Research Question

Available online at www.sciencedirect.com

ScienceDirect Behavior Therapy 49 (2018) 866–880

www.elsevier.com/locate/bt

I Did OK, but Did I Like It? Using Ecological Momentary Assessment to Examine Perceptions of Social Interactions Associated With Severity of Social Anxiety and Depression

Emily C. Geyer Karl C. Fua

Katharine E. Daniel⁎

Philip I. Chow Wes Bonelli Yu Huang

Laura E. Barnes Bethany A. Teachman University of Virginia

Socially anxious and depressed individuals tend to evaluate their social interactions negatively, but little is known about the specific real-time contributors to these negative percep- tions. The current study examined how affect ratings during social interactions predict later perceptions of those interactions, and whether this differs by social anxiety and depression severity. Undergraduate participants (N = 60) responded to a smartphone application that prompted participants to answer short questions about their current affect and social context up to 6 times a day for 2 weeks. At the end of each day, participants answered questions about their perceptions of their social interactions from that day. Results indicated that the link between negative affective

This research was supported by a University of Virginia Hobby Postdoctoral Fellowship and Predoctoral Fellowships in Computa- tional Science. We thank the University of Virginia’s Program for Anxiety, Cognition, and Treatment Lab and the Barnes Lab, and would especially like to extend our appreciation to Sarai Arbus, Somil Chugh, Virginia Clemo, Sekar Novika, Bryana Schantz, and Kaitlin Wray, who worked as research assistants on this study. We are also grateful to our colleagues from the UVA engineering department—Matthew Gerber and Haoyi Xiong—who helped develop the smartphone application, Sensus.

0005-7894/© 2018 Association for Behavioral and Cognitive Therapies. Published by Elsevier Ltd. All rights reserved.

experiences reported during social interactions and the end- of-day report of enjoyment (but not effectiveness) of those experiences was more negative when social anxiety was more severe. The link between negative affective experiences rated during social interactions and the end-of-day report of effectiveness (but not enjoyment) during those social encounters was more negative when depression was more severe. These findings demonstrate the importance of examining self-perceptions of social interactions based both on the extent to which individuals think that they met the objective demands of an interaction (i.e., effective- ness, mastery) and the extent to which they liked or disliked that interaction (i.e., enjoyment, pleasure). These findings also highlight how real-time assessments of daily social interactions may reveal the key experiences that contribute to negative self-evaluations across disorders, potentially identifying critical targets for therapy.

Keywords: Ecological Momentary Assessment; social anxiety; depression; social interaction; affect

SOCIAL ANXIETY AND MAJOR DEPRESSIVE DISORDER are highly prevalent (with lifetime prevalence rates of 12% and 17%, respectively, in the U.S. population; Kessler et al., 2005) and have insidious conse- quences across many domains of functioning. Even

867percept ions of soc i al int eract ions

subthreshold levels of social anxiety and depression can have damaging effects (Fergusson, Horwood, Ridder, & Beautrais, 2005). For individuals with symptoms of social anxiety and depression, there often seems to be a disconnect between their in-the- moment social experiences and their memory of those experiences, presumably due to biases that affect their memory and judgment (e.g. Amir, Bower, Briks, & Freshman, 2003). For instance, although socially anxious individuals often perform objectively as well as nonanxious individuals in social situations, they are likely to retrospectively rate their own performance more negatively (Norton & Hope, 2001). This disconnect is problematic because it means that, even when things seem to go well in the moment, impairments in the ability to process positive affective informa- tion (e.g., Kashdan, 2007) can hinder the creation or reinforcement of a positive self-concept. This may reinforce a negative self-concept over time, making it unlikely socially anxious and depressed individuals will be open to taking on new social challenges, thereby perpetuating maladaptive social withdrawal patterns. To date, there has been little research on the relationship between in-the-mo- ment positive and negative affect ratings during social experiences and later, retrospective ratings of interpersonal effectiveness during, and enjoyment of, those same experiences. Using advances in technology-assisted data collection methodologies (e.g., Ecological Momentary Assessment; EMA), we can more directly assess in what ways daily experiences do or do not inform memories for social interactions, ultimately providing insight into the mechanisms underlying judgment and memory biases in social anxiety and depression. The current study tracked self-reports of positive

and negative affect experienced by participants during social interactions throughout the day using randomly timed questionnaire prompts deployed via a mobile phone application (Sensus; Xiong, Huang, Barnes, & Gerber, 2016) installed on participants’ personal phones. Sensus also admin- istered questionnaires at the end of each day to gain perspective on participants’ retrospective reports of the day. We examined how severity of social anxiety and depression each moderated the associ- ation between the in-the-moment affect ratings obtained throughout the day and participants’ nightly retrospective reports of their perceived effectiveness and enjoyment of their social interac- tions that day. This daily monitoring approach can provide a more nuanced and ecologically valid understanding of how real-time affect predicts later perceptions of social interactions. This approach also lays the groundwork for future refinements to

assessments and interventions for highly prevalent mood and anxiety experiences.

Negative cognitive biases Cognitive biases influence the way socially anxious and depressed individuals process their social experiences and how these experiences are subse- quently incorporated into their positive and nega- tive beliefs about themselves. During social interactions, socially anxious (vs. non–socially anxious) individuals tend to exhibit a range of negative cognitive biases, such as interpreting ambiguous stimuli more negatively (Amir, Beard, & Bower, 2005), reporting less positive affect (Kashdan, 2007), and activating more negative beliefs about themselves (Hofmann, 2007). Nota- bly, there are mixed findings about whether socially anxious individuals display more negative memory biases than do nonanxious individuals. While some research suggests that after social interactions, socially anxious individuals recall these experiences as being more negative (Mellings & Alden, 2000), other studies were unable to establish a reliable relationship between social anxiety and memory biases (e.g., Heinrichs & Hofmann, 2001), and effect sizes vary considerably based on encoding procedures and retention intervals used (see review inMitte, 2008). It is important to note that memory bias research with mixed outcomes generally utilizes lab-based memory tasks involving words (e.g., valenced or threatening vs. nonthreatening words). Research on post-event processing that has examined recall of social anxiety-relevant events (e.g., event details, performance feedback, self- judgments), which is arguably more ecologically valid than word recall, has somewhat more reliably shown that socially anxious individuals’ negatively biased self-concepts, interpretations, and memories are maintained and perpetuated by post-event processing (Brozovich & Heimberg, 2008). Likewise, research has indicated that depressed

individuals exhibit negative memory biases, espe- cially on tests of explicit memory examining preferential recall for negative information (see MacLeod & Mathews, 2004). For instance, de- pressed (vs. nondepressed) individuals are more likely to recall negative feelings from their experi- ences (Dalgleish & Watts, 1990), lack normative preferential encoding of positive cues (Gotlib, Jonides, Buschkuehl, & Joormann, 2011), and are also more likely to provide overgeneral, nonspecific reports of their memories (Sumner, Griffith, & Mineka, 2010). Formerly (vs. never) depressed patients are also more likely to recall negative information about themselves (Romero, Sanchez, & Vazquez, 2014). Even so, as is the case for social

868 geyer et al .

