti- vates people’s intrapsychic and social behavior because the finite nature of human beings can be a difficult reality to accept (Greenberg et al., 1986; Pyszczynski et al., 1999). Death awareness—when people become aware of their inevitable death—can be particularly anxiety provoking because humans are biologically ingrained with desires to sustain their lives. Yet all people know they must eventually perish and strongly
desire that death would not be a fundamental human condition. All human beings therefore have intrapsy- chic conflicts, or existential anxiety, as they struggle to reconcile their desire for immortality with the know- ledge of inevitable death (Greenberg et al., 1986; Pyszczynski et al., 1999). As Hayes, Schimel, Arndt, and Faucher (2010) stated, “The knowledge of death, coupled with a basic desire for continued life, creates an existential dilemma capable of producing poten- tially paralyzing anxiety, or terror” (p. 700).
According to TMT, people do not consciously experience this intrapsychic conflict for most of their waking life because humans have devised elaborate ways to suppress death awareness (Greenberg et al., 1986; Pyszczynski et al., 1999). For the times people are reminded of death, humans have developed a dual process component of proximal and distal defenses to suppress death-thoughts into the unconscious (Greenberg et al., 1986; Hayes et al., 2010; Pyszczynski et al., 1999). Proximal defenses are rational responses to death awareness that involve thought suppression, distraction, or underestimation of current death vul- nerability (Hayes et al., 2010). Upon reminders of death, people initially use proximal defenses to sup- press death awareness outside of their focal attention into the unconscious (Hayes et al., 2010; Pyszczynski et al., 1999). However, proximal defenses require sus- tained cognitive resources, and over time unconscious death-thoughts presumably become more readily accessible just outside of conscious awareness (Hayes et al., 2010; Pyszczynski et al., 1999), or “on the fringes of consciousness—highly accessible but not in current focal attention” (Pyszczynski, Solomon, & Greenberg, 2015, p. 15). In response to this death- thought suppression and delayed rebound effect (Trafimow & Hughes, 2012), people use distal defenses to further buffer against increasingly salient preconscious levels of death-thoughts as proximal defenses diminish (Hayes et al., 2010; Pyszczynski et al., 1999).
The proponents of TMT believe that humans devel- oped cultural worldviews and self-esteem as distal defenses against death awareness. First, cultural world- views buffer against existential anxiety by convincing humans that there is order in a world of chaos, pro- viding conscious and time-consuming activities, and ensuring them that there is purpose and meaning to their lives if they live up to certain expectations (Greenberg et al., 1986; Pyszczynski et al., 1999). In other words, culture provides humans illusory senses of immediate biological safety, distraction, and value. Second, self-esteem provides reassurance to
342 J. M. MCCONNELL
people that they are valuable and have lived up to their cultural expectations (Greenberg et al., 1986; Pyszczynski et al., 1999). According to TMT, cultural worldviews and self-esteem work in combination to convince humans that despite biological death they will persist on with real (e.g., heaven, reincarnation) or symbolic (e.g., fame, societal contribution) everlast- ing life (Greenberg et al., 1986). If biological death is not the end, then the terror it evokes has a dimin- ished psychological impact consciously.
TMT has amassed an impressive amount of empir- ical support over three decades with more than 500 studies to date (Darrell & Pyszczynski, 2016), includ- ing multiple meta-analyses (Burke, Kosloff, & Landau, 2013; Burke, Martens, & Faucher, 2010; Martens, Burke, Schimel, & Faucher, 2011; Steinman & Updegraff, 2015; vanDellen, Campbell, Hoyle, & Bradfield, 2011). Scholars elsewhere have extensively reviewed TMT’s theoretical modifications and experi- mental evidence (e.g., Burke et al., 2010; Hayes et al., 2010; Pyszczynski, Greenberg, & Solomon, 1997; Pyszczynski et al., 1999; Pyszczynski et al., 2015; Solomon, Greenberg, & Pyszczynski, 2004; Steinman & Updegraff, 2015; Vail et al., 2012). Despite TMT’s remarkable number of studies that have provided empirical support and its popularity in social psych- ology, at times it has been contradicted or unsup- ported by empirical evidence (e.g., Proulx & Heine, 2008; Trafimow & Hughes, 2012). TMT also has received considerable criticism (e.g., Martin & van den Bos, 2014; see also Psychological Inquiry special issues: Pyszczynski et al., 1997 [Vol. 8, Issue 1], and Martin & Erber, 2006 [Vol. 17, Issue 4]). For instance, two meta-analyses suggested that TMT’s tenets are not culturally universal (Yen & Cheng, 2010) and research effect sizes have been partially a by-product of experimental bias (Yen & Cheng, 2013).
Although Pyszczynski et al. (1997) claimed that TMT is a parsimonious theory to explain virtually all human behavior, and that it explains the findings of diverse theories of social motivation, other scholars have made equal attempts with competing theories to explain the social-cognitive phenomena and findings in the purview of TMT (e.g., Heine, Proulx, & Vohs, 2006; Proulx, Inzlicht, & Harmon-Jones, 2012; see also Martin & Erber, 2006 [Psychological Inquiry, Vol. 17, Issue 4], and Proulx, 2012 [Social Cognition, Vol. 30, Issue 6], for special issues). However, some evidence has indicated that death may be a unique psycho- logical threat. For example, Martens et al.’s (2011) meta-analysis found that death awareness elicited higher levels of defensiveness compared to meaning
or certainty threats when measurement occurred after delays from the experimental manipulation. Recently, scholars have attempted to integrate many theories of psychological defense, including TMT, into one coher- ent theoretical framework (e.g., George & Park, 2016; Hart, 2014), or worked to go beyond integration to unification in psychological theory (e.g., Henriques, 2011; Trafimow, 2012b; see also Gaj, 2016). Despite these attempts, TMT currently remains a mainstay in social psychology, and researchers have continued to investigate three main hypotheses derived from the theory: the mortality-salience (TMT Hy1), death- thought accessibility (TMT Hy2), and anxiety-buffer (TMT Hy3) hypotheses (Darrell & Pyszczynski, 2016; Pyszczynski et al., 2015).
Mortality salience hypothesis The TMT Hy1 posits that increased death awareness— operationalized as mortality salience (MS)—will cause people to cognitively suppress their death-related thoughts, defend their own unique cultural world- views, and bolster their self-esteem to buffer against conscious awareness of death (Pyszczynski et al., 1999; Pyszczynski et al., 2015). This hypothesis assumes that MS will initially increase proximal defenses and even- tually increase distal defenses once proximal ones diminish. In other words, people will display increased amounts of in-group identification, out-group dispar- agement, and self-esteem enhancement due to MS as cognitive suppression diminishes over time. For instance, people have initially suppressed thoughts of death following death awareness (Arndt, Cook, Goldenberg, & Cox, 2007), but after a delay distributed excessive hot sauce to other cultural group members’ food (McGregor et al., 1998) or displayed self-serving attributions (Mikulincer & Florian, 2002). McConnell (2015) indicated that people may engage in the process of self-forgiveness following MS as a method to reaffirm cultural values and repair self-esteem. If TMT is an explanatory framework for self-forgiveness, then MS would lead to increased self- forgiveness after a delay. No known work has tested the MS hypothesis in relation to self-forgiveness.
Burke et al.’s (2010) meta-analysis of 277 experi- ments found that increased initial MS had a so-called moderate effect on increased distal defense behaviors. Neutral (e.g., television) or negative (e.g., dental pain) comparison conditions and gender negligibly moder- ated the effect MS had on distal defenses, but being a United States citizen or an undergraduate increased the MS effect. When Yen and Cheng (2013) reana- lyzed the literature in another meta-analysis,
BASIC AND APPLIED SOCIAL PSYCHOLOGY 343
experimental team moderated the MS effect as well. The MS effect also increased with longer delays fol- lowing initial reminders of death (Burke et al., 2010). Time lapse from initial MS is an important theoretical concept in TMT. Distal defenses are supposedly not activated without a sufficient delay for increasingly salient nonconscious death-thoughts to become accessible (cf. Burke et al., 2010; Hayes et al., 2010; Pyszczynski et al., 1999; Steinman & Updegraff, 2015; Trafimow & Hughes, 2012).
Death-thought accessibility hypothesis The TMT Hy2 posits thoughts of death just outside of focal awareness—operationalized as death-thought accessibility (DTA)—mediate between MS and distal defenses (Hayes et al., 2010). In other words, TMT theorists believe that initial MS has an effect on distal defenses through DTA. Therefore, if TMT is the the- oretical foundation for self-forgiveness, then initial MS would lead to increased self-forgiveness after a delay as DTA becomes more salient. No known work has tested the DTA hypothesis in relation to self- forgiveness.
A growing body of literature has supported that MS, death-reminders, and other threats (e.g., cultural value, self-esteem) lead to increased DTA following a delay. Steinman and Updegraff’s (2015) meta-analysis of 99 experiments found a so-called moderate effect size on DTA following existentially related threats, and more dis- tractor tasks or longer delays increased the effect. Neutral or negative comparison conditions and the measurement type of DTA (e.g., word-fragment comple- tion, reaction time) negligibly moderated the effect. Despite the large body of evidence for the DTA hypoth- esis, there also has been conflicting evidence. For instance, in a series of six experiments, Trafimow and Hughes (2012) found the exact opposite of TMT Hy2. DTA actually decreased with delays. Similarly, other researchers have found that DTA negligibly changed with experimental manipulations in multiple unpublished studies (e.g., Hart, 2014; Hughes, as cited in Hart, 2014).
Given TMT researchers believe DTA mediates exist- ential threats and distal defenses, it is surprising at first glance that there is a paucity of studies within TMT research that directly test DTA as a mediator (Hayes et al., 2010). Researchers have begun such work, and multiple studies have implied that DTA either mediated or moderated existential threat and defense (e.g., Cox, Eaton, Ekas, & Van Enkevort, 2015; Cox, Reid-Arndt, Arndt, & Moser, 2012; Das, Bushman, Bezemer, Kerkhof, & Vermeulen, 2009; Echebarria-Echabe, 2013; Fransen, Fennis, Pruyn, & Das, 2008; Hayes et al., 2015;
Motyl et al., 2013; Vail, Arndt, Motyl, & Pyszczynski, 2012). Yet there are inherent challenges in demonstrat- ing DTA mediation for three main reasons. First, meas- urement-of-mediation designs cannot rationally provide direct support of mediation (Kline, 2015; Spencer et al., 2005; Tate, 2015; see Trafimow, 2015 [Basic and Applied Social Psychology, Vol. 37, Issue 4], for a special issue on mediation). Second, pure experimental-causal- chain designs are difficult to establish in TMT because manipulating implicit processes such as DTA are diffi- cult, if not impossible, without evoking MS or anxiety- buffer threats (Hayes et al., 2010; Hayes & Schimel, 2018). Third, a measurement-of-mediation design may inadvertently expose a comparison group to death- related stimuli while filling in a DTA measure. Recently, Hayes and Schimel (2018) demonstrated that the measurement of DTA in comparison groups obfus- cated MS threat effects on distal defenses, but there was a rebound effect after a post-DTA measurement delay. Hayes and Schimel (2018) indicated that DTA measure- ment itself becomes a MS threat, and therefore meas- urement artifacts minimize the chances of empirically demonstrating the mediational aspect of the DTA hypothesis in one experiment.
Anxiety-buffer hypothesis The TMT Hy3 posits strengthening self-esteem pro- tects against anxiety with subsequent existential threats. Inversely, weakening self-esteem makes people vulnerable to existential anxiety (Hayes et al., 2010; Pyszczynski, Greenberg, Solomon, Arndt, & Schimel, 2004; Pyszczynski et al., 1999, 2015). According to the proponents of TMT, the anxiety-buffer regulates exist- ential anxiety, DTA, and worldview defenses with and without explicit reminders of death. Thus, if TMT is a theoretical underpinning of self-forgiveness, then MS and subsequent DTA would lead to self-forgiveness when self-esteem is low. In addition, even in the absence of MS threats, decreased self-esteem would lead to increased DTA and, in turn, increased self-for- giveness. No known work has tested the anxiety-buffer hypothesis in relation to self-forgiveness.
A large body of literature has supported that self- esteem serves as an anxiety buffer against MS, world- view threats, DTA, and other existentially related phe- nomena (Pyszczynski et al., 2004; Hayes et al., 2010; Schimel, Landau, & Hayes, 2008; Steinman & Updegraff, 2015). Research has supported that self- esteem serves as a buffer against general and existential anxiety and that people bolster their self-esteem to pro- tect against existential angst (Pyszczynski et al., 2004; Schimel et al., 2008). For instance, Greenberg et al.
344 J. M. MCCONNELL
(1992) found increased self-esteem through positive per- sonality feedback decreased anxiety about a death- related video.
Hypotheses summary TMT Hy1, TMT Hy2, and TMT Hy3 each represent a unique facet of TMT’s theoretical underpinnings and empirical research. Hayes et al. (2010) offered an illus- trative theoretical process model of TMT that depicts the temporal sequence these hypotheses claim. In short, MS leads to activation of the anxiety buffer either immediately or following a delay if proximal defenses are initially active. If the anxiety buffer is weak, then people will experience high levels of DTA and subsequent distal defenses to decrease DTA. If the anxiety buffer is strong, then people will experi- ence low levels of DTA (Hayes et al., 2010). Interested readers can turn to Hayes et al. (2010) for a compre- hensive review and description of their TMT process model, including various nuance processes. Although researchers often have isolated these hypotheses, the processes they describe are inextricably interrelated. Therefore, I tested all three hypotheses in the current series of studies to comprehensively evaluate if TMT is an explanatory framework for self-forgiveness.
Self-forgiveness: A self-esteem distal defense process?
According to TMT, self-esteem is a fundamental psy- chological mechanism that buffers against existential anxiety by convincing people they are worthy of either real or symbolic eternal life because they have lived up to their cultural standards (Greenberg et al., 1986; Pyszczynski et al., 1999, 2004). People can fail to meet cultural expectations in many ways, and modern lexi- con includes numerous words to describe such wrong- doings. For instance, people can commit wrongdoings that are immoral (e.g., wicked deed, evil doing, atro- city), unethical (e.g., violation, infraction, misconduct), illegal (e.g., offense, crime, misdemeanor, felony), interpersonal (e.g., let down, betrayal, harm, hurt, wound, injure), self-inflicted (e.g., self-harm, self- injury, self-handicap, mistake, blunder), and religious/ spiritual (e.g., sin, transgression, bad karma, p�apa, dhanb) in nature. McConnell (2015) theorized that guilt and shame for cultural worldview violations were the nexus among self-esteem, death anxiety, and exist- ential motivations for self-forgiveness. If Solomon, Greenberg, and Pyszczynski’s (1997, p. 70) fundamen- tal proposition that death is central to “virtually every- thing that we think, feel, say, and do” is true, then it
would follow: (a) culture-based rule violations cause guilt and shame, (b) which reduce self-esteem because of their death-related implications, and (c) self-for- giveness is an attempt to mollify existential angst based in such self-conscious emotions.
To expand on this theoretical extrapolation of TMT, people experience guilt or shame when they have violated cultural rules and they have taken responsibility for their actions (McConnell, 2015). Guilt and shame are self-conscious signifiers to indi- viduals that they have violated cultural expectations required for real or symbolic eternal life. As a result, the self-conscious nature of guilt and shame reduce self-esteem. With the anxiety buffer weakened, MS and DTA will cause people to experience existential anxiety as they struggle to reconcile their finite lives and imperfect character. Existential anxiety will be amplified to the extent that they consciously and unconsciously believe cultural norm violators are unworthy to inherent real or symbolic everlasting life. In this context, death has frightening implications when self-esteem does not provide reassurance that biological death will simply be a new existential awak- ening. Faced with this existential anxiety, people are compelled to repair their self-esteem and existential well-being through self-forgiveness. Overall, if the anxiety buffer is weak (TMT Hy3), MS should lead to DTA after a delay (TMT Hy2), and in turn the distal defense of self-forgiveness (TMT Hy1).
If these assumptions were true, then TMT would shed an interesting light on why guilt, shame, and decreased self-forgiveness have consistently related to depression, anxiety, and decreased overall well-being (e.g., Davis et al., 2015; Kim, Thibodeau, & Jorgensen, 2011; McConnell, 2015; Tangney & Dearing, 2002) and why death anxiety has played a unique role in psychopathology (Iverach, Menzies, & Menzies, 2014). Several studies have indicated that people experienced guilt following MS (e.g., Arndt, Greenberg, Solomon, Pyszczynski, & Schimel, 1999; Fergus & Valentiner, 2012; Goldenberg, Heflick, & Cooper, 2008; Harrison & Mallet, 2013; Hirschberger, Florian, & Mikulincer, 2002), and one qualitative study revealed that some TMT participants wrote about guilt during MS manipulations (Kastenbaum & Heflick, 2011). In add- ition, anecdotal evidence indicated that some people have sought self-forgiveness (Ferrell et al., 2014) and forgiveness (Heflick, 2005) near the end of their lives. Researchers have studied forgiveness and recon- ciliation in the context of TMT (Anglin, 2014; Schimel, Wohl, & Williams, 2006; Van Tongeren, Green, Davis, Worthington, & Reid, 2013; Wilson &
BASIC AND APPLIED SOCIAL PSYCHOLOGY 345
Bernas, 2011), but the current studies are the first known attempts to evaluate if TMT is an explanatory framework for self-forgiveness.
General method
In a series of four experiments, I examined the valid- ity of TMT as a theoretical foundation for self-forgive- ness by systematically testing TMT Hy1, TMT Hy2, and TMT Hy3. I expected that MS would lead to the distal defense of self-forgiveness (TMT Hy1) through DTA after a delay (TMT Hy2) in the context of a weakened anxiety buffer from recalling an interper- sonal offense (TMT Hy3). I considered TMT’s meth- odological designs, procedures, and measurements known to elicit the highest effect sizes (cf. Burke et al., 2010, 2013; Hayes et al., 2010; Martens et al., 2011; Steinman & Updegraff, 2015; Yen & Cheng, 2010, 2013) in the conceptualization of the current studies, although they were designed and conducted before the online publication of Hayes and Schimel’s (2018) study. The following is an overview of an ini- tial attempt to test all three of TMT’s experimental hypotheses in one study and an experimental-causal- chain procedure designed to test the sequence in a step-by-step manner across three follow-up studies. All studies received prior approval from the author’s Institutional Research Review Boards.
Experiment 1
In Experiment 1, I tested an interactional-moderated- mediation model among MS, DTA, and self- forgiveness. Specifically, following an anxiety-buffer threat of a recalled recent interpersonal offense, I tested if DTA mediated the relation between MS and state self-forgiveness after a delay, and if this relation was moderated by offense severity, conciliatory behav- iors, perceived forgiveness, or effort. I selected these four moderators to control for their statistical variance due to their consistent relation to the process of self- forgiveness. To maximize construct validity, I utilized state measures of guilt and self-forgiveness to create the latent outcome variable “self-forgiveness” and assessed participants’ level of responsibility to rule out pseudo self-forgiveness. I excluded shame because of its inconsistent relation to self-forgiveness (see McConnell, 2015, for a full rationale of contextual variables and construct validity). I used a randomized, post-test-only, double-blind, comparison-group design (MS vs. dental pain salience [DP]) combined with a measurement-of-mediation procedure (Shadish, Cook,
& Campbell, 2002; Spencer et al., 2005). I tested the following hypotheses in a single study:
1. Mediation hypothesis. Mediation can be minim- ally alluded to when an experimentally manipu- lated independent/predictor variable (MS) affects a dependent/criterion variable (self-forgiveness) but only through a measured third variable (DTA). There also must be a conceptually based temporal sequence (Kline, 2015; Spencer et al., 2005; Tate, 2015). Partial mediation is suggested when the relation between the predictor and cri- terion variables is reduced in part after account- ing for the mediating effect, whereas full mediation is implied when the relation is nullified (Baron & Kenny, 1986; Kline, 2015; Preacher, Rucker, & Hayes, 2007; Sobel, 1982). Because TMT theorists posit MS produces distal defenses through emerging preconscious death-thoughts (e.g., Hayes et al., 2010), I predicted that DTA (TMT Hy2) would fully mediate between MS and self-forgiveness (TMT Hy1) following an anxiety- buffer threat of recalling an interpersonal offense (TMT Hy3).
