Back to Index

Valid entries and size of groups

There is a total of 353 entries in the dataset. 18 participants failed the attention check or the comprehension check. The valid number of entries is n = 335.

training n percentage
yes_ai 67 20.00%
no_ai 69 20.60%
yes_human 71 21.19%
no_human 66 19.70%
control 62 18.51%

Planned pairwise comparisons

Per the preregistration, 6 of the 10 possible pairs among the 5 training groups are tied to specific hypotheses. A 7th comparison (yes_ai vs. yes_human, labeled “Extra” below) was added afterward, at request, to test whether AI help and human help are viewed differently – it is not part of the preregistration. These are the only pairwise comparisons tested below (Bonferroni corrected within this set of 7, per DV – note this deviates from the preregistration’s specified Holm-Bonferroni method; see S3-analysis.Rmd for the preregistered version); the omnibus one-way ANOVA in each DV section still uses all 5 groups.

group1 group2 hypothesis description
yes_ai no_ai H1a AI-trained vs. non-AI-trained
yes_ai control H1b AI-trained vs. unknown AI-use history
no_ai control H2 Non-AI-trained vs. unknown AI-use history
yes_human no_human H3a Peer-tutor-trained vs. non-peer-tutor-trained
yes_human control H3b Peer-tutor-trained vs. unknown tutor-use history
no_human control Note Non-peer-tutor-trained vs. unknown tutor-use history (expected null)
yes_ai yes_human Extra AI-trained vs. peer-tutor-trained (not preregistered)

Note (from preregistration section 2): the no_human vs. control comparison is included for completeness/symmetry with H2, but the preregistration explicitly does not expect this one to be significant.

Creating composites

Per the preregistration, multi-item DVs (ability, benevolence, integrity, autonomy) are composited if Cronbach’s alpha >= .80; otherwise they should be analyzed as individual items.

ability

α = 0.944 95% CI [ 0.934 , 0.954 ]

✓ Composite ability created (α meets preregistered threshold of 0.8 )

benevolence

α = 0.892 95% CI [ 0.872 , 0.913 ]

✓ Composite benevolence created (α meets preregistered threshold of 0.8 )

integrity

α = 0.879 95% CI [ 0.857 , 0.902 ]

✓ Composite integrity created (α meets preregistered threshold of 0.8 )

autonomy

α = 0.837 95% CI [ 0.808 , 0.866 ]

✓ Composite autonomy created (α meets preregistered threshold of 0.8 )

Effect of condition on DVs (one-way ANOVAs + planned comparisons)

Comfort

Single item, 0-6 scale (0 = “Not comfortable at all”, 6 = “Very comfortable”).

“On a scale of 0 to 6, where 0 = Not comfortable at all and 6 = Very comfortable, how comfortable would you be having Dr. Wilson as your doctor?”

Summary

training variable n mean sd se ci
yes_ai comfort 66 3.970 1.559 0.192 0.383
no_ai comfort 69 5.565 0.675 0.081 0.162
yes_human comfort 71 4.803 1.272 0.151 0.301
no_human comfort 66 4.288 1.356 0.167 0.333
control comfort 61 4.820 1.133 0.145 0.290

ANOVA

## $output1
## Anova Table (Type II tests)
## 
## Response: comfort
##           Sum Sq  Df F value    Pr(>F)    
## training   99.68   4  16.391 2.969e-12 ***
## Residuals 498.68 328                      
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## $output2
## # Effect Size for ANOVA
## 
## Parameter | Eta2 |       95% CI
## -------------------------------
## training  | 0.17 | [0.10, 1.00]
## 
## - One-sided CIs: upper bound fixed at [1.00].
## $output3
## # Effect Size for ANOVA
## 
## Parameter | Cohen's f |      95% CI
## -----------------------------------
## training  |      0.45 | [0.34, Inf]
## 
## - One-sided CIs: upper bound fixed at [Inf].

