PCSI Gender Intervention Study Pilot Analyses
Descriptive Statistics (reported by condition)
control:
Table continues below vars n mean sd median trimmed interesting 1 14 4.357 1.499 4.5 4.333 trustworthy 2 14 3.786 1.424 4 3.75 easy 3 14 5.643 1.598 6 5.833 believable 4 14 4.5 1.345 4.5 4.5 biased 5 14 3.286 1.267 3.5 3.25 general_interesting 6 14 4.571 1.453 4.5 4.583 general_trustworthy 7 14 5.286 0.7263 5 5.333 general_easy 8 14 3.357 0.9288 3 3.333 general_believable 9 14 5.357 1.008 5.5 5.417 general_biased 10 14 3.5 1.401 3.5 3.417 comprehension_errors 11 14 0.5 0.6504 0 0.4167 gender_identify_cis 12 14 38.93 29.72 33 35.17 gender_identify_trans 13 14 110.9 70.2 140 112.7 gender_identify_comp 14 14 -72 67.29 -69.5 -68.58 Table continues below mad min max range skew kurtosis interesting 1.483 2 7 5 0.3193 -0.9789 trustworthy 1.483 2 6 4 0.0515 -1.375 easy 1.483 2 7 5 -1.028 -0.2711 believable 0.7413 2 7 5 0 -0.8387 biased 2.224 2 5 3 0.1334 -1.797 general_interesting 0.7413 2 7 5 0.1484 -0.9106 general_trustworthy 1.483 4 6 2 -0.411 -1.186 general_easy 0.7413 2 5 3 0.3875 -0.8902 general_believable 0.7413 3 7 4 -0.6828 -0.05134 general_biased 2.224 2 6 4 0.234 -1.517 comprehension_errors 0 0 2 2 0.7787 -0.6555 gender_identify_cis 21.5 3 120 117 1.254 1.369 gender_identify_trans 71.16 2 199 197 -0.2878 -1.72 gender_identify_comp 103 -186 1 187 -0.1975 -1.599 se interesting 0.4006 trustworthy 0.3806 easy 0.4272 believable 0.3593 biased 0.3385 general_interesting 0.3882 general_trustworthy 0.1941 general_easy 0.2482 general_believable 0.2695 general_biased 0.3743 comprehension_errors 0.1738 gender_identify_cis 7.943 gender_identify_trans 18.76 gender_identify_comp 17.98 experimental:
Table continues below vars n mean sd median trimmed interesting 1 14 4.286 1.541 4 4.25 trustworthy 2 14 3.929 1.141 4 3.917 easy 3 14 5.857 0.8644 6 5.833 believable 4 14 4.571 1.284 5 4.667 biased 5 14 4.071 0.8287 4 4.167 general_interesting 6 14 5.143 1.406 5 5.25 general_trustworthy 7 14 4.857 1.406 5 4.917 general_easy 8 14 4.5 1.653 4.5 4.5 general_believable 9 14 4.714 1.139 5 4.833 general_biased 10 14 4.143 1.099 4 4.333 comprehension_errors 11 14 0.1429 0.3631 0 0.08333 gender_identify_cis 12 14 32.07 21.85 36 31.42 gender_identify_trans 13 14 34.21 15.02 36 33.67 gender_identify_comp 14 14 -2.143 14.77 0 -2.5 Table continues below mad min max range skew kurtosis interesting 1.483 2 7 5 -0.09328 -1.044 trustworthy 1.483 2 6 4 0.415 -0.6854 easy 1.483 5 7 2 0.2437 -1.713 believable 1.483 2 6 4 -0.4505 -1.08 biased 0 2 5 3 -0.8682 0.4378 general_interesting 1.483 2 7 5 -0.6948 -0.329 general_trustworthy 1.483 2 7 5 -0.3836 -0.9862 general_easy 2.224 2 7 5 0 -1.306 general_believable 1.483 2 6 4 -0.6397 -0.1735 general_biased 1.483 1 5 4 -1.547 2.011 comprehension_errors 0 0 1 1 1.826 1.455 gender_identify_cis 8.896 0 72 72 0.2873 -0.5909 gender_identify_trans 0 3 72 69 0.2661 1.537 gender_identify_comp 0 -36 36 72 0.3354 2.172 se interesting 0.4118 trustworthy 0.305 easy 0.231 believable 0.3431 biased 0.2215 general_interesting 0.3759 general_trustworthy 0.3759 general_easy 0.4417 general_believable 0.3043 general_biased 0.2938 comprehension_errors 0.09705 gender_identify_cis 5.839 gender_identify_trans 4.014 gender_identify_comp 3.947
T-test examining difference between conditions for all variables
Article specific questions – we want these to be as close as possible
No significant differences here. However, the biased question is marginal in that people appear to think the scientists in the experimental article are more biased than those in the control article. This makes a lot of sense to me and I am okay with this difference I think.
