Why regression model coefficients can be different when they are plotted/displayed in the forest plot

Viewed 26

I am trying to carry out a subgroup analysis by plotting the mixed effects model results in a forest plot using metafor in r. The forest plot shows the mixed effect model coefficients (estimates and confidence interval), however, the estimates and confidence interval values in the forest plot are different than the model results.

For example the model shows the estimates and confidence intervals as follows: g1 = 0.19 CI = [0.1228 ; 0.2679] and g2 = 0.22 CI = [0.1704 ; 0.2778] while in the forest plot the values differ: g1 = 0.24 CI = [-0.14 ; 0.61] and g2 = -0.54, CI = [-0.67 ; -0.41] Could anyone explain why this can happen?

0 Answers
Related