All Planned Comparisons

All 7 planned pairwise comparisons – the 6 preregistered ones plus one added afterward (AI-trained vs. peer-tutor-trained) – Bonferroni corrected within this set of 7; see ‘Planned pairwise comparisons’ section above for the full hypothesis mapping. Individual hypothesis tabs (isolated table + plot per comparison) follow.

hypothesis description group1 group2 n1 n2 p p.signif p.adj p.adj.signif effsize conf.low conf.high Cohens’ d
H1a AI-trained vs. non-AI-trained yes_ai no_ai 67 69 0.0000 **** 0.0000 **** -1.3388 -1.72 -1.01 large
H1b AI-trained vs. unknown AI-use history yes_ai control 67 62 0.0001 *** 0.0009 *** -0.6200 -0.98 -0.28 moderate
H2 Non-AI-trained vs. unknown AI-use history no_ai control 69 62 0.0007 *** 0.0046 ** 0.8117 0.49 1.17 large
H3a Peer-tutor-trained vs. non-peer-tutor-trained yes_human no_human 71 66 0.0151 * 0.1057 ns 0.3922 0.06 0.75 small
H3b Peer-tutor-trained vs. unknown tutor-use history yes_human control 71 62 0.9380 ns 1.0000 ns -0.0139 -0.34 0.37 negligible
Note Non-peer-tutor-trained vs. unknown tutor-use history (expected null) no_human control 66 62 0.0157 * 0.1099 ns -0.4241 -0.78 -0.09 small
Extra AI-trained vs. peer-tutor-trained (not preregistered) yes_ai yes_human 67 71 0.0001 **** 0.0007 *** -0.5878 -0.97 -0.21 moderate

H1a

AI-trained vs. non-AI-trained (yes_ai vs. no_ai)

hypothesis description group1 group2 n1 n2 p p.signif p.adj p.adj.signif effsize conf.low conf.high Cohens’ d
H1a AI-trained vs. non-AI-trained yes_ai no_ai 67 69 0 **** 0 **** -1.3388 -1.72 -1.01 large

H1b

AI-trained vs. unknown AI-use history (yes_ai vs. control)

hypothesis description group1 group2 n1 n2 p p.signif p.adj p.adj.signif effsize conf.low conf.high Cohens’ d
H1b AI-trained vs. unknown AI-use history yes_ai control 67 62 1e-04 *** 9e-04 *** -0.62 -0.98 -0.28 moderate

H2

Non-AI-trained vs. unknown AI-use history (no_ai vs. control)

hypothesis description group1 group2 n1 n2 p p.signif p.adj p.adj.signif effsize conf.low conf.high Cohens’ d
H2 Non-AI-trained vs. unknown AI-use history no_ai control 69 62 7e-04 *** 0.0046 ** 0.8117 0.49 1.17 large

H3a

Peer-tutor-trained vs. non-peer-tutor-trained (yes_human vs. no_human)

hypothesis description group1 group2 n1 n2 p p.signif p.adj p.adj.signif effsize conf.low conf.high Cohens’ d
H3a Peer-tutor-trained vs. non-peer-tutor-trained yes_human no_human 71 66 0.0151 * 0.1057 ns 0.3922 0.06 0.75 small

H3b

Peer-tutor-trained vs. unknown tutor-use history (yes_human vs. control)

hypothesis description group1 group2 n1 n2 p p.signif p.adj p.adj.signif effsize conf.low conf.high Cohens’ d
H3b Peer-tutor-trained vs. unknown tutor-use history yes_human control 71 62 0.938 ns 1 ns -0.0139 -0.34 0.37 negligible

Note

Non-peer-tutor-trained vs. unknown tutor-use history (expected null) (no_human vs. control)

hypothesis description group1 group2 n1 n2 p p.signif p.adj p.adj.signif effsize conf.low conf.high Cohens’ d
Note Non-peer-tutor-trained vs. unknown tutor-use history (expected null) no_human control 66 62 0.0157 * 0.1099 ns -0.4241 -0.78 -0.09 small

Extra

AI-trained vs. peer-tutor-trained (not preregistered) (yes_ai vs. yes_human)

hypothesis description group1 group2 n1 n2 p p.signif p.adj p.adj.signif effsize conf.low conf.high Cohens’ d
Extra AI-trained vs. peer-tutor-trained (not preregistered) yes_ai yes_human 67 71 1e-04 **** 7e-04 *** -0.5878 -0.97 -0.21 moderate

Ability

3-item composite. Agree-disagree scale (0 = Strongly disagree, 6 = Strongly agree).

