Paper two

Counterfactual Thinking: Appointing Blame Comment by Ryan Winter: The title page here is essentially the same one from Paper I. It has the title (in APA format), author name, and university affiliation. Want my advice? If you did well on the Paper I title page, reuse it!

Former Student

Florida International University

Comment by Ryan Winter: The good news is that this example paper is on the same topic as the example paper from Paper I. I’m going to show you the progress of the paper throughout the semester, so you can see how you will eventually combine Papers I, II, III, and IV into Paper V. Let’s continue looking at counterfactual thinking! But again, this is an EXAMPLE paper. The topic here (counterfactual thinking) differs from your study. Do NOT discuss counterfactuals or any of the variables in this example in your paper unless they are relevant to your own topic.

Methods Comment by Ryan Winter: The word Method here is centered and bolded, as is recommended by the APA

Participants Comment by Ryan Winter: Participant (also bolded) is flush left

One hundred and twenty six students from Florida International University were randomly selected to participate in our study. Of these 126 participants, 37% (n = 47) were male and 63% (n = 79) were female. Ages ranged from a minimum of 17 to a maximum of 58 with an average of 22.32 years (SD = 6.30). The sample population consisted of 68.3% Hispanic Americans (n = 86), 8.7% African Americans (n = 11), 19% Caucasians (n = 24), 1.6% Asians (n = 2), and 2.4% who did not specify their ethnicity (n = 3). See Table 1. Comment by Ryan Winter: When a number starts a sentence, spell out the number Comment by Ryan Winter: Note the mean and standard deviation here, which is helpful for knowing about the makeup of the sample. The mean, of course, is the average

Materials and Procedure Comment by Ryan Winter: Also bolded and flush left. You will notice that this author combined materials and procedures, which was good for this simple study. She could have separated them, though, and talked about the taxi scenario and questionnaires in a “materials” section and the procedure separately in the “procedure” section. I like this combined choice, though, for this design.

In accordance with the standardized guidelines for informed consent, prospective participants were notified of the potential risks and benefits of participating in the study before being introduced to the research material. If the student verbally agreed to participate, he or she was given one of three different documents, each of which consisted of four parts or sections. In part one of the study, the participant read a short scenario concerning a paraplegic couple, Tina and Eugene, who requested a taxi for a night out with friends. Each of the three documents depicted the same initial situation with alternate conditions (changeable, unchangeable, or neutral). Comment by Ryan Winter: Noting the IV helps a lot. You can tell the author knows what his IV is. There is only one, with three levels

In the changeable condition, the taxi driver arrived to pick up the couple, only to promptly decline their fare upon seeing that they were both paraplegic. Without enough time to call for another taxi, Tina and Eugene decided to take Tina’s car, which was handicap equipped. In order to reach their destination, they had to cross a bridge that had been weakened the night before due to a severe storm. The damaged bridge collapsed mere minutes before the couple reached it. Unable to see the missing portion of the bridge in the night, Tina and Eugene drove off the road, into the river below, and drowned. The taxi driver, who had left 15 minutes earlier, managed to make it safely across, before the collapse. In the unchangeable condition, the situation remained mostly the same with the exception that the taxi driver arrived at the bridge after it had collapsed and plummeted into the water as well. He managed to make it out of the car and swim to safety, but Tina and Eugene drowned. In the neutral condition, the taxi arrived to pick up the couple but promptly refused their fare as soon as he realized that they were both paraplegic. In this condition, the taxi driver did eventually agree to take Tina and Eugene to their destination downtown, albeit after much argument. Due to the recently collapsed bridge, the taxi driver drove his passengers and himself off the road and into the river below. He barely managed to make it out of the car before drowning. Tina and Eugene’s outcome remained the same. Comment by Ryan Winter: Notice how thorough the description of the scenario is here. If you wanted to replicate this study, you would know exactly what to do because the author tells you exactly what she did. Make sure the description of your IV is equally clear.

After reading one of the scenarios described above, the participant continued on to the remainder of the study, which was composed of a series of open, partially open, and close-ended questions.

In part two, the student participating in the study was asked to procure 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.

