pap4

SCENARIO MUTABILITY & COGNITION 1

4

COUNTERFACTUAL THINKING: APPOINTING BLAME

14

SCENARIO MUTABILITY & COGNITION

Comment by Ryan Winter: Title pages should be a piece of cake by now!

Scenario Mutability and Need for Cognition: Appointing Blame When A Car Accident Seems Avoidable

Former Student

Florida International University

Method Study Two Comment by Ryan Winter: The good news here is that the set up for the methods for study two is similar to the set up for study one. You can copy and paste a bit, but there are going to be some really IMPORTANT differences that I will highlight throughout

Participants

One hundred and sixty subjects, 90% ( n = 144) university students, were recruited to participate in study two. Of these 160 participants, 33% ( n = 52) were male and 67% ( n = 108) were female. Ages ranged from a minimum of 17 to a maximum of 64 with an average of 22.38 years ( SD = 5.14). Our sample population consisted of 76% Hispanic Americans ( n = 122), 9% African Americans ( n = 15), 9% Caucasians ( n = 14), 3% Asian American ( n = 5), and 3% Others ( n = 4). See Table 6. Comment by Ryan Winter: The participant section in Paper IV is similar to that in Paper II, but your participants will differ. IMPORTANT: Include your sample and numbers for study two, as these participants are not the same ones you used in study one Comment by Ryan Winter: Note that this student started her tables with Table 6. She had five tables from Study One, so it makes sense for her to start with Table 6 for study two. For you, I am pretty sure you will continue listing tables starting with Table 5. Comment by Ryan Winter: Make sure to have a callout (“Table 6”) followed immediately by the table. You can group all demographics into the same table (Include the “Statistics”, “Gender”, and “Ethnicity” tables all under the general “Table 6” phrase)

Table 6

Demographics – Study Two

Materials and Procedure

Prospective participants were asked to take part in an online study being conducted for research purposes. If the subject agreed to participate, verbally or otherwise, he or she was directed to the survey developed through Qualtrics software. In accordance with the standardized guidelines for informed consent, subjects were first notified of the potential risks and benefits of participating in the study before being introduced to the research material. Once the participant confirmed their approval, they were eligible to continue with the rest of the survey, which consisted of six different parts or sections. Comment by Ryan Winter: It’s a new study (online this time!) so make sure to note it accordingly

In Part One of the study, we manipulated the subject’s Need for Cognition (NFC) by randomly assigning them to one of two possible groups. Depending on which group the subject was appointed to, they were presented with either five low-NFC or five high-NFC statements procured from the 18-item NFC scale developed by Caccioppo et al. (1984). After reading each statement, the participant was then asked to rate how much they agreed with each remark on a number scale. The numbers on the scale ranged from one (somewhat agree) to seven (completely agree). For example, a participant presented with a set of high-NFC statements was asked to rate the statement “I prefer complex to simple problems,” while a participant presented with a set of low-NFC statements was asked to rate the statement “I only think as hard as I have to” on the previously mentioned number scale. Comment by Ryan Winter: For your study, make sure to include all new relevant information. This student author used Need for Cognition as her second independent variable, so it was important to elaborate on how they was manipulated and measured in her study

In Part Two of the study, participants read one of two short scenarios concerning a paraplegic couple, Tina and Eugene, who requested a taxi for a night out with friends. These scenarios were identical to the ones used in study one. Here, however, we omitted the neutral condition since it rarely differed from the unchangeable condition in study one. Once again, and similar to study one, participants continued on to Part Three of the study, which asked them to provide as many ‘If Only’ statements as possible, meaning that they had to list all the factors they could think of that could have possibly changed the outcome of the event. Subjects were able to complete an unlimited number of statements, though the most any participant completed was ten statements. Comment by Ryan Winter: You’ll notice a very brief mention of the scenarios here. Since this author talked about them in detail in the methods section for study one, there is no need to say it all again. He readers already know what the study involves now, so no need to go into detail about it once again