  • ability1: “Dr. Wilson is well qualified to treat patients.”
  • ability2: “I feel confident in Dr. Wilson’s clinical skills.”
  • ability3: “Dr. Wilson is highly capable of performing his medical duties.”

Summary

training variable n mean sd se ci
yes_ai ability 67 4.164 1.355 0.166 0.331
no_ai ability 69 5.493 0.559 0.067 0.134
yes_human ability 71 4.892 1.019 0.121 0.241
no_human ability 66 4.641 1.048 0.129 0.258
control ability 62 4.981 0.782 0.099 0.199

ANOVA

## $output1
## Anova Table (Type II tests)
## 
## Response: ability
##           Sum Sq  Df F value    Pr(>F)    
## training   64.03   4  16.314 3.317e-12 ***
## Residuals 323.80 330                      
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## $output2
## # Effect Size for ANOVA
## 
## Parameter | Eta2 |       95% CI
## -------------------------------
## training  | 0.17 | [0.10, 1.00]
## 
## - One-sided CIs: upper bound fixed at [1.00].
## $output3
## # Effect Size for ANOVA
## 
## Parameter | Cohen's f |      95% CI
## -----------------------------------
## training  |      0.44 | [0.34, Inf]
## 
## - One-sided CIs: upper bound fixed at [Inf].

All Planned Comparisons

All 7 planned pairwise comparisons – the 6 preregistered ones plus one added afterward (AI-trained vs. peer-tutor-trained) – Bonferroni corrected within this set of 7; see ‘Planned pairwise comparisons’ section above for the full hypothesis mapping. Individual hypothesis tabs (isolated table + plot per comparison) follow.

hypothesis description group1 group2 n1 n2 p p.signif p.adj p.adj.signif effsize conf.low conf.high Cohens’ d
H1a AI-trained vs. non-AI-trained yes_ai no_ai 67 69 0.0000 **** 0.0000 **** -1.2886 -1.62 -1.03 large
H1b AI-trained vs. unknown AI-use history yes_ai control 67 62 0.0000 **** 0.0000 **** -0.7312 -1.08 -0.44 moderate
H2 Non-AI-trained vs. unknown AI-use history no_ai control 69 62 0.0034 ** 0.0237 * 0.7591 0.42 1.13 moderate
H3a Peer-tutor-trained vs. non-peer-tutor-trained yes_human no_human 71 66 0.1400 ns 0.9800 ns 0.2426 -0.13 0.61 small
H3b Peer-tutor-trained vs. unknown tutor-use history yes_human control 71 62 0.6050 ns 1.0000 ns -0.0973 -0.42 0.25 negligible
Note Non-peer-tutor-trained vs. unknown tutor-use history (expected null) no_human control 66 62 0.0533 ns 0.3731 ns -0.3657 -0.70 -0.03 small
Extra AI-trained vs. peer-tutor-trained (not preregistered) yes_ai yes_human 67 71 0.0000 **** 0.0001 *** -0.6097 -0.99 -0.27 moderate

H1a

AI-trained vs. non-AI-trained (yes_ai vs. no_ai)

hypothesis description group1 group2 n1 n2 p p.signif p.adj p.adj.signif effsize conf.low conf.high Cohens’ d
H1a AI-trained vs. non-AI-trained yes_ai no_ai 67 69 0 **** 0 **** -1.2886 -1.62 -1.03 large