2. Moderated-mediation hypothesis. A moderated mediation is suggested when the strength or dir- ection of a mediated relation is contingent on one or more ancillary variables (Preacher et al., 2007). In other words, the presence of other variables strengthens or weakens the relation of the medi- ated process. MS and DTA may help explain why people strive for self-forgiveness, although offense severity, conciliatory behaviors, perceived forgive- ness, and effort to reduce emotions play import- ant roles in how easily people accomplish self- forgiveness. Therefore, I predicted that higher lev- els of conciliatory behaviors, perceived forgive- ness, and effort would increase the relation between DTA and self-forgiveness, whereas higher levels of offense severity would decrease the rela- tion (McConnell, 2015).
3. Interactional-moderated-mediation hypothesis. Interactional-moderated-mediation is implied when a moderated-mediation relation is not apparent at all levels of the independent variable (Hayes et al., 2010; Preacher et al., 2007). In other words, the experimental group would exhibit a moderated-mediation relation, but the compari- son group would not. Participants in a TMT com- parison group are less likely to have sufficient DTA to exhibit any distal defenses because they are not explicitly reminded of death, although
346 J. M. MCCONNELL
they may have inadvertent implicit exposure (Burke et al., 2010; Hayes et al., 2010; Hayes & Schimel, 2018; Steinman & Updegraff, 2015). Therefore, I predicted that the MS experimental group would have more DTA and self-forgiveness than the DP comparison group, and in turn DTA would positively relate to self-forgiveness for the MS group but not for the DP group.
Method
Participants Participants solicited from introductory to psychology (n ¼ 175; 77.8%) or interpersonal relations (n ¼ 50; 22.2%) classes at a public university were recruited via a posting on the psychology department’s website and were allowed to participate if they were at least 18 years old. An initial 226 participants participated; however, one person was not included in the analyses because he admitted to random responding. See Table 1 for a summary of demographic characteristics.
Procedure Participants registered for the experiment using an online scheduling program. Similar to previous research, potential participants were recruited to enroll in the study if they felt poorly about hurting one person close to them with their words or actions within the last month (Exline et al., 2007, 2011; Fisher & Exline, 2006; Hall & Fincham, 2008). Upon arriving at an experimental room within 3 days after recruit- ment (Hall & Fincham, 2008), participants were greeted by my research assistant who was blind to the experimental conditions. Participants first completed an informed consent and then an offense packet while seated privately in a classroom setting. The first task in the offense packet was to recall and briefly write about their recent interpersonal wrongdoing in order to induce an anxiety-buffer threat. Participants then filled in counterbalanced measures of responsibility, offense severity, conciliatory behavior, and perceived forgiveness. (See Figure 1 for a visual depiction of the experimental sequence and the Materials section for a description of packets used in Experiment 1.)
Participants then completed the death awareness packet, which consisted of several personality meas- ures administered under the guise of collecting data for another researcher who “asked us to collect data for them” as other TMT researchers have done (see Burke et al., 2010). I included this deception to reduce suspicion about filling in the criterion and effort measures after the experimental manipulation rather
than directly after the offense recall (Burke et al., 2010; Hayes et al., 2010). Embedded at the end of this packet, I randomly assigned participants to one of two conditions blind to the research assistant. Participants were asked to respond to two open-ended questions designed to increase either (a) MS (n ¼ 112; 49.8%) or (b) DP (n ¼ 113; 50.2%; Rosenblatt, Greenberg, Solomon, Pyszczynski, & Lyon, 1989). My research assistant then gave them a number of additional coun- terbalanced measures to delay time after the experi- mental manipulation. This distractor-delay aimed to increase the MS effect on DTA and dependent meas- ures, as well as rule out effects of social desirability and affect (cf. Burke et al., 2010; Hayes et al., 2010; Steinman & Updegraff, 2015; Trafimow & Hughes, 2012).
Next, both the experimental (MS) and comparison (DP) groups filled in a DTA word-fragment comple- tion task. The research assistant then told participants that he forgot to administer some measures from the prior study. All participants then filled in state self- forgiveness and guilt scales in counterbalanced orders followed by an effort to reduce emotions measure. Participants subsequently completed a demographic questionnaire. Next, I personally interviewed partici- pants about any emotional distress, demand character- istics, and hypotheses guessing. Participants reported no adverse events. Interviews revealed zero partici- pants, including those who were suspicious of poten- tial deception (n ¼ 54), were able to guess the rationale for the study’s methodology or its hypothe- ses, and no one was even aware of TMT itself. Finally, I debriefed them about the deception and gave them research credit through an online database.
Materials I used the following offense and death awareness packets, as well as the distractor-delay, DTA, and cri- terion measures. See Table 2 for Experiment 1 scale properties.
Offense packet. The following offense recall instruc- tions and counterbalanced measures were given to participants to assess their wrongdoing and related social-cognitive phenomena.
Offense recall. Participants received the following instructions:
Please recall a recent event in the past month in which you hurt someone close to you by something you said or did. Recall an event in which you currently—right now—feel blameworthy and
BASIC AND APPLIED SOCIAL PSYCHOLOGY 347
Ta b le
1. D em
og ra p h ic ch ar ac te ri st ic s fo r Ex p er im en ts
1– 4.
M (S D )
Ra n g e
1 2
3 4
1 2
3 4
A g e
19 .7 7 (1 .8 1)
39 .5 6 (1 5. 42 )
42 .5 7 (1 6. 19 )
39 .6 8 (1 5. 67 )
18 /3 0
18 /8 0
18 /8 8
18 /8 4
Ed uc at io n
13 .1 2 (1 .3 8)
13 .6 2 (3 .2 5)
13 .7 8 (3 .6 3)
13 .7 2 (3 .5 8)
12 /2 1
0/ 25
0/ 30
0/ 30
D ay s si n ce
of fe n se
17 .3 8 (1 0. 30 )
98 .5 3 (1 11 .1 6)
– 87 .6 2 (1 05 .1 1)
1/ 31
1/ 36 5
– 1/ 36 5
Ti m es
co m m it te d
2. 93
(8 .8 8)
15 78 .2 4 (3 37 11 .2 9)
– 13 31 .4 1 (3 36 50 .3 3)
1/ 10 0
1/ 99 99 99
– 1/ 10 00 00 0
G en d er
Re lig io n
1 2
3 4
1 2
3 4
M an
79 (3 5. 1%
) 38 1 (4 0. 9%
) 16 8 (3 5. 3%
) 38 3 (4 2. 9%
) C at h ol ic
55 (2 4. 4%
) 16 8 (1 8. 0%
) 86
(1 8. 1%
) 15 8 (1 7. 7%
) W om
an 14 5 (6 4. 4%
) 55 0 (5 9. 0%
) 30 6 (6 4. 3%
) 50 7 (5 6. 8%
) Pr ot es ta n t
86 (3 8. 2%
) 35 5 (3 8. 1%
) 16 6 (3 4. 8%
) 33 7 (3 7. 8%
) O th er
1 (0 .4 % )
1 (0 .1 % )
2 (0 .4 % )
2 (0 .4 % )
O rt h od
ox 0 (0 % )
7 (0 .8 % )
2 (0 .4 % )
5 (0 .6 % )
Je w is h
5 (2 .2 % )
16 (1 .7 % )
8 (1 .7 % )
19 (2 .1 % )
M us lim
3 (1 .3 % )
6 (0 .6 % )
2 (0 .4 % )
4 (0 .4 % )
Bu d d h is t
3 (1 .3 % )
15 (1 .6 % )
9 (1 .9 % )
15 (1 .7 % )
H in d u
1 (0 .4 % )
4 (0 .4 % )
6 (1 .3 % )
4 (0 .4 % )
Et h n ic it y
A g n os ti c
25 (1 1. 1%
) 65
(7 .0 % )
33 (6 .9 % )
49 (5 .5 % )
1 2
3 4
A th ei st
20 (8 .9 % )
55 (5 .9 % )
19 (4 .0 % )
53 (5 .9 % )
W h it e
18 4 (8 1. 8%
) 72 0 (7 7. 3%
) 38 0 (7 9. 8%
) 67 9 (7 6. 1%
) N on
e 27
(1 2%
) 24 1 (2 5. 9%
) 14 5 (3 0. 4%
) 24 8 (2 7. 8%
) A fr ic an
A m er ic an /B la ck
20 (8 .9 % )
78 (8 .4 % )
36 (7 .6 % )
85 (9 .9 % )
La ti n o/ a
2 (0 .9 % )
67 (7 .2 % )
18 (3 .8 % )
46 (5 .2 % )
M ar it al St at us
A si an
or A si an
A m er ic an
5 (2 .2 % )
33 (3 .5 % )
22 (4 .7 % )
38 (4 .2 % )
1 2
3 4
N at iv e A m er ic an
0 (0 .0 % )
6 (0 .6 % )
7 (1 .5 % )
11 (1 .2 % )
Si n g le
20 3 (9 0. 2%
) 27 2 (2 9. 2%
) 11 7 (2 4. 6%
) 28 7 (3 2. 2%
) Pa ci fic
Is la n d er
0 (0 .0 % )
1 (0 .1 % )
2 (0 .4 % )
4 (0 .4 % )
C oh
ab it at in g
16 (7 .1 % )
15 9 (1 7. 1%
) 49
(1 0. 3%
) 13 6 (1 5. 2%
) N . A fr ic an
an d M . Ea st er n
0 (0 .0 % )
1 (0 .1 % )
2 (0 .4 % )
4 (0 .4 % )
M ar ri ed
4 (1 .8 % )
36 5 (3 9. 2%
) 23 0 (4 8. 3%
) 33 7 (3 7. 8%
) Bi ra ci al
8 (3 .6 % )
15 (1 .6 % )
2 (0 .4 % )
6 (0 .7 % )
Se p ar at ed
2 (0 .9 % )
16 (1 .7 % )
11 (2 .3 % )
28 (3 .1 % )
M ul ti ra ci al
5 (2 .2 % )
7 (0 .8 % )
2 (0 .4 % )
12 (1 .3 5%
) D iv or ce d
0 (0 .0 % )
91 (9 .8 % )
45 (9 .5 % )
76 (8 .5 % )
U n sp ec ifi ed
1 (0 .4 % )
4 (0 .4 % )
5 (1 .1 % )
4 (0 .4 % )
W id ow
ed 0 (0 .0 % )
29 (3 .1 % )
24 (5 .0 % )
28 (3 .1 % )
N ot e. Ex p er im en ta l co n d it io n s co m b in ed
w it h in
st ud
ie s, 1:
n ¼ 22 5;
2: n ¼ 93 2;
3: n ¼ 47 6;
4: n ¼ 89 2.
348 J. M. MCCONNELL
personally upset about what you did. For a moment, visualize in your mind the events and the interactions you may have had with the person you offended. Try to visualize the person and recall what happened. Now please briefly describe the event in five to seven sentences.
Forgiveness and self-forgiveness researchers have used similar recalls in previous research (Exline et al., 2007; Fisher & Exline, 2006; Hall & Fincham, 2008; McCullough et al., 1998). My research assistant and I reviewed offense recalls to ensure that participants adequately responded to the prompt with an interper- sonal offense.
Responsibility. The Revised Causal Dimension Scale (CDSII; McAuley, Duncan, & Russell, 1992) consists of four separate subscales—Causality, External Control, Personal Control, and Stability—which all have demonstrated adequate levels of internal consist- ency reliabilities in a previous sample. All nine items are rated with a 9-point bipolar-type scale (e.g., Is the cause(s) something: permanent vs. temporary; inside of you vs. outside of you). The CDSII also has dem- onstrated construct validity as evidenced by content validity and confirmatory factor analysis (McAuley et al., 1992). Although the internal consistencies for the external control (a ¼ .73; M ¼ 11.34, SD ¼ 5.92) and personal control (a ¼ .78; M ¼ 20.78, SD ¼ 5.56) were adequate, they were not for causality (a ¼ .58; M ¼ 16.51, SD ¼ 5.46) and stability (a ¼ .35; M ¼ 11.67, SD ¼ 4.93) in Experiment 1.
Perceived offense severity. The Perceived Offense Severity Scale (POS; Hall & Fincham, 2008) is a three-
item measure on a 7-point Likert-type scale, from 1 (very positively) to 7 (very negatively), that assesses how people believe they affected the victim, themselves, and their relationship with the victim (e.g., How did your behavior affect the victim?). The scale has demonstrated an adequate level of internal consistency reliability in a previous sample (Hall & Fincham, 2008). Hall and Fincham (2008) found the POS had a negative correl- ation with self-forgiveness, which supported its conver- gent validity (Hall & Fincham, 2008). The internal consistency was less than adequate (a ¼ .53; M¼ 17.25, SD ¼ 2.53) in Experiment 1.
Conciliatory behaviors. A modified version of the CBS (McCullough et al., 1997) developed by McConnell, Dixon, and Finch (2012) is a five-item measure (e.g., I tried/will try to make amends or compensations) rated on a 7-point Likert-type scale from 1 (strongly dis- agree) to 7 (strongly agree). The modified CBS demon- strated a desirable level of internal consistency reliability in a previous sample (McConnell et al., 2012). The modified CBS has good content validity because it includes all components of an effective apology (Olshtain, 1989; Weiner et al., 1991). The modified CBS also has demonstrated construct validity as evidenced by confirmatory factor analysis (McConnell et al., 2012). McConnell et al. (2012) found that the CBS had a positive correlation with perceived forgiveness, which supported its convergent validity. The internal consistency was adequate (a ¼ .85; M ¼ 27.28, SD ¼ 7.75) in Experiment 1.
Perceived forgiveness. To quantify participants’ levels of perceived forgiveness, I used two items developed
Additional MaterialsOffense Packet
Offense Recall
Counterbalanced
Responsibility (CDS-II)
Perceived Offense Severity (POS)
Conciliatory Behaviors
(CBS)
Perceived Forgiveness (PF-Victim)
Death Awareness Packet
Distractor-Delay Measures
Mediating Measure Criterion Measures
Cover Story
Narcissism (NPI-40)
Personality (BFI-44)
M O R T A L I T Y
S A L I E N C E
Counterbalanced
Social Desirability (M-C SDS)
Mood (PANAS)
Puzzle Task
Death-Thought Accessibility (DTA Word Completion
Task)
Counterbalanced
Self-Forgiveness (SSFS)
Guilt (SSGS)
Effort (ES)
Demographic Questionnaire
D E N T A L
P A I N
Counterbalanced
Social Desirability (M-C SDS)
Mood (PANAS)
Puzzle Task
Death-Thought Accessibility (DTA Word Completion
Task)
Counterbalanced
Self-Forgiveness (SSFS)
Guilt (SSGS)
Effort (ES)
Demographic Questionnaire
Threat Condition (MAPS)
Mortality Salience
OR
Dental Pain
Salience
D E C E P T I O N
Figure 1. Graphical representation of the experimental procedure for Experiment 1.
BASIC AND APPLIED SOCIAL PSYCHOLOGY 349
Ta b le
2. Sc al e p ro p er ti es
fo r Ex p er im en ts
1– 4.
a M
SD Ra n g e
VI F
Sk ew
n es s
Ku rt os is
1 2
3 4
1 2
3 4
1 2
3 4
1 2
3 4
1 2
3 4
1 2
3 4
1 2
3 4
D TA
†� –
– –
– 1. 66
– 1. 88
1. 79
1. 09
– 0. 99
1. 02
0/ 5
– 0/ 5
0/ 6
1. 12
– 1. 07
1. 05
0. 39
– �0
.8 1
0. 27
�0 .3 4
– �0
.3 4
0. 26
SF FA
.8 7
.8 7
– .8 6
19 .0 6
21 .3 5
– 21 .7 2
5. 75
5. 72
– 5. 66
8/ 32
8/ 32
– 8/ 32
3. 15
2. 83
– 3. 09
0. 16
�0 .1 9
– �0
.2 0
�0 .5 0
�0 .4 2
– �0
.3 1
SF B
.9 0
.9 1
– .9 2
26 .8 0
27 .7 8
– 28 .0 0
6. 18
6. 56
– 6. 59
10 /3 6
9/ 36
– 9/ 36
2. 69
2. 73
– 3. 33
�0 .5 6
�0 .7 9
– �0
.9 0
�0 .3 8
0. 07
– 0. 36
H FS
– –
.7 1
– –
– 28 .6 3
– –
– 6. 63
– –
– 6/ 42
– –
– 2. 02
– –
– 0. 00
– –
– �0
.2 7
– PP RS
– .8 1
– .8 0
– 36 .6 7
– 36 .5 4
– 11 .2 5
– 11 .2 4
– 5/ 50
– 5/ 50
– 1. 28
– 1. 30
– �0
.5 4
– �0
.5 6
– �0
.4 8
– �0
.4 2
C D SI I: Ex te rn al
.7 3
– –
– 11 .3 4
– –
– 5. 92
– –
– 3/ 27
– –
– 1. 44
– –
– 0. 40
– –
– �0
.6 7
– –
– C D SI I: Pe rs on
al .7 8
– –
– 20 .7 8
– –
– 5. 56
– –
– 3/ 27
– –
– 1. 39
– –
– �1
.0 2
– –
– 0. 65
– –
– PO
S .5 3
.6 6
– .7 5
17 .2 5
15 .2 6
– 15 .0 7
2. 53
3. 36
– 3. 61
5/ 21
3/ 21
– 3/ 21
1. 50
1. 43
– 1. 41
�0 .8 7
�0 .2 4
– �0
.3 2
2. 01
0. 21
– 0. 28
SS G S: G ui lt (R )
.7 9
.8 7
– .8 9
11 .6 8
16 .2 4
– 16 .0 2
4. 33
5. 84
– 5. 89
5/ 25
5/ 25
– 5/ 25
1. 99
1. 81
– 1. 97
0. 73
�0 .3 6
– �0
.2 4
0. 53
�0 .9 2
– �1
.0 2
C BS
.8 5
– –
– 27 .2 8
– –
– 7. 75
– –
– 5/ 35
– –
– 1. 58
– –
– �0
.9 4
– –
– �0
.1 7
– –
– PF -V ic ti m †
– –
– –
3. 41
– –
– 1. 27
– –
– 1/ 5
– –
– 1. 59
– –
– �0
.4 6
– –
– �0
.7 4
– –
– ES
.7 5
– –
– 10 .4 3
– –
– 5. 09
– –
– 0/ 20
– –
– 1. 35
– –
– �0
.2 9
– –
– �0
.6 8
– –
– RS ES
– .8 8
.9 0
.8 9
– 18 .4 9
19 .7 5
18 .2 8
– 6. 27
6. 25
6. 55
– 0/ 30
0/ 30
0/ 30
– 2. 09
2. 31
2. 44
– �0
.0 1
�0 .2 0
�0 .1 7
– �0
.5 2
�0 .4 8
�0 .3 5
M -C
SD S( K R )
.7 7
.8 1
.7 8
.7 9
13 .7 1
17 .0 8
17 .9 0
16 .7 5
5. 02
5. 67
5. 47
5. 51
2/ 27
2/ 32
0/ 33
3/ 32
1. 83
1. 31
1. 36
1. 24
�0 .0 3
0. 05
0. 24
0. 11
�0 .5 0
�0 .3 1
0. 05
�0 .2 6
PA N A S- Po si ti ve
.8 6
.9 1
.9 2
.9 1
28 .8 7
30 .9 7
30 .0 2
29 .7 0
8. 41
9. 42
9. 58
9. 43
10 /4 8
10 /5 0
10 /5 0
10 /5 0
1. 56
1. 38
1. 61
1. 40
�0 .0 4
�0 .0 4
�0 .0 6
0. 07
�0 .6 5
�0 .6 4
�0 .6 4
�0 .7 3
PA N A S- N eg at iv e
.8 5
.9 3
.9 3
.9 3
19 .6 2
19 .2 8
18 .1 6
19 .8 7
7. 83
9. 34
9. 14
9. 44
10 /4 4
10 /5 0
10 /4 9
10 /4 8
1. 56
1. 41
1. 41
1. 61
0. 89
0. 94
1. 22
0. 85
0. 09
0. 01
0. 85
�0 .2 3
N ot e.