Similar to study one, Part Four presented participants with a series of 12 questions or statements about their general thoughts regarding the specific situation they read about. The questions included “How avoidable do you think the accident was (1 = not at all avoidable, 9 = very avoidable)?”, “How causal was the role of the taxi driver in the couple’s death (1 = not at all causal, 9 = the most important cause)”?, “How much control do you think the taxi driver had (1 = no control, 9 = complete control)?”, “How negligent was the taxi driver (1 = not at all negligent, 9 = completely negligent)?”, “How much money for damages should the taxi driver be responsible for (1 = no money, 9 = as much as possible)?”, “How foreseeable was the couple’s death (1 = not at all foreseeable, 9 = completely foreseeable)?”, “How guilty should the taxi driver feel about the accident (1 = no guilt at all, 9 = as much guilt as possible)”, and “How much blame does the taxi driver deserve for the accident (1 = no blame at all, 9 = total blame)?” Remaining items focused on a series of four statements about the taxi drive, all rated on scales ranging from 1 (Strongly Disagree) to 9 (Strongly Agree). These statements included, “The taxi driver was reckless”, “The taxi driver was patient”, “The taxi driver was careful”, and “The taxi driver was hasty”. Comment by Ryan Winter: You know exactly what the DVs are here, and you know the range for each scale. This is VERY important. If you tell me the scale was 1 to 9 but that is it, I won’t know if 1 is a good score or a bad score. Does 9 mean they could avoid it or they could not avoid it? I need to see both the scale AND the labels for the DV to make sense Comment by Ryan Winter: Since these four questions all use the same 1 (Strongly Disagree) to 9 (Strongly Agree) scale, the student only provide the scale once. It gets repetitive if you add the same scale after each question.

Part Five of the survey consisted of the remaining eight manipulation check questions for NFC. Similar to section one of the study, the participant was asked to rate eight dispositional statements on a scale of one to seven. For example, the statement “Thinking is not my idea of fun” would be rated from a scale of one (extremely uncharacteristic) to seven (extremely characteristic). Comment by Ryan Winter: These DVs are new (not present in study one) and are important to describe as new variables in study two

Part Six of the study asked for the participant’s demographic information, including gender, age, ethnicity, country of birth, their first language, whether they are a student at Florida International University, etc. Several questions asked about information directly relevant to the scenario such as if the subject had ever been in a major car accident or if he or she knew anyone who was paraplegic.

Finally, in Part Seven we asked “Did the taxi driver agree to drive the couple?” The answer options were yes or no, or a nominal response. This final question served as an attention check, which informed us if the participant was attentive to the study and allowed us to exclude potentially misrepresentative responses from our data. Concluding the study, the participant was debriefed.

Although we had several dependent variables, our primary focus involved the perceived blameworthiness of the taxi driver and the number of ‘If Only’ statements the participants could create. We also analyzed the interaction between scenario mutability and NFC for both dependent variables.

Results Study Two Comment by Ryan Winter: Results are again brand new for this paper, so make sure they reflect your new Factorial design

Using condition (changeable versus unchangeable) as the independent variable and whether the taxi driver agreed to drive the couple as the dependent variable, the manipulation check was not significant. That is, most participants in both the changeable (95%) and the unchangeable (97.5%) conditions said the taxi driver failed to pick up the paraplegic couple, X2(1) = 0.69, p = .405. Phi, which is most appropriate for a 2 X 2 chi square, showed a weak effect. This is not surprising, as we eliminated the neutral condition (in study one, this was the only condition where the taxi did, in fact, pick up the couple). Thus participants did pay attention to their condition in study two. See Table 7. Comment by Ryan Winter: Again, include a callout to the table and then immediately provide the table

Table 7

Crosstabs and Chi Square – Study Two

To test our first dependent variable, we ran a 2 X 2 factorial ANOVA with NFC (high vs. low) and scenario condition (changeable vs. unchangeable) as our independent variables and participant’s responses to the question, “How much blame does the taxi driver deserve for the accident?” as our dependent variable. Results demonstrated no significant main effect for NFC on perceived blame, F(1, 152) = 1.69, p = .196. This means that there was no meaningful differences in the assignment of blame between the high-NFC ( M = 3.72, SD = 2.44) and low-NFC groups ( M = 4.12, SD = 2.49). There was, however, a significant main effect for scenario condition, F(1, 152) = 3.98, p = .048. Participants in the changeable condition ( M = 4.27, SD = 2.35) perceived the taxi driver to be more blameworthy for the couple’s death than participants in the unchangeable condition ( M = 3.56, SD = 2.47). Unfortunately, there was no interaction of NFC and scenario, F(1, 152) = 0.00, p = .196, indicating that perceived culpability did not significantly differ among high NFC changeable participants ( M = 4.04, SD = 2.28), high NFC unchangeable participants ( M = 3.27, SD = 2.44), low NFC changeable participants ( M = 4.56, SD = 2.44), and low NFC unchangeable participants ( M = 3.77, SD = 2.51). See Table 8

Table 8

ANOVA Perceived Blameworthiness – Study Two Comment by Ryan Winter: She changed her title here a bit from Paper II, but still looked at a similar variable (blame) for Paper III