anxiety, results in the depression literature are mixed (Barry, Naus, & Rehm, 2004). While the current study does not directly assess memory biases for social interactions (because we have no “ground truth” for how effective the day’s interac- tions were), this mixed literature nonetheless provides important background and motivation for the current study, because our approach may help clarify under what conditions negative mem- ories for social interactions will occur by examining how real-time affect predicts these later memories. This is important because understanding the specific real-time contributors to negative percep- tions of social interactions can be used to develop interventions that target those factors in situ. Affect is likely a strong predictor of how events

will be remembered, given that affect influences how deeply events are processed (Clore, Gasper, & Garvin, 2001), and activates and reinforces pre- existing beliefs about the self (Petersen, Stahlberg, & Dauenheimer, 2000; Robinson & Clore, 2002). Given that positive and negative affect can be considered as two correlated but separable dimen- sions (e.g., Cacioppo & Berntson, 1994), it is important to consider the impact of both dimen- sions on individuals’ recollection of experiences. Negative affect during social interactions seems likely to have a greater negative impact on individuals’ later recollections of the events the more severe their social anxiety and depression are, given that socially anxious individuals fear negative evaluations from others and depressed individuals have heightened feelings of worthlessness (McGirr et al., 2007) and expectations of undesirable events (Strunk, Lopez, & DeRubeis, 2006). Positive affect seems likely to have a relatively smaller impact on the relationship between an individual’s reported social anxiety or depression severity and that individual’s recollection of their social interactions, given socially anxious and depressed individuals typically experience impairment when processing positive experiences (e.g., Kashdan, 2007; Werner- Seidler, Banks, Dunn, & Moulds, 2013), which may limit their processing of the positive affective information. Alternatively, cognitive biases associ- ated with social anxiety or depression may increase individuals’ sensitivity to the lack of positive affect, highlighting differences between expected feelings of positivity and actual affective experience during an event. This discrepancy between expected and actual experience could, in turn, strengthen nega- tive beliefs about the self (e.g., Wood, Anthony, & Foddis, 2006). Thus, negative affect, and possibly lack of positive affect, during social experiences are expected to be particularly influential predictors of subsequent memories when social anxiety and

depression are more severe, because these affective experiences are more congruent with the negative self-concepts associated with social anxiety and depression, and will therefore be processed more deeply and preferentially recalled when social anxiety and depression are more severe.

PERCEIVED SOCIAL EFFECTIVENESS AND ENJOYMENT

In terms of which aspects of the memories for social interactions would be predicted by affect, we focused on those features that are emphasized in research concerning whether people are experienc- ing reinforcement from their environment—name- ly, mastery and pleasure—due to the centrality of these features for maintaining social withdrawal and dysphoric mood (see discussion of behavioral activation principles in Dimidjian, Martell, Herman-Dunn, & Hubley, 2014). Socially anxious (compared to nonanxious) individuals are more likely to engage in post-event processing of social events during which perceived negative information about their prior performance is frequently replayed (Edwards, Rapee, & Franklin, 2003). This emphasis on (biased) evaluations of effective- ness has been a major focus in past research (e.g., Hofmann, 2007; Moscovitch, 2009) for good reason, given prominent fears about negative evaluation in social anxiety. Importantly, we expect that state anxiety and other forms of negative affect during social situations will also influence later perceptions of whether social events were enjoy- able, given that affect plays a key role in individ- uals’ assessment of both liking (i.e., enjoyment) and effectiveness (Clore et al., 2001). Consistent with this idea, socially anxious individuals experience both increased negative affect during social inter- actions and also during post-event processing (e.g., Kashdan & Roberts, 2007), which seems likely to impact evaluations of both effectiveness and enjoyment of earlier social interactions. Recollec- tions of enjoyment have been understudied in social anxiety but are important to better understand because they likely influence decisions about subsequent social engagement vs. avoidance. These outcomes are central for understanding depressive biases and withdrawal behaviors as well. Depressed (vs. nondepressed) individuals tend to rate their social competence or effectiveness more negatively (Gable & Shean, 2000), and also regularly experience anhedonia (a loss of pleasur- able affect; Watson et al., 1995), which would likely impact depressed individuals’ evaluations of their enjoyment of social interactions. Together, this suggests both perceived ineffec-

tiveness and lack of enjoyment are likely more salient during and after social interactions when

869percept ions of soc i al int eract ions

social anxiety and depression are more severe, and both may be strong indicators of the ways in which social interactions are later remembered (and misremembered).

Ecological momentary assessment: overview and hypotheses

EMA is increasingly being used in studies as a way to track people’s thoughts, behaviors, emotions, and contexts within their natural environment. Given that up to 77% of Americans own a smartphone (Pew Research Center, 2018), EMA studies are uniquely positioned to capture rich data in an efficient and relatively more ecologically valid way. EMA designs have been successfully used to study different forms of psychopathologies (Ben- Zeev, Young, & Madsen, 2009; Kashdan & Collins, 2010) and track treatment outcomes over time (Shiffman, Stone, & Hufford, 2008). For instance, Kashdan and Collins (2010) utilized an EMA design to track individuals’ everyday feelings and found that higher levels of social anxiety were associated with less frequent reports of positive emotions and more frequent reports of anger throughout the day. In another study, Ben-Zeev and colleagues (2009) used EMA to examine the relationship between depressed individuals’ in-the- moment self-reports of affect and their retrospective summaries of those moments, and showed that depressed (vs. nondepressed) individuals were more inaccurate in recalling negative affect. This suggests that EMA designs can provide a useful means to more naturalistically track how severity of social anxiety or depression moderates how individuals’ daily affective experiences predict their subsequent perceptions of their earlier social interactions. With prompted questions using EMA, we can get much closer in time to reporting on events than is possible with typical questionnaires. However, it remains unrealistic to require participants to always imme- diately stop their daily activities to answer ques- tionnaires throughout the day. Though this means that affect ratings during social interactions might not truly be “in-the-moment” (e.g., in our study, the average elapsed time before a participant responds to a random time prompt is 16 minutes), these ratings bring us closer to true in situ reports and minimize recall bias. The present study used an EMA design to track

college students’ daily affect during social interac- tions over 2 weeks using participant’s smartphones. The study aimed to examine the impact of baseline social anxiety and depression severity on the relationship between individuals’ in-the-moment positive and negative affect ratings during their social interactions and their retrospective percep-

tions of their effectiveness during, and enjoyment of, those social interactions at the end of the same day. First, we hypothesized that, across all levels of

social anxiety and depression severity, self-reports of negative affect during social interactions would be negatively related to retrospective ratings of overall effectiveness and enjoyment at the end of that day, and positive affect would be positively related to those same retrospective ratings. Second, we hypothesized that the negative relationship between in-the-moment negative affect during social interactions and individuals’ retrospective ratings of perceived effectiveness within, and enjoyment of, these interactions would be stronger when social anxiety or depression was more severe. Third, it was not clear how increased severity of social anxiety or depression would impact reactions to positive experiences, because high symptom levels across social anxiety and depression are associated with both pervasive negative biases and with nonreactivity to the presence of positive affect. Thus, it was plausible that even in-the-moment positive affect during social interactions would more strongly predict negative retrospective ratings when social anxiety or depression was more severe, given that positive affect may still activate fears of evaluation. Alternatively, pervasive negative biases could also make situations where positive affect was absent more salient, resulting in stronger associations between lack of positive affect and negative retrospective ratings. Yet another possibil- ity is that severity of social anxiety or depression may not particularly influence the relationship between in-the-moment positive affect and later memories of social interactions. This could occur given that negative interactions are typically very salient in social anxiety and depression while positive cues are often not remembered as well (Mansell, Ehlers, Clark, & Chen, 2002; Duque & Vázquez, 2015; Levens & Gotlib, 2010). Hence, our tests for effects of positive affective experiences were exploratory. We did not have specific hypotheses pertaining to

differences in how social anxiety vs. depression severity were expected to moderate same-day recall of social interactions. Social anxiety and depression are highly comorbid, and given the high similarity in the nature of cognitive biases present in each condition, it was expected that the impact of social anxiety and depression severity on the relationships between in-the-moment affect and retrospective ratings would be similar. Thus, comparisons between social anxiety and depression in our analyses were exploratory. Notwithstanding, given there are both some interesting similarities

870 geyer et al .

between social anxiety and depression (e.g., beliefs about inadequacy and tendency for social with- drawal) and some interesting differences (e.g., different findings for memory biases and motiva- tion), we felt it was valuable to examine both as moderators.