In part three, the participant was presented with a series of questions about their thoughts regarding the specific situation they read about. After reading each question, the participant was asked to record his or her response in a scale of one to nine. These questions included how avoidable they thought the accident was (1 = not at all avoidable, 9 = very avoidable), the causal role of the taxi driver in the couple’s death (1 = not at all causal, 9 = the most important cause), their thoughts on how much control the taxi driver had (1 = no control, 9 = complete control), the negligence of the taxi driver (1 = not at all negligent, 9 = completely negligent), how much money for damages the taxi driver was responsible for (1 = no money, 9 = as much as possible), the foreseeability of the couple’s death (1 = not at all foreseeable, 9 = completely foreseeable), and how much blame the taxi driver deserved for the event (1 = no blame at all, 9 = total blame). Remaining questions focused on a series of 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”. The last question of part three was a yes or no question that asked the participant whether the taxi driver agreed to drive the couple or not. 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. 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 four asked for the participant’s demographic information, including gender, age, ethnicity, their first language, and whether they were a student at Florida International University. Concluding the study, the participant was debriefed on his or her contribution to the study as well as our insights on counterfactual thinking and our main hypothesis. Comment by Ryan Winter: You can see her procedure, right! Very clear, very step-by-step

Although we had several dependent variables, our primary focus involved the perceived blameworthiness of the taxi driver, the number of ‘If Only’ statements the participants could create, and the manipulation check regarding whether the driver agreed to take the couple. As such, these are the only three dependent variables that we analyzed.

Results Comment by Ryan Winter: Results is centered and bold. The results section comes right after the methods – there is no page break

Using survey condition (changeable vs. unchangeable vs. neutral) as our independent variable and whether participants recalled whether the taxi driver picked up the paraplegic couple as the dependent variable, we ran a manipulation check in which we saw a significant effect, X2(2) = 93.95, p < .001. Participants in the changeable and unchangeable conditions correctly said the taxi did not pick up the couple (95.2% and 90.5%, respectively) while few participants in the neutral condition said the driver picked up the couple (4.8%). Phi showed a large effect. This indicates that participants did pay attention to whether the taxi driver picked up the couple. See Table 2. Comment by Ryan Winter: The chi square here is useful for data that is nominal in nature (that is, there is no numerical difference between factors). Here, they either read about a taxi picking up the couple or they didn’t. We cannot look at a mean or average value here (what is the average between yes and no?), so the chi square looks at the number of people who say yes and the number who say no. Here, we want the participants in some conditions to say yes (if the taxi picked up the couple) and no (if he didn’t pick them up)

For our main analysis, our first One-Way ANOVA test revealed significant differences among our independent variable, the scenario conditions (changeable, unchangeable, or neutral) and our dependent variable, perceived blameworthiness of the taxi driver, F(2, 122) = 3.55, p = .032. A subsequent Tukey post hoc test supported our hypothesis by demonstrating that participants were more likely to blame the taxi driver in the changeable condition (M = 4.51, SD = 2.06) than in the unchangeable condition (M = 3.38, SD = 2.14).. However, there were no significant difference for perceived blame between the neutral condition (M = 4.36, SD = 2.11) and either the changeable or unchangeable conditions. These results indicate that in situations where the outcome is perceived as mutable (changeable), individuals are more likely to assign blame to the actor who could have acted differently (unchangeable). See Table 3. Comment by Ryan Winter: A One Way ANOVA is appropriate here since there are three levels to the single IV and the DV is on an interval scale (it ranges from 1 to 9) Comment by Ryan Winter: The student here provided an exact p value. This is acceptable, though you can also use p < .05, p > .05, or p < .01 where appropriate Comment by Ryan Winter: As you can see, this student did find significance, so she ran post hoc tests on the ANOVA using Tukey. But what if there was no significance,? Well, look what happens in the next ANOVA!

We were also interested in the number of ‘If Only’ statements generated for each condition. We ran a One-Way ANOVA test using the different conditions (changeable, unchangeable, or neutral) as our independent variable, and the number of counterfactuals produced as our dependent variable. The results revealed that the relationship between condition and number of ‘If Only’ statements produced was not significant, F(2, 123) = 1.79, p = .171. Our initial prediction that participants would develop more counterfactuals in the changeable condition was not supported since the number of counterfactuals generated in the changeable condition (M = 5.41, SD = 2.21), the unchangeable condition (M = 4.57, SD = 2.04), and the neutral condition (M = 4.88, SD = 1.85) did not differ. Since the p-value for the ANOVA test was not significant, there was no need to run post hoc tests. See Table 4. Comment by Ryan Winter: So this student ran a second ANOVA, which I think is best. But since the dependent variable used here was scaled (confidence, which is on a 1 to 9 scale), the student could have just as easily run a t-Test focusing on only two levels of the IV. Let me show you what that might look like. “We ran a t-Test looking only at the changeable and unchangeable conditions as our independent variable and number of If Only statements generated as our dependent variable. The t-Test was not significant, t(72) = 1.76, p > .05. Participants did not generate any more counterfactuals in the changeable condition (M = 5.56, SD = 2.76) than in the unchangeable condition (M = 4.36, SD = 2.06).” I could do something similar comparing the changeable and neutral conditions with a t-Test or comparing the neutral and unchangeable conditions, but running three t-Tests is a lot. Much easier to do it with one ANOVA, which looks at all three comparisons at the same time! Comment by Ryan Winter: Even though the ANOVA was not significant, I’d still like you to provide the means and standard deviations for the analysis