H1b

AI-trained vs. unknown AI-use history (yes_ai vs. control)

hypothesis description group1 group2 n1 n2 p p.signif p.adj p.adj.signif effsize conf.low conf.high Cohens’ d
H1b AI-trained vs. unknown AI-use history yes_ai control 67 62 0 **** 0 **** -0.7312 -1.08 -0.44 moderate

H2

Non-AI-trained vs. unknown AI-use history (no_ai vs. control)

hypothesis description group1 group2 n1 n2 p p.signif p.adj p.adj.signif effsize conf.low conf.high Cohens’ d
H2 Non-AI-trained vs. unknown AI-use history no_ai control 69 62 0.0034 ** 0.0237 * 0.7591 0.42 1.13 moderate

H3a

Peer-tutor-trained vs. non-peer-tutor-trained (yes_human vs. no_human)

hypothesis description group1 group2 n1 n2 p p.signif p.adj p.adj.signif effsize conf.low conf.high Cohens’ d
H3a Peer-tutor-trained vs. non-peer-tutor-trained yes_human no_human 71 66 0.14 ns 0.98 ns 0.2426 -0.13 0.61 small

H3b

Peer-tutor-trained vs. unknown tutor-use history (yes_human vs. control)

hypothesis description group1 group2 n1 n2 p p.signif p.adj p.adj.signif effsize conf.low conf.high Cohens’ d
H3b Peer-tutor-trained vs. unknown tutor-use history yes_human control 71 62 0.605 ns 1 ns -0.0973 -0.42 0.25 negligible

Note

Non-peer-tutor-trained vs. unknown tutor-use history (expected null) (no_human vs. control)

hypothesis description group1 group2 n1 n2 p p.signif p.adj p.adj.signif effsize conf.low conf.high Cohens’ d
Note Non-peer-tutor-trained vs. unknown tutor-use history (expected null) no_human control 66 62 0.0533 ns 0.3731 ns -0.3657 -0.7 -0.03 small

Extra

AI-trained vs. peer-tutor-trained (not preregistered) (yes_ai vs. yes_human)

hypothesis description group1 group2 n1 n2 p p.signif p.adj p.adj.signif effsize conf.low conf.high Cohens’ d
Extra AI-trained vs. peer-tutor-trained (not preregistered) yes_ai yes_human 67 71 0 **** 1e-04 *** -0.6097 -0.99 -0.27 moderate

Benevolence

3-item composite. Agree-disagree scale (0 = Strongly disagree, 6 = Strongly agree).

  • benevolence1: “Dr. Wilson is genuinely concerned about his patients’ welfare.”
  • benevolence2: “Dr. Wilson would never knowingly do anything to harm a patient.”
  • benevolence3: “Dr. Wilson truly looks out for what matters most to his patients.”

Summary

training variable n mean sd se ci
yes_ai benevolence 67 4.418 1.078 0.132 0.263
no_ai benevolence 69 4.870 0.934 0.112 0.224
yes_human benevolence 71 4.793 0.980 0.116 0.232
no_human benevolence 66 4.586 0.861 0.106 0.212
control benevolence 62 4.602 1.004 0.128 0.255

ANOVA

## $output1
## Anova Table (Type II tests)
## 
## Response: benevolence
##           Sum Sq  Df F value  Pr(>F)  
## training    8.79   4  2.3162 0.05711 .
## Residuals 313.07 330                  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## $output2
## # Effect Size for ANOVA
## 
## Parameter | Eta2 |       95% CI
## -------------------------------
## training  | 0.03 | [0.00, 1.00]
## 
## - One-sided CIs: upper bound fixed at [1.00].
## $output3
## # Effect Size for ANOVA
## 
## Parameter | Cohen's f |      95% CI
## -----------------------------------
## training  |      0.17 | [0.00, Inf]
## 
## - One-sided CIs: upper bound fixed at [Inf].