Ex p er im en ta l co n d it io n s co m b in ed
w it h in
st ud
ie s,
1: n ¼ 22 5;
2: n ¼ 93 2;
3: n ¼ 47 6;
4: n ¼ 89 2.
a: C ro n b ac h ’s
al p h a;
KR : Ku
d er -R ic h ar d so n -2 0
A lp h a;
VI F:
V ar ia n ce
In fla ti on
Fa ct or
(M ul ti co lli n ea ri ty );
D TA
: D ea th
Th ou
g h t A cc es si b ili ty ; SF FA
: St at e Se lf- Fo rg iv in g Fe el in g s an d A ct io n s; SF B:
Se lf- Fo rg iv in g Be lie fs ; H FS : H ea rt la n d Fo rg iv en es s Sc al e;
PP RS : Pe rc ei ve d Pe rs on
al Re sp on
si b ili ty
Sc al e;
C D SI I: Re vi se d C au sa l
D im en si on
Sc al e;
PO S:
Pe rc ei ve d
O ff en se
Se ve ri ty ; SS G S:
St at e Sh am
e an d
G ui lt
Sc al e:
G ui lt
Su b sc al e (R : G ui lt
su b sc al e re ve rs ed
in Ex p er im en t 1) ; C BS : C on
ci lia to ry
Be h av io rs
Sc al e;
PF -V ic ti m : Pe rc ei ve d
Fo rg iv en es s fr om
V ic ti m ; ES : Ef fo rt Sc al e; RS ES : Ro se n b er g Se lf- Es te em
Sc al e; M -C
SD S: M ar lo w -C ro w n e So ci al D es ir ab ili ty
Sc al e; PA
N A S: Po si ti ve
an d N eg at iv e A ff ec t Sc h ed ul e.
†S in g le -i te m
sc al e.
� E xp er im en ts
3 an d 4 N eu tr al D TA
co n d it io n s ex cl ud
ed in
D TA
p ro p er ti es
p re se n te d .
350 J. M. MCCONNELL
by Hall and Fincham (2008; i.e., The victim has for- given me and A higher power has forgiven me) rated on a 5-point Likert-type scale from 1 (not at all) to 5 (completely). These two items are treated as two separ- ate single-item measures. I used only the perceived victim forgiveness (PF-Victim) item in the analyses because Experiment 1 was concerned primarily with self-forgiveness for interpersonal offenses. PF-Victim has positively correlated to self-forgiveness, which supported its convergent validity (Hall & Fincham, 2008). PF-Victim had a mean of 3.41 (SD ¼ 1.27) in Experiment 1.
Death awareness packet. The following packet was given to participants ostensibly as data for another research project in the order presented. These meas- ures and the experimental manipulation were not counterbalanced to ensure that each participant had consistent times between the experimental manipula- tion and filling in the remaining measures.
Cover story instructions. To conceal the experimental manipulation, the research assistant told the partici- pants, “Another researcher has asked us to collect data for them because they are under a time crunch. The following is a number of measures the researcher asked us to have you fill in. … Please fill in this packet as well.”
Narcissism. The 40-item Narcissistic Personality Inventory (NPI-40; Raskin & Terry, 1988) was used as filler for the cover story, and it was not used in the main statistical analyses. The NPI-40 has demon- strated adequate levels of internal consistency, 13- week test–retest reliability, and construct validity as evidenced by confirmatory factor analysis in previous samples (del Rosario & White, 2005; Raskin & Terry, 1988). The NPI-40 is a forced-choice questionnaire whereby the more narcissistic item increases partic- ipants’ total score (e.g., I am much like everybody else vs. I am an extraordinary person). The internal con- sistency was adequate (Kuder-Richardson-20 ¼ .81; M ¼ 15.98, SD ¼ 6.35) in Experiment 1.
Personality. The 44-item Big Five Inventory (John, Donahue, & Kentle, 1991) also was used as filler for the cover story, and it too was not used in the main statistical analyses. Each of the 44-item Big Five Inventory’s five domains of personality—Openness to Experience, Conscientiousness, Extraversion,
Agreeableness, and Neuroticism—has had adequate internal consistency, test–retest reliabilities, and con- vergent validity with the NEO Five-Factor Inventory and the Revised NEO Personality Inventory in previ- ous samples (John, Naumann, & Soto, 2008; Rammstedt & John, 2007). Each item is measured on a 5-point Likert-type scale from 1 (disagree strongly) to 5 (agree strongly). The internal consistency for Openness (a ¼ .77; M¼ 36.29, SD¼ 6.50), Conscientiousness (a ¼ .75; M¼ 31.19, SD¼ 5.26), Extraversion (a ¼ .88; M¼ 28.05, SD¼ 6.86), Agreeableness (a ¼ .82; M¼ 34.91, SD¼ 6.13), and Neuroticism (a ¼ .87; M¼ 23.58, SD¼ 7.04) were adequate in Experiment 1.
Threat Condition. The threat condition manipulation consisted of the Mortality Attitudes Personality Survey (Rosenblatt et al., 1989). This manipulation asks par- ticipants to respond to two open-ended questions regarding either death (experimental) or dental pain (comparison): “Please briefly describe the emotions that the thought of (your own death | dental pain) arouses in you” and “Jot down, as specifically as you can, what you think will happen to you as you physic- ally (die | experience dental pain) and once you (are physically dead | have experienced dental pain).” The Mortality Attitudes Personality Survey is reportedly an effective manipulation and is the most widely used in TMT research (Burke et al., 2010; Hayes et al., 2010). It also yields equivalent effect sizes to other MS manipulations such as death videos and subliminal messages (Burke et al., 2010). A neutral condition was not included in this study because it does not have larger effect sizes compared to a negative stimulus such as dental pain (Burke et al., 2010; Steinman & Updegraff, 2015).
Distractor-delay measures. The following three meas- ures were given as distractor-delays in counterbal- anced orders, and I used the first two to rule out confounding variables.
Social desirability. The Marlowe-Crowne Social Desirability Scale (M-C SDS; Crowne & Marlowe, 1960) contains 33 true–false items (e.g., I never resent being asked to return a favor). The M-C SDS total score increases as participants endorse socially desirable responses. The M-C SDS has demonstrated desirable levels of internal consistency reliability and 4-week test–retest reliability in a previous sample (Crowne & Marlowe, 1960). The M-C SDS has posi- tively correlated to the Lie validity scale of the
BASIC AND APPLIED SOCIAL PSYCHOLOGY 351
Minnesota Multiphasic Personality Inventory, which indicated convergent validity (Crowne & Marlowe, 1960). Experiment 1’s sample had a similar mean and standard deviation (M ¼ 13.71, SD ¼ 5.02) as other undergraduate samples (see Andrews & Meyer, 2003). The internal consistency also was adequate (Kuder–Richardson–20 ¼ .77) in Experiment 1.
Affect. The Positive and Negative Affective Schedule (PANAS; Watson, Clark, & Tellegen, 1988) has 20 items (e.g., excited, scared) rated on a 5-point Likert- type scale, from 1 (very slightly or not at all) to 5 (extremely), that assesses positive and negative affect in the present moment. The Positive and Negative sub- scales have each demonstrated adequate internal con- sistency in a previous sample (Watson et al., 1988). The Positive and Negative subscales have positively correlated with mood inventories assessing positive or negative affect, respectively, which supported the con- vergent validity of the PANAS (Watson et al., 1988). The internal consistency for Positive Affect (a ¼ .86; M ¼ 28.87, SD¼ 8.41) and Negative Affect (a ¼ .85; M ¼ 19.62, SD¼ 7.83) were adequate in Experiment 1.
Puzzle task. Participants were given 3 min to fill in a word search puzzle used in TMT research to serve as a distractor-delay (Greenberg, Pyszczynski, Solomon, Simon, & Breus, 1994).
Mediating measure: DTA. To assess participants’ DTA, I used a word-fragment completion task (Greenberg et al., 1994). The task consists of 25 words, each missing two letters. Six of these words can be completed as either death-related (i.e., buried, dead, grave, killed, skull, coffin) or neutral words (e.g., skill). The word-fragment completion task is scored by tallying up the number of words that participants complete as death oriented (i.e., 0–6). Putting aside the complications in assuming nonconscious death activations even exist, the various convergent findings in multiple languages support the construct validity of the word-completion DTA task (Hayes et al., 2010). DTA had an overall mean of 1.66 (SD ¼ 1.09) in Experiment 1 with similar means for the MS experi- mental group (M ¼ 1.65, SD ¼ 1.05) and the DP com- parison group (M ¼ 1.67, SD ¼ 1.12).
Deception. To reintroduce the remaining criterion and moderating variables related to self-forgiveness, the research assistant stated, “Whoops. I messed up the packets for today. These were supposed to be with
the first study about the event you recalled. Please fill in these as well.”
Criterion measures. The following two measures made up the criterion latent-variable “self-forgiveness” and were given in counterbalanced orders.
Self-forgiveness. The State Self-Forgiveness Scale (SSFS; Wohl, DeShea, & Wahkinney, 2008) consists of two subscales—Self-Forgiving Feelings and Actions (SFFA) and Self-Forgiving Beliefs (SFB)—that have both demonstrated desirable levels of internal consist- ency reliabilities in a previous sample (Wohl et al., 2008). All 17 items are quantified with a 4-point Likert-type scale, that is, 1 (not at all) to 4 (com- pletely). The SSFS positively correlated to depression in a previous study, which supported its convergent validity (Wohl et al., 2008). The internal consistencies for SFFA (a ¼ .87; M ¼ 19.06, SD ¼ 5.75) and SFB (a ¼ .90; M ¼ 26.80, SD ¼ 6.18) were adequate in Experiment 1.
Guilt. The State Shame and Guilt Scale (SSGS; Marschall, Sanfter, & Tangney, 1994) consists of 15 items assessing guilt (e.g., I feel remorse, regret), shame (e.g., I feel worthless, powerless), and pride (e.g., I feel proud) quantified on a 5-point Likert-type scale, that is, 1 (not feeling this way at all) to 5 (feeling this way very strongly). Each subscale of the SSGS has demonstrated desirable levels of internal consistency reliabilities in a previous sample (Tangney & Dearing, 2002). The SSGS has good content validity because it is derived from a large body of theoretical and empir- ical literature on self-conscious emotions (Tangney & Dearing, 2002). McConnell et al. (2012) found that the Guilt subscale was positively correlated to concili- atory behavior and negatively correlated to self-for- giveness, which supported its convergent validity. I excluded administration of the pride subscale in all experiments. In Experiment 1, I reversed scored the Guilt scale (GuiltR) to conceptually pair it with state self-forgiveness. The internal consistency for GuiltR
was adequate (a ¼ .79; M ¼ 11.68, SD ¼ 4.33) in Experiment 1.
Additional moderating measure: Effort. Although effort was conceptually a moderator in Experiment 1, I chose to administer it last because of its relevance to the process of accomplishing self-forgiveness (McConnell, 2015). In other words, I gave the meas- ure last to retrospectively assess the extent that
352 J. M. MCCONNELL
participants felt they had put forth effort toward self- forgiveness. I used two items modeled after Fisher and Exline (2006; i.e., How much effort did you take to reduce your negative feelings? How much time did you take to reduce your negative feelings?). The Effort Scale (ES) is quantified on an 11-point Likert-type scale (0 [no effort] to 10 [great effort]), and the two items are averaged together to quantify participants’ effort. The ES has demonstrated adequate intercorrela- tions and internal consistency in a previous sample (Fisher & Exline, 2006). The ES has positively corre- lated to repentance, humbling change, and self-for- giveness, which supported its convergent validity (Exline et al., 2011; Fisher & Exline, 2006). The internal consistency was adequate (a ¼ .75; M ¼ 10.43, SD ¼ 5.09) in Experiment 1.
Demographic questionnaire. I used a demographic questionnaire asking about participants’ age, gender, ethnicity, religious affiliation, marital status, level of education, offense date, and repeat offending.
Results and discussion
Missing data Excluding cases with missing data is problematic because it reduces sample size and introduces biases by modifying the sample means, standard errors, and standard deviations (Rubin, 1987). There were 51 (.001%) missing datum points within the SSFS, CDSII, SSGS, M-C SDS, PANAS, and NPI–40 with no more than two missing datum points per item. Therefore, I concluded that multiple imputation would introduce minimal statistical biases. I used AMOS 16.0 (Arbuckle, 2007) to conduct multiple imputation with 10 iterations on each separate subscale to generate val- ues for the missing data. I then aggregated the 10 iter- ations of generated values into one data set for analyses.
Responsibility Minimizing responsibility (e.g., denying, externalizing) can masquerade as self-forgiveness on the guilt and self-forgiveness measures I used in Experiment 1. To make sure minimizing was a minimal extraneous fac- tor, I classified the participants’ levels of taking responsibility with a two-step cluster analysis, one- way multivariate analysis of variance (MANOVA), and distribution outliers with SPSS 23.0. Taking responsibility was operationally defined as having higher degrees of personal control and lower degrees of external control. Due to unsatisfactory internal
consistency, I dropped Causality (a ¼ .58) and Stability (a ¼ .35) from analyses. Using External Control and Personal Control, two-step cluster ana- lysis elicited three clusters with good cluster quality as determined by cohesion and separation (see Table 3). The lowest responsibility cluster (External: M ¼ 15.54, SD ¼ 5.97; Personal: M ¼ 12.70, SD ¼ 4.01) had 54 members. The moderate responsibility cluster (External: M ¼ 14.19, SD ¼ 3.75; Personal: M ¼ 21.80, SD ¼ 2.53) had 86 members. The highest responsibility cluster (External: M ¼ 5.79, SD ¼ 2.44; Personal: M ¼ 24.87, SD ¼ 2.45) had 85 members. A one-way MANOVA Wilks’s Lambda indicated that clusters had different levels of external and personal con- trol (gp
2 ¼ .61). On external control and personal control, there
were four (1.78%) and 10 (4.44%) participant outliers beyond 2 SD, respectively. Although there were 54 (24.00%) people included in the lowest responsibility cluster, only three of these (5.56%) were considered 2 SD outliers on both external and personal control. Moreover, responsibility was very minimally related to self-forgiveness (SFFA, External: r ¼ .02; SFB, External: r ¼ �.08; SFFA, Personal: r ¼ �.03; SFB, Personal: r ¼ .07) and GuiltR (External: r ¼ �.03; Personal: r ¼ .02). I concluded that retention of outliers would not bias the results any more than their exclusion.
Intercorrelation matrices I created intercorrelation matrices with SPSS 23.0 for the entire sample (see Table 4), the experimental group (see Table 5), and the comparison group (see Table 6).
Main analyses I used Mplus 6.11 and SPSS 23.0 to test the medi- ation, moderated-mediation, and interactional-moder- ated-mediation hypotheses.
Mediation hypothesis. To test the mediation hypoth- esis of threat condition predicting the latent variable “self-forgiveness” (i.e., SFFA, SFB, and GuiltR) through DTA (see Figure 2), I used Mplus 6.11 to perform a bootstrapped path analysis. The
Table 3. Clusters for entire sample in Experiment 1. n M SD Range Skewness Kurtosis
Low responsibility cluster: External 54 15.54 5.97 3/27 �0.10 �0.59 Low responsibility cluster: Personal 54 12.70 4.01 3/18 �0.86 0.07 Mod responsibility cluster: External 86 14.19 3.75 8/23 0.35 �0.72 Mod responsibility cluster: Personal 86 21.80 2.53 17/27 0.52 �0.50 High responsibility cluster: External 85 5.79 2.44 3/11 0.61 �0.51 High responsibility cluster: Personal 85 24.87 2.45 17/27 �1.16 0.68 Note. n ¼ 225.
BASIC AND APPLIED SOCIAL PSYCHOLOGY 353
bootstrapped mediation model with 5,000 bootstrap draws elicited adequate fit, v2(4) ¼ 2.80, v2/df ¼ .70; comparative fit index =1.00, Tucker–Lewis index =1.01, root mean square error of approximation ¼ .00; Akaike information criterion =4601.48, Bayesian infor- mation criterion =4649.30, but had relations of min- imal effect sizes (see Table 7): (a) Threat condition very minimally impacted DTA in the unexpected dir- ection (b ¼ �.02, b ¼ �.01, R2 = .00), (b) threat condi- tion minimally increased self-forgiveness (b ¼ 1.66, b ¼ .15, R2 = .02), and (c) DTA very minimally impacted self-forgiveness in the unexpected direction (b ¼ �.04, b ¼ �.01, R2 = .00). The M-C SDS (gp
2 ¼ � .00) and PANAS–Positive and Negative (gp
2 ¼ � .00) had near-zero differences between groups, which indicated neither social desirability nor affect impacted the results of the mediation model.
Overall, there was a minimal direct effect of MS on self-forgiveness, but mediation through DTA was near-zero and in the opposite of the hypothesized dir- ection. Mediation was not supported.
Moderated-mediation hypothesis. To test the moder- ated-mediation hypothesis (see Figure 3), I used Mplus 6.11 to perform a bootstrapped path analysis. The bootstrapped mediation model with 5,000 boot- strap draws elicited marginal fit, v2(20) ¼ 43.22, v2/ df ¼ 2.16; comparative fit index ¼ .98, Tucker–Lewis index ¼ .97, root mean square error of approx- imation ¼ .07; Akaike information criterion =14335.49, Bayesian information criterion =14622.44, and had relations of minimal effect sizes. Again, threat condi- tion very minimally impacted DTA in the unexpected direction (b ¼ �.02, b ¼ �.01, R2 = .00) and threat condition minimally increased self-forgiveness when the four moderators were considered in the model (b ¼ 1.54, b ¼ .15, R2 = .02). Individual interactions on DTA predicting self-forgiveness were small or very minimal (see Table 7): POS (b ¼ .23, b ¼ .04, R2 = .00), CBS (b ¼ �.05, b ¼ �.01, R2 = .00), PF-Victim (b ¼ .88, b ¼ .17, R2 = .03), and ES (b ¼ .10, b ¼ .02, R2 = .00). Post hoc moderation analyses conducted with Hayes’s (2013) SPSS 23.0 PROCESS Macro also revealed interactions that yielded approximately zero or minimal changes to DTA predicting SFFA, SFB, or GuiltR individually. In cases of minimal moderation, the direct effect of DTA on self-forgiveness remained near-zero and/or in the unexpected direction. There were no more than 3.89 raw data point changes on SFFA, SFB, or GuiltR across all values of DTA and individual moderators (see Table 7). Although
interactions were small or very minimal, they are triv- ial given the range of scores on SFFA, SFB, and GuiltR (see Table 2).