Finally, we ran an independent samples t-Test with the changeable and unchangeable conditions only and “How avoidable was the accident” as the dependent variable, which was significant, t(82) = 2.71, p < .01. Participants thought the accident was more avoidable in the changeable condition (M = 5.31, SD = 1.77) than in the unchangeable condition (M = 4.21, SD = 1.85). See Table 5.

Discussion Comment by Ryan Winter: Your discussion does not need to be extensive, but I do want you to note whether you supported or did not support your hypothesis and provide some possible reasons for your findings. You can make some educated guesses about what might be going on, but make them reasonable!

We predicted that participants would place more blame on an actor whose behavior led to an undesirable outcome (death) when that actor could have acted differently primarily because these participants would generate more “If Only” counterfactual statements that would lead them to see the outcome could have been avoided. Conversely, we predicted that participants who read about an undesirable outcome that could not have been avoided would assign less blame to the actor and would think of fewer counterfactual “If Only” statements. Results partially supported these predictions, as we did find more blame for in the changeable condition compared to the unchangeable (though neither differed from the neutral condition), and they thought the accident was more avoidable in the changeable condition than in the unchangeable condition. However, the number of counterfactual statements that participants generated did not differ among our three conditions. It could be that participants were unfamiliar with the counterfactual task, which requires some deep thinking, though on a more unconscious level they could have seen the changeable condition as evidencing more elements of blame. This begs the question: what if participants were forced to think deeper? This is the focus of our second study. Comment by Ryan Winter: This question here is actually a lead-in to the student’s next study. Your own methods, results, and discussion paper can end here, but keep in mind that your final paper is only halfway done right now! In Paper III, IV, and V, you will help design a follow-up study to your first study, so as you write this paper try to think about what you would do differently and what you might add in a follow-up study.

Table 1 Comment by Ryan Winter: You will have at least four tables for Study One. Label them in terms of table number (and make sure to provide a callout for the table in the results section). Tables are numbered sequentially, with the word Table flush left and in bold.

Demographics – Study One Comment by Ryan Winter: The table title is right above the table itself. It is flush left and is in italics. For Table 1, include all of your demographics (the statistics table, the gender table, and the ethnicity table). Note: We do not need to see the age table, which focuses on the age frequencies. It is better to use the mean age in the statistics table (rather than the age frequency in the age table). Make sure each table is flush left

Comment by Ryan Winter: To add tables, simply go into your SPSS output. You can right-click on the table and then copy it. Then just paste it into your table page! Alternatively, you can use the “Snipping tool” function available on most computers. (Do a search for it!). This allows you to draw a virtual box around text and then copy it like a picture. Then just paste the picture into the table page Finally, your last option is to do the work by hand. Insert a table with rows and columns and transfer over the information. This is the hard way, though. Both of the options above took me less than a minute. Recreating a table manually will take a much longer time!

Table 2

Crosstabs and Chi Square – Study One

Table 3

ANOVA Blame – Study One Comment by Ryan Winter: Make sure to give a good description of YOUR dependent variable. In this paper, she looked at blame as a DV, so she put that word here. Use YOUR dependent variable in the description

Table 4

ANOVA Number of Counterfactuals – Study One

Table 5

t-Test “Was the accident avoidable?” – Study One Comment by Ryan Winter: Note that you may not run a t-Test in your study. If you do, make sure to include both the group statistics and the independent samples t-Test tables! Comment by Ryan Winter: If your t-Table goes onto multiple lines, that is okay. This student just deleted a few columns from the t-Test to make it fit the page, but if your t-Table goes over into other rows, that is okay.