Method PARTICIPANTS

A total of 72 undergraduate participants enrolled in the study. Twelve participants were excluded from analyses due to software bugs and compatibility issues, leaving N = 60 participants between 17 and 36 years old (mean = 19.9, SD = 2.6; 55% female; only 1 participant was older than 22). Participants reported their race and ethnicity as 40%White, 5% Black, 5% Hispanic, 37% Asian, 8% Mixed, and 5% unknown/unreported. This research was con- ducted using a novel, customized mobile phone application developed at the University of Virginia (Xiong et al., 2016). Participants were eligible to participate if they had access to an iPhone (version 5 or newer) or an Android mobile phone running at least an Android 5.0 operating system (98.6% of all students in the psychology participant pool had either an Android or iPhone that was a sufficiently new model that they were eligible to participate). To assess for moderation effects by social anxiety and depression severity, no cutoffs were used during recruitment, and social anxiety and depres- sion severity were examined on a continuous scale to allow for examination of the impact of symptom heterogeneity (see recommendations by U.S. Na- tional Institute of Mental Health Research Domain Criteria: Interim Guidance, Notice Number NOT- MH-11-005). Participants recruited via the Depart- ment of Psychology participant pool were reim- bursed 4 course credits, and participants recruited via advertisements were paid $40.

materials

Baseline Self-Reported Measures Social Anxiety Severity. The Social Interaction

Anxiety Scale (SIAS; Mattick & Clarke, 1998) assesses distress in social situations and was administered at a baseline laboratory session. Participants rated the degree to which they agreed with 20 statements (e.g. “I have difficulty talking with other people”), from 0 (“not at all character- istic or true of me”) to 4 (“extremely characteristic or true of me”). SIAS was examined as a continuous variable (M = 29.17, SD = 8.84) to assess moderation effects of social anxiety severity. Approximately 16.6% of our sample would score at or above the mean of a sample diagnosed with

social anxiety disorder (M = 40.0; SD = 16.0; Mattick & Clarke, 1998), suggesting good repre- sentation of highly anxious individuals in the sample. Internal consistency of the SIAS in this sample was good (alpha = .75).

Depression Severity. The 7-item depression sub- scale from the Depression, Anxiety, and Stress Scales (DASS-21; Lovibond & Lovibond, 1995) assessed depression severity at baseline. Partici- pants rated the degree to which they agreed with each statement (e.g. “I felt that life was meaning- less”) from 0 (“did not apply to me at all”) to 3 (“applied to me very much or most of the time”). This measure was used as a continuous variable (M = 3.06, SD = 2.56) to assess moderation effects of depression severity. Approximately 7% of our sample would qualify as mildly depressed (scoring between 10 and 13), and none were moderately depressed (above 13), using established cutoffs (Lovibond & Lovibond, 1995). Internal consisten- cy of the depression subscale of DASS-21 in this sample was good (alpha = .84).

EMA Mobile Application The customized phone app (Sensus) was scheduled to prompt participants with up to 6 randomly timed surveys and 1 end-of-day survey each day during the 2-week study. For the current study, we used the ratings of state affect, social context, and end-of- day retrospective ratings of effectiveness within, and enjoyment of, the day’s social situations. All study procedures and data have been made publicly available. Please refer to Boukhechba, Daros, et al. (2018) to access the full list of measures (including all actively and passively sensed data from Sensus) that were administered as part of a larger study.

State Affect. Each day, participants provided up to 6 separate ratings for current positive affect (“How positive are you feeling?”) and negative affect (“How negative are you feeling?”) using a continuous sliding scale from 0 (“not at all”) to 100 (“very positive” or “very negative”, depending on the question). Sliding scales were always presented with the response toggle initially situated at the midpoint to reduce biased reporting.

Social Context. Participants responded to the question, “Who are you interacting with?”, during each randomly timed survey to identify those prompts that occurred during a social interaction. Participants were asked to select all options that applied to their current situation: I am alone, with a family member, classmate/coworker, romantic partner/close friend, acquaintance/stranger, other (with optional input).

871percept ions of soc i al int eract ions

End-of-Day Retrospective Ratings of Effective- ness. Each night, participants responded to the question, “Overall, how effectively did you interact with others today?”, using a continuous sliding scale from 0 (“not at all effectively”) to 100 (“really effectively”).

End-of-Day Retrospective Ratings of Enjoy- ment. Each night, participants responded to the question, “Overall, how much did you enjoy or dislike your social interactions today?”, using a continuous sliding scale from 0 (“really disliked”) to 100 (“really enjoyed”).

PROCEDURE

Participants were told that the study examined thoughts and feelings as people interacted in their daily environment. They were eligible to participate if they had access to an iPhone 5 or later iOS model, or an Android mobile phone that had the Android 5.0 or later mobile operating system. Participants attended two laboratory sessions, approximately 2 weeks apart, each lasting approximately 60 mi- nutes. During the first laboratory session, partici- pants provided informed consent and completed measures of social anxiety and depression severity. At the end of this baseline session, research staff installed Sensus on participants’ personal smart- phones. Sensus prompted participants up to 6 times per day between 9:00 A.M. and 9:00 P.M., with each prompt randomly firing once within each 2-hour window (i.e., once between 9:00 A.M. and 11:00 A. M., once between 11:00 A.M. and 1:00 P.M., etc.). Prompts remained open for participants to respond to until the next prompt was fired. Sensus also prompted participants once at the end of each day (at 10:00 P.M.) to answer a short questionnaire about their retrospective perceived overall effec- tiveness and enjoyment of their social interactions that day. Prompts fired at the end of the day remained open until answered, or closed automat- ically at 2:00 A.M.the next morning. After the 2- week EMA period, participants returned for a second laboratory session, which included mea- sures not directly related to the current hypotheses, and were then fully debriefed.

Plan for Analyses All analyses were conducted using linear mixed- effects models that can account for changes over time and are more robust against missing data, compared to repeated-measures analysis of vari- ance (Nich & Carroll, 1997), and allowed for modeling of within-person and within-day variabil- ity. The models were fitted using the “lme4” package in the R statistical analysis environment

(Bates, Maechler, Bolker, & Walker, 2014). Given that it is difficult to obtain standard effect sizes for mixed-effects models, we report standardized beta- weights. Due to software bugs, an average of 2.5 prompts

(42% of the expected 6) were responded to at random times during the day per participant, and approximately 8 end-of-day prompts (57% of the expected 14) were responded to by each participant over the course of the study. Given that we are interested in affective experiences during and retrospective evaluations after social interactions, only random time responses in which participants indicated they were “not alone” were considered for inclusion in the current analysis. To be included, all qualifying random time responses must have had the corresponding end-of-day response completed on that same day. Across all participants, 45% of all responded prompts occurred when a participant reported not being alone. Each participant contrib- uted an average of 1.7 random time prompts per day in which they reported not being alone (28% of the expected 6). Overall, each participant contrib- uted an average of 6.9 days (49% of the expected 14) of qualifying data over the entire study. Participants yielded a total of 732 observations from pairs of random time and end of day prompts across all days. To examine how the relationship between affect

experienced during daily social interactions and end-of-day retrospective ratings of effectiveness during, and enjoyment of, those social interactions might vary at different levels of social anxiety or depression severity, separate mixed-effects models were computed with end-of-day retrospective re- ports of effectiveness and of enjoyment as criterion variables. The predictor variables in the models were either negative or positive affect experienced during social interactions on a particular day, and either SIAS or DASS-D (as measures of social anxiety or depression severity). Interactions be- tween predictor variables (affect and severity of either social anxiety or depression, depending on the analysis) were interpreted with SIAS or DASS-D as the moderator. All predictors (i.e., negative and positive affect, SIAS, DASS-D) in our models were analyzed as continuous variables and were stan- dardized before they were entered into the models. In all models, Subject was entered as a random intercept to account for differences in mean responses for each individual, and Day was also entered as a random intercept to account for mean differences across different days of the study. Outliers were identified as residuals that were greater than 2.5 standard deviations from the estimated model and were removed. Due to the

872 geyer et al .

limited number of random prompts per day (average of 1.7) during which participants reported not being alone, the models examined how aggregated negative or positive affect on a day would predict end-of-day ratings of effectiveness or enjoyment, and how this effect is moderated by social anxiety or depression severity. There were not sufficient data points to examine data at the trial by trial level.