Overall, the direct effect of DTA on self-forgive- ness was minimal and in the opposite of the hypothesized direction, and individual variables played little role in moderating this unexpected dir- ect effect. The results of the moderated-mediation model further supported that DTA did not meditate the minimal direct effect of MS on self-forgiveness despite inclusion of four variables known to usually effect the self-forgiveness process (see McConnell, 2015).
Interactional-moderated-mediation hypothesis. To evidence interactional-moderated-mediation, the experi- mental group must have more DTA and self-forgiveness than the comparison group. A one-way analysis of vari- ance (ANOVA) with threat condition predicting DTA indicated there was a near-zero difference between MS (M= 1.65, SD ¼ 1.05) and DP (M = 1.67, SD ¼ 1.12; gp
2 ¼ � .00). A one-way MANOVA with threat condi- tion predicting a linear combination of SFFA, SFB, and GuiltR indicated there was a minimal increase on self- forgiveness (gp
2 ¼ .02). Post hoc discriminant function analysis revealed one function (k ¼ .023) that did not effectively discriminate between groups (56.40% cor- rectly classified, 53.30% cross-validated). Post hoc one- way ANOVAs with threat condition predicting each dependent variable separately again revealed minimal increases on SFFA (gp
2 ¼ .02), SFB (gp2 ¼ .01), and GuiltR (gp
2 ¼ .01). These small effects represented no more than 1.65 raw data point mean differences between MS (see Table 8; SFFA: M =19.89, SD ¼ 5.30; SFB: M =27.42, SD ¼ 5.64; GuiltR: M =12.17, SD ¼ 4.33) and DP (SFFA: M = 18.24, SD ¼ 6.08; SFB: M =26.19, SD ¼ 6.63; GuiltR: M =11.19, SD ¼ 4.30) on self-forgive- ness measures. Because there was a near-zero difference between experimental groups for DTA, I did not test separate interactional-moderated-mediation models for each experimental condition. Of note, there were small or very minimal intercorrelations among MS, DTA, and self-forgiveness within and between experimental condi- tions (see Tables 4, 5, and 6).
Overall, there was a near-zero difference between threat conditions on DTA and MS minimally increased self-forgiveness. Near-zero effects for DTA are unsup- portive of interactional-moderated-mediation. These results replicate the unsupportive findings of the medi- ation and the moderated-mediation models.
354 J. M. MCCONNELL
Ta b le
4. St an d ar d iz ed
Pe ar so n an d Sp ea rm
an ’s Rh
o in te rc or re la ti on
s in
Ex p er im en t 1 fo r b ot h ex p er im en ta l co n d it io n s co m b in ed .
(1 )
(2 )
(3 )
(4 )
(5 )
(6 )
(7 )
(8 )
(9 )
(1 0)
(1 1)
(1 2)
(1 3)
(1 4)
(1 5)
(1 6)
(1 7)
(1 8)
(1 9)
(2 0)
(2 1)
(2 2)
(2 3)
(2 4)
C on
d it io n p
(1 )
1 D TA
(2 )
�. 01
1 SF FA
(3 )
.1 5
.0 0
1 SF B
(4 )
.0 8
�. 04
.7 4
1 G ui lt R
(5 )
.1 3
�. 07
.5 9
.4 6
1 PO
S (6 )
�. 03
�. 02
�. 31
�. 30
�. 38
1 C BS
(7 )
.0 0
.0 4
�. 30
�. 22
�. 33
.1 3
1 PF -V ic ti m
(8 )
.1 0
.1 4
.0 0
.0 5
.0 4
�. 36
.2 9
1 ES
(9 )
.1 3
.0 6
�. 16
�. 15
�. 31
.1 3
.3 2
�. 02
1 M -C
SD S
(1 0)
.0 5
�. 14
.2 4
.1 8
.1 6
�. 09
.0 4
.1 0
.0 5
1 PA
N A S- Po si ti ve
(1 1)
.0 1
�. 03
.0 9
.0 6
�. 06
�. 10
�. 02
.0 5
.2 5
.2 4
1 PA
N A S- N eg at iv e
(1 2)
.0 4
�. 05
�. 26
�. 26
�. 27
.2 2
.0 1
�. 22
.1 8
�. 13
.2 6
1 C D SI I: C au sa lit y
(1 3)
.0 8
.0 6
�. 10
�. 18
.0 2
�. 01
.0 4
.0 2
�. 07
�. 14
�. 07
.0 6
1 C D SI I: Ex te rn al
(1 4)
.0 1
�. 07
.0 2
�. 08
�. 03
�. 02
�. 09
�. 15
�. 01
.0 1
�. 00
.0 5
�. 28
1 C D SI I: Pe rs on
al (1 5)
�. 06
.0 9
�. 03
.0 7
.0 2
�. 12
.1 1
.2 3
�. 01
�. 05
�. 01
�. 10
.2 4
�. 44
1 C D SI I: St ab ili ty
(1 6)
.0 3
�. 08
�. 05
�. 02
�. 12
.1 0
�. 01
�. 21
.1 1
�. 08
.0 4
.2 0
�. 10
�. 01
�. 07
1 N PI -4 0
(1 7)
�. 16
�. 06
�. 01
.0 7
.0 2
�. 01
�. 03
�. 06
.1 4
.0 1
.1 8
.0 2
�. 10
�. 05
.0 5
�. 01
1 BF I- 44 : O p en n es s
(1 8)
�. 02
�. 01
.0 5
.0 4
.0 7
�. 04
.0 9
�. 02
.0 4
.0 7
.1 1
�. 03
.2 1
�. 10
.1 2
�. 04
.1 6
1 BF I- 44 : C on
sc ie n ti ou
sn es s
(1 9)
.0 0
�. 03
.0 6
.1 1
.0 3
�. 01
.1 3
.1 1
.0 9
.3 7
.2 1
�. 13
�. 12
�. 01
.0 6
.0 8
.1 0
�. 04
1 BF I- 44 : Ex tr av er si on
(2 0)
.0 2
�. 03
�. 04
.0 8
�. 12
.0 7
.1 5
�. 04
.1 5
.1 3
.3 0
.0 1
�. 07
�. 15
.1 1
.0 5
.4 8
.1 2
.1 6
1 BF I- 44 : A g re ea b le n es s
(2 1)
�. 02
�. 07
.1 9
.2 1
�. 01
�. 07
.2 2
.2 0
.1 5
.5 3
.2 4
�. 17
�. 13
�. 09
.0 7
�. 03
�. 10
.1 7
.2 3
.2 7
1 BF I- 44 : N eu ro ti ci sm
(2 2)
.0 5
.0 3
�. 31
�. 22
�. 19
.2 5
.0 2
�. 10
�. 04
�. 48
�. 24
.3 5
.1 6
.0 8
�. 07
.0 0
�. 10
�. 10
�. 29
�. 19
�. 40
1 D ay s Si n ce
O ff en se
(2 3)
.1 2
.1 2
.0 0
�. 01
.0 0
�. 07
.0 3
.1 4
.0 1
�. 03
.0 5
.0 0
.0 9
�. 01
�. 06
.0 4
�. 06
�. 06
�. 08
.0 4
.0 9
.0 3
1 Ti m es
C om
m it te d
(2 4)
�. 03
�. 02
.0 1
.0 3
.0 6
�. 05
.0 3
.0 5
.0 6
�. 11
.0 6
�. 08
.1 0
.0 4
.0 4
�. 03
.0 4
.0 8
.0 9
.0 4
.0 0
�. 08
�. 01
1
N ot e. n ¼ 22 5.
p : Sp ea rm
an ’s Rh
o C or re la ti on
; D TA
: D ea th
Th ou
g h t A cc es si b ili ty ; SF FA
: St at e Se lf- Fo rg iv in g Fe el in g s an d A ct io n s; SF B:
Se lf- Fo rg iv in g Be lie fs ; G ui lt R : SS G S: G ui lt it em
s re ve rs ed ; PO
S: Pe rc ei ve d O ff en se
Se ve ri ty ; C BS : C on
ci lia to ry
Be h av io rs
Sc al e;
PF -V ic ti m : Pe rc ei ve d
Fo rg iv en es s fr om
V ic ti m ; ES : Ef fo rt
Sc al e;
M -C
SD S:
M ar lo w -C ro w n e So ci al
D es ir ab ili ty
Sc al e;
PA N A S:
Po si ti ve
an d
N eg at iv e A ff ec t Sc h ed ul e;
C D SI I: Re vi se d C au sa l D im en si on
Sc al e; N PI -4 0:
N ar ci ss is ti c Pe rs on
al it y In ve n to ry ; BF I- 44 : Bi g Fi ve
In ve n to ry .
BASIC AND APPLIED SOCIAL PSYCHOLOGY 355
Ta b le
5. St an d ar d iz ed
Pe ar so n in te rc or re la ti on
s in
Ex p er im en t 1 fo r m or ta lit y sa lie n ce
co n d it io n .
(1 )
(2 )
(3 )
(4 )
(5 )
(6 )
(7 )
(8 )
(9 )
(1 0)
(1 1)
(1 2)
(1 3)
(1 4)
(1 5)
(1 6)
(1 7)
(1 8)
(1 9)
(2 0)
(2 1)
(2 2)
(2 3)
D TA
(1 )
1 SF FA
(2 )
.0 1
1 SF B
(3 )
�. 03
.7 1
1 G ui lt R
(4 )
�. 05
.4 7
.3 8
1 PO
S (5 )
�. 02
�. 39
�. 28
�. 40
1 C BS
(6 )
.1 0
�. 18
�. 18
�. 32
.0 6
1 PF -V ic ti m
(7 )
.2 2
.1 3
.0 5
.1 6
�. 45
.2 6
1 ES
(8 )
�. 03
.0 8
.0 9
�. 24
.0 7
.3 6
�. 04
1 M -C
SD S
(9 )
�. 12
.2 9
.1 9
.1 0
�. 22
.0 9
.1 8
.2 8
1 PA
N A S- Po si ti ve
(1 0)
�. 02
.2 7
.1 7
.0 5
�. 22
�. 01
.0 5
.2 7
.2 4
1 PA
N A S- N eg at iv e
(1 1)
�. 13
�. 26
�. 27
�. 25
.1 2
.0 3
�. 19
.1 2
�. 13
.3 3
1 C D SI I: C au sa lit y
(1 2)
.0 9
�. 19
�. 19
�. 10
.0 3
.1 2
.0 0
�. 17
�. 19
.0 2
.1 0
1 C D SI I: Ex te rn al
(1 3)
.0 0
.0 7
�. 12
�. 07
�. 03
�. 06
�. 05
.0 9
.0 3
.0 5
.0 6
�. 34
1 C D SI I: Pe rs on
al (1 4)
.0 7
�. 06
.0 6
.0 8
�. 12
.0 9
.1 9
�. 18
�. 10
.0 3
�. 16
.2 7
�. 49
1 C D SI I: St ab ili ty
(1 5)
�. 07
�. 11
�. 02
�. 09
.1 9
.0 9
�. 20
.1 7
�. 06
�. 02
.0 7
�. 13
�. 05
�. 19
1 N PI -4 0
(1 6)
.0 4
.1 4
.2 3
.0 8
.0 2
�. 14
�. 10
.1 5
.0 4
.2 4
.0 1
�. 10
�. 02
.0 0
�. 01
1 BF I- 44 : O p en n es s
(1 7)
.0 4
.1 7
.0 8
.1 5
�. 14
.0 2
.1 5
�. 09
.1 5
.1 6
�. 09
.2 9
�. 23
.2 7
�. 07
.1 1
1 BF I- 44 : C on
sc ie n ti ou
sn es s
(1 8)
�. 05
.1 2
.1 4
�. 01
.0 2
.1 4
�. 02
.1 9
.3 7
.2 1
�. 18
�. 15
.0 9
�. 09
.0 6
.1 9
�. 01
1 BF I- 44 : Ex tr av er si on
(1 9)
.0 7
.0 4
.1 9
�. 08
.0 4
.0 7
�. 12
.1 5
.1 1
.2 6
.0 2
�. 01
�. 11
.1 1
.0 8
.4 8
.0 5
.1 2
1 BF I- 44 : A g re ea b le n es s
(2 0)
.0 0
.2 6
.2 4
�. 05
�. 14
.2 5
.2 5
.3 3
.6 1
.2 6
�. 19
�. 12
�. 09
.0 6
.0 8
�. 06
.2 2
.1 9
.2 8
1 BF I- 44 : N eu ro ti ci sm
(2 1)
�. 06
�. 45
�. 30
�. 24
.4 1
.0 2
�. 13
�. 13
�. 46
�. 21
.3 5
.2 5
.0 2
�. 01
�. 05
�. 04
�. 23
�. 32
�. 11
�. 38
1 D ay s Si n ce
O ff en se
(2 2)
.0 9
�. 04
�. 03
�. 11
.0 4
.0 0
.0 1
.0 4
.0 1
.0 4
.1 0
�. 01
.0 4
�. 16
.0 6
�. 13
�. 12
�. 10
.0 0
.1 2
�. 03
1 Ti m es
C om
m it te d
(2 3)
.2 1
.0 2
�. 06
�. 02
�. 03
�� .1 6
.0 9
�. 08
.0 0
.1 7
.0 4
.1 3
.1 4
.0 5
�. 22
.2 2
.1 8
.0 3
�. 02
�� .1 1
.0 7
.0 3
1
N ot e.
n ¼ 11 2.
D TA
: D ea th
Th ou
g h t A cc es si b ili ty ; SF FA
: St at e Se lf- Fo rg iv in g Fe el in g s an d A ct io n s;
SF B:
Se lf- Fo rg iv in g Be lie fs ; G ui lt R : SS G S: G ui lt it em
s re ve rs ed ; PO
S: Pe rc ei ve d O ff en se
Se ve ri ty ; C BS : C on
ci lia to ry
Be h av io rs
Sc al e;
PF -V ic ti m : Pe rc ei ve d Fo rg iv en es s fr om
V ic ti m ; ES : Ef fo rt Sc al e;
M -C
SD S:
M ar lo w -C ro w n e So ci al
D es ir ab ili ty
Sc al e;
PA N A S:
Po si ti ve
an d N eg at iv e A ff ec t Sc h ed ul e;
C D SI I: Re vi se d C au sa l D im en si on
Sc al e; N PI -4 0:
N ar ci ss is ti c Pe rs on
al it y In ve n to ry ; BF I- 44 : Bi g Fi ve
In ve n to ry .
356 J. M. MCCONNELL
Ta b le
6. St an d ar d iz ed
Pe ar so n in te rc or re la ti on
s in
Ex p er im en t 1 fo r d en ta l p ai n co n d it io n .
(1 )
(2 )
(3 )
(4 )
(5 )
(6 )
(7 )
(8 )
(9 )
(1 0)
(1 1)
(1 2)
(1 3)
(1 4)
(1 5)
(1 6)
(1 7)
(1 8)
(1 9)
(2 0)
(2 1)
(2 2)
(2 3)
D TA
(1 )
1 SF FA
(2 )
�. 01
1 SF B
(3 )
�. 05
.7 5
1 G ui lt R
(4 )
�. 10
.6 9
.5 1
1 PO
S (5 )
�. 03
�. 24
�. 31
�. 35
1 C BS
(6 )
�. 01
�. 40
�. 27
�. 35
.1 9
1 PF -V ic ti m
(7 )
.0 7
�. 13
.0 4
�. 11
�. 27
.3 1
1 ES
(8 )
.1 4
�. 32
�. 32
�. 37
.1 8
.3 1
.0 2
1 M -C
SD S
(9 )
�. 17
.1 9
.1 6
.2 0
.0 3
.0 0
.0 1
�. 15
1 PA
N A S- Po si ti ve
(1 0)
�. 03
�. 08
�. 04
�. 18
.0 1
�. 04
.0 4
.2 3
.2 4
1 PA
N A S- N eg at iv e
(1 1)
.0 3
�. 28
�. 26
�. 30
.3 2
�. 02
�. 25
.2 7
�. 14
.1 9
1 C D SI I: C au sa lit y
(1 2)
.0 3
�. 04
�. 19
.1 2
�. 04
�. 03
.0 4
.0 2
�. 11
�. 16
.0 2
1 C D SI I: Ex te rn al
(1 3)
�. 13
�. 03
�. 05
.0 0
�. 01
�. 12
�. 27
�. 10
�. 02
�. 06
.0 4
�. 22
1 C D SI I: Pe rs on
al (1 4)
.1 0
�. 01
.0 8
�. 02
�. 12
.1 3
.2 6
.1 3
�. 01
�. 04
�. 04
.2 2
�. 39
1 C D SI I: St ab ili ty
(1 5)
�. 09
.0 0
�. 02
�. 14
.0 2
�. 10
�. 22
.0 7
�. 11
.1 0
.2 7
�. 07
.0 4
.0 3
1 N PI -4 0
(1 6)
�. 15
�. 08
�. 03
.0 0
�. 04
.0 7
.0 0
.0 9
�. 01
.1 3
.0 5
�. 08
�. 08
.0 9
�. 01
1 BF I- 44 : O p en n es s
(1 7)
�. 06
�. 06
.0 1
.0 1
.0 5
.1 5
�. 18
.1 6
.0 0
.0 5
.0 4
.1 4
.0 4
�. 02
�. 02
.2 1
1 BF I- 44 : C on
sc ie n ti ou
sn es s
(1 8)
�. 01
.0 0
.0 9
.0 7
�. 03
.1 3
.2 3
.0 1
.3 7
.2 2
�. 08
�. 10
�. 10
.1 9
.1 0
.0 3
�. 07
1 BF I- 44 : Ex tr av er si on
(1 9)
�. 12
�. 11
�. 01
�. 17
.0 9
.2 1
.0 3
.1 7
.1 5
.3 4
.0 0
�. 12
�. 20
.1 1
.0 2
.5 0
.1 9
.1 8
1 BF I- 44 : A g re ea b le n es s
(2 0)
�. 13
.1 4
.1 9
.0 3
�. 02
.1 9
.1 5
.0 1
.4 6
.1 8
�. 15
�. 14
�. 08
.0 7
�. 13
�. 14
.1 2
.2 6
.2 6
1 BF I- 44 : N eu ro ti ci sm
(2 1)
.1 1
�. 21
�. 17
�. 17
.1 3
.0 1
�. 09
.0 4
�. 50
�. 26
.3 5
.0 8
.1 3
�. 11
.0 4
�. 14
.0 1
�. 27
�. 26
�. 41
1 D ay s Si n ce
O ff en se
(2 2)
.1 5
.0 2
�. 01
.0 8
�. 17
.0 6
.2 6
.0 1
�. 07
.0 6
�. 12
.0 5
�. 06
.0 4
.0 3
.0 3
.0 2
�. 06
.0 6
.0 7
.0 7
1 Ti m es
C om
m it te d
(2 3)
�. 07
.0 2
.0 6
.1 0
�. 07
.0 7
.0 6
.0 9
�. 15
.0 5
�. 12
.1 2
.0 3
.0 4
.0 0
.0 0
.0 8
.1 2
.0 5
.0 2
�. 11
�. 01
1
N ot e.
n ¼ 11 3.
D TA
: D ea th
Th ou
g h t A cc es si b ili ty ; SF FA
: St at e Se lf- Fo rg iv in g Fe el in g s an d A ct io n s;
SF B:
Se lf- Fo rg iv in g Be lie fs ; G ui lt R : SS G S: G ui lt it em
s re ve rs ed ; PO
S: Pe rc ei ve d O ff en se
Se ve ri ty ; C BS : C on
ci lia to ry
Be h av io rs
Sc al e;
PF -V ic ti m : Pe rc ei ve d Fo rg iv en es s fr om
V ic ti m ; ES : Ef fo rt Sc al e;
M -C
SD S:
M ar lo w -C ro w n e So ci al
D es ir ab ili ty
Sc al e;
PA N A S:
Po si ti ve
an d N eg at iv e A ff ec t Sc h ed ul e;
C D SI I: Re vi se d C au sa l D im en si on
Sc al e; N PI -4 0:
N ar ci ss is ti c Pe rs on
al it y In ve n to ry ; BF I- 44 : Bi g Fi ve
In ve n to ry .