All Planned Comparisons

All 7 planned pairwise comparisons – the 6 preregistered ones plus one added afterward (AI-trained vs. peer-tutor-trained) – Bonferroni corrected within this set of 7; see ‘Planned pairwise comparisons’ section above for the full hypothesis mapping. Individual hypothesis tabs (isolated table + plot per comparison) follow.

hypothesis description group1 group2 n1 n2 p p.signif p.adj p.adj.signif effsize conf.low conf.high Cohens’ d
H1a AI-trained vs. non-AI-trained yes_ai no_ai 67 69 0.0072 ** 0.0505 ns -0.4481 -0.79 -0.1100 small
H1b AI-trained vs. unknown AI-use history yes_ai control 67 62 0.2840 ns 1.0000 ns -0.1766 -0.53 0.1800 negligible
H2 Non-AI-trained vs. unknown AI-use history no_ai control 69 62 0.1180 ns 0.8260 ns 0.2762 -0.08 0.6400 small
H3a Peer-tutor-trained vs. non-peer-tutor-trained yes_human no_human 71 66 0.2140 ns 1.0000 ns 0.2245 -0.11 0.5900 small
H3b Peer-tutor-trained vs. unknown tutor-use history yes_human control 71 62 0.2590 ns 1.0000 ns 0.1930 -0.12 0.5500 negligible
Note Non-peer-tutor-trained vs. unknown tutor-use history (expected null) no_human control 66 62 0.9250 ns 1.0000 ns -0.0175 -0.37 0.3300 negligible
Extra AI-trained vs. peer-tutor-trained (not preregistered) yes_ai yes_human 67 71 0.0243 * 0.1701 ns -0.3650 -0.70 -0.0006 small

H1a

AI-trained vs. non-AI-trained (yes_ai vs. no_ai)

hypothesis description group1 group2 n1 n2 p p.signif p.adj p.adj.signif effsize conf.low conf.high Cohens’ d
H1a AI-trained vs. non-AI-trained yes_ai no_ai 67 69 0.0072 ** 0.0505 ns -0.4481 -0.79 -0.11 small

H1b

AI-trained vs. unknown AI-use history (yes_ai vs. control)

hypothesis description group1 group2 n1 n2 p p.signif p.adj p.adj.signif effsize conf.low conf.high Cohens’ d
H1b AI-trained vs. unknown AI-use history yes_ai control 67 62 0.284 ns 1 ns -0.1766 -0.53 0.18 negligible

H2

Non-AI-trained vs. unknown AI-use history (no_ai vs. control)

hypothesis description group1 group2 n1 n2 p p.signif p.adj p.adj.signif effsize conf.low conf.high Cohens’ d
H2 Non-AI-trained vs. unknown AI-use history no_ai control 69 62 0.118 ns 0.826 ns 0.2762 -0.08 0.64 small

H3a

Peer-tutor-trained vs. non-peer-tutor-trained (yes_human vs. no_human)

hypothesis description group1 group2 n1 n2 p p.signif p.adj p.adj.signif effsize conf.low conf.high Cohens’ d
H3a Peer-tutor-trained vs. non-peer-tutor-trained yes_human no_human 71 66 0.214 ns 1 ns 0.2245 -0.11 0.59 small

H3b

Peer-tutor-trained vs. unknown tutor-use history (yes_human vs. control)

hypothesis description group1 group2 n1 n2 p p.signif p.adj p.adj.signif effsize conf.low conf.high Cohens’ d
H3b Peer-tutor-trained vs. unknown tutor-use history yes_human control 71 62 0.259 ns 1 ns 0.193 -0.12 0.55 negligible

Note

Non-peer-tutor-trained vs. unknown tutor-use history (expected null) (no_human vs. control)

hypothesis description group1 group2 n1 n2 p p.signif p.adj p.adj.signif effsize conf.low conf.high Cohens’ d
Note Non-peer-tutor-trained vs. unknown tutor-use history (expected null) no_human control 66 62 0.925 ns 1 ns -0.0175 -0.37 0.33 negligible

Extra

AI-trained vs. peer-tutor-trained (not preregistered) (yes_ai vs. yes_human)

hypothesis description group1 group2 n1 n2 p p.signif p.adj p.adj.signif effsize conf.low conf.high Cohens’ d
Extra AI-trained vs. peer-tutor-trained (not preregistered) yes_ai yes_human 67 71 0.0243 * 0.1701 ns -0.365 -0.7 -6e-04 small

Integrity

3-item composite. Agree-disagree scale (0 = Strongly disagree, 6 = Strongly agree).