Results Mean, standard deviation, maximum, and mini- mum values of SIAS, DASS-D, effectiveness, enjoy- ment, negative affect, and positive affect ratings, and correlations between the measures are present- ed in Table 1. Male versus female participants did not show

significant differences in effectiveness (Standardized B = - .14, p = .11), enjoyment (Standardized B = -.017, p = .85), negative affect (Standardized B = -.14, p = .36), and positive affect ratings (Standard- ized B = -.009, p = .96). When in social situations (vs. alone), on average across the full sample, participants reported significantly higher positive affect (Standardized B = .23, p = .003) and lower negative affect (standardized B = -.12, p = .008). The interactions between social situation (alone vs. not alone) and severity of depression or social anxiety in predicting positive or negative affect were not significant (all p N .05). That is, there were no significant changes in the relationship between severity of depression or social anxiety and positive or negative affect as a function of social situation. Finally, as expected, severity of social anxiety was significantly negatively associated with positive affect (standardized B = -.17, p = .02) and significantly positively associated with negative affect (standardized B = .23, p = .001). Likewise, severity of depression was significantly negatively associated with negative affect (standardized B =

Table 1 Descriptive Statistics and Correlations for Anxiety and Depressio Participants

Descriptive Statistics

Measure Mean SD Median Max Min

SIAS 29.17 8.84 29.00 52.00 11.0 DASS-D 3.06 2.56 3.00 12.00 0.00 NA 29.57 20.86 25.00 100.00 1.00 PA 66.42 21.40 69.00 100.00 1.00 Eff 68.69 20.27 71.00 100.00 9.00 Enj 69.67 20.28 72.00 100.00 1.00

Note. SIAS = Social Interaction Anxiety Scale; DASS-D = Depression, Affect (in-the-moment); PA = Positive Affect (in-the-moment); Eff = Effe

.14, p = .04). Contrary to expectations, severity of depression was not significantly associated with positive affect (standardized B = -.07, p = .35).

predicting retrospective perceived

effectiveness and enjoyment of so-

cial interactions

Interestingly, there was no reliable relationship between mean end-of-day ratings of perceived effectiveness or enjoyment and the measures of social anxiety or depression severity (Mean Effec- tiveness and Social Anxiety: Standardized B = -.096, p = .33; Mean Effectiveness and Depression: Standardized B = -.16, p = .09; Mean Enjoyment and Social Anxiety: Standardized B = -.061, p = .57; Mean Enjoyment and Depression: Standardized B = -.045, p = .65). We then checked if this was due to lack of variability (i.e., the variance, computed as the squared-deviation of the variable from its mean, of an individual’s end-of-day ratings of effectiveness or enjoyment over the course of the study with each variable measured on a 0–100 scale) in more socially anxious or depressed individuals’ (i.e., in a subset of the sample containing individuals who scored more than 1 SD above full sample mean on either SIAS or DASS-D) end-of day ratings, but there was no reliable relationship between the extent of variability across the end-of-day ratings of enjoyment or effectiveness and the measures of social anxiety or depression severity (Variability of Effectiveness ratings and Social Anxiety: Standard- ized B = .11, p = .37; Variability of Effectiveness ratings and Depression: Standardized B = .12, p = .33; Variability of Enjoyment ratings and Social Anxiety: Standardized B = .11, p = .37; Variability of Enjoyment ratings and Depression: Standardized B = -.021, p = .87), indicating that the subset of high socially anxious or depressed individuals did not exhibit an unusually restricted response style.

n Severity, Affect, Effectiveness, and Enjoyment Aross All

Correlations

SIAS DASS-D NA PA Eff Enj

0 - .58 - .24 .11 - -.22 -.04 -.77 - -.25 -.19 -.41 .45 - -.21 -.06 -.38 .44 .80 -

Anxiety, and Stress Scales-Depression subscale; NA = Negative ctiveness (end-of-day); Enj = Enjoyment (end-of-day)

873percept ions of soc i al int eract ions

EFFECTIVENESS OF SOCIAL INTERACTIONS

As expected, there were main effects of negative and positive affect reported during social interactions on retrospective, end-of-day ratings of perceived effectiveness during social interactions. Specifically, there was a significant negative relationship be- tween negative affect reported during the day’s social interactions and end-of-day ratings of effec- tiveness (Standardized B = -.23; p b .001) and a significant positive relationship between positive affect and end-of-day ratings of effectiveness (Standardized B = .20; p b .001). Contrary to hypotheses, there were no significant interactions between negative or positive affect and SIAS in predicting end-of-day ratings of effectiveness (neg- ative affect: Standardized B = -.06; p = .10; positive affect: Standardized B = .03; p = .38), meaning that severity of social anxiety did not moderate the extent to which daily affect predicted end-of-day perceived effectiveness. However, in line with hypotheses, significant interactions between affect (both negative and positive) and DASS-D were observed. Specifically, the relationship between negative affect experienced during social interac- tions throughout the day and end-of-day reports of perceived effectiveness was more negative when

FIGURE 1 The relationship between negative throughout the day and end-of-day reports of perc more negative when depression was more severe. were analyzed as continuous variables; however, fo are plotted for depression severity more than 1 Depression, Anxiety, and Stress Scales-Depression

depression was more severe (Standardized B = -0.12, p = .002; see Figure 1). Analogously, when depression was more severe, the negative impact on the relationship between lack of positive affect and end-of-day ratings of effectiveness was greater (Standardized B = .13, p = .001), indicating that severity of depression moderated the extent to which the absence of positive affect during social interactions negatively predicted retrospective per- ceptions of effectiveness during these interactions. All models were repeated while controlling for either DASS-D (when SIAS was the moderator) or SIAS (when DASS-D was the moderator). All patterns and significant interactions held after entering the control variables. Taken together, results indicate that the negative relationship between negative affect and perceived effectiveness (and likewise, the negative relationship between lack of positive affect and perceived effectiveness) was stronger when depression was more severe, but was unchanged across all levels of social anxiety severity.

ENJOYMENT OF SOCIAL INTERACTIONS

As expected, there was a significant negative relationship between negative affect experienced

affect experienced during social interactions eived effectiveness (but not enjoyment) was Note: Negative affect and depressive severity r purposes of illustration, the regression lines SD above and below the mean; DASS-D: subscale.

874 geyer et al .

during daily social interactions and overall rating of enjoyment at the end of that day (Standardized B = -.20, p b .001). Likewise, there was a significant positive relationship between positive affect reports during a day’s social interactions and retrospective rating of enjoyment at the end of the day (Standardized B = .20, p b .001). Consistent with hypotheses, a significant interaction was observed between negative affect and SIAS in predicting end- of-day ratings of enjoyment. Specifically, the negative relationship between negative affect re- ported during social interactions and end-of-day ratings of enjoyment was stronger when social anxiety was more severe (Standardized B = -.08, p = 0.03; see Figure 2). No significant interaction was observed between positive affect and SIAS in predicting ratings of enjoyment (Standardized B = .027; p = .46). Contrary to hypotheses, no significant interactions were observed between negative or positive affect and DASS-D in predict- ing end-of-day ratings of enjoyment (negative affect: Standardized B = -.047; p = .24; positive affect: Standardized B = .05; p = .23). Like before, all patterns and significant interactions held after controlling for either DASS-D or SIAS, depending on the analysis. Taken together, results indicate that

FIGURE 2 The negative relationship betwee interactions and end-of-day ratings of enjoyment social anxiety was more severe. Note: Negative aff as continuous variables; however, for purposes of i social anxiety severity more than 1 SD above an Anxiety Scale.

more negative (but not less positive) affect experi- enced during social interactions was a stronger predictor of retrospective reports of lower perceived enjoyment when social anxiety was more severe, but was unchanged across all levels of depression severity.

