Hello. I am using the user-written command "KHB" (Karlson/Holm/Breen) to run a mediation analysis (Stata 15, SE). Using data from the web just as an example (below), I have a binary outcome (union), a continuous predictor (wage), and three mediators (south industry tenure).

sysuse nlsw88

khb logit union (wage) (south industry tenure), disentangle

When I run this code, the KHB program provides a nice breakdown of the unique contributions of each mediator to the overall indirect effect (via the "disentangle" option).

Here is where I am having difficulty: I am able to run the command line below that includes generating bootstrapped standard errors, but when I add the ", disentangle" option to the end of the command line, the program will not separate each of the indirect effects.

sysuse nlsw88

bootstrap _b[Diff], reps(50): khb logit union (wage) (south industry tenure), disentangle

I reached out to one of the authors of the program regarding this issue (who was incredibly helpful with a related issue), but I have not heard back regarding the question posed here. Does anyone have any thoughts on how to handle this issue?

Many thanks in advance!