Course: PSY 2205 – Psychology II | Source: Perez, J. C. – "Social Loafing: The Motivation and Attitude Behind It" (FIU)
Difficulty: Intermediate
Prerequisites: Read the companion notes on Social Loafing core theory first. You will also need a basic understanding of ANOVA, chi-square tests, and what independent/dependent variables mean.
This set of notes covers the empirical study conducted at Florida International University that tested social loafing directly. Where the theory notes explain why people loaf, these notes walk through how researchers measured it, what they found, and how they extended the design with an anchoring manipulation in Study Two. You will see real ANOVA output, Tukey post hoc comparisons, and chi-square results, so this is also good revision for research methods alongside the social psychology content.
The FIU study assigned 195 participants to one of three conditions (Individual Total, Group Total, Group Average) and had them solve maths problems. Participants in the Individual condition solved significantly more problems, confirming social loafing. However, all participants believed they had outperformed the average, regardless of condition, showing a self-perception bias. Study Two added an anchoring manipulation (high vs. low anchor) to test whether providing a reference point shifts both effort and self-assessment.
One-Way ANOVA
A statistical test that compares means across three or more groups to determine whether at least one group mean differs significantly from the others. Think of it as asking: "Are these groups performing differently enough that it is unlikely to be chance?"
Tukey post hoc test (Tukey HSD)
A follow-up test run after a significant ANOVA result. It compares every pair of group means to identify which specific groups differ from each other. In simple terms, ANOVA tells you "something is different"; Tukey tells you "which pairs are different."
Chi-square test
A statistical test that examines whether there is a significant association between two categorical variables by comparing observed frequencies to expected frequencies. Think of it as checking whether the pattern you see in the data is a real relationship or just random variation.
Cramer's V
A measure of the strength of association between two categorical variables, used alongside chi-square. Ranges from 0 (no association) to 1 (perfect association). In simple terms, chi-square tells you whether a relationship exists; Cramer's V tells you how strong it is.
Manipulation check
A step in an experiment where researchers verify that participants understood and correctly identified the condition they were assigned to. Think of it as a comprehension check to make sure the experimental setup worked as intended.
Anchoring heuristic
A cognitive bias where people rely too heavily on the first piece of information they receive (the "anchor") when making subsequent judgments (Epley & Gilovich, 2006). In simple terms, the first number you hear pulls your estimate toward it, even if it is irrelevant.
High anchor / low anchor
In this study, the high anchor was telling participants that a prior pilot group solved 30 maths problems; the low anchor was telling them the pilot group solved 10. These reference points were expected to shift both effort and self-assessment.
Independent variable (IV)
The factor the researcher manipulates. In Study One, the IV was the scoring condition (Individual Total vs. Group Total vs. Group Average).
Dependent variable (DV)
The outcome the researcher measures. In Study One, the DVs were total maths problems solved and self-reported belief about performance relative to average.
195 FIU students, ages 19 to 79 (M = 31.15, SD = 12.45).
Gender: 44.6% male, 50.8% female, 0.5% non-binary, 4.1% other.
Race: 66.2% Latino/a, 11.3% Black, 8.7% White, 5.6% Asian, 3.1% MENA, 5.1% other.
Each participant was randomly assigned to one of three conditions:
Individual Total: your own score determines "best performance."
Group Total: your score is pooled with two others into a total group score.
Group Average: your score is pooled with two others into an average group score.
The IV was the scoring condition. The DV was the number of maths problems solved (out of 42 multiplication problems).
Participants gave consent, then read condition-specific instructions.
They solved up to 42 multiplication problems (told it was acceptable not to finish all of them).
Afterwards, they rated six statements on a 1 (strongly disagree) to 7 (strongly agree) scale about how they felt during the task, including whether they believed they had completed more problems than average.
A manipulation check confirmed participants could correctly identify their assigned condition.
One-Way ANOVA: F(2, 192) = 6.50, p < .001. The effect was significant.
Tukey post hoc results:
Individual Total (M = 28.68, SD = 11.18) solved significantly more problems than both Group Total (M = 22.32, SD = 11.00) and Group Average (M = 22.98, SD = 10.98).
Group Total and Group Average did not differ from each other.
This supports the first prediction: identifiable individual contributions lead to higher effort.
One-Way ANOVA: F(2, 192) = 0.60, p = .55. Not significant.
All three conditions rated themselves similarly: IT (M = 5.80), GT (M = 5.70), GA (M = 5.66).
All groups believed they had outperformed the average, regardless of their actual performance.
This means the second prediction (that everyone would think they did well) was supported, but there was no difference between conditions, so the post hoc tests were not needed.
Chi-square: X2(4) = 240.15, p < .001. Significant.
99% of Individual condition participants correctly recalled their condition.
75% of Group Total and 79% of Group Average participants correctly recalled theirs.
Cramer's V = .785, indicating a strong association.
This confirms the manipulation worked as intended.
The first hypothesis was supported: participants whose contributions were individually identifiable solved more problems.
The second hypothesis was partially supported: everyone believed they outperformed average, but there was no difference across conditions in that belief.
This raised a follow-up question: what if you anchor participants by telling them how many problems a prior group solved? Would that shift their self-assessment?
Study One showed that everyone believed they outperformed average, with no difference across conditions. Researchers wanted to see whether providing a reference point (an anchor) would shift both actual performance and self-assessment.
Anchoring is a cognitive bias where an initial piece of information (the anchor) pulls subsequent judgments toward it (Epley & Gilovich, 2006).
People use the anchor as a starting point and adjust from it, but adjustments are typically insufficient, meaning the final judgment stays closer to the anchor than it should.