  • integrity1: “Dr. Wilson has values I share.”
  • integrity2: “Dr. Wilson’s decisions and actions are guided by sound principles.”
  • integrity3: “Dr. Wilson can be counted on to be honest with his patients.”

Summary

training variable n mean sd se ci
yes_ai integrity 67 3.853 1.241 0.152 0.303
no_ai integrity 69 4.807 0.915 0.110 0.220
yes_human integrity 71 4.603 0.947 0.112 0.224
no_human integrity 66 4.152 0.927 0.114 0.228
control integrity 62 4.349 0.961 0.122 0.244

ANOVA

## $output1
## Anova Table (Type II tests)
## 
## Response: integrity
##           Sum Sq  Df F value    Pr(>F)    
## training   38.06   4  9.4128 3.241e-07 ***
## Residuals 333.55 330                      
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## $output2
## # Effect Size for ANOVA
## 
## Parameter | Eta2 |       95% CI
## -------------------------------
## training  | 0.10 | [0.05, 1.00]
## 
## - One-sided CIs: upper bound fixed at [1.00].
## $output3
## # Effect Size for ANOVA
## 
## Parameter | Cohen's f |      95% CI
## -----------------------------------
## training  |      0.34 | [0.23, Inf]
## 
## - One-sided CIs: upper bound fixed at [Inf].

All Planned Comparisons

All 7 planned pairwise comparisons – the 6 preregistered ones plus one added afterward (AI-trained vs. peer-tutor-trained) – Bonferroni corrected within this set of 7; see ‘Planned pairwise comparisons’ section above for the full hypothesis mapping. Individual hypothesis tabs (isolated table + plot per comparison) follow.

hypothesis description group1 group2 n1 n2 p p.signif p.adj p.adj.signif effsize conf.low conf.high Cohens’ d
H1a AI-trained vs. non-AI-trained yes_ai no_ai 67 69 0.0000 **** 0.0000 **** -0.8763 -1.24 -0.55 large
H1b AI-trained vs. unknown AI-use history yes_ai control 67 62 0.0054 ** 0.0378 * -0.4449 -0.79 -0.12 small
H2 Non-AI-trained vs. unknown AI-use history no_ai control 69 62 0.0098 ** 0.0683 ns 0.4880 0.15 0.86 small
H3a Peer-tutor-trained vs. non-peer-tutor-trained yes_human no_human 71 66 0.0090 ** 0.0629 ns 0.4821 0.15 0.84 small
H3b Peer-tutor-trained vs. unknown tutor-use history yes_human control 71 62 0.1470 ns 1.0000 ns 0.2662 -0.05 0.61 small
Note Non-peer-tutor-trained vs. unknown tutor-use history (expected null) no_human control 66 62 0.2660 ns 1.0000 ns -0.2098 -0.56 0.16 small
Extra AI-trained vs. peer-tutor-trained (not preregistered) yes_ai yes_human 67 71 0.0000 **** 0.0001 *** -0.6821 -1.02 -0.35 moderate

H1a

AI-trained vs. non-AI-trained (yes_ai vs. no_ai)

hypothesis description group1 group2 n1 n2 p p.signif p.adj p.adj.signif effsize conf.low conf.high Cohens’ d
H1a AI-trained vs. non-AI-trained yes_ai no_ai 67 69 0 **** 0 **** -0.8763 -1.24 -0.55 large

H1b

AI-trained vs. unknown AI-use history (yes_ai vs. control)

hypothesis description group1 group2 n1 n2 p p.signif p.adj p.adj.signif effsize conf.low conf.high Cohens’ d
H1b AI-trained vs. unknown AI-use history yes_ai control 67 62 0.0054 ** 0.0378 * -0.4449 -0.79 -0.12 small

