Revising an Overclaimed Lab Discussion Section
You are writing the discussion section of a 1,500-word lab report for a second-year psychology methods course. Your group ran a reaction-time experiment with 36 participants: 18 in a quiet room, 18 with background speech at 65 dB. The quiet group mean was 412 ms (SD 38); the speech group mean was 447 ms (SD 51). A t-test gave p = .04, Cohen's d = 0.71. Your lab manual says the department's replication threshold is p < .01 for a 'robust' finding. Your group's draft says, 'Background speech significantly impairs reaction time, confirming the distraction hypothesis.' Your TA's feedback says: 'Do not overclaim; note the small sample and the threshold; also consider that the speech group had two participants who reported sleep deprivation.' You need to revise the discussion in 25 minutes.
1. Rewrite the draft's claim in one or two sentences that are accurate for the report. Explain what changes you made and why.
Read the worked answer
Revised claim: 'Background speech was associated with slower mean reaction time (447 ms vs. 412 ms), a difference that reached conventional significance (p = .04, d = 0.71) but not the department's stricter p < .01 threshold for a robust finding.' I made four changes. I replaced 'significantly impairs' with 'was associated with slower' because the design is a between-groups comparison, not a causal manipulation check; 'impairs' implies a mechanism the study did not test. I added the means and effect size so readers can judge practical size, not just the p value. I replaced 'confirming' with 'reached conventional significance but not the department's threshold' because the lab manual defines 'robust' as p < .01, and p = .04 does not meet it. I also kept the finding visible rather than dismissing it; a .71 effect size is not trivial. In the limitations, I would add the two speech-group participants who reported sleep deprivation, because that alternative explanation weakens the causal claim and should be addressed before the conclusion.
2. Your groupmate argues, 'p = .04 is still significant, so we can say the hypothesis is supported.' How would you respond in the discussion, and what wording would you use to keep the finding useful without overclaiming?
Read the worked answer
I would respond that 'significant' is not a yes/no license. At p = .04, the result passes the conventional .05 line, but the department's manual sets p < .01 for a robust finding, so the group cannot use the stronger label without ignoring its own criteria. I would also point out that p = .04 is sensitive to small changes: with 36 participants and two sleep-deprived cases in the speech group, the result could shift. The useful wording is 'consistent with' or 'provides preliminary support for,' not 'confirms' or 'proves.' For example: 'The results provide preliminary support for the distraction hypothesis: the speech group was slower by 35 ms, with a moderate-to-large effect size, but the finding falls short of the department's robustness threshold and should be replicated with a larger, better-controlled sample.' This keeps the direction and size of the effect, acknowledges the threshold, and turns the limitation into a reason for replication rather than a reason to hide the result.