Self-biases are pervasive in decision-making (Scheller, 2022). When people see a reference point for how others performed, they tend to follow it.
Same basic task as Study One (42 multiplication problems, same three scoring conditions).
Added a second IV: anchoring condition.
High anchor: participants were told a prior pilot group solved around 30 problems.
Low anchor: participants were told a prior pilot group solved around 10 problems.
This creates a 3 (condition) x 2 (anchor) factorial design.
Main effect of condition: Individual Total participants would solve more problems than Group conditions (same as Study One).
Main effect of anchor: high-anchor participants would solve more problems than low-anchor participants.
Interaction: the combination of Individual Total + high anchor would produce the highest scores, while Group Average + low anchor would produce the lowest.
Participants in the Individual Total condition rating above 5 would agree more strongly that they outperformed average.
High-anchor participants would rate themselves higher than low-anchor participants.
Interaction: Individual Total + high anchor would produce the highest self-ratings, while Group Average + low anchor would produce the lowest, as those participants lose their sense of individuality and trail behind the reference point.
"A non-significant ANOVA means nothing happened." The non-significant result for "I completed more problems than average" (p = .55) does not mean participants did not believe they outperformed others. It means they all believed it equally, regardless of condition. The finding is the uniformity of the bias, not an absence of effect.
"Group Total and Group Average are the same thing." They are different conditions: Group Total pools individual scores into a summed group score, while Group Average divides that sum. The study found no significant difference between these two, but they are conceptually distinct manipulations.
"The manipulation check is a dependent variable." The manipulation check (chi-square) is not testing the hypothesis. It is verifying that the experimental setup worked, i.e., that participants understood which condition they were in. It is a validity check on the method.
"Anchoring is the same as priming." Anchoring specifically involves a numerical reference point that pulls subsequent quantitative judgments toward it. Priming is broader and can involve any stimulus that influences subsequent processing. They overlap but are not interchangeable.
Be able to read and interpret ANOVA output: F-statistic, degrees of freedom, p-value. Know that F(2, 192) = 6.50 with p < .001 means at least one group mean is significantly different.
Understand when and why a Tukey post hoc test is run (only after a significant ANOVA).
Know the difference between a significant and non-significant ANOVA result and what each means substantively in this study.
The chi-square manipulation check is a separate analysis from the main hypothesis tests. Be clear on its purpose.
For Study Two, be able to describe a factorial design (3 x 2), identify the IVs and DVs, and explain what an interaction effect would look like.
The anchoring heuristic is likely to appear in exam questions about cognitive biases. Be able to define it and give an example from this study.
True or False: The ANOVA for "I completed more problems than average" was significant. (False: F(2, 192) = 0.60, p = .55.)
Fill in the blank: Participants in the ________ condition solved the most maths problems. (Individual Total.)
True or False: A Tukey post hoc test is needed after every ANOVA. (False: only after a significant ANOVA.)
Fill in the blank: The anchoring heuristic predicts that people who are given a _____ anchor will attempt to solve more problems. (High.)
True or False: Cramer's V of .785 indicates a weak association. (False: it indicates a strong association.)
Q: In Study One, what were the three experimental conditions and what was the independent variable?
A: The three conditions were Individual Total, Group Total, and Group Average. The independent variable was the scoring condition, which determined whether participants believed their score would be evaluated individually or pooled with others.
Q: Report the ANOVA result for total maths score in Study One and explain what it means.
A: F(2, 192) = 6.50, p < .001. This means there was a statistically significant difference in the number of maths problems solved across the three conditions. Tukey post hoc tests showed that the Individual Total condition (M = 28.68) solved significantly more problems than both the Group Total (M = 22.32) and Group Average (M = 22.98) conditions, which did not differ from each other.
Q: Why was the ANOVA for "I completed more problems than average" not significant, and what does that tell us?
A: F(2, 192) = 0.60, p = .55. All three groups rated themselves similarly (around 5.7 on a 7-point scale), meaning participants in every condition believed they had outperformed the average. The finding is that self-perception bias was uniform across conditions, not that participants did not hold the belief.
Q: What was the purpose of the chi-square test in Study One, and what did it show?
A: The chi-square tested whether participants correctly recalled their assigned condition (manipulation check). X2(4) = 240.15, p < .001, with Cramer's V = .785. This confirmed that the vast majority of participants understood which condition they were in, validating the experimental setup.
Q: Describe how anchoring was introduced in Study Two and what the researchers predicted would happen.
A: Participants were told that a prior pilot group had solved either 30 problems (high anchor) or 10 problems (low anchor). Researchers predicted a main effect of anchor (high-anchor participants solving more problems), a main effect of condition (Individual Total solving more), and an interaction where Individual Total + high anchor would produce the highest scores and Group Average + low anchor the lowest.
The anchoring heuristic connects to the broader cognitive biases unit. If your course covers Tversky and Kahneman's heuristics and biases framework, anchoring is one of the core examples.
The self-perception bias (everyone thinks they did better than average) connects to the better-than-average effect and illusory superiority, covered in the self-concept and social cognition sections of most social psychology courses.
The research methods in this study (ANOVA, chi-square, Tukey post hoc, manipulation checks) are directly transferable to your statistics and research methods coursework. Reviewing this study doubles as stats revision.
Social loafing study, FIU psychology, PSY 2205, one-way ANOVA, Tukey HSD, Tukey post hoc, chi-square test, Cramer's V, manipulation check, anchoring heuristic, anchoring and adjustment, high anchor low anchor, factorial design, 3x2 design, independent variable, dependent variable, motivational loss, identifiability, self-perception bias, better-than-average effect, illusory superiority, group dynamics experiment, Epley and Gilovich, Perez FIU