Skip to main content
. 2025 Feb 1;15:3980. doi: 10.1038/s41598-025-86623-6

Table 5.

Between-participant comparisons for advice evaluations (Study 4). The columns represent the mean evaluations for ChatGPT and human-generated advice (“M”) and their corresponding standard deviations (“SD”). The right-hand side shaded portion of the table presents the results of within-participant t-tests with the mean difference between the variables, with associated t-statistic, p-value, Cohen’s d effect sizes, and 95% confidence interval (“Lower CI” and “Upper CI”).

ChatGPT-Ratings ORDER Mean diff. t p-value Cohen’s d Lower CI Upper CI
ChatGPT First Self
First
Variable M SD M SD
Effectiveness 5.55 1.12 5.68 1.24 0.13 −1.04 .300 −0.11 −0.37 0.12
Quality 5.46 1.29 5.74 1.26 0.28 −2.11 .035 −0.22 −0.55 −0.02
Authenticity 5.16 1.37 5.34 1.51 0.19 −1.23 .218 −0.13 −0.48 0.11
Self-Ratings ORDER Mean diff. t p-value Cohen’s d Lower CI Upper CI
ChatGPT First Self
First
Variable M SD M SD
Effectiveness 5.67 0.88 5.62 1.03 −0.05 0.51 .607 0.05 −0.15 0.25
Quality 5.47 1.19 5.48 1.20 0.01 −0.08 .934 −0.01 −0.26 0.24
Authenticity 6.18 0.83 6.45 0.65 0.26 −3.36 .001 −0.35 −0.42 −0.11