BASIC AND APPLIED SOCIAL PSYCHOLOGY 357
Summary The results of Experiment 1 were unsupportive of the mediation, moderated-mediation, and interactional- moderated-mediation hypotheses. First, MS did not lead to greater DTA. In fact, DTA was on average slightly larger in the DP comparison group. Second, more DTA negligibly led to less self-forgiveness, and moderations by perceived offense severity, conciliatory behaviors, perceived forgiveness, and effort were near- zero or inconsequently small. Third, although there was a minimal direct effect of MS on self-forgiveness, the effect accounted for only 2% of the variance in self-forgiveness and approximately 1 raw data point increases. This small effect is negligibly supportive of existential motivations for self-forgiveness in context of TMT’s all-encompassing claims.
The unsupportive and negligible findings of Experiment 1 could be explained by experimental or measurement artifacts. First, because the offense recall threat did not experimentally manipulate perceived offense severity and guilt intensity or measure effect on
self-esteem, participants’ anxiety-buffers may have still remained strong prior to the MS manipulation. Second, several measurement artifacts of the experimental sequence may have impacted the results. The use of conciliatory behaviors, perceived forgiveness, and nar- cissism measures may have unintentionally strength- ened the participants’ anxiety buffer. Also, the measurement-of-mediation design may have inadvert- ently exposed the comparison group to death-related stimuli while filling in the DTA measure. Third, using a convenience sample of young adults who self-selected to speak about wrongdoings may have biased the results. Finally, because I did not use a model-building approach to investigate an experimental-causal-chain, the findings of Experiment 1 cannot directly evaluate mediation due to its measurement-of-mediation design.
For these reasons, I dismantled the experimental sequence into three follow-up experimental-causal- chain studies that addressed the aforementioned con- founding factors and more fully isolated TMT Hy1, TMT Hy2, and TMT Hy3 where feasible. When
Figure 2. Mediation path analysis model. Note. Standardized beta weights presented. SFFA: Self-Forgiving Feelings and Actions; SFB: Self-Forgiving Beliefs; GuiltR: State Shame and Guilt Scale: Guilt (reversed).
Table 7. Summary table of mediation and moderated-mediation model results for Experiment 1. Mediation Moderated-Mediation
Predictor Moderator Criterion b b R2 b b R2 DR2 Draw
MS j DP on DTA �.02 �.01 .00 �.02 �.01 .00 – – MS j DP on SF 1.66 .15 .02 1.54 .15 .02 – – DTA on SF �.04 �.01 .00 �.04 �.01 .00 – – DTA with POS on SF – – – .23 .04 – .00 –
CBS on SF – – – �.05 �.01 – .00 – PF-Victim on SF – – – .88 .17 – .03 – ES on SF – – – .10 .02 – .00 –
DTA with POS on SFFA – – – – – – .00 – SFB – – – – – – .00 – GuiltR – – – – – – .00 –
CBS on SFFA – – – – – – .00 – SFB – – – – – – .01 3.89 GuiltR – – – – – – .00 –
PF-Victim on SFFA – – – – – – .02 1.96 SFB – – – – – – .01 2.30 GuiltR – – – – – – .00 –
ES on SFFA – – – – – – .00 – SFB – – – – – – .02 2.15 GuiltR – – – – – – .01 3.57
Note. n ¼ 225. b: unstandardized beta; b: standardized beta; Draw: max raw data change in interaction on criterion variable; MS: Mortality Salience; DP: Dental Pain Salience; SFFA: State Self-Forgiving Feelings and Actions; SFB: State Self-Forgiving Beliefs; GuiltR: SSGS:Guilt items reversed; SF: “Self-for- giveness” latent variable (SFFA, SFB, and GuiltR); POS: Perceived Offense Severity; CBS: Conciliatory Behaviors Scale; PF-Victim: Perceived Forgiveness from Victim; ES: Effort Scale.
358 J. M. MCCONNELL
relevant, subsequent studies included two levels of the anxiety-buffer threat, a self-esteem measure, and a neutral word-completion measure. In addition, I excluded the conciliatory behaviors, perceived forgive- ness, and narcissism measures and replaced the CDSII with a more parsimonious measure. Moreover, I con- ducted the follow-up studies via online modules to use a more generalizable population and remove any self-selection bias. Of note, participants were randomly assigned to one of three experiments without repetitive participation, and a postexperiment questionnaire affirmed they had never before participated in a self- forgiveness or death-related study. Finally, responsibil- ity, perceived offense severity, and guilt were concep- tualized both as part of the anxiety-buffer manipulation itself and as manipulation checks rather than moderating or criterion variables.
Experiment 2
To address alternative explanations of Experiment 1, I conducted a study designed to test TMT Hy1 while controlling for TMT Hy3. Specifically, I provided two levels of the anxiety-buffer threat condition (mild vs. bad) along with manipulation checks that were trailed by an MS manipulation. I utilized a mild threat com- parison group because a no-offense threat control group would be conceptually unrelated to state self- forgiveness. I used a randomized, posttest-only, dou- ble-blind four-group factorial design (mild MS vs. mild DP vs. bad MS vs. bad DP). I expected that MS would predict greater levels of self-forgiveness than
DP salience (TMT Hy1) and that this effect would occur with the bad offense recall to a larger degree than the mild offense recall (TMT Hy3).
Method
Participants A convenience sample of 932 U.S. citizens who were at least 18 years old participated online via Qualtrics in exchange for approximately $1.5 USD compensa- tion. See Table 1 for a summary of demographic characteristics.
Procedure Participants first completed an informed consent and then were randomly assigned to one of four condi- tions (mild MS vs. mild DP vs. bad MS vs. bad DP; n ¼ 233 each). Participants then recalled and briefly wrote about a mild or bad interpersonal wrongdoing committed by them within the past year. As part of the anxiety-buffer manipulation, participants filled in counterbalanced measures of responsibility, perceived offense severity, and guilt, and then finished with a self-esteem manipulation check. Participants then completed the death awareness packet with distractor- delays all identical to Experiment 1 but without the NPI-40 and the overlooked-measures deception. Last, participants filled in a state self-forgiveness scale and subsequently completed a demographic questionnaire. See Figure 4 for a visual depiction of the experimen- tal sequence.
Figure 3. Moderated-mediation path analysis model. Note. Standardized beta weights presented. SFFA: Self-Forgiving Feelings and Actions; SFB: Self-Forgiving Beliefs; GuiltR: State Shame and Guilt Scale: Guilt (reversed).
BASIC AND APPLIED SOCIAL PSYCHOLOGY 359
Materials I excluded the CDSII, CBS, PF-Victim, NPI-40, DTA, ES, and deception from Experiment 2 (cf. Figures 1 and 4) but retained the distractor-delay measures, SSFS, and demographic questionnaire. The following highlight other major modifications or additions. See Table 2 for Experiment 2 scale properties.
Offense packet. The offense packet was modified by including a two-level offense recall, a different meas- ure of responsibility along with the POS and SSGS in counterbalanced orders, and a self-esteem manipula- tion check.
Offense recall. Participants were instructed,
Please recall a recent event in the past year in which you did something (mildly wrong | wrong) to someone. Recall something you did that was (fairly minor | pretty bad). Only recall something that most people would believe is really [not that bad (e.g., borrowing a pen and not returning it) | a pretty bad thing to do (e.g., betraying someone, neglecting something very important, physical altercation) and NOT something that is [a pretty bad thing to do (e.g., betraying someone, neglecting something very important, physical altercation) | fairly minor (e.g., borrowing a pen and not returning it)]. This should be something that you take responsibility for doing. For a moment, visualize in your mind the events and the interactions you may have had in this situation. Try to visualize the person and recall what happened. Now please briefly describe the event in three to five sentences.
My research assistant and I reviewed offense recalls to ensure participants adequately responded to the prompt with a mild or bad interpersonal offense.
Responsibility. The CDSII was replaced with the more parsimonious Perceived Personal Responsibility Scale (PPRS; Fisher & Exline, 2006). The PPRS is a five-item measure on a 10-point Likert-type scale from 1 (com- pletely disagree) to 10 (completely agree) that assesses the extent to which people believe they are responsible for an offense (e.g., “This was clearly my fault”). The scale has demonstrated adequate levels of internal consistency reliability in a previous sample (Fisher & Exline, 2006).
Perceived offense severity. The offense packet retained the POS.
Guilt. The offense packet now included the SSGS Guilt and Shame subscales in Experiment 2. The Guilt sub- scale was used for analyses.
Self-esteem. The Rosenberg Self-Esteem Scale (RSES; Rosenberg, 1965) was added as a manipulation check at the end of the offense packet. The RSES is a 10-item measure on a 4-point Likert-type scale from 0 (strongly disagree) to 3 (strongly agree) that quantifies self-esteem (e.g., “I feel that I have a number of good qualities”). The RSES is a widely used measure of self-esteem, and
Table 8. Summary table of experimental results. Experiment 1 Experiment 2
Mild Bad
MS DP MS DP MS DP
M SD M SD g2p k % M SD M SD M SD M SD g 2 p
DTA 1.65 1.05 1.67 1.12 .00 – – SF – – – – .02 .02 56.40 SFFA 21.51 5.82 22.53 5.24 20.71 5.59 20.67 6.02 .02 SFFA 19.89 5.30 18.24 6.08 .02 – – Offense Recall .01 SFB 27.42 5.64 26.19 6.63 .01 – – Death Threat .00 GuiltR 12.17 4.33 11.19 4.30 .01 – – Interaction .00
SFB 28.09 6.46 29.14 5.58 27.46 6.57 26.42 7.26 .02 Offense Recall .02 Death Threat .00 Interaction .00
Experiment 3 (DTA conditions only) Experiment 4 (DTA conditions only)
Predictor Criterion Mild DTA Bad DTA Mild Neutral Bad Neutral
M SD M SD b b� R2 M SD M SD M SD M SD g2p RSES on DTA 19.92 6.53 1.88 .99 �.02 �.14 .02 DTA 1.63 .99 1.94 1.04 – – – – .02 DTA on HFS 1.88 .99 29.00 7.28 �1.20 �.16 .03 SFFA 23.09 5.18 20.75 5.98 – – – – .04
SFB 29.57 5.71 26.95 6.87 – – – – .04 Predictor Criterion
M SD M SD b b� R2 DTA on SFFA 1.79 1.02 21.92 5.71 �.70 �.13 .02 DTA on SFB 1.79 1.02 28.26 6.44 �.95 �.15 .02
Note. gp 2: partial eta-squared; k: discriminant function eigenvalue; %: percentage classified in post hoc discriminant analysis; b: unstandardized beta;
b: standardized beta; SFFA: State Self-Forgiving Feelings and Actions; SFB: State Self-Forgiving Beliefs; GuiltR: SSGS:Guilt items reversed; SF: all linear combination variables in multivariate analysis of variance (SFFA, SFB, and GuiltR); RSES: Rosenberg Self-Esteem Scale; DTA: Death Thought Accessibility.
360 J. M. MCCONNELL
the scale has demonstrated adequate levels of internal consistency reliability and construct validity in previous samples (Robins, Hendin, & Trzesniewski, 2001).
Death awareness packet The death awareness packet was identical to Experiment 1 except for the exclusion of the NPI-40.
Results and discussion
Manipulation check I used SPSS 23.0 for all analyses in Experiment 2. There were no missing data. A factorial ANOVA indi- cated the PPRS was different among mild MS (M ¼ 35.75, SD¼ 11.53), mild DP (M ¼ 35.73, SD¼ 10.92), bad MS (M ¼ 37.04, SD¼ 11.80), and bad DP (M ¼ 38.15, SD¼ 10.62; gp2 ¼ .01) with a small main effect for offense recall type (gp
2 ¼ .01) and a near-zero effect for death-threat condition (gp
2 ¼ � .00). A factorial ANOVA revealed that the POS was different among mild MS (M ¼ 14.58, SD¼ 2.92), mild DP (M ¼ 14.43, SD¼ 3.00), bad MS (M ¼ 15.93, SD¼ 3.61), and bad DP (M ¼ 16.09, SD¼ 3.54; gp2 ¼ .05) with a medium main effect for offense recall type (gp
2 ¼ .05) and a near-zero effect for death-threat condition (gp
2 ¼ � .00). A factorial ANOVA indicated that the SSGS Guilt subscale was different among mild MS (M ¼ 15.03, SD¼ 5.78), mild DP (M ¼ 14.52, SD¼ 5.76), bad MS (M ¼ 17.61, SD¼ 5.57), and bad DP (M ¼ 17.79, SD¼ 5.52; gp
2 ¼ .06) with a medium main effect for offense recall type (gp
2 ¼ .06) and a near-zero effect for death-threat condition (gp
2 ¼ � .00). A factorial ANOVA specified
the RSES was different among mild MS (M ¼ 18.73, SD¼ 6.56), mild DP (M ¼ 19.70, SD¼ 6.08), bad MS (M ¼ 18.12, SD¼ 6.25), and bad DP (M ¼ 17.40, SD¼ 6.01; gp2 ¼ .02) with a small main effect for offense recall type (gp
2 ¼ .01) and a near-zero effect for death-threat condition (gp
2 ¼ � .00). As expected, the offense recall manipulation increased levels of responsi- bility, perceived offense severity, and guilt, as well as decreased self-esteem for the bad offense recall group.
Main analyses A factorial ANOVA revealed the SFFA was different among mild MS (M ¼ 21.51, SD ¼ 5.82), mild DP (M ¼ 22.53, SD ¼ 5.24), bad MS (M ¼ 20.71, SD ¼ 5.59), and bad DP (M ¼ 20.67, SD ¼ 6.02; gp
2 ¼ .02); unexpectedly, there was a near-zero effect for death-threat condition (gp
2 ¼ � .00), but there was a small main effect for offense recall type (gp
2 ¼ .01). There was a near-zero interaction effect (gp
2 ¼ � .00). A factorial ANOVA found that the SFB was different among mild MS (M ¼ 28.09, SD ¼ 6.46), mild DP (M ¼ 29.14, SD ¼ 5.58), bad MS (M ¼ 27.46, SD ¼ 6.57), and bad DP (M ¼ 26.42, SD ¼ 7.26; gp2 ¼ .02) with the same unexpected find- ings for death-threat (gp
2 ¼ �.00) and offense recall type conditions (gp
2 ¼ .02). There was a near-zero interaction effect (gp
2 ¼ � .00). The M-C SDS had a near-zero difference among groups (gp
2 ¼ � .00). Although PANAS-Positive had a near-zero differ- ence among groups (gp
2 ¼ � .00), PANAS-Negative had a small difference among mild MS (M ¼ 18.60, SD ¼ 8.95), mild DP (M ¼ 18.36, SD ¼ 8.94), bad MS (M ¼ 19.61, SD ¼ 9.31), and bad DP (M ¼ 20.58,
Additional MaterialsOffense Packet Death Awareness
Packet Distractor-Delay
Measures Dependent Variable
B A D
M I L D
M I L D
B A D
M O R T A L I T Y
S A L I E N C E
Counterbalanced
Social Desirability (M-C SDS)
Mood (PANAS)
Puzzle Task
Self-Forgiveness (SSFS)
Demographic Questionnaire
D E N T A L
P A I N
Counterbalanced
Social Desirability (M-C SDS)
Mood (PANAS)
Puzzle Task
Cover Story
Personality (BFI-44)
Self-Forgiveness (SSFS)
Counterbalanced
Responsibility (PPRS)
Perceived Offense Severity (POS)
Guilt (SSGS)
Self-Esteem
(RSES)
Demographic Questionnaire
Threat Condition (MAPS)
Mortality Salience
OR
Dental Pain
Salience
Threat Condition
Mild Offense Recall
OR
Bad Offense Recall
Figure 4. Graphical representation of the experimental procedure for Experiment 2.
BASIC AND APPLIED SOCIAL PSYCHOLOGY 361
SD ¼ 10.03, gp2 ¼ .01). Understandably, the small main effect on PANAS-Negative resulted from offense recall type (gp
2 ¼ .01) and not from death threat (gp
2 ¼ � .00). Overall, threat condition had a near-zero effect on self-forgiveness regardless of offense recall conditions.
Exploratory analyses I simultaneously entered PPRS (b ¼ �.04, b ¼ �.08), POS (b ¼ �.08, b ¼ �.05), SSGS Guilt (b ¼ �.20, b ¼ �.20), and RSES (b ¼ .48, b ¼ .52) into a multiple linear regression model predicting SFFA (R2 = .43). I also simultaneously entered PPRS (b ¼ �.04, b ¼ �.07), POS (b ¼ �.08, b ¼ �.04), SSGS Guilt (b ¼ �.15, b ¼ �.14), and RSES (b ¼ .57, b ¼ .54) into a multiple linear regression model predicting SFB (R2 ¼ .39). Although PPRS and POS were minimally related, SSGS guilt was negatively associated with self- forgiveness to a small to moderate degree and RSES was positively related to a large degree.
Summary The results of Experiment 2 suggested that the min- imal direct effect observed in Experiment 1 was likely spurious. In Experiment 2, MS had a near-zero effect on self-forgiveness even in conditions with higher responsibility taking, more severe offenses, greater guilt, and lower self-esteem. These results did not sup- port TMT Hy1 when controlling for TMT Hy3. Although there were small to moderate differences between offense recall groups on manipulation checks, MS should have had an effect on self-forgiveness at least commensurate with self-esteem levels to support TMT Hy1. In addition, findings were actually the exact opposite of TMT Hy3 as self-forgiveness was higher in the low anxiety-buffer threat condition rather than the high anxiety-buffer threat condition. Exploratory analyses revealed that guilt was negatively associated with self-forgiveness and that self-esteem was positively related.
Experiment 3
The measurement-of-mediation design of Experiment 1 cannot infer causation. Experimental-causal-chains more rationally test the temporal sequence implied in mediation (Kline, 2015; Spencer et al., 2005; Tate, 2015; Trafimow, 2015). Because there is no known way to directly manipulate DTA without MS or anxiety-buffer threats, baseline DTA studies are a second-best alternative (Hayes et al., 2010). To more rationally investigate the mediational aspect of TMT
Hy2, I conducted a dispositional self-esteem, baseline DTA, and trait self-forgiveness study. To account for unintended death-related cognition, I also included a neutral word-fragment completion task condition. I used a randomized, double-blind comparison group design (DTA vs. neutral). I predicted that baseline DTA would positively relate to trait self-forgiveness (TMT Hy2) and that dispositional self-esteem would negatively relate to baseline DTA (TMT Hy3) for the participants in the DTA word-fragment comple- tion condition.
Method
Participants A convenience sample of 476 U.S. citizens who were at least 18 years old participated online via Qualtrics in exchange for approximately $1.5 USD remuner- ation. See Table 1 for a summary of demographic characteristics.
Procedure Participants first completed an informed consent and then were randomly assigned to one of two conditions (DTA vs. neutral; n ¼ 238 each). Participants then completed a self-esteem measure, followed by a DTA or neutral word-fragment completion task and then a trait self-forgiveness measure. Subsequently, partici- pants filled in social desirability and affect measures in counterbalanced orders. Finally, they completed a demographic questionnaire. See Figure 5 for a visual depiction of the experimental sequence.