Discussion This EMA study used self-reported affect during social interactions to predict participants’ later overall ratings of their effectiveness during, and enjoyment of, that day’s interactions. As expected, across all levels of social anxiety and depression severity, negative (and lack of positive) affective experiences during social interactions were signif- icantly negatively associated with end-of-day, retrospective perceptions of effectiveness and en- joyment. Interestingly, severity of social anxiety and depression moderated the association between in- the-moment affect and retrospective ratings of effectiveness and enjoyment differently. In particu- lar, as expected, in-the-moment negative affect in social situations was especially negatively predictive of retrospective reports of enjoyment of that day’s social interactions when social anxiety was more severe, but contrary to expectations, not when

n negative affect reported during social (but not effectiveness) was stronger when ect and social anxiety severity were analyzed llustration, the regression lines are plotted for d below the mean; SIAS: Social Interaction

875percept ions of soc i al int eract ions

depression was more severe. In contrast, in-the- moment negative (and lack of positive affect) in social situations, were especially negatively predic- tive of retrospective reports of effectiveness during that day’s social interactions when depression was more severe, but contrary to expectations, not when social anxiety was more severe. It is also notable that severity of social anxiety and depression did not predict daily retrospective ratings of effective- ness and enjoyment overall. Instead, the observed effects of social anxiety and depression severity predicting more negative effects of affect on retrospective social interaction perceptions was especially strong under conditions of relatively higher negative affect (and, in the case of depres- sion, lower positive affect). These findings suggest that negative retrospective self-evaluations of affect and performance during social interactions are not pervasive across all dimensions of affect, but occur mainly when negative affect is activated. Thus, biased beliefs about performance in social interac- tions may derive mostly from situations perceived to be stressful (and/or nonrewarding, in the case of depression), especially when social anxiety or depression is more severe.

differential prediction of effective-

ness and enjoyment by social anxiety

and depression severity

The negative relationship between individuals’ perception of how much they enjoyed the day’s interactions and real-time negative affect was stronger when social anxiety was more severe, while the negative relationship between how effectively individuals believed they performed during that day’s interactions and real-time nega- tive affect was stronger when depression was more severe. Thus, the nature of negative memories predicted by affective experiences may differ (i.e., mastery/effectiveness vs. pleasure/enjoyment) de- pending on the severity of social anxiety or depression that an individual experiences. This distinct pattern for the moderation effects of social anxiety severity vs. depression severity was not predicted and needs to be replicated, but assuming it is a reliable finding, it raises intriguing questions about why negative judgments were ‘spared’ for effectiveness when social anxiety was more severe and for enjoyment when depression was more severe. High socially anxious individuals experience

deficits in reward processing specific to social situations (Richey et al., 2013) rather than in general, so the negative affective experiences in social situations may stand out and the lack of enjoyment will be preferentially remembered

(hence, the stronger predictive relationship for in- the-moment affect and later reports of enjoyment when social anxiety was more severe). In contrast, due to more pervasive anhedonia for depressed individuals (e.g., Rizvi, Pizzagalli, Sproule, & Kennedy, 2016), negative affective experiences in social situations may stand out less as a distinctive cue that a situation was unenjoyable, so the predictive relationship is weaker when depression was more severe. Depressed individuals’ tendency to report overgeneral memories may also be a factor in that the particular variations in affect may not be well recalled. The finding that severity of depression is associ-

ated with a stronger link between in-the-moment affect and more negative perceptions of effective- ness during social interactions was expected and is not surprising given that lower self-mastery beliefs have previously been associated with more severe depression (Marshall & Lang, 1990). What is harder to explain is why this same moderation effect was not evident for social anxiety severity, given prior studies have found that positive mastery beliefs, including perceptions of social competence, are generally associated with lower anxiety (Felsten & Wilcox, 1992). One speculation is that individ- uals with more severe social anxiety expect to feel miserable during social interactions in which they fear evaluation; so, affective information does not provide unique information about whether they performed effectively or not. Perhaps for individ- uals with more severe social anxiety, other cues are considered more useful in determining effectiveness (e.g., behaviors exhibited by the interaction partner, or the anxious individual’s blushing or stuttering speech). It will be interesting in future work to determine what in-the-moment indicators (other than affect) are especially predictive of retrospective ratings of social effectiveness. It is also important to consider the strong

correlation observed between effectiveness and enjoyment (.8, as presented in Table 1). In many ways, the strong relationship was expected given the similarity between the constructs, in that both are associated with an individual’s perception of how well a social interaction went. As laid out in the hypotheses section, we expected similar effects of social anxiety and depression severity on effectiveness and enjoyment. It was thus surprising that results indicated that social anxiety and depression severity differentially shift the impact of positive or negative affect on effectiveness and enjoyment. These differences highlight the impor- tance of taking a more nuanced approach to examining how well an individual believes a social interaction went, and making sure precise language

876 geyer et al .

is used in clinical assessments and interventions. Given the high correlation between effectiveness and enjoyment, yet unique findings for each construct, it will be interesting in future work to determine what about the nonshared variance between effectiveness and enjoyment may be contributing to their unique findings. Adding additional measures tied to competence and plea- sure may help in this regard. Clearly, these explanations for the different

moderation findings are speculative and post hoc, so it will be important to both replicate the current findings and directly test the different possible explanations. Nonetheless, these results clearly point to the value of directly assessing and examining both perceptions of effectiveness and enjoyment. For instance, it would be interesting to ask for descriptive reports of the day’s social interactions at the end of the day to see whether the memories are more general for those with more severe depression, and whether this overgeneral effect can account for the different moderation findings. Analogously, assessing for deficits in social vs. nonsocial reward processing may help explain the discrepant findings across symptom domains. In turn, it would be helpful to track additional in-the-moment social interaction indica- tors, such as heart rate, skin conductance, or shaking/jittering, that may more strongly predict effectiveness ratings for those with more severe social anxiety. An additional open question concerns when low

positive affect (vs. only high negative affect) will interact with severity of social anxiety or depression to predict negative social memories. Different possible relationships were laid out regarding this question because it was not clear whether, due to the importance of social interactions for individuals experiencing more severe social anxiety or depres- sion, in-the-moment low positive affect during social interactions would more strongly predict retrospective ratings, or whether greater concerns specific to negative interactions would mean positive affect would not be especially predictive of later social memories. In the current study, low positive affect did interact with severity of depres- sion—but not severity of social anxiety—to predict later, more negative social interaction perceptions of effectiveness. The presence of anhedonia in depression may be one reason for this particular association between positive affect and severity of depression. However, there is some evidence that social anxiety is also associated with low positive affect (Kashdan, 2007), so this explanation is not wholly satisfying. It is also possible that socially anxious individuals’ fear of even positive evalua-

tions (Weeks, Heimberg, Rodebaugh, & Norton, 2008) could mean that positive affect provides mixed signals, in that it both indicates that one is enjoying the experience and feels it is going well, but also that a positive evaluation may be occurring, which can be frightening. Again, it will be interesting in future work to more comprehen- sively assess anhedonia and fears of positive evaluation to determine if these factors can help explain the unique role of low positive affect in the formation of negative social memories.

ema and sampling considerations

Researchers seek to capture in-the-moment experi- ences as accurately and as frequently as possible through the use of both active and passive sensing, and progress in that direction is being made in the fields of psychology, computer science, and engi- neering. Yet, there continues to be methodological challenges in EMA research to achieve high response rates and balance the desire for more complete information with the need to reduce participant burden. Due to technical issues and imperfect participant compliance, the average sampling rate in the current study was 2–3 random-time prompts per day (out of a possible 6). Therefore, the obtained “snapshots” of partic- ipants’ daily experiences in the current study are likely to provide a biased sample of daily life. That said, it is recommended that the maximum sam- pling rate for EMA prompts does not exceed 5–6 surveys per day (Burke et al., 2017), so, even with perfect compliance, EMA data are selected snap- shots. We believe that while clear limitations in our data exist, especially concerning the likely nonran- dom nature of at least some of the missed prompts, the data still provide valuable information about participant experiences and in far more contexts than can typically be obtained from studies conducted solely in the laboratory. For instance, we obtained a good sampling of situations in which participants were alone (55% of prompts) vs. not alone (45% of prompts), suggesting that the current sampling rate did not exclude social situations (when responding can be more challenging). This increases confidence in the validity of the current findings. That said, it is important that researchers continue to examine the conditions under which participant compliance in EMA studies can be increased, such as factors that impact the amount of time participants take to respond to prompts (see Boukhechba, Cai, et al., 2018). This work can provide insight into what participants might be doing when they are not responding to prompts and when different participants might choose to re- spond to prompts, which may ultimately help