H2

Non-AI-trained vs. unknown AI-use history (no_ai vs. control)

hypothesis description group1 group2 n1 n2 p p.signif p.adj p.adj.signif effsize conf.low conf.high Cohens’ d
H2 Non-AI-trained vs. unknown AI-use history no_ai control 69 62 0.0098 ** 0.0683 ns 0.488 0.15 0.86 small

H3a

Peer-tutor-trained vs. non-peer-tutor-trained (yes_human vs. no_human)

hypothesis description group1 group2 n1 n2 p p.signif p.adj p.adj.signif effsize conf.low conf.high Cohens’ d
H3a Peer-tutor-trained vs. non-peer-tutor-trained yes_human no_human 71 66 0.009 ** 0.0629 ns 0.4821 0.15 0.84 small

H3b

Peer-tutor-trained vs. unknown tutor-use history (yes_human vs. control)

hypothesis description group1 group2 n1 n2 p p.signif p.adj p.adj.signif effsize conf.low conf.high Cohens’ d
H3b Peer-tutor-trained vs. unknown tutor-use history yes_human control 71 62 0.147 ns 1 ns 0.2662 -0.05 0.61 small

Note

Non-peer-tutor-trained vs. unknown tutor-use history (expected null) (no_human vs. control)

hypothesis description group1 group2 n1 n2 p p.signif p.adj p.adj.signif effsize conf.low conf.high Cohens’ d
Note Non-peer-tutor-trained vs. unknown tutor-use history (expected null) no_human control 66 62 0.266 ns 1 ns -0.2098 -0.56 0.16 small

Extra

AI-trained vs. peer-tutor-trained (not preregistered) (yes_ai vs. yes_human)

hypothesis description group1 group2 n1 n2 p p.signif p.adj p.adj.signif effsize conf.low conf.high Cohens’ d
Extra AI-trained vs. peer-tutor-trained (not preregistered) yes_ai yes_human 67 71 0 **** 1e-04 *** -0.6821 -1.02 -0.35 moderate

Autonomy

3-item composite. Agree-disagree scale (0 = Strongly disagree, 6 = Strongly agree).

  • autonomy1: “Dr. Wilson exercises independent judgment in his clinical decisions.”
  • autonomy2: “Dr. Wilson makes his own decisions rather than relying on external sources to make decisions for him.”
  • autonomy3: “Dr. Wilson takes personal responsibility for the consequences of his decisions.”

Summary

training variable n mean sd se ci
yes_ai autonomy 67 3.540 1.342 0.164 0.327
no_ai autonomy 69 5.101 0.772 0.093 0.185
yes_human autonomy 71 4.183 1.054 0.125 0.249
no_human autonomy 66 4.836 0.837 0.103 0.206
control autonomy 62 4.315 0.982 0.125 0.249

ANOVA

## $output1
## Anova Table (Type II tests)
## 
## Response: autonomy
##           Sum Sq  Df F value    Pr(>F)    
## training   99.85   4  24.119 < 2.2e-16 ***
## Residuals 341.56 330                      
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## $output2
## # Effect Size for ANOVA
## 
## Parameter | Eta2 |       95% CI
## -------------------------------
## training  | 0.23 | [0.16, 1.00]
## 
## - One-sided CIs: upper bound fixed at [1.00].
## $output3
## # Effect Size for ANOVA
## 
## Parameter | Cohen's f |      95% CI
## -----------------------------------
## training  |      0.54 | [0.43, Inf]
## 
## - One-sided CIs: upper bound fixed at [Inf].