To test our second dependent variable, we ran another 2 X 2 factorial ANOVA with NFC (high vs. low) and scenario condition (changeable vs. unchangeable) as our independent variables and number of “If Only” counterfactual statements as our dependent variable. There was no main effect for NFC on the number of “If Only” thoughts generated, F(1, 156) = .001, p = .975. This means that there was no difference in the number of counterfactual thoughts generated between the high-NFC group ( M = 3.87, SD = 1.77) and low-NFC group ( M = 3.81, SD = 2.46). Similarly, there was no main effect between for scenario, F(1, 56) = 2.05, p = .154. That is, there was no significant difference in the number of “If Only” statements generated between the changeable condition ( M = 4.09, SD = 2.28) and the unchangeable condition ( M = 3.60, SD = 1.97). We also examined the interaction between the two independent variables (high vs. low-NFC and unchangeable vs. changeable scenario) and the dependent variable. We found no interaction of NFC and scenario condition, F(1, 156) = 1.04, p = .310, indicating that the number of “If Only” thoughts created did not vary between the high NFC unchangeable condition ( M = 3.79, SD = 1.92), high NFC changeable condition ( M = 3.93, SD = 1.68), low NFC unchangeable condition ( M = 3.46, SD = 2.01), or low NFC changeable condition ( M = 4.29, SD = 2.93). See Table 9. Comment by Ryan Winter: If you look back at the example paper for Paper III, you’ll see the hypotheses for this results section. The results once again focus on two different dependent variables, but in a factorial design each DV has two main effects and one interaction associated with it. Comment by Ryan Winter: So no main effects and no interaction for this first variable. I still want to see the means, though, so include them.

Table 9

ANOVA Number of Counterfactuals – Study Two

Finally, we were interested in seeing if participants differed in the amount of guilt they felt the taxi driver should be feeling. Thus we ran another 2 X 2 factorial ANOVA with NFC (high vs. low) and scenario condition (changeable vs. unchangeable) as our independent variables and the “How guilty should the taxi driver feel about the accident” as our dependent variable. Results showed no significant main effect for need for cognition, F(1, 156) = 0.97, p = .326. That is, those high in need for cognition ( M = 7.26, SD = 1.03) did not differ in their assessments of the taxi driver’s guilt compared to those low in need for cognition ( M = 7.05, SD = 0.93). There was a significant main effect for scenario, F(1, 156) = 8.78, p = .004. Participants felt the taxi driver should feel more guilt in the changeable condition ( M = 7.39, SD = 0.92) than in the unchangeable condition ( M = 6.93, SD = 0.99). This main effect, however, was qualified by a significant need for cognition X scenario interaction, F(1, 156) = 10.27, p = .002. This required four additional simple effects tests.

First, for participants low in need for cognition only, there was no difference in guilt perception between participants in the changeable condition ( M = 7.03, SD = 1.06) and participants in the unchangeable condition ( M = 7.07, SD = 0.83), F(1, 78) = 0.03, p = .326. Second, for participants high in need for cognition only, participants felt the taxi driver should feel significantly more guilt in the changeable condition ( M = 7.65, SD = 0.71) than participants in the unchangeable condition ( M = 6.74, SD = 1.16), F(1, 78) = 19.12, p < .001. Third, for participants in the changeable condition only, participants felt the taxi driver should feel significantly more guilt if the participant was high in need for cognition ( M = 7.65, SD = 0.71) rather than low in need for cognition ( M = 7.03, SD = 1.06), F(1, 78) = 9.96, p = .002. Finally, for participants in the unchangeable condition, there was no difference in guilt perception between participants high in need for cognition ( M = 6.74, SD = 1.16) and participants low in need for cognition ( M = 7.07, SD = 0.83), F(1, 78) = 2.20, p = .142. See Table 10.

Table 10

2 X 2 ANOVA Perceived Guilt of the Taxi Driver – Study Two

Discussion Study Two

Although study two posited that Need for Cognition would impact participants and their generation of counterfactual statements and their assessment of blame, results did not support this contention. For both dependent variables, Need for Cognition did not result in main effects. Despite predictions to the contrary, those high in NFC did not generate any more counterfactuals than those low in NFC, and those high in NFC did not blame the taxi driver any more than those low in NFC. Nor did NCF interact with scenario, despite our prediction that those high NFC would generate the most counterfactuals and find the most blame when given the changeable scenario compared to other conditions. However, scenario did show a significant main effect such that participants found more blame for the taxi driver in the changeable condition than the unchangeable condition. The more interesting results occurred for perceptions about how guilty the taxi driver should feel. When the situation was changeable (that is, preventable in some way), those high in need for cognition thought the taxi driver should feel more guilt. If the situation was unchangeable or participants were low in need for cognition, the participants did not feel the taxi driver should feel guilty. This supports our predictions.