877percept ions of soc i al int eract ions

researchers determine the meaning of missing data in EMA studies. Additionally, it is important to consider that the

sample used in this study is composed of college undergraduates, and that the study was conducted during an academic semester. As such, it is foreseeable that the majority of social interactions took place in a more structured environment (e.g., students will likely be interacting in their classes), which may reduce some of the typical ambiguity about the social performance expectations that can add to fears of negative evaluation. It would be interesting to learn if and how the results differ if the study were conducted in a less structured environment or with a sample at a different developmental stage.

clinical implications

The current study used a nonclinical sample in which severity of social anxiety and depression were considered on a continuum, which might limit the generalizability of our results to a clinical sample. Nonetheless, this design allowed us to effectively examine moderation effects of social anxiety and depression severity, and also highlight- ed some unique potential treatment emphases. This is intriguing given social anxiety and depression symptoms are highly comorbid, and recent cogni- tive-behavioral treatments targeting both psycho- pathologies simultaneously can be efficacious (Hofmann, Asnaani, Vonk, Sawyer, & Fang, 2012). Yet, memories of those who have more severe social anxiety versus depression appear to be differentially influenced by the valence and intensity of affect experienced during social interactions. Our results suggest that it will be especially important to help prevent socially anxious individuals from exaggerating how miserable they were during a social interaction, and to try to focus their memories more on the presence of positive cues and less on the experienced negative affect. One approach may be to normalize and encourage nonjudgmental acceptance of in-the-moment nega- tive affect to make it less salient, such as by telling clients they should assume they will feel negatively during social interactions for a while until their social anxiety symptoms reduce, so instead they should focus on evidence of their ability to stay in the situation despite feeling negatively. For more depressed individuals, it will be key to break the strong association between feeling high negative affect and low positive affect in-the-moment with later assumptions of low effectiveness during the interactions; for instance, examining evidence that makes it clear that they can be competent and achieve their goals despite the unpleasant emotions

they experience. Common across the suggested approaches for social anxiety and depression is the recognition that in-the-moment affect does not need to be judged as a paramount or inevitable determinant of later memories for the interaction. Integrating many sources of information to form a later evaluation may help to reduce a focal bias on the affective cues. We expect that traditional cognitive behavioral approaches, as well as more recent acceptance and mindfulness approaches (Hofmann, Sawyer, Witt, & Oh, 2010) and emotion-focused techniques (Greenberg, 2004) could all be helpful in this regard. In addition, being able to identify contributors to

negative memories in real-time means we can then think about targeting these events when they occur instead of waiting for clients to come to the therapist’s office after extensive, maladaptive post-event process- ing has occurred. Just-In-Time interventions that target stressful events is still in its infancy (Nahum- Shani, Hekler, & Spruijt-Metz, 2015) but raises exciting possibilities to develop highly targeted, personalized interventions that “catch” early signs of developing negative biases before they become deeply instantiated and more resistant to change.

limitations and conclusion

It is important that these results be considered in the context of the study’s limitations. First, the current analyses focused only on in-the-moment affect as a predictor, and there are many other contextual factors that could be predictive of later negative social memories. For example, retrospective effec- tiveness ratings may vary by whether socially anxious individuals were with a close other vs. a group of strangers, while this factor may be less central for depressed individuals. Future work could examine additional features of the social context (e.g., location, number of people present, extent of verbal communication, etc.) to obtain greater clarity regarding which features will predict retrospective effectiveness and enjoyment ratings and for which populations (socially anxious, depressed, or other groups that are prone to biased recall, such as persons with anger management issues). Second, we worked with a nonclinical sample of

college students that endorsed broad ranges of social anxiety and depression symptom severity. This was intentional and we believe still clinically valuable, given subthreshold symptoms are often associated with significant impairments in func- tional domains (e.g., Karsten et al., 2010; Stein, Torgrud, & Walker, 2000) and these impairments increase continuously as symptom severity in- creases (e.g., Lewinsohn et al., 2000). Our

878 geyer et al .

approach thus means findings may be applicable to a larger group of people (i.e., not only diagnosed individuals). That said, there was a good represen- tation of high symptomatic individuals for social anxiety in our sample (though less so for depres- sion), and we recognize the importance of investi- gating and replicating the effects at the extreme ends of the spectrum. Thus, it will be important for future studies to replicate these findings with clinically diagnosed, community samples (though working with college students was intentional, we recognize that college students’ daily experiences can be very different from those of most of the population) and examine samples with comorbid social anxiety and depression. Furthermore, it is notable that our sample exhibited higher mean levels of social anxiety severity compared to reported undergraduate means in other samples (e.g.,M = 19.0; SD = 10.1; as reported inMattick& Clarke, 1998). Though higher than the levels reported by Mattick and Clarke in the late 1990s, given the known cohort increases in anxiety levels on college campuses over the past few decades (Center for Collegiate Mental Health, 2016), it is possible that our sample’s mean SIAS score is not particularly elevated relative to current levels of social anxiety across college samples. It would be helpful to have updated norms on these measures to better reflect possible cohort and history effects. Third, due to technical difficulties that limited the

number of random prompts we obtained per day, our analyses considered the association between aggre- gated levels of negative and positive affect across social interactions on a given day and retrospective effectiveness and enjoyment ratings for that day. We are currently conducting a follow-up, larger scale study that will more reliably administer the full 6 random time prompts each day. With this richer dataset, an interesting direction will be to examine the effects of having variable affective experiences throughout a day; for instance, if having one highly negative social experience and fivemoderatelypositive experiences on a day has a different impact on perceptions of effectiveness and enjoyment than having five moderately negative experiences and one highly positive experience that day. Fourth, though the active sensing questionnaires

administered throughout the day were able to capture ratings of affect and interactions close in time to the events, given limits of current technol- ogy, we recognize that these ratings are still subject to some recall biases, given participants could foreseeably respond to questionnaires after an interaction event (instead of interrupting their interactions to answer the questionnaires). It is thus important for research to explore reliable

markers from passively collected data (e.g., using accelerometers, gyroscopes, heart rate monitors) that could track state affect and behavior during interactions as these interactions unfold. Fifth, participant noncompliance during the study

was inflated by technical issues that resulted in some EMA prompts not being delivered at scheduled times. This was primarily caused by our use of a native app to deliver EMA prompts. As a result, the current sample size is relatively small. However, we were able to use 732 distinct observations (i.e., pairs of random time and end of day prompts) in our analyses. Researchers have long documented the obstacles to conducting EMA studies (Burke et al., 2017), which include both participant noncompli- ance as well as technical issues. This highlights the importance of selecting and developing reliable methods of EMA delivery for future studies, as well as addressing participant compliance and attrition rates. To start addressing these issues, in a current study, we have implemented weekly compliance monitoring protocols and provide a sliding scale compensation schedule to participants based on their individual completion rate. Finally, despite the wide prevalence of mobile

phones, a meaningful proportion of the population still does not own a smartphone. Thus, EMA studies might not capture the behaviors of this particular subset of the population, which tend to be older (N65 years old), less highly educated, and/ or have lower incomes (Pew Research Center, 2018). Despite this, a large proportion of these demographics do use smartphones, so future studies could focus on oversampling, or do targeted recruiting of, individuals from these demographics. Despite these limitations, the study provided the

first examination of how in-the-moment affective experiences predict retrospective perceptions of performance and enjoyment of social interactions, and how these predictions vary by severity of social anxiety and depression. The study further high- lights the value of EMA methodologies in elucidat- ing important real-time characteristics of psychopathologies that can ultimately allow for the tailoring of treatments to address the idiograph- ic concerns and negative memories of individuals experiencing social anxiety and depression.

Conflict of Interest Statement The authors declare that there are no conflicts of interest.

