Skip to Main Content
Table 5. Predictive Bias for Ethnicity
  Holdout Data  Training and Holdout Combined 
  b β 95% CI p  b β 95% CI p 
Depression predicted 3.30 .26 .01 .51 .039  4.66 .74 .64 .84 < .001 
 ethnicity 0.08 .08 -.33 .50 .693  0.02 .02 -.16 .20 .841 
 predicted X ethnicity 0.72 .06 -.44 .56 .821  0.43 .07 -.13 .27 .502 
Anxiety predicted 2.67 .27 -.03 .57 .082  4.45 .76 .64 .89 < .001 
 ethnicity 0.16 .18 -.26 .62 .422  0.11 .12 -.07 .31 .228 
 predicted X ethnicity 1.33 .13 -.47 .73 .663  1.32 .23 -.02 .48 .074 
Post-Traumatic Stress predicted 3.50 .28 .00 .55 .051  4.58 .75 .65 .85 < .001 
 ethnicity 0.09 .34 -.10 .77 .126  0.03 .11 -.07 .29 .235 
 predicted X ethnicity 2.41 .19 -.36 .74 .499  0.77 .13 -.08 .33 .226 
Anger predicted 2.89 .17 -.11 .45 .233  5.43 .75 .65 .86 < .001 
 ethnicity -0.13 -.17 -.59 .26 .440  -0.01 -.01 -.20 .17 .875 
 predicted X ethnicity -2.19 -.13 -.69 .43 .651  0.99 .14 -.08 .35 .212 
  Holdout Data  Training and Holdout Combined 
  b β 95% CI p  b β 95% CI p 
Depression predicted 3.30 .26 .01 .51 .039  4.66 .74 .64 .84 < .001 
 ethnicity 0.08 .08 -.33 .50 .693  0.02 .02 -.16 .20 .841 
 predicted X ethnicity 0.72 .06 -.44 .56 .821  0.43 .07 -.13 .27 .502 
Anxiety predicted 2.67 .27 -.03 .57 .082  4.45 .76 .64 .89 < .001 
 ethnicity 0.16 .18 -.26 .62 .422  0.11 .12 -.07 .31 .228 
 predicted X ethnicity 1.33 .13 -.47 .73 .663  1.32 .23 -.02 .48 .074 
Post-Traumatic Stress predicted 3.50 .28 .00 .55 .051  4.58 .75 .65 .85 < .001 
 ethnicity 0.09 .34 -.10 .77 .126  0.03 .11 -.07 .29 .235 
 predicted X ethnicity 2.41 .19 -.36 .74 .499  0.77 .13 -.08 .33 .226 
Anger predicted 2.89 .17 -.11 .45 .233  5.43 .75 .65 .86 < .001 
 ethnicity -0.13 -.17 -.59 .26 .440  -0.01 -.01 -.20 .17 .875 
 predicted X ethnicity -2.19 -.13 -.69 .43 .651  0.99 .14 -.08 .35 .212 

Note.b refers to the unstandardized estimate; β refers to the standardized estimate; 95% CI refers to the 95% confidence interval around β. In these analyses, ethnicity corresponds to identifying as Hispanic/Latino (nholdout = 24, ncombined = 68) or not (nholdout = 198, ncombined = 590); ethnicity was contrast coded (Not Hispanic/Latino = -0.5; Hispanic/Latino = 0.5) so positive values for ethnicity indicate a higher intercept for Hispanic/Latino participants (relative to Non-Hispanic/Latino participants) and positive values for the predicted by ethnicity interaction indicate a steeper slope for Hispanic/Latino participants (relative to Non-Hispanic/Latino participants).

Close Modal

or Create an Account

Close Modal
Close Modal