Materials I used the RSES, DTA word-completion task, M-C SDS, PANAS, demographic questionnaire, and the fol- lowing measures in Experiment 3. See Table 2 for Experiment 3 scale properties.
Neutral word-fragment completion task. To account for unintended death-related cognition, I used a neu- tral word-fragment completion task that consisted of 25 words. Each word was missing two letters and could not be completed in death-related ways.
Trait Self-Forgiveness. The Heartland Forgiveness Scale (HFS; Thompson et al., 2005) is an 18-item measure on a 7-point Likert-type scale from 1 (almost always false of me) to 7 (almost always true of me) that assesses dispositional forgiveness of self, others, and situations. I used the six-item HFS self-forgive- ness subscale (e.g., “With time I am understanding of
362 J. M. MCCONNELL
myself for mistakes I’ve made”) to measure trait self- forgiveness. The HFS has demonstrated desirable lev- els of internal consistency reliability and convergent validity in a previous sample (Thompson et al., 2005).
Results and discussion
Main analyses I used SPSS 23.0 for all analyses in Experiment 3. There were no missing data. To explore the relations among dispositional self-esteem, baseline DTA, and trait self- forgiveness, I analyzed data only from the DTA word- fragment completion condition, as the neutral condition did not have a chance to fill in death-related words. There was a small negative relation with RSES predicting baseline DTA in a linear regression model (b¼ �.02, b ¼ �.14, R2 = .02). Unexpectedly, there was a small negative relation rather than a positive relation when baseline DTA predicted HFS in a linear regression model (b ¼ �1.20, b ¼ �.16, R2 = .03). A one-way ANOVA indicated there was a near-zero difference in HFS across word-fragment completion task conditions (gp
2 ¼ � .00), which indicated that filling in the DTA measure in and of itself minimally impacted trait self- forgiveness measurement. In addition, M-C SDS (gp
2 ¼ � .00), PANAS-Positive (gp2 ¼ � .00), and PANAS-Negative (gp
2 ¼ � .00) had near-zero differences across word-fragment completion task conditions. Overall, the results are mixed. The RSES had a small negative relation with baseline DTA, but unexpectedly
baseline DTA was negatively related to HFS to a small degree.
Exploratory analysis RSES positively related to HFS in linear regression model to a large degree (b ¼ .73, b ¼ .69, R2 ¼ .47).
Summary The results of Experiment 3 are consistent with the unsupportive findings of Experiment 1 and 2. Although Experiment 3 found a small negative rela- tion between dispositional self-esteem and baseline DTA, self-esteem only accounted for 2% of the vari- ance in DTA. This finding is negligibly supportive of TMT Hy3 in consideration that TMT theorists believe the anxiety buffer is the main regulator of DTA. Alternatively, dispositional self-esteem may not have been low enough to impact baseline DTA more. Yet the small relation between self-esteem and DTA did not correspond with the predicted relation between baseline DTA and trait self-forgiveness. Experiment 3 actually found the opposite of TMT Hy2. Higher levels of baseline DTA were associated with less trait self- forgiveness. Exploratory analyses revealed self-esteem was positively related with self-forgiveness.
Experiment 4
To experimentally test TMT Hy3 independent of TMT Hy1, I provided an identical anxiety-buffer threat as in
Additional MaterialsCriterion VariableBaseline Measure Additional MaterialsBaseline Measure
D T A
Counterbalanced
Social Desirability (M-C SDS)
Mood (PANAS)
Death-Thought Accessibility (DTA Word Completion
Task)
Demographic Questionnaire
N E U T R A L
Counterbalanced
Social Desirability (M-C SDS)
Mood (PANAS)
Neutral Word Completion Task
Self-Esteem
(RSES)
Demographic Questionnaire
Trait Self-Forgiveness
(HFS)
Trait Self-Forgiveness
(HFS)
Figure 5. Graphical representation of the experimental procedure for Experiment 3.
BASIC AND APPLIED SOCIAL PSYCHOLOGY 363
Experiment 2, excluded the MS manipulation, and measured DTA after a delay. Some TMT Hy1 research has supported that DTA effects deteriorate over time when threats are not explicitly death related (Steinman & Updegraff, 2015); however, few to no studies have systematically examined distractor-delay effects in the context of an anxiety-buffer threat (Webber, Zhang, Schimel, & Blatter, 2016). Therefore, I chose to keep a distractor-delay to keep methodo- logical consistency between Experiments 1 and 4, but I measured the time delay in seconds to account for its effect. To speculate about DTA mediation, I also included a self-forgiveness measure and neutral word- fragment completion task conditions to account for unintended death-related cognition. I used a random- ized, posttest-only, double-blind four-group factorial design (mild DTA vs. mild neutral vs. bad DTA vs. bad neutral) combined with a measurement-of-medi- ation procedure. I expected that the bad offense recall would predict greater DTA than the mild offense recall (TMT Hy3). I also predicted that DTA would be positively related to self-forgiveness in the DTA con- ditions (TMT Hy2) and self-forgiveness levels would be equivalent across the DTA and neutral conditions.
Method
Participants A convenience sample of 892 U.S. citizens who were at least 18 years old participated online via Qualtrics in exchange for approximately $1.5 USD compensa- tion. See Table 1 for a summary of demographic characteristics.
Procedure and materials Participants first completed an informed consent and then were randomly assigned to one of four condi- tions (mild DTA vs. mild neutral vs. bad DTA vs. bad neutral; n ¼ 223 each). They then completed the exact same offense packet as Experiment 2 followed by the same distracter-delay measures in counterbalanced orders. Next, they filled in either the DTA or the neu- tral word-fragment completion task, trailed by the SSFS, and finally a demographic questionnaire. See Figure 6 for a visual depiction of the experimental sequence, and see Table 2 for Experiment 4 scale properties.
Results and discussion
Manipulation check I used SPSS 23.0 for all analyses in Experiment 4. There were no missing data. A factorial ANOVA indi- cated that the PPRS was different among mild DTA (M ¼ 34.85, SD ¼ 11.09), mild neutral (M ¼ 36.26, SD ¼ 11.77), bad DTA (M ¼ 37.81, SD ¼ 11.10), and bad neutral (M ¼ 37.22, SD ¼ 10.82, gp2 ¼ .01) with a small main effect for offense recall type (gp
2 ¼ .01) and a near-zero effect for word-fragment completion condition (gp
2 ¼ � .00). A factorial ANOVA revealed that the POS was different among mild DTA (M ¼ 13.73, SD ¼ 3.35), mild neutral (M ¼ 14.25, SD ¼ 3.30), bad DTA (M ¼ 15.87, SD ¼ 3.64), and bad neutral (M ¼ 16.43, SD ¼ 3.47; gp2 ¼ .10) with a medium main effect for offense recall type (gp
2 ¼ .09) and a near-zero effect for word-fragment completion condition (gp
2 ¼ � .00). A factorial ANOVA indicated
Additional MaterialsDependent VariableOffense Packet Distractor-Delay
Measures Dependent Variable
B A D
M I L D
M I L D
B A D
D T A
Counterbalanced
Social Desirability (M-C SDS)
Mood (PANAS)
Puzzle Task
Death-Thought Accessibility (DTA Word Completion
Task)
Demographic Questionnaire
N E U T R A L
Counterbalanced
Social Desirability (M-C SDS)
Mood (PANAS)
Puzzle Task
Neutral Word Completion Task
Counterbalanced
Responsibility (PPRS)
Perceived Offense Severity (POS)
Guilt (SSGS)
Self-Esteem
(RSES)
Demographic Questionnaire
Threat Condition
Mild Offense Recall
OR
Bad Offense Recall
Self-Forgiveness (SSFS)
Self-Forgiveness (SSFS)
Figure 6. Graphical representation of the experimental procedure for Experiment 4.
364 J. M. MCCONNELL
that the SSGS Guilt subscale was different among mild DTA (M ¼ 14.31, SD ¼ 5.67), mild neutral (M ¼ 14.32, SD ¼ 5.84), bad DTA (M ¼ 17.61, SD ¼ 5.59), and bad neutral (M ¼ 17.86, SD ¼ 5.49, gp
2 ¼ .08) with a medium main effect for offense recall type (gp
2 ¼ .08) and a near-zero effect for word-frag- ment completion condition (gp
2 ¼ � .00). A factorial ANOVA specified that the RSES was different among mild DTA (M ¼ 19.51, SD ¼ 6.17), mild neutral (M ¼ 19.67, SD ¼ 6.62), bad DTA (M ¼ 16.85, SD ¼ 6.84), and bad neutral (M ¼ 17.10, SD ¼ 6.04, gp
2 ¼ .04) with a small to medium main effect for offense recall type (gp
2 ¼ .04) and a near-zero effect for word-fragment completion condition (gp
2 ¼ � .00). As with Experiment 2, the offense recall manipulation had the intended effect on responsibility, perceived offense severity, guilt, and self-esteem for the bad offense recall group.
Main analyses As with Experiment 3, I analyzed only data from the DTA word-fragment completion condition for the main analyses. A one-way ANOVA revealed that the offense recall had a minimal effect on DTA for the mild DTA (M ¼ 1.63, SD ¼ .99) and bad DTA conditions (M ¼ 1.94, SD ¼ 1.04, gp2 ¼ .02). There was a near-zero relation between distractor- delay seconds (M ¼ 477.44, SD ¼ 321.62, min ¼ 208.58, max ¼ 3903.15) and DTA in a linear regression model (b ¼ .00, b ¼ .03, R2 = .00). The M-C SDS had a near- zero difference among groups (gp
2 ¼ � .00). Although PANAS-Positive had a near-zero difference among groups (gp
2 ¼ � .00), PANAS-Negative had a small difference among mild DTA (M ¼ 18.53, SD ¼ 8.54), mild neutral (M ¼ 18.19, SD ¼ 9.28), bad DTA (M ¼ 21.06, SD ¼ 9.94), and bad neutral (M ¼ 21.71, SD ¼ 9.52, gp2 ¼ .03). Understandably, the small main effect on PANAS-Negative resulted from offense recall type (gp
2 ¼ .03) and not from word-fragment comple- tion condition (gp
2 ¼ � .00). Overall, the offense recall had a minimal effect on DTA.
Exploratory analyses To explore if DTA measurement had an effect on self-forgiveness measurement, I collapsed offense recall conditions within word-fragment completion task conditions and then conducted a one-way ANOVA that indicated SFFA had a near-zero differ- ence between word-fragment completion task condi- tions (gp
2 ¼ �.00). Also, a one-way ANOVA indicated that SFB had a near-zero difference between word- fragment completion task conditions (gp
2 ¼ �.00).
These findings indicated that filling in the DTA meas- ure in and of itself minimally impacted self-forgive- ness measurement.
Within the DTA word-fragment completion task conditions, a one-way ANOVA indicated SFFA was greater in the mild offense recall condition (M ¼ 23.09, SD ¼ 5.18) than the bad offense recall condition (M ¼ 20.75, SD ¼ 5.98, gp2 ¼ .04). Also, a one-way ANOVA indicated SFB was greater in the mild offense recall condition (M ¼ 29.57, SD ¼ 5.71) than the bad offense recall condition (M ¼ 26.95, SD ¼ 6.87, gp2 ¼ .04). Unexpectedly, DTA was nega- tively related to both SFFA (b ¼ �.70, b ¼ �.13, R2 = .02) and SFB (b ¼ �.95, b ¼ �.15, R2 = .02) to a small degree in separate linear regression models. Within the DTA word-fragment completion task conditions, I simultaneously entered PPRS (b ¼ �.04, b ¼ �.07), POS (b ¼ �.06, b ¼ �.04), SSGS Guilt (b ¼ �.16, b ¼ �.16), and RSES (b ¼ .47, b ¼ .55) into a multiple linear regression model predicting SFFA (R2 = .42). I also simultaneously entered PPRS (b ¼ .00, b ¼ .00), POS (b ¼ .00, b ¼ .00), SSGS Guilt (b ¼ �.18, b ¼ �.16), and RSES (b ¼ .60, b ¼ .62) into a multiple linear regression model predicting SFB (R2 = .47). Although PPRS and POS were minimally related or very minimally related, SSGS Guilt was negatively associated with self-forgiveness to a small degree and RSES was positively related to a large degree. Overall, unexpectedly, DTA was negatively related to self-for- giveness to a small degree.
Summary The results of Experiment 4 are consistent with the unsupportive findings of Experiments 1, 2, and 3. Although Experiment 4 found a small relation between offense recall type and DTA, offense recall only accounted for 2% of the variance in DTA. This corre- sponded with a mean difference of 0.31 death-related word completions between the mild and bad offense recall conditions. Again, this is negligibly supportive for TMT Hy3 given the central nature of self-esteem as a buffer against DTA in TMT. Alternatively, the offense recalls perhaps did not reduce the anxiety buffer enough for DTA to become preconscious to a higher degree. Nonetheless, the small relation between offense recall and DTA did not correspond with the predicted relation between DTA and self-forgiveness in explora- tory components of Experiment 4. Measurement of DTA had a near-zero impact on self-forgiveness meas- urement, but even if it did inadvertently change self- forgiveness, higher levels of DTA were associated with less self-forgiveness. In addition, a more severe anxiety-
BASIC AND APPLIED SOCIAL PSYCHOLOGY 365
buffer threat actually predicted less rather than more self-forgiveness. These findings are the opposite of TMT Hy2 and TMT Hy3. Exploratory analyses revealed that guilt was negatively associated with self-forgiveness and that self-esteem was positively related.
General discussion
Review of findings
A large body of literature has indicated that self-for- giveness is related to overall well-being, including existential well-being (Davis et al., 2015; McConnell, 2015). As proposed by McConnell’s (2015) C-T-E framework, TMT could potentially explain the basis for self-forgiveness. With theory and research in both TMT and self-forgiveness as a rationale, I positioned self-forgiveness within the mortality-salience (Hy1), death-thought accessibility (Hy2), and anxiety-buffer hypotheses of TMT (Hy3).
In Experiment 1, I hypothesized an interactional- moderated-mediation model among MS, DTA, and self-forgiveness would be supported. Specifically, I hypothesized that after an anxiety-buffer threat to induce guilt for a recent interpersonal offense, experimentally manipulated MS would increase self-forgiveness through DTA after a delay, and the relation between DTA and self-forgiveness would be moderated by offense severity, conciliatory behav- iors, perceived forgiveness, and effort to reduce emotions. Experiment 1 did not support TMT Hy1, TMT Hy2, or TMT Hy3. Although the mediation, moderated-mediation, and interactional-moderated- mediation models were not supported, there was a minimal direct effect of MS on self-forgiveness that was negligibly supportive of existential motivations for self-forgiveness. Because the results of Experiment 1 may have been disturbed by experi- mental or measurement artifacts, I followed up with a three-study experimental-causal-chain. In Experiment 2, an expanded four-group factorial experimental manipulation without measurement of DTA mediation did not support TMT Hy1 while controlling for TMT Hy3. The MS manipulation had a near-zero effect on self-forgiveness at both levels of the anxiety-buffer threat. This suggested that the small direct effect found in Experiment 1 was spuri- ous. In Experiment 3, TMT Hy3 was negligibly sup- ported when higher dispositional self-esteem predicted lower baseline DTA to a minimal degree, and the opposite of TMT Hy2 was found when higher baseline DTA predicted lower trait self-for- giveness. In Experiment 4, a two-level anxiety buffer
threat with an exploratory measurement of DTA mediation procedure negligibly supported TMT Hy3
and found the inverse of TMT Hy2. More severe offense recalls led to minimal increases in DTA and DTA was associated with less self-forgiveness.
Eternal sunshine of the spotless mind or spotty empirical findings?
Most perplexing is that in two experiments MS min- imally increased or very minimally impacted partic- ipants’ levels of a well-reasoned “distal defense” to self-esteem (i.e., self-forgiveness) after a delay. Although MS effects in Experiment 1 may have been obfuscated by unintended MS exposure in the com- parison group while measuring DTA (Hayes et al., 2010; Hayes & Schimel, 2018), the negligible findings actually were further reduced in Experiment 2 with a two-level anxiety-buffer threat and no measurement of DTA. The current studies did not support TMT Hy1. Also puzzling is that initial MS did not increase participants’ levels of DTA after a delay as indicated by TMT and much of its published research (Hayes et al., 2010; Pyszczynski et al., 1999; Steinman & Updegraff, 2015). Although baseline and experimen- tally induced DTA was minimally associated with dis- positional self-esteem and self-esteem threats in two experiments, DTA was never associated with more self-forgiveness. In fact, higher DTA was associated with less self-forgiveness in three studies. The current studies along with others directly contradicted TMT Hy2 (e.g., Hart, 2014; Hughes, 2013; Trafimow & Hughes, 2012). Likewise confusing is that one-level and two-level anxiety-buffer threats in multiple experiments did not result in MS, DTA, or distal defense effects consistent with TMT and its research (e.g., Hayes et al., 2010; Pyszczynski et al., 1999, 2015). In experiments with two-level offense recalls, there was actually higher self-forgiveness with less severe anxiety-buffer threats. The current studies failed to support TMT’s anxiety-buffer hypothesis. Overall, the negligible and unexpected findings of the current studies have challenged the central assump- tions of TMT, its most widely used methodologies, and McConnell’s (2015) C-T-E model of self-forgive- ness. Existential motivations for self-forgiveness were unsupported, whereas emotional, social-cognitive, and offense-related determinants were associated with self- forgiveness in exploratory zero-order correlation and multiple linear regression analyses. Specifically, exploratory correlation and regression analyses revealed attributions, perceived offense severity, guilt,
366 J. M. MCCONNELL
perceived forgiveness, effort to reduce emotions, neur- oticism, and self-esteem all had associations with self- forgiveness consistent with prior research (McConnell, 2015).
These unsupportive findings have importance for future experimental existential psychology studies as well as theoretical refinement of McConnell’s (2015) C-T-E model of self-forgiveness. Scholars in psych- ology have a history of ignoring or trivializing negli- gible findings and confirmatory approaches to publication. A biased publication approach is not sci- entific. A truly scientific approach gives equal weight to negligible findings, disconfimatory evidence, and alternative explanations (Popper, 1963; Tracey & Glidden-Tracey, 1999; Trafimow, 2014). The paradox of scientific progress is that important theoretical developments follow the dismantling of commonly accepted and seemingly valid ideas (Kuhn, 1962). These scientific revolutions begin from falsification (Kuhn, 1962; Popper, 1963) and from negligible find- ings (Tracey & Glidden-Tracey, 1999; Trafimow, 2014). Reviewing the history of physics is helpful in illustrating the importance of valuing negligible find- ings in psychology and in science more general (Trafimow, 2014). Toward this aim, Trafimow (2014) highlighted a famous failed experiment by Michelson and Morley (1887). Michelson and Morley were unable to support the previously popular idea that luminiferous ether carried light waves. They failed to detect any luminiferous ether, and this gave way for physicists to accept Einstein’s (1905) groundbreaking theory of special relativity. In addition, observing black swans is an apt analogy for scientific falsifica- tion. How do you know all swans are not white, or death awareness is not all-encompassing? You falsify the theory by finding one or more black swans, or one or more negligible TMT findings. There are black swans, and there are negligible TMT findings from experiments with internal validity. The results of the current studies found no compelling evidence to sup- port existential motivations for self-forgiveness or decisive support for general anxiety-buffer or DTA effects. Rather, the current studies along with prior experiments (e.g., Hart, 2014; Hughes, 2013; Trafimow & Hughes, 2012) indicated that TMT’s body of evidence is inconsistent or even contradictory to the theory. Either the theory is not truthful or the theory is not as robust as its proponents have suggested.