References

Amir, N., Beard, C., & Bower, E. (2005). Interpretation bias and social anxiety. Cognitive Therapy and Research, 29(4), 433–443. https://doi.org/10.1007/s10608-005-2834-5

879percept ions of soc i al int eract ions

Amir, N., Bower, E., Briks, J., & Freshman, M. (2003). Implicit memory for negative and positive social information in individuals with and without social anxiety. Cognition & Emotion, 17(4), 567–583. https://doi.org/10.1080/ 02699930302300

Barry, E. S., Naus, M. J., & Rehm, L. P. (2004). Depression and implicit memory: Understanding mood congruent memory bias. Cognitive Therapy and Research, 28(3), 387–414. https://doi.org/10.1023/B:COTR.0000031808.00502.2e

Bates, D., Maechler, M., Bolker, B., &Walker, S. (2014). lme4: Linear mixed-effects models using Eigen and S4. R package version 1.0 – 6. Retrieved from. http://CRAN.R-project.org/ packagelme4.

Ben-Zeev, D., Young, M. A., & Madsen, J. W. (2009). Retrospective recall of affect in clinically depressed individ- uals and controls. Cognition and Emotion, 23(5), 1021–1040. https://doi.org/10.1080/02699930802607937

Boukhechba, M., Daros, A., Chow, P., Fua, K., Teachman, B., & Barnes, L. (2018, March 15). DemonicSalmon. https:// doi.org/10.17605/OSF.IO/WDUK6

Boukhechba, M., Cai, L., Chow, P. I., Fua, K., Gerber, M. S., Teachman, B. A., & Barnes, L. E. (2018). Contextual Analysis to Understand Compliance with Smartphone-based Ecological Momentary Assessment. Proceedings of the 2018 ACM PervasiveHealth. https://doi.org/10.475/123_4

Brozovich, F., & Heimberg, R. G. (2008). An analysis of post- event processing in social anxiety disorder. Clinical Psy- chology Review, 28(6), 891–903. https://doi.org/10.1016/j. cpr.2008.01.002

Burke, L. E., Shiffman, S., Music, E., Styn, M. A., Kriska, A., Smailagic, A., & Rathbun, S. L. (2017). Ecological momentary assessment in behavioral research: addressing technological and human participant challenges.Journal of Medical Internet Research, 19(3)e77 https://dx.doi.org/10. 2196%2Fjmir.7138

Cacioppo, J. T., & Berntson, G. G. (1994). Relationship between attitudes and evaluative space: A critical review, with emphasis on the separability of positive and negative substrates. Psychological Bulletin, 115(3), 401–423. https:// doi.org/10.1037//0033-2909.115.3.401

Center for Collegiate Mental Health (2016). 2015 Annual Report. Retrieved from. https://ccmh.psu.edu/files/2017/10/ 2015_CCMH_Report_1-18-2015-yq3vik.pdf.

Clore, G. L., Gasper, K., & Garvin, E. (2001). Affect as information. In J. P. Forgas (Ed.), Handbook of affect and social cognition (pp. 121–144). Mahwah, NJ: Lawrence Erlbaum Associates.

Dalgleish, T., & Watts, F. N. (1990). Biases of attention and memory in disorders of anxiety and depression. Clinical Psychology Review, 10(5), 589–604. https://doi.org/10. 1016/0272-7358(90)90098-U

Dimidjian, S., Martell, C. R., Herman-Dunn, R., & Hubley, S. (2014). Behavioral activation for depression. In D. Barlow (Ed.), Clinical handbook of psychological disorders (pp. 353–393). (5th ed.). New York, NY: Guilford Press.

Duque, A., & Vázquez, C. (2015). Double attention bias for positive and negative emotional faces in clinical depression: Evidence from an eye-tracking study. Journal of Behavior Therapy and Experimental Psychiatry, 46, 107–114. https:// doi.org/10.1016/j.jbtep.2014.09.005

Edwards, S. L., Rapee, R. M., & Franklin, J. (2003). Postevent rumination and recall bias for a social performance event in high and low socially anxious individuals. Cognitive Therapy and Research, 27(6), 603–617. https://doi.org/10. 1023/A:1026395526858

Felsten, G., & Wilcox, K. (1992). Influences of stress and situation-specific mastery beliefs and satisfaction with social

support on well-being and academic performance. Psycho- logical Reports, 70(1), 291–303. https://doi.org/10.2466/ pr0.1992.70.1.291

Fergusson, D. M., Horwood, L. J., Ridder, E. M., & Beautrais, A. L. (2005). Subthreshold depression in adolescence and mental health outcomes in adulthood. Archives of General Psychiatry, 62(1), 66–72. https://doi.org/10.1001/archpsyc. 62.1.66

Gable, S. L., & Shean, G. D. (2000). Perceived social competence and depression. Journal of Social and Personal Relationships, 17(1), 139–150. https://doi.org/10.1177/ 0265407500171007

Gotlib, I. H., Jonides, J., Buschkuehl, M., & Joormann, J. (2011). Memory for affectively valenced and neutral stimuli in depression: evidence from a novel matching task. Cognition and Emotion, 25(7), 1246–1254. https://doi. org/10.1080/02699931.2010.538374

Greenberg, L. S. (2004). Emotion–focused therapy. Clinical Psychology & Psychotherapy, 11(1), 3–16. https://doi.org/ 10.1002/cpp.388

Heinrichs, N., & Hofmann, S. G. (2001). Information processing in social phobia: A critical review. Clinical Psychology Review, 21(5), 751–770. https://doi.org/10. 1016/S0272-7358(00)00067-2

Hofmann, S. G. (2007). Cognitive factors that maintain social anxiety disorder: A comprehensive model and its treatment implications. Cognitive Behaviour Therapy, 36 ( 4 ) , 1 9 3 – 2 0 9 . h t t p s : / / d o i . o r g / 1 0 . 1 0 8 0 / 16506070701421313

Hofmann, S. G., Asnaani, A., Vonk, I. J. J., Sawyer, A. T., & Fang, A. (2012). The efficacy of cognitive behavioral therapy: A review of meta-analyses. Cognitive Therapy and Research, 36(5), 427–440. https://doi.org/10.1007/ s10608-012-9476-1

Hofmann, S. G., Sawyer, A. T., Witt, A. A., & Oh, D. (2010). The effect of mindfulness-based therapy on anxiety and depression: A meta-analytic review. Journal of Consulting and Clinical Psychology, 78(2), 169–183. https://doi.org/10. 1037/a0018555

Karsten, J., Hartman, C. A., Ormel, J., Nolen, W. A., & Penninx, B. W. J. H. (2010). Subthreshold depression based on functional impairment better defined by symptom severity than by number of DSM-IV symptoms. Journal of Affective disorders, 123(1), 230–237. https://doi.org/10. 1016/j.jad.2009.10.013

Kashdan, T. B. (2007). Social anxiety spectrum and diminished positive experiences: Theoretical synthesis and meta-analy- sis. Clinical Psychology Review, 27(3), 348–365. https://doi. org/10.1016/j.cpr.2006.12.003

Kashdan, T. B., & Collins, R. L. (2010). Social anxiety and the experience of positive emotion and anger in everyday life: an ecological momentary assessment approach.Anxiety, Stress, and Coping, 23(3), 259–272. https://doi.org/10.1080/ 10615800802641950

Kashdan, T. B., & Roberts, J. E. (2007). Social anxiety, depressive symptoms, and post-event rumination: Affective consequences and social contextual influences. Journal of Anxiety Disorders, 21(3), 284–301. https://doi.org/10. 1016/j.janxdis.2006.05.009

Kessler, R. C., Berglund, P., Demler, O., Jin, R., Merikangas, K. R., & Walters, E. E. (2005). Lifetime prevalence and age-of- onset distributions of DSM-IV disorders in the National Comorbidity Survey Replication. Archives of General Psychiatry, 62(6), 593–602. https://doi.org/10.1001/ archpsyc.62.6.593

Lewinsohn, P. M., Solomon, A., Seeley, J. R., & Zeiss, A. (2000). Clinical implications of “subthreshold” depressive

880 geyer et al .

symptoms. Journal of Abnormal Psychology, 109(2), 345–351.

