Source: "Facebook Rudeness: What Is an Individual's Reaction to Negative Feedback?" – Florida International University
Difficulty: Intermediate | Prerequisites: Part 1 of these study notes (background concepts); basic understanding of one-way ANOVA and chi-square tests.
Study One is the first of two experiments in this paper. It tests a straightforward question: if a Facebook user responds rudely to people calling their shared news story "fake," do observers judge that user more negatively than if the response is polite? This is the foundation the whole paper builds on. Study Two (covered in Part 3) adds age as a second variable. You need to understand this study's design, conditions, and results before Study Two makes sense.
Participants read a fictional Facebook page where a user named Corey shared a dubious news story and then responded to criticism in one of three ways: rudely disagreeing, politely disagreeing, or politely agreeing the story was fake. Those who saw the rude response rated Corey as significantly ruder and felt angrier than those who saw either polite response. The two polite conditions did not differ meaningfully from each other.
Independent variable (IV)
The variable the researcher manipulates. In Study One, the IV is the rudeness condition: Rude Disagree (RD), Polite Disagree (PD), or Polite Agree (PA). Think of it as the thing the experimenters changed between groups.
Dependent variable (DV)
The variable the researcher measures to see the effect of the IV. Here, the main DVs are agreement with "Corey seems like a rude person" and "Reading Corey's last post made me feel angry," both on a 1-to-6 Likert scale.
One-way ANOVA
A statistical test that compares the means of three or more groups to determine whether at least one group mean differs significantly from the others. In simple terms, it answers: "Did the rudeness condition make a real difference in how people rated Corey?"
Tukey HSD (Honestly Significant Difference)
A post hoc test used after a significant ANOVA to identify which specific pairs of group means differ from each other. Think of it as the follow-up that tells you which conditions are driving the overall effect.
Chi-square test (χ²)
A statistical test used to determine whether there is a significant association between two categorical variables. In this study, it was used for the manipulation check: did participants correctly recall Corey's response type?
Cramer's V
An effect-size measure for chi-square tests with tables larger than 2×2. A strong Cramer's V here meant participants were reliably remembering which condition they were in.
Manipulation check
A procedure to confirm that participants perceived the experimental manipulation as intended. Participants were asked to recall whether Corey rudely disagreed, politely disagreed, or politely agreed.
Likert scale
A rating scale (here 1 to 6) used to measure the degree of agreement with a statement. 1 = strongly disagree, 6 = strongly agree.
N = 187 (100 male, 87 female)
Age range: 16 to 59, mean 25.35 (SD = 9.109)
Ethnicity: 47.6% Hispanic, 25.1% Caucasian, 16.6% African American, 4.3% Asian American, 1.6% Native American, 4.8% Other
Recruited at Florida International University
Between-subjects design with one IV (rudeness condition) at three levels: Rude Disagree (RD), Polite Disagree (PD), Polite Agree (PA)
Participants were randomly assigned to one condition
The Facebook page for a fictional user, Corey McMillan, was kept identical across conditions except for Corey's reply to commenters
Corey's profile was deliberately gender-neutral (ambiguous profile photo, gender-neutral name) to avoid confounding gender expectations with rudeness judgments
Corey shared a story claiming "FCC Forces CNN to change its accreditation from 'news' to 'entertainment'"
Two commenters (Peyton Halliburton and Riley Anderson) both called the story fake
Corey's response varied by condition:
RD condition: Corey calls commenters "liberal idiots" and morons, insists the story is real
PD condition: Corey disagrees politely, suggests commenters do their own research
PA condition: Corey admits being wrong, agrees the story is fake after further research
"Corey seems like a rude person" (1-to-6 Likert scale)
"Reading Corey's last post made me feel angry" (1-to-6 Likert scale)
Manipulation check: participants selected which response type they remembered Corey giving
χ²(4) = 149.11, p < .05, with a strong Cramer's V (.631)
69.7% of RD participants correctly recalled rude disagreement
66.7% of PD participants correctly recalled polite disagreement
80.0% of PA participants correctly recalled polite agreement
Conclusion: the manipulation worked. Participants registered the intended condition.
One-way ANOVA: F(2, 184) = 13.797, p < 0.05
RD condition: M = 3.38, SD = 1.187 (highest rudeness rating)
PD condition: M = 2.98, SD = 0.920
PA condition: M = 2.71, SD = 0.786 (lowest rudeness rating)
Tukey HSD: the significant difference was between RD and PA (mean difference = 0.670, p = .001). The PD condition fell between the two but did not differ significantly from either.
One-way ANOVA: F(2, 184) = 26.238, p < 0.05
RD condition: M = 3.09, SD = 1.160 (highest anger)
PA condition: M = 2.45, SD = 0.741
PD condition: M = 2.23, SD = 0.760 (lowest anger)
Tukey HSD: RD differed significantly from both PD (p = .000) and PA (p = .001). PD and PA did not differ from each other (p = .365).