Given the results of the current studies, scholars can reconcile the extant TMT and self-forgiveness literature in one of three ways: (a) TMT is not true,
(b) TMT does not reliably elicit DTA and distal defenses experimentally, or (c) self-forgiveness is not a distal defense despite TMT declaring death explains virtually all human cognition, emotion, and behavior. First, despite its impressive empirical sup- port, TMT has considerable theoretical and empir- ical flaws that need to be reconciled (see Martin & van den Bos, 2014; see also Psychological Inquiry special issues: Pyszczynski et al., 1997 [Vol. 8, Issue 1] and Martin & Erber, 2006 [Vol. 17, Issue 4]). For instance, Martin and van den Bos (2014) clarified that TMT uses inconsistent evidence in confirmatory ways and ignores alternative explanations. Scholars will have to complete further conceptual, theoretical, and empirical work to understand more fully the veracity of TMT. Second, effect sizes summarized in meta-analytic reviews have indicated that TMT does not reliably produce supportive evidence. Although Burke and colleagues’ (2010) and Steinman and Updegraff’s (2015) reports provided meta-analytic support for the MS and DTA hypotheses, researchers must consider common misconceptions of meta- analytic reviews (Aguinis, Pierce, Bosco, Dalton, & Dalton, 2011). For instance, single-effect size sum- maries cannot provide a full picture of TMT’s publi- cation biases (cf. Yen & Cheng, 2013), its file drawer problems (cf. Hart, 2014; Hughes, 2013), and its effect size ranges (cf. Trafimow & Hughes, 2012). Indeed, TMT meta-analyses include studies with near-zero, minimal, and contradictory effect sizes (cf. Burke et al., 2010; Hart, 2014; Hughes, 2013; Proulx & Heine, 2008; Steinman & Updegraff, 2015; Trafimow & Hughes, 2012; Yen & Cheng, 2010, 2013). As TMT’s research findings continue to accu- mulate in, I hope, an unbiased and scientific fashion, scholars may be able to come to a deeper under- standing of the robustness of TMT. Finally, if TMT is indeed true, then its assumptions seem to suggest that self-forgiveness should be a distal defense for self-esteem reductions secondary to guilt. Although this first attempt to empirically investigate this line of reasoning was unsupportive and contradicted TMT in multiple studies, the findings of the current studies are somewhat inconclusive about TMT’s pos- sible role in self-forgiveness because MS and anx- iety-buffer threats did not have robust effects on self-esteem and DTA, the fundamental mechanisms believed to bring about distal defenses. On one hand, perhaps MS and anxiety-buffer threats did not produce the expected DTA and self-forgiveness effects because TMT is not true or because TMT effects are more circumscribed and less powerful
BASIC AND APPLIED SOCIAL PSYCHOLOGY 367
than proposed by its proponents. On the other hand, perhaps terror management motivations for self-for- giveness are nonexistent or they were simply not detected in the current studies because manipula- tions were not robust enough.
As self-forgiveness research using TMT accumu- lates, scholars may come to a more definitive conclusion about existential motivations for self- forgiveness. However, perhaps a number of alterative theories to TMT may better explain the origin of self-forgiveness. Three potential theoretical frame- works are (a) the stress and coping model (Lazarus & Folkman, 1984), (b) the integrative theory of psy- chological defense (Hart, 2014), and (c) the tripartite conceptual view of meaning in life (George & Park, 2016). The stress and coping model proposed people use problem-focused and emotion-focused coping to bring about resolutions to stressful events (Lazarus & Folkman, 1984). The integrative theory of psycho- logical defense claims threats to insecurity (death related or otherwise) activates the security system comprising self-esteem, attachment, and worldviews (Hart, 2014). The tripartite conceptual view of meaning in life integrates various meaning frame- works, such as the meaning maintenance model (Heine et al., 2006), with conceptualizations of meaning in life (i.e., comprehension, purpose, and mattering; George & Park, 2016). Alternatively, cur- rent or future attempts at a unifying theory in psychology may be fruitful (see Trafimow, 2012b; see also Gaj, 2016). Future research studies can clar- ify what theoretical frameworks have explanatory power in why people experience self-forgiveness, and thereby continue to clarify the C-T-E model of self- forgiveness proposed by McConnell (2015).
Strengths
There are several methodological strengths worthy of mention. First, I used all three of TMT’s main hypotheses in the design of multiple studies. Although no one study can fully rule out extraneous factors, an omnibus initial study with a follow-up experimen- tal-causal-chain dismantling procedure minimizes alternative explanations. Second, the randomized, posttest-only, double-blind comparison group and fac- torial designs alongside counterbalancing of measures protects well against internal threats to validity (Shadish et al., 2002). The current series of studies rules out threats to internal validity such as history, maturation, testing, selection, mortality, experimenter bias, and hypotheses guessing. Finally, I utilized three
distractor-delays following the experimental manipula- tions. Given that similar distractor-delays are what have supposedly caused the highest effect sizes in DTA and distal defenses (Burke et al., 2010; Steinman & Updegraff, 2015), the current studies’ results cannot be explained by insufficient distractor-delays.
There also are a number of statistical conclusion validity strengths. First, the breadth of variables accounted for in Experiment 1 was an important strength in attempting to account for unexplained variance and interactions. By using important varia- bles in the extant self-forgiveness literature, the study attempted to clarify how interpersonal, offense-spe- cific, and behavioral variables modify potential exist- ential motivations for self-forgiveness. Second, a number of confounds were ruled out, such as con- struct invalidity (e.g., minimizing offenses), socioemo- tional variables (e.g., social desirability & affect), nonrecent offenses, and ceiling or flooring effects. Third, counterbalancing measures improves the math- ematical rationale for using mediation approaches that have experimental manipulations, measurement-of- mediation, and conceptually based temporal sequences (Kline, 2015; Preacher et al., 2007; Spencer et al., 2005; Tate, 2015). Finally, utilization of a large sample size increases the statistical conclusion validity related to negligible findings.
Limitations
Although the current series of studies protects well against traditional threats to statistical and internal validity, they had some limitations. First, low psycho- metric internal consistency reliability of perceived offense severity partially reduced the statistical conclu- sion validity of the moderation results in Experiment 1 (Kline, 2015). Thus, the results concerning offense severity should remain cautionary.
Second, there are a number of psychometric and methodological problems with measuring DTA. Although word-fragment completion tasks are widely used in TMT research, questions can be raised about the extent they actually measure non- conscious death thoughts, or if nonconscious death thoughts ever could be measured by anything. Operationalization of nonobservable constructs must include sound specification of auxiliary assumptions that rationally bridge constructs with observable measures (Trafimow, 2012a). Perhaps DTA meas- ures, such as word-fragment completion tasks, do not meet this auxiliary specification requirement. If this is the case, then the current studies may have
368 J. M. MCCONNELL
actually missed measuring the DTA that TMT claims. In addition, measuring DTA in Experiment 1 may have contaminated the comparison group with death-related cognition if participants responded in death-related ways on the word-fragment completion task (Hayes et al., 2010; Hayes & Schimel, 2018). However, death-thought contamination was ruled out as an extraneous factor when I included neutral word- fragment completion task conditions in Experiment 4. In addition, because the measurement-of-mediation component in Experiments 1 and 4 did not manipulate DTA, these studies do not directly allow for causal inferences between DTA and self-forgiveness (Kline, 2015; Spencer et al., 2005; Tate, 2015; Trafimow, 2015). Measurement-of-mediation designs can only suggest possible causal mediation at best. Regardless, the inabil- ity to make causal inferences does not change the cur- rent studies’ conclusions. After all, findings of Experiments 1 and 4 did not suggest possible or actual mechanisms of mediation but actually supported a lack of mediation. Measurement-of-mediation results also were reinforced through the unsupportive findings of the experimental-causal-chain series. If a mediation between MS and self-forgiveness through DTA existed, then the results from the hypothetical mediation mod- els and the experimental-causal-chain would have at least suggested the hypothesized relations.
Third, Experiment 4 had a distractor-delay that may have obscured anxiety-buffer threat effects on DTA. Although there was not a relation between dis- tractor-delay time and DTA for the time frames observed, there was no one who had a zero distractor- delay time. On average, participants experienced a 477.44-s delay across three distracting measures with a minimum of 208.58-s within the sample. If anxiety- buffer effects on DTA completely disappear with any time delay, Experiment 4 would have not detected an effect. Future TMT research should systematically investigate various distractor-delay times with offense- related anxiety-buffer threats.
Fourth, an interaction of selection and treatment in Experiment 1 limits generalizability because an under- graduate sample consisting mostly of young, unmarried White Christians is different than the general popula- tion in meaningful ways, such as how relevant death is developmentally. However, three follow-up experiments using a more generalizable population had similar find- ings. Nonetheless, future ecologically valid research may explore TMT and self-forgiveness in populations such as young people in hospice care and death-row inmates to further clarify generalizability. After all, the terror of
death may be most relevant for young individuals whose lives may be cut short (Pyszczynski et al., 2015).
Conclusion
Ethical debates concerning whether self-forgiveness is warranted are controversial, but correlational research has suggested that people do seem to benefit from self-forgiveness in various biopsychosocial domains of well-being (Davis et al., 2015; McConnell, 2015). Yet scholars have not fully clarified the nature of why people experience self-forgiveness. McConnell (2015) proposed that terror management theory may be a cogent explanation for the motivation behind self-for- giveness. However, the results of the current studies do not support the assumption that MS leads to increased self-forgiveness, either directly or indirectly through DTA and various moderators, in the context of a self-esteem threat of recalling recent interpersonal offenses. This suggested self-forgiveness is not an existentially motivated experience. Alternatively, these studies and other studies (e.g., Hart, 2014; Hughes, 2013; Proulx & Heine, 2008; Trafimow & Hughes, 2012; Yen & Cheng, 2013) challenge the veracity of TMT itself, and other existentially oriented theories may explain the origin of self-forgiveness. Future studies, including replications and extensions, as well as alternative theoretical frameworks may help eluci- date why people strive for self-forgiveness and provide further refinement to McConnell’s (2015) C-T-E model of self-forgiveness.
Acknowledgments
Portions of this article are from the author’s doctoral disser- tation completed at Ball State University. Dr. McConnell expresses his gratitude to Paul M. Spengler, PhD, Stefan�ıa Ægisd�ottir, PhD, W. Holmes Finch, PhD, Thomas M. Holtgraves, PhD, David N. Dixon, PhD, Felicia A. Dixon, PhD, Kristopher J. Preacher, PhD, and David Trafimow, PhD for their helpful consultation. He also expresses grati- tude to his research assistants, Tully Roll and Donna Azcuna.
ORCID
John M. McConnell http://orcid.org/0000-0002- 4742-9854
References
Aguinis, H., Pierce, C. A., Bosco, F. A., Dalton, D. R., & Dalton, C. M. (2011). Debunking myths and urban
BASIC AND APPLIED SOCIAL PSYCHOLOGY 369
legends about meta-analysis. Organizational Research Methods, 14, 306–331. doi:10.1177/1094428110375720
Andrews, P., & Meyer, R. G. (2003). Marlowe-Crowne social desirability scale and short form C: Forensic norms. Journal of Clinical Psychology, 59, 483–492. doi: 10.1002/jclp.10136
Anglin, S. M. (2014). From avoidance to approach: The effects of mortality salience and attachment on the motiv- ation to repair troubled relationships. Personality and Individual Differences, 66, 86–91. doi:10.1016/ j.paid.2014.03.012
Arbuckle, J. L. (2007). Amos 16.0 user’s guide. Chicago, IL: Amos Development Corporation.
Arndt, J., Cook, A., Goldenberg, J. L., & Cox, C. R. (2007). Cancer and the threat of death: The cognitive dynamics of death thought suppression and its impact on behav- ioral health intentions. Journal of Personality and Social Psychology, 92, 12–29. doi:10.1037/0022-3514.92.1.12
Arndt, J., Greenberg, J., Solomon, S., Pyszczynski, T., & Schimel, J. (1999). Creativity and terror management: Evidence that creative activity increases guilt and social projection following mortality salience. Journal of Personality and Social Psychology, 77, 19–32. doi:10.1037/ 0022-3514.77.1.19
Baron, R. M., & Kenny, D. A. (1986). The moderator–me- diator variable distinction in social psychological research: Conceptual, strategic, and statistical considera- tions. Journal of Personality and Social Psychology, 51, 1173–1182. doi:10.1037/0022-3514.51.6.1173
Burke, B. L., Kosloff, S., & Landau, M. J. (2013). Death goes to the polls: A meta-analysis of mortality salience effects no political attitudes. Political Psychology, 34, 183–200. doi:10.1111/pops.12005
Burke, B. L., Martens, A., & Faucher, E. H. (2010). Two decades of terror management theory: A meta-analysis of mortality salience research. Personality and Social Psychology Review, 14, 155–195. doi:10.1177/ 1088868309352321
Cox, C. R., Eaton, S., Ekas, N. V., & Van Enkevort, E. A. (2015). Death concerns and psychological well-being in mothers of children with autism spectrum disorder. Research in Developmental Disabilities, 45-46, 229–238. doi:10.1016/j.ridd.2015.07.029
Cox, C. R., Reid-Arndt, S. A., Arndt, J., & Moser, R. P. (2012). Considering the unspoken: The role of death cog- nition in quality of life among women with and without breast cancer. Journal of Psychosocial Oncology, 30, 128–139. doi:doi:10.1080/07347332.2011.633980.
Crowne, D. P., & Marlowe, D. (1960). A new scale of social desirability independent of psychopathology. Journal of Consulting and Clinical Psychology, 24, 349–354. doi: 10.1037/h0047358
Darrell, A., & Pyszczynski, T. (2016). Terror management theory: Exploring the role of death in life. In L. A. Harvell & G. S. Nisbett (Eds.), Denying death: An inter- disciplinary approach to terror management theory (pp. 1–15). New York, NY: Routledge.
Das, E., Bushman, B. J., Bezemer, M. D., Kerkhof, P., & Vermeulen, I. E. (2009). How terrorism news reports increase prejudice against outgroups: A terror manage- ment account. Journal of Experimental Social Psychology, 45, 453–459. doi:10.1016/j.jesp.2008.12.001
Davis, D. E., Ho, M. Y., Griffin, B. J., Bell, C., Hook, J. N., Van Tongeren, D. R., … Westbrook, C. J. (2015). Forgiving the self and physical and mental health corre- lates: A meta-analytic review. Journal of Counseling Psychology, 62, 329–335. doi:10.1037/cou0000063
Davis, D. E., Worthington, E. L., Jr., Hook, J. N., & Hill, P. C. (2013). Research on religion/spirituality and forgive- ness: A meta-analytic review. Psychology of Religion and Spirituality, 5, 233–241. doi:10.1037/a0033637
del Rosario, P. M., & White, R. M. (2005). The Narcissistic Personality Inventory: Test-retest stability and internal consistency. Personality and Individual Differences, 39, 1075–1081. doi:10.1016/j.paid.2005.08.001
Echebarria-Echabe, A. (2013). Mortality salience and uncer- tainty: Similar effects but different processes?. European Journal of Social Psychology, 43, 185–191. doi:10.1002/ ejsp.1938
Einstein, A. (1905). Zur Elektrodynamik bewegter K€orper. Annalen Der Physik, 17, 890–921. doi:10.1002/ andp.19053221004
Exline, J. J., DeShea, L., & Holeman, V. T. (2007). Is apology worth the risk? Predictors, outcomes, and ways to avoid regret. Journal of Social and Clinical Psychology, 26, 479–504. doi:10.1521/jscp.2007.26.4.479
Exline, J. J., Root, B. L., Suhrida, Y., Martin, A. M., & Fisher, M. L. (2011). Reparative behaviors and self-for- giveness: Effects of a laboratory-based exercise. Self and Identity, 10, 101–126. doi:10.1080/15298861003669565
Fawcett, J. (1993). Analysis and evaluation of nursing theo- ries. Philadelphia: F. A. Davis.
Fergus, T. A., & Valentiner, D. P. (2012). Terror manage- ment theory and scrupulosity: An experimental investiga- tion. Journal of Obsessive-Compulsive and Related Disorders, 1, 104–111. doi:10.1016/j.jocrd.2012.01.003
Ferrell, B., Otis-Green, S., Baird, P., & Garcia, A. (2014). Nurses’ responses to requests for forgiveness and the end of life. Journal of Pain and Symptom Management, 47, 631–641. doi:10.1016/j.jpainsymman.2013.05.009
Fisher, M. L., & Exline, J. J. (2006). Self-forgiveness versus excusing: The roles of remorse, effort, and acceptance of responsibility. Self and Identity, 5, 127–146. doi:10.1080/ 15298860600586123
Fisher, M. L., & Exline, J. J. (2010). Moving toward self-for- giveness: Removing barriers related to shame, guilt, and regret. Social and Personality Psychology Compass, 4, 548–558. doi:10.1111/j.1751-9004.2010.00276.x
Fransen, M. L., Fennis, B. M., Pruyn, A. T. H., & Das, E. (2008). Rest in peace? Brand-induced mortality salience and consumer behavior. Journal of Business Research, 61, 1053–1061. doi:10.1016/j.jbusres.2007.09.020
Gaj, N. (2016). Unity and fragmentation in psychology: The philosophical and methodological roots of the discipline. New York, NY: Routledge
George, L. S., & Park, C. L. (2016). Meaning in life as com- prehension, purpose, and mattering: Toward integration and new research. Review of General Psychology, 20, 205–220. doi:10.1037/gpr0000077
Goldenberg, J. L., Heflick, N. A., & Cooper, D. P. (2008). The thrust of the problem: Bodily inhibitions and guilt as a function of mortality salience and neuroticism. Journal of Personality, 76, 1055–1080. doi:10.1111/j.1467- 6494.2008.00513.x
370 J. M. MCCONNELL
Greenberg, J., Pyszczynski, T., & Solomon, S. (1986). The causes and consequences of a need for self-esteem: A ter- ror management theory. In R. F. Baumeister (Ed.), Public self and private self (pp. 189–212). New York: Springer- Verlag.
Greenberg, J., Pyszczynski, T., Solomon, S., Simon, L., & Breus, M. (1994). The role of consciousness and accessi- bility of death related thoughts in mortality salience effects. Journal of Personality and Social Psychology, 67, 627–637. doi:10.1037/0022-3514.67.4.627
Greenberg, J., Solomon, S., Pyszczynski, T., Rosenblatt, A., Burling, J., Lyon, D., … Pinel, E. (1992). Why do people need self-esteem? Converging evidence that self-esteem serves as an anxiety-buffering function. Journal of Personality and Social Psychology, 63, 913–922. doi: 10.1037/0022-3514.63.6.913
Hall, J. H., & Fincham, F. D. (2005). Self-forgiveness: The stepchild of forgiveness research. Journal of Social and Clinical Psychology, 24, 621–637. doi:10.1521/ jscp.2005.24.5.621
Hall, J. H., & Fincham, F. D. (2008). The temporal course of self-forgiveness. Journal of Social and Clinical Psychology, 27, 174–202. doi:10.1521/jscp.2008.27.2.174
Harrison, P. R., & Mallet, R. K. (2013). Mortality salience motivates the defense of environmental values and increases collective ecoguilt. Ecopsychology, 5, 36–43. doi: 10.1089/eco.2012.0070
Hart, J. (2014). Toward an integrated theory of psycho- logical defense. Perspectives on Psychological Science, 9, 19–39. doi:10.1177/1745691613506018
Hayes, A. F. (2013). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. New York, NY: The Guilford Press.