Levens, S. M., & Gotlib, I. H. (2010). Updating positive and negative stimuli in working memory in depression.Journal of Experimental Psychology: General, 139(4), 654–664. https://dx.doi.org/10.1037%2Fa0020283

Lovibond, P. F., & Lovibond, S. H. (1995). The structure of negative emotional states: Comparison of the Depression Anxiety Stress Scales (DASS) with the Beck Depression and Anxiety Inventories. Behaviour Research and Therapy, 33(3), 335–343. https://doi.org/10.1016/0005-7967(94)00075-U

MacLeod, C., & Mathews, A. (2004). Selective memory effects in anxiety disorders: An overview of research findings and their implications. In D. Reisberg, & P. Hartel (Eds.), Memory and Emotion (pp. 155–185). New York: Oxford University Press.

Mansell, W., Ehlers, A., Clark, D. M., & Chen, Y. -P. (2002). Attention to positive and negative social-evaluative words: Investigating the effects of social anxiety, trait anxiety, and social threat. Anxiety, Stress, and Coping: An International Journal, 15(1), 19–29. https://doi.org/10.1080/10615800290007263

Marshall, G. N., & Lang, E. L. (1990). Optimism, self-mastery, and symptoms of depression in women professionals. Journal of Personality and Social Psychology, 59(1), 132–139. https://doi.org/10.1037/0022-3514.59.1.132

Mattick, R. P.,&Clarke, J. C. (1998). Development and validation of measures of social phobia scrutiny fear and social interaction anxiety. Behaviour Research and Therapy, 36(4), 455–470. https://doi.org/10.1016/S0005-7967(97)10031-6

McGirr, A., Renaud, J., Seguin, M., Alda, M., Benkelfat, C., Lesage, A., & Turecki, G. (2007). An examination of DSM- IV depressive symptoms and risk for suicide completion in major depressive disorder: a psychological autopsy study. Journal of Affective Disorders, 97(1), 203–209. https://doi. org/10.1016/j.jad.2006.06.016

Mellings, T. M., & Alden, L. E. (2000). Cognitive processes in social anxiety: The effects of self-focus, rumination and anticipatory processing. Behaviour Research and Therapy, 38 (3), 243–257. https://doi.org/10.1016/S0005-7967(99)00040-6

Mitte, K. (2008). Memory bias for threatening information in anxiety and anxiety disorders: A meta-analytic review. Psychological Bulletin, 134, 886–911. https://doi.org/10. 1037/a0013343

Moscovitch, D. A. (2009).What is the core fear in social phobia? A new model to facilitate individualized case conceptualization and treatment. Cognitive and Behavioral Practice, 16(2), 123–134. https://doi.org/10.1016/j.cbpra.2008.04.002

Nahum-Shani, I.,Hekler, E.B.,&Spruijt-Metz,D. (2015). Building health behaviormodels to guide the development of just-in-time adaptive interventions: A pragmatic framework. Health Psy- chology, 34, 1209–1219. https://doi.org/10.1037/hea0000306

Nich, C., & Carroll, K. (1997). Now you see it, now you don’t: a comparison of traditional versus random-effects regression models in the analysis of longitudinal follow-up data from a clinical trial. Journal of Consulting and Clinical Psychology, 65 (2), 252–261. https://doi.org/10.1037/0022-006X.65.2.252

Norton, P. J., & Hope, D. A. (2001). Kernels of truth or distorted perceptions: Self and observer ratings of social anxiety and performance. Behavior Therapy, 32(4), 765–786. https://doi.org/10.1016/S0005-7894(01)80020-4

Petersen, L. E., Stahlberg, D., & Dauenheimer, D. (2000). Effects of self-schema elaboration on affective and cognitive reactions to self-relevant information.Genetic, Social, and General Psychology Monographs, 126(1), 25–42. Retrieved from https://www.ncbi.nlm.nih.gov/pubmed/10713900

Pew Research Center (2018, February 5). “Mobile Fact Sheet,” Washington, D.C. Retrieved from. http://www.pewinternet. org/fact-sheet/mobile/.

Richey, J. A., Rittenberg, A., Hughes, L., Damiano, C. R., Sabatino, A., Miller, S., & Dichter, G. S. (2013). Common and distinct neural features of social and non-social reward processing in autism and social anxiety disorder. Social Cognitive and Affective Neuroscience, 9(3), 367–377. https://doi.org/10.1093/scan/nss146

Rizvi, S. J., Pizzagalli, D. A., Sproule, B. A., & Kennedy, S. H. (2016). Assessing anhedonia in depression: Potentionals and pitfalls. Neuroscience and Biobehavioral Reviews, 65, 21–35. https://doi.org/10.1016/j.neubiorev.2016.03.004

Robinson, M. D., & Clore, G. L. (2002). Belief and feeling: evidence for an accessibility model of emotional self-report. Psychological Bulletin, 128(6), 934–960. https://doi.org/10. 1037/0033-2909.128.6.934

Romero, N., Sanchez, A., & Vazquez, C. (2014). Memory biases in remitted depression: The role of negative cognitions at explicit and automatic processing levels. Journal of Behavior Therapy and Experimental Psychiatry, 45(1), 128–135. https://doi.org/10.1016/j.jbtep.2013.09.008

Shiffman, S., Stone, A. A., & Hufford, M. R. (2008). Ecological momentary assessment.Annual Review of Clinical Psychol- ogy, 4, 1–32. Retrieved from https://www.ncbi.nlm.nih. gov/pubmed/18509902

Stein, M. B., Torgrud, L. J., & Walker, J. R. (2000). Social phobia symptoms, subtypes, and severity: findings from a community survey. Archives of General Psychiatry, 57(11), 1046–1052. https://doi.org/10.1001/archpsyc.57.11.1046

Strunk, D. R., Lopez, H., & DeRubeis, R. J. (2006). Depressive symptoms are associated with unrealistic negative predic- tions of future life events. Behaviour Research and Therapy, 44(6), 861–882. https://doi.org/10.1016/j.brat.2005.07.001

Sumner, J., Griffith, J. W., & Mineka, S. (2010). Overgeneral autobiographical memory as a predictor of the course of depression: Ameta-analysis. Behaviour Research and Therapy, 48(7), 614–625. https://doi.org/10.1016/j.brat.2010.03.013

Watson,D.,Weber, K., Assenheimer, J. S., Clark, L. A., Strauss,M. E., & McCormick, R. A. (1995). Testing a tripartite model: I. Evaluating the convergent and discriminant validity of anxiety and depression symptom scales. Journal of Abnormal Psychol- ogy, 104(1), 3–14. https://doi.org/10.1037/0021-843X.104.1.3

Weeks, J. W., Heimberg, R. G., Rodebaugh, T. L., &Norton, P. J. (2008). Exploring the relationship between fear of positive evaluation and social anxiety. Journal of Anxiety Disorders, 22 (3), 386–400. https://doi.org/10.1016/j.janxdis.2007.04.009

Werner-Seidler, A., Banks, R., Dunn, B. D., & Moulds, M. L. (2013). An investigation of the relationship between positive affect regulation and depression. Behaviour Research and Therapy, 51(1), 46–56. https://doi.org/10.1016/j.brat.2012. 11.001

Wood, J. V., Anthony, D. B., & Foddis, W. F. (2006). Should people with low self-esteem strive for high self-esteem? InM. H. Kernis (Ed.), Self-esteem issues and answers: A source- book of current perspectives (pp. 288–296). New York: Psychology Press.

Xiong, H., Huang, Y., Barnes, L., & Gerber, M. (2016, September). Sensus: A cross-platform, general-purpose system for mobile crowdsensing in human-subject studies. ACM International Joint Conference on Pervasive and Ubiquitous Computing; September 12-16, 2016; Heidel- berg, Germany (pp. 415–426).

RECEIVED: January 8, 2018 ACCEPTED: July 30, 2018 AVAILABLE ONLINE: 3 August 2018