All Planned Comparisons

All 7 planned pairwise comparisons – the 6 preregistered ones plus one added afterward (AI-trained vs. peer-tutor-trained) – Bonferroni corrected within this set of 7; see ‘Planned pairwise comparisons’ section above for the full hypothesis mapping. Individual hypothesis tabs (isolated table + plot per comparison) follow.

hypothesis description group1 group2 n1 n2 p p.signif p.adj p.adj.signif effsize conf.low conf.high Cohens’ d
H1a AI-trained vs. non-AI-trained yes_ai no_ai 67 69 0.0000 **** 0.0000 **** -1.4316 -1.82 -1.10 large
H1b AI-trained vs. unknown AI-use history yes_ai control 67 62 0.0000 **** 0.0001 *** -0.6549 -1.02 -0.33 moderate
H2 Non-AI-trained vs. unknown AI-use history no_ai control 69 62 0.0000 **** 0.0001 **** 0.8969 0.53 1.31 large
H3a Peer-tutor-trained vs. non-peer-tutor-trained yes_human no_human 71 66 0.0002 *** 0.0014 ** -0.6830 -1.08 -0.34 moderate
H3b Peer-tutor-trained vs. unknown tutor-use history yes_human control 71 62 0.4580 ns 1.0000 ns -0.1287 -0.47 0.19 negligible
Note Non-peer-tutor-trained vs. unknown tutor-use history (expected null) no_human control 66 62 0.0040 ** 0.0281 * 0.5728 0.22 0.94 moderate
Extra AI-trained vs. peer-tutor-trained (not preregistered) yes_ai yes_human 67 71 0.0002 *** 0.0017 ** -0.5349 -0.90 -0.20 moderate

H1a

AI-trained vs. non-AI-trained (yes_ai vs. no_ai)

hypothesis description group1 group2 n1 n2 p p.signif p.adj p.adj.signif effsize conf.low conf.high Cohens’ d
H1a AI-trained vs. non-AI-trained yes_ai no_ai 67 69 0 **** 0 **** -1.4316 -1.82 -1.1 large

H1b

AI-trained vs. unknown AI-use history (yes_ai vs. control)

hypothesis description group1 group2 n1 n2 p p.signif p.adj p.adj.signif effsize conf.low conf.high Cohens’ d
H1b AI-trained vs. unknown AI-use history yes_ai control 67 62 0 **** 1e-04 *** -0.6549 -1.02 -0.33 moderate

H2

Non-AI-trained vs. unknown AI-use history (no_ai vs. control)

hypothesis description group1 group2 n1 n2 p p.signif p.adj p.adj.signif effsize conf.low conf.high Cohens’ d
H2 Non-AI-trained vs. unknown AI-use history no_ai control 69 62 0 **** 1e-04 **** 0.8969 0.53 1.31 large

H3a

Peer-tutor-trained vs. non-peer-tutor-trained (yes_human vs. no_human)

hypothesis description group1 group2 n1 n2 p p.signif p.adj p.adj.signif effsize conf.low conf.high Cohens’ d
H3a Peer-tutor-trained vs. non-peer-tutor-trained yes_human no_human 71 66 2e-04 *** 0.0014 ** -0.683 -1.08 -0.34 moderate

H3b

Peer-tutor-trained vs. unknown tutor-use history (yes_human vs. control)

hypothesis description group1 group2 n1 n2 p p.signif p.adj p.adj.signif effsize conf.low conf.high Cohens’ d
H3b Peer-tutor-trained vs. unknown tutor-use history yes_human control 71 62 0.458 ns 1 ns -0.1287 -0.47 0.19 negligible

Note

Non-peer-tutor-trained vs. unknown tutor-use history (expected null) (no_human vs. control)

hypothesis description group1 group2 n1 n2 p p.signif p.adj p.adj.signif effsize conf.low conf.high Cohens’ d
Note Non-peer-tutor-trained vs. unknown tutor-use history (expected null) no_human control 66 62 0.004 ** 0.0281 * 0.5728 0.22 0.94 moderate

Extra

AI-trained vs. peer-tutor-trained (not preregistered) (yes_ai vs. yes_human)

hypothesis description group1 group2 n1 n2 p p.signif p.adj p.adj.signif effsize conf.low conf.high Cohens’ d
Extra AI-trained vs. peer-tutor-trained (not preregistered) yes_ai yes_human 67 71 2e-04 *** 0.0017 ** -0.5349 -0.9 -0.2 moderate