Interesting
| Test statistic | df | P value | Alternative hypothesis |
|---|---|---|---|
| 0.1243 | 25.98 | 0.902 | two.sided |
| mean in group control | mean in group experimental |
|---|---|
| 4.357 | 4.286 |
Trustworthy
| Test statistic | df | P value | Alternative hypothesis |
|---|---|---|---|
| -0.2929 | 24.82 | 0.772 | two.sided |
| mean in group control | mean in group experimental |
|---|---|
| 3.786 | 3.929 |
Easy to read
| Test statistic | df | P value | Alternative hypothesis |
|---|---|---|---|
| -0.4412 | 20.01 | 0.6638 | two.sided |
| mean in group control | mean in group experimental |
|---|---|
| 5.643 | 5.857 |
Believable
| Test statistic | df | P value | Alternative hypothesis |
|---|---|---|---|
| -0.1438 | 25.94 | 0.8868 | two.sided |
| mean in group control | mean in group experimental |
|---|---|
| 4.5 | 4.571 |
Biased
| Test statistic | df | P value | Alternative hypothesis |
|---|---|---|---|
| -1.942 | 22.41 | 0.06479 | two.sided |
| mean in group control | mean in group experimental |
|---|---|
| 3.286 | 4.071 |
General questions – we also want these to be as close as possible but mostly for the purposes of confirming the samples between conditions are similar
For the most part, there are no differences here. However, there is a significant difference in how easy people think articles about research are to read (in general) between conditions. Not exactly sure what to make of this difference and I don’t think it is the most critical.
Interesting
| Test statistic | df | P value | Alternative hypothesis |
|---|---|---|---|
| -1.057 | 25.97 | 0.3 | two.sided |
| mean in group control | mean in group experimental |
|---|---|
| 4.571 | 5.143 |
Trustworthy
| Test statistic | df | P value | Alternative hypothesis |
|---|---|---|---|
| 1.013 | 19.47 | 0.3234 | two.sided |
| mean in group control | mean in group experimental |
|---|---|
| 5.286 | 4.857 |
Easy to read
| Test statistic | df | P value | Alternative hypothesis |
|---|---|---|---|
| -2.256 | 20.47 | 0.03517 * | two.sided |
| mean in group control | mean in group experimental |
|---|---|
| 3.357 | 4.5 |
Believable
| Test statistic | df | P value | Alternative hypothesis |
|---|---|---|---|
| 1.582 | 25.62 | 0.126 | two.sided |
| mean in group control | mean in group experimental |
|---|---|
| 5.357 | 4.714 |
Biased
| Test statistic | df | P value | Alternative hypothesis |
|---|---|---|---|
| -1.351 | 24.61 | 0.189 | two.sided |
| mean in group control | mean in group experimental |
|---|---|
| 3.5 | 4.143 |
Comprehension Errors – We want these to be as close as possible as well
No significant difference here, though it is trending. People made slightly more comprehension errors for the control conditions. It is important to note here that the errors were very low in general and this is driven by just a handful of people who made errors.
| Test statistic | df | P value | Alternative hypothesis |
|---|---|---|---|
| 1.794 | 20.39 | 0.08769 | two.sided |
| mean in group control | mean in group experimental |
|---|---|
| 0.5 | 0.1429 |
Gender Development Sliders – We don’t always want these to be the same!
Cisgender gender identification age
No significant difference here, which I think is okay since we are mostly targeting transgender gender development age relative to cis.
| Test statistic | df | P value | Alternative hypothesis |
|---|---|---|---|
| 0.6956 | 23.87 | 0.4934 | two.sided |
| mean in group control | mean in group experimental |
|---|---|
| 38.93 | 32.07 |
gender_identify_trans”, “gender_identify_comp”
Transgender gender identification age
A big difference here! People very clearly are stating that trans kids identify their gender at much later ages in the control group compared to the experimental group.
| Test statistic | df | P value | Alternative hypothesis |
|---|---|---|---|
| 3.999 | 14.19 | 0.001288 * * | two.sided |
| mean in group control | mean in group experimental |
|---|---|
| 110.9 | 34.21 |
Composite gender identification age (a negative number here means that people think trans kids identify their gender later than cis kids)
Another big difference in the expected direction!
| Test statistic | df | P value | Alternative hypothesis |
|---|---|---|---|
| -3.794 | 14.25 | 0.001917 * * | two.sided |
| mean in group control | mean in group experimental |
|---|---|
| -72 | -2.143 |