Hayes, J., & Schimel, J. (2018). Unintended effects of meas- uring implicit processes: The case of death-thought acces- sibility in mortality salience studies. Journal of Experimental Social Psychology, 74, 257–269. doi:10.1016/ j.jesp.2017.10.005
Hayes, J., Schimel, J., Arndt, J., & Faucher, E. H. (2010). A theoretical and empirical review of death-thought accessi- bility concept in terror management research. Psychological Bulletin, 136, 699–739. doi:10.1037/ a0020524
Hayes, J., Schimel, J., Williams, T. J., Howard, A. L., Webber, D., & Faucher, E. H. (2015). Worldview accom- modation: Selectively modifying committed beliefs pro- vides defense against worldview threat. Self and Identity, 14, 521–548. doi:10.1080/15298868.2015.1036919
Heflick, N. A. (2005). Sentenced to die: Last statements and dying on death row. Omega - Journal of Death and Dying, 51, 323–336. doi:10.2190/96X8-FLUT-TCLH-EL71
Heine, S. J., Proulx, T., & Vohs, K. D. (2006). The meaning maintenance model: On the coherence of social motiva- tions. Personality and Social Psychology Review, 10, 88–110. doi:10.1207/s15327957pspr1002_1
Henriques, G. (2011). A new unified theory of psychology. New York, NY: Springer
Hirschberger, G., Florian, V., & Mikulincer, M. (2002). The anxiety buffering function of close relationships: Mortality salience effects on the readiness to compromise mate selection standards. European Journal of Social Psychology, 32, 609–625. doi:10.1002/ejsp.110
Hughes, J. S. (2013). [Tests of TMT death thought suppres- sion and rebound hypothesis]. Unpublished raw data.
Iverach, L., Menzies, R. G., & Menzies, R. E. (2014). Death anxiety and its role in psychopathology: Reviewing the status of a transdiagnostic construct. Clinical Psychology Review, 34, 580–593. doi:10.1016/j.cpr.2014.09.002
John, O. P., Donahue, E. M., & Kentle, R. L. (1991). The Big Five Inventory–Versions 4a and 54. Berkeley, CA: University of California, Berkeley, Institute of Personality and Social Research.
John, O. P., Naumann, L. P., & Soto, C. J. (2008). Paradigm shift to the integrative Big Five trait taxonomy: History, measurement, and conceptual issues. In O. P. John, R. W. Robins, & L. A. Pervin (Eds.), Handbook of personal- ity: Theory and research (pp. 114–158). New York, NY: Guilford Press.
Kastenbaum, R., & Heflick, N. A. (2011). Sad to say: Is it time for sorrow management theory?. Omega - Journal of Death and Dying, 62, 305–327. doi:10.2190/OM.62.4.a
Kim, S., Thibodeau, R., & Jorgensen, R. S. (2011). Shame, guilt, and depressive symptoms: A meta-analytic review. Psychological Bulletin, 137, 68–96. doi:10.1037/a0021466
Kline, R. (2015). The mediation myth. Basic and Applied Social Psychology, 37, 202–213. doi:10.1080/ 01973533.2015.1049349
Kuhn, T. S. (1962). The structure of scientific revolutions. Chicago: University of Chicago Press.
Lavelock, C. R., Griffin, B. J., & Worthington, E. L. Jr. (2013). Forgiveness, religiousness, spirituality, and health in people with physical challenges: A review and a model. In R. L. Piedmont & A. Village (Eds.), Research in the social scientific study of religion (Vol. 24, pp. 53–92). Boston: Brill.
Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal, and coping. New York: Springer.
Marschall, D., Sanfter, J., & Tangney, J. P. (1994). The state shame and guilt scale. Fairfax, VA: George Mason University.
Martens, A., Burke, B. L., Schimel, J., & Faucher, E. H. (2011). Same but different: Meta-analytically examining the uniqueness of mortality salience effects. European Journal of Social Psychology, 41, 6–10. doi:10.1002/ ejsp.767
Martin, L. L., & Erber, R. (2006). Introduction [Special issue]. Psychological Inquiry, 17, 277. doi:10.1080/ 10478400701366928
Martin, L. L., & van den Bos, K. (2014). Beyond terror: Towards a paradigm shift in the study of threat and cul- ture. European Review of Social Psychology, 25, 32–70. doi:10.1080/10463283.2014.923144
McAuley, E., Duncan, T. E., & Russell, D. W. (1992). Measuring causal attributions: The revised causal dimen- sion scale (CDS-II). Personality and Social Psychology Bulletin, 18, 566–573. doi:10.1177/0146167292185006
McConnell, J. M. (2015). A conceptual-theoretical-empirical framework for self-forgiveness: Implications for research and practice. Basic and Applied Social Psychology, 37, 143–164. doi:/10.1080/01973533.2015.1016160
McConnell, J. M., Dixon, D. N., & Finch, W. H. (2012). An alternative model of self- forgiveness. The New School Psychology Bulletin, 9, 35–51.
BASIC AND APPLIED SOCIAL PSYCHOLOGY 371
McCullough, M. E., Rachal, K. C., Sandage, S. J., Worthington, E. L., Jr., Brown, S. W., & Hight, T. L. (1998). Interpersonal forgiving in close relationships: II. Theoretical elaboration and measurement. Journal of Personality and Social Psychology, 75, 1586–1603. doi: 10.1037/0022-3514.75.6.1586
McCullough, M. E., Worthington, E. L., Jr., & Rachal, K. C. (1997). Interpersonal forgiving in close relationships. Journal of Personality and Social Psychology, 73, 321–336. doi:10.1037/0022-3514.73.2.321
McGregor, H., Lieberman, J. D., Solomon, S., Greenberg, J., Arndt, J., Simon, L., & Pyszczynski, T. (1998). Terror management and aggression: Evidence that mortality sali- ence motivates aggression against worldview threatening others. Journal of Personality and Social Psychology, 74, 590–605. doi:10.1037/0022-3514.74.3.590
Michelson, A. A., & Morley, E. W. (1887). On the relative motion of the earth and the luminiferous ether. American Journal of Science, s3–34, 333–345. doi:10.2475/ajs.s3- 34.203.333
Mikulincer, M., & Florian, V. (2002). The effect of mortality salience on self-serving attributions—Evidence for the function of self-esteem as a terror management mechan- ism. Basic and Applied Social Psychology, 24, 261–271. doi:10.1207/S15324834BASP2404_2
Motyl, M., Hart, J., Cooper, D. P., Heflick, N., Goldenberg, J., & Pyszczynski, T. (2013). Creatureliness priming reduces aggression and support for war. British Journal of Social Psychology, 52, 648–666. doi:10.1111/j.2044- 8309.2012.02115.x
Mullet, E., Neto, F., & Rivi�ere, S. (2005). Personality and its effects on revenge, resentment, forgiveness, and self-forgiveness. In E. L. Worthington Jr. (Ed.), Handbook of forgiveness (pp. 159–181). New York: Routledge.
Olshtain, E. (1989). Apologies across languages. In S. Blum- Kulka, J. House, & G. Kasper (Eds.), Cross-cultural prag- matics: Requests and apologies (pp. 157). Norwood, NJ: Ablex.
Pope, A. (1796). A select collection of poems. New London: Springer. (Original work published 1717)
Popper, K. (1963). Conjectures and refutations: The growth of scientific knowledge. New York, NY: Routledge & Kegan Paul.
Preacher, K. J., Rucker, D. D., & Hayes, A. F. (2007). Addressing moderated mediation hypotheses: Theory, methods, and prescriptions. Multivariate Behavioral Research, 42, 185–227. doi:10.1080/00273170701341316.
Proulx, T. (2012). Threat-compensation in social psych- ology: Is there a core motivation? [Special issue]. Social Cognition, 30, 643–651. doi:10.1521/soco.2012.30.6.643
Proulx, T., & Heine, S. J. (2008). The case of the transmog- rifying experimenter: Affirmation of a moral schema fol- lowing implicit change detection. Psychological Science, 19, 1294–1300. doi:10.1111/j.1467-9280.2008.02238.x
Proulx, T., Inzlicht, M., & Harmon-Jones, E. (2012). Understanding all inconsistency compensation as a pallia- tive response to violated expectations. Trends in Cognitive Sciences, 16, 285–291. doi:10.1016/j.tics.2012.04.002
Pyszczynski, T., Greenberg, J., & Solomon, S. (1997). Why do we need what we need? A terror management per- spective on the roots of human social motivation [Special
issue]. Psychological Inquiry, 8, 1–20. doi:10.1207/ s15327965pli0801_1
Pyszczynski, T., Greenberg, J., & Solomon, S. (1999). A dual-process model of defense against conscious and unconscious death-related thoughts: An extension of ter- ror management theory. Psychological Review, 106, 835–845. doi:10.1037/0033-295X.106.4.835
Pyszczynski, T., Greenberg, J., Solomon, S., Arndt, J., & Schimel, J. (2004). Why do people need self-esteem? A theoretical and empirical review. Psychological Bulletin, 130, 435–468. doi:10.1037/0033-2909.130.3.435
Pyszczynski, T., Solomon, S., & Greenberg, J. (2015). Thirty years of terror management theory: From genesis to reve- lation. Advances in Experimental Social Psychology, 52, 1–70. doi:10.1016/bs.aesp.2015.03.001
Rammstedt, B., & John, O. P. (2007). Measuring personality in one minute or less: A 10-item short version of the Big Five Inventory in English and German. Journal of Research in Personality, 41, 203–212. doi:10.1016/ j.jrp.2006.02.001
Raskin, R., & Terry, H. (1988). Principal-components ana- lysis of the Narcissistic Personality Inventory and further evidence of its construct validity. Journal of Personality and Social Psychology, 54, 890–902. doi:10.1037/0022- 3514.54.5.890
Robins, R. W., Hendin, H. M., & Trzesniewski, K. H. (2001). Measuring global self-esteem: Construct valid- ation of a single-item measure and the Rosenberg self- esteem scale. Personality and Social Psychology Bulletin, 27, 151–161. doi:10.1177/0146167201272002
Rosenberg, M. (1965). Society and the adolescent self-image. Princeton: Princeton University Press.
Rosenblatt, A., Greenberg, J., Solomon, S., Pyszczynski, T., & Lyon, D. (1989). Evidence for terror management the- ory: I. The effects of mortality salience on reactions to those who violate or uphold cultural values. Journal of Personality and Social Psychology, 57, 681–690.
Rubin, D. B. (1987). Multiple imputation for nonresponse in surveys. New York: John Wiley & Sons.
Schimel, J., Landau, M., & Hayes, J. (2008). Self-esteem: A human solution to the problem of death. Social and Personality Psychology Compass, 2, 1218–1234. doi: 10.1111/j.1751-9004.2008.00104.x
Schimel, J., Wohl, M. J. A., & Williams, T. (2006). Terror management and trait empathy: Evidence that mortality salience promotes reactions of forgiveness among people with high (vs. low) trait empathy. Motivation and Emotion, 30, 214–227. doi:10.1007/s11031-006-9040-y
Shadish, W. R., Cook, T. D., & Campbell, D. T. (2002). Experimental and quasi-experimental designs for general- ized causal inference. Boston: Houghton Mifflin.
Shan, J., & Xu, G. (2008). Self-forgiveness–A new way of mental counseling. Chinese Journal of Clinical Psychology, 16, 92–94.
Sobel, M. E. (1982). Asymptotic confidence intervals for indirect effects in structural equation models. In S. Leinhardt (Ed.), Sociological Methodology (pp. 290–312). Washington, DC: American Sociological Association.
Solomon, S., Greenberg, J., & Pyszczynski, T. (2004). The cultural animal: Twenty years of terror management the- ory and research. In J. Greenberg, S. L. Koole, & T.
372 J. M. MCCONNELL
Pyszczynski (Eds.), Handbook of Experimental Existential Psychology (pp. 13–34). New York, NY: Guilford Press.
Solomon, S., Greenberg, J., & Pyszczynski, T. (1997). Return of the living dead. Psychological Inquiry, 8, 59–71. doi: 10.1207/s15327965pli0801_13
Spencer, S. J., Zanna, M. P., & Fong, G. T. (2005). Establishing a causal chain: Why experiments are often more effective than meditational analyses in examining psychological processes. Journal of Personality and Social Psychology, 89, 845–851. doi:10.1037/0022-3514.89.6.845
Steinman, C. T., & Updegraff, J. A. (2015). Delay and death-thought accessibility: A meta-analysis. Personality and Social Psychology Bulletin, 41, 1682–1696. doi: 10.1177/0146167215607843
Tangney, J. P., Boone, A. L., & Dearing, R. L. (2005). Forgiving the self: Conceptual issues and empirical find- ings. In E. L. Worthington, Jr. (Ed.), Handbook of forgive- ness (pp.143–158). New York, NY: Routledge.
Tangney, J. P., & Dearing, R. L. (2002). Shame and guilt. New York, NY: The Guilford Press.
Tate, C. U. (2015). On the overuse and misuse of mediation analysis: It may be a matter of timing. Basic and Applied Social Psychology, 37, 235–246. doi:10.1080/ 01973533.2015.1062380
Thompson, L. Y., Snyder, C. R., Hoffman, L., Michael, S. T., Rasmussen, H. N., Billings, L. S., … Roberts, D. E. (2005). Dispositional forgiveness of self, others, and situa- tions. Journal of Personality, 73, 313–360. doi:10.1111/ j.1467-6494.2005.00311.x
Tracey, T. J. G., & Glidden-Tracey, C. E. (1999). Integration of theory, research design, measurement, and analysis: Toward a reasoned argument. The Counseling Psychologist, 27, 299–324. doi:10.1177/0011000099273002
Trafimow, D. (2012). The role of auxiliary assumption for the validity of manipulations and measures. Theory & Psychology, 22, 486–498. doi:10.1177/0959354311429996
Trafimow, D. (2012). The role of mechanisms, integration, and unification in science and psychology. Theory & Psychology, 22, 697–703. doi:10.1177/0959354311433929
Trafimow, D. (2014). Editorial. Basic and Applied Social Psychology, 36, 1–2. doi:10.1080/01973533.2014.865505
Trafimow, D. (2015). Introduction to the special issue on mediation analyses: What if planetary scientists used mediation analysis to infer causation? [Special issue]. Basic and Applied Social Psychology, 37, 197–201. doi: 10.1080/01973533.2015.1064290
Trafimow, D., & Hughes, J. S. (2012). Testing the death thought suppression and rebound hypothesis: Death thought accessibility following mortality salience decreases during a delay. Social Psychological and Personality Science, 3, 622–629. doi:10.1177/ 1948550611432938
Vail, K. I., Arndt, J., Motyl, M., & Pyszczynski, T. (2012). The aftermath of destruction: Images of destroyed build- ings increase support for war, dogmatism, and death thought accessibility. Journal of Experimental Social Psychology, 48, 1069–1081. doi:10.1016/j.jesp.2012.05.004
Vail, K. E., III, Juhl, J., Arndt, J., Vess, M., Routledge, C., & Rutjens, B. T. (2012). When death is good for life: Considering the positive trajectories of terror
management. Personality and Social Psychology Review, 16, 303–329. doi:10.1177/1088868312440046
vanDellen, M. R., Campbell, W. K., Hoyle, R. H., & Bradfield, E. K. (2011). Compensating, resisting, and breaking: A meta-analytic examination of reactions to self-esteem threat. Personality and Social Psychology Review, 15, 51–74. doi:10.1177/1088868310372950
Van Tongeren, D. R., Green, J. D., Davis, D. E., Worthington, E. L., & Reid, C. A. (2013). Till death do us part: Terror management and forgiveness in close rela- tionships. Personal Relationships, 20, 755–768. doi: 10.1111/pere.12013
Watson, D., Clark, L. E., & Tellegen, A. (1988). Development and validation of brief measures of positive and negative affect: The PANAS scales. Journal of Personality and Social Psychology, 54, 1063–1070. doi: 10.1037/0022-3514.54.6.1063
Webb, J. R., Hirsch, J. K., & Toussaint, L. (2011). Forgiveness and alcohol problems: A review of the litera- ture and a call for intervention-based research. Alcoholism Treatment Quarterly, 29, 245–273. doi: 10.1080/07347324.2011.585922
Webber, D., Zhang, R., Schimel, J., & Blatter, J. (2016). Finding death in meaninglessness: Evidence that death- thought accessibility increases in response to meaning threats. British Journal of Social Psychology, 55, 144–161. doi:10.1111/bjso.12118
Weiner, B., Graham, S., Peter, O., & Zmuidinas, M. (1991). Public confession and forgiveness. Journal of Personality, 59, 281–312. doi:10.1111/j.1467-6494.1991.tb00777.x
Wilson, K. M., & Bernas, R. (2011). A good man is hard to find: Forgiveness, terror management, and religiosity. Research in the Social Scientific Study of Religion, 22, 126–140. doi:10.1163/ej.9789004207271.i-360.18
Wohl, M. J. A., DeShea, L., & Wahkinney, R. L. (2008). Looking within: Measuring state self-forgiveness and its rela- tionship to psychological wellbeing. Canadian Journal of Behavioural Science, 40, 1–10. doi:10.1037/0008-400x.40.1.1.1
Wohl, M. J. A., & McLaughlin, K. J. (2014). Self-forgiveness: The good, the bad, and the ugly. Social and Personality Compass, 8, 422–425. doi:10.1111/spc3.12119
Woodyatt, L., Worthington, E. L. Jr., Wenzel, M., & Griffin, B. J. (Eds.). (2017). Handbook of the psychology of self-for- giveness. Cham, Switzerland: Springer International Publishing AG.
Worthington, E. L., Jr., Witvliet, C. V. O., Pietrini, P., & Miller, A. J. (2007). Forgiveness, health, and well-being: Review of evidence for emotional versus decisional for- giveness, dispositional forgiveness, and reduced unfor- giveness. Journal of Behavioral Medicine, 30, 291–302. doi:10.1007/s10865-007-9105-8
Yen, C. L., & Cheng, C. P. (2010). Terror management among Taiwanese: Worldview defence or resigning to fate? Asian Journal of Social Psychology, 13, 185–194. doi: 10.1111/j.1467- 839X.2010.01328.x
Yen, C. L., & Cheng, C. P. (2013). Researcher effects on mortality salience research: A meta- analytic moderator analysis. Death Studies, 37, 636–652. doi:10.1080/ 07481187.2012.682290
BASIC AND APPLIED SOCIAL PSYCHOLOGY 373
Copyright of Basic & Applied Social Psychology is the property of Taylor & Francis Ltd and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
- Abstract
- Psychological origins of self-forgiveness
- Terror management theory
- Mortality salience hypothesis
- Death-thought accessibility hypothesis
- Anxiety-buffer hypothesis
- Hypotheses summary
- Self-forgiveness: A self-esteem distal defense process?
- General method
- Experiment 1
- Method
- Participants
- Procedure
- Materials
- Offense packet
- Death awareness packet
- Distractor-delay measures
- Mediating measure: DTA
- Deception
- Criterion measures
- Additional moderating measure: Effort
- Demographic questionnaire
- Results and discussion
- Missing data
- Responsibility
- Intercorrelation matrices
- Main analyses
- Mediation hypothesis
- Moderated-mediation hypothesis
- Interactional-moderated-mediation hypothesis
- Summary
- Experiment 2
- Method
- Participants
- Procedure
- Materials
- Offense packet
- Death awareness packet
- Results and discussion
- Manipulation check
- Main analyses
- Exploratory analyses
- Summary
- Experiment 3
- Method
- Participants
- Procedure
- Materials
- Neutral word-fragment completion task
- Trait Self-Forgiveness
- Results and discussion
- Main analyses
- Exploratory analysis
- Summary
- Experiment 4
- Method
- Participants
- Procedure and materials
- Results and discussion
- Manipulation check
- Main analyses
- Exploratory analyses
- Summary
- General discussion
- Review of findings
- Eternal sunshine of the spotless mind or spotty empirical findings?
- Strengths
- Limitations
- Conclusion
- Acknowledgments
- References