The hypothesis was supported: rudely disagreeing led to higher rudeness ratings and more anger than politely responding.
The PD condition was ambiguous, sitting between the other two without reaching significance against either. This may reflect individual variation in what people consider "rude" vs. merely disagreeable.
The finding that PD and PA did not differ supports the idea that politeness (regardless of whether the person agrees or disagrees) is the key factor, not the content of the reply itself.
Anyone managing a public-facing social media account should note: a rude response to criticism damages your perceived character more than the original criticism damages your credibility. This applies equally to brands, politicians, and individuals. For content moderators, the finding supports prioritising tone-based interventions (e.g. prompts that encourage civil responses) over simply removing critical comments.
A polite disagreement is perceived the same as a rude one. It is not. In this study, politely disagreeing was rated similarly to politely agreeing, and both were rated as far less rude than the rude condition.
The content of the reply (agree vs. disagree) matters more than the tone. The opposite is true here. Whether Corey agreed or disagreed with the criticism made little difference; what mattered was whether the reply was polite or rude.
A strong manipulation check means strong results on the DV. These are separate things. A manipulation check confirms participants noticed the IV. The ANOVA on the DV tells you whether the IV had an effect. Both can succeed or fail independently.
⚠️ Know the three conditions (RD, PD, PA) and what each one manipulated. This is the backbone of the entire paper.
⚠️ Be able to interpret the ANOVA results: F(2, 184) = 13.797 for rudeness and F(2, 184) = 26.238 for anger. Know that both are significant at p < .05.
⚠️ Understand why the Tukey HSD matters: the ANOVA tells you "at least one group differs," but Tukey tells you which specific pairs differ.
⚠️ The PD condition's lack of significant difference from either RD or PA is a discussion point, not a failure. Be ready to explain why this might happen (individual variation in rudeness thresholds).
True or False: The study used a within-subjects design.
False. It was a between-subjects design; each participant saw only one condition.
Fill in the blank: The three rudeness conditions were RD, PD, and ______.
PA (Polite Agree)
True or False: The Tukey HSD showed a significant difference between the PD and PA conditions on the rudeness DV.
False. PD and PA did not differ significantly.
Fill in the blank: A Cramer's V of .631 indicates a ______ effect size.
Strong
True or False: Participants in the RD condition felt significantly angrier than those in both the PD and PA conditions.
True.
Q: Describe the experimental design of Study One, including the IV, DV, and conditions.
A: Study One used a between-subjects design with one IV (rudeness condition) at three levels: Rude Disagree, Polite Disagree, and Polite Agree. The DVs were participants' ratings of "Corey seems like a rude person" and "Reading Corey's last post made me feel angry," both measured on a 1-to-6 Likert scale. Participants were randomly assigned to one condition.
Q: What was the purpose of the manipulation check, and what did it reveal?
A: The manipulation check verified that participants correctly perceived which condition they were in. A chi-square test, χ²(4) = 149.11, p < .05, with a strong Cramer's V (.631), confirmed that most participants accurately recalled Corey's response type, indicating the experimental manipulation was effective.
Q: Why did the researchers keep Corey's profile gender-neutral?
A: To prevent gender-based expectations from confounding the results. If participants assumed Corey was male or female, pre-existing stereotypes about gender and rudeness could have influenced their impressions independently of the rudeness manipulation.
Q: Interpret the Tukey HSD results for the "Corey seems like a rude person" DV.
A: The only significant pairwise difference was between RD (M = 3.38) and PA (M = 2.71), with a mean difference of 0.670 (p = .001). The PD condition (M = 2.98) did not differ significantly from either RD or PA. This tells us the clear contrast is between rude and polite-agreeing responses, while polite disagreement occupies an ambiguous middle ground.
Q: Why might the PD condition not have differed significantly from the other two conditions?
A: Participants may have varying personal thresholds for what counts as rude. A polite disagreement might strike some people as slightly confrontational and others as entirely civil. The researchers suggest that individual differences in how people define rudeness explain the PD condition's middle position.
This study connects to impression formation and person perception in social psychology, specifically how peripheral cues (tone, word choice) can outweigh central content in shaping judgments. It also relates to the broader literature on incivility effects, where rude behaviour in one domain (online) shapes perceptions that carry into other domains (offline character judgments). The statistical methods here (one-way ANOVA, Tukey HSD, chi-square with Cramer's V) are standard tools in PSY 3215 and worth revising through this concrete example.
One-way ANOVA, Tukey HSD post hoc, chi-square manipulation check, Cramer's V, between-subjects design, Likert scale, Facebook rudeness experiment, impression formation, online incivility, rudeness perception, polite disagreement, negative feedback response, PSY 3215 Study One, FIU psychology, social media experiment