how to calculate the experience diversity of panels?

[Fred Lee]

Every expert panel has about 5 experts, every expert has different experience such as service, R&D, management. How to calculate the experience diversity of panels? Maybe something like Herfindahl index?

Here is the tips for calculating method:Criscuolo, P., Dahlander, L., Grohsjean, T., & Salter, A. (2017). Evaluating novelty: the role of panels in the selection of R&D projects. Academy of Management Journal, 60(2), 433–460. https://doi.org/10.5465/amj.2014.0861

* Example generated by -dataex-. To install: ssc install dataex
clear
input int(expID panelID) byte(exp_state exp_manager exp_RD exp_adm exp_teacher exp_designer exp_lawyer exp_reporter exp_medical exp_engineer exp_foreign exp_service)
137 1 . . . . . . . . . . . .
 70 1 0 1 0 0 0 0 0 0 0 0 0 0
141 1 . . . . . . . . . . . .
 41 1 . . . . . . . . . . . .
 35 1 . . . . . . . . . . . .
  8 2 0 0 1 0 0 0 0 0 0 1 0 0
112 2 . . . . . . . . . . . .
111 2 . . . . . . . . . . . .
187 2 0 1 0 1 0 0 0 0 0 0 0 1
 98 2 . . . . . . . . . . . .
170 3 . . . . . . . . . . . .
254 3 1 1 0 0 0 0 0 1 0 0 0 0
  1 3 . . . . . . . . . . . .
149 3 . . . . . . . . . . . .
174 3 0 0 0 0 0 0 0 0 0 0 0 0
 73 4 . . . . . . . . . . . .
 99 4 . . . . . . . . . . . .
121 4 0 1 1 0 0 0 0 0 0 1 0 0
110 4 . . . . . . . . . . . .
263 4 0 1 0 0 0 0 0 0 0 0 0 0
end

[Fred Lee]

Could anyone help me figure out,please? Thanks vety mcuh!

[Nick Cox]

I don't see that this has anything to do with the Herfindahl measure (used earlier than Herfindahl by many others). I haven't read your linked article.

As I understand it, you are concerned with the range of experience of panels, and is that anything but the number of 1s in each observation in your byte variables?

See

help egen

for row functions.

[Phil Bromiley]

Actually, since experience is 0/1 with lots of zeros, you can calculate the double summation based simply on the number of 1's in the observation. You can use egen with rowtotal. The number of pairs from a given number of entities has a standard equation solution- I think it is N(N-1)/2. [But check this with some examples] That is, with three 1's, you want the number of distinct pairs that can be formed. Likewise for four 1's etc. Then, you use egen by panel to get panel totals.

[Fred Lee]

I don't see that this has anything to do with the Herfindahl measure (used earlier than Herfindahl by many others). I haven't read your linked article.

As I understand it, you are concerned with the range of experience of panels, and is that anything but the number of 1s in each observation in your byte variables?

See

help egen

for row functions.

Thanks, Nick! All the exp_* variables are dummy variables, indicating that whether the experts have experience of management, RD etc.
I need to first capture the different experiences each expert have, and then calculate the diversity of each panel.

[Fred Lee]

Actually, since experience is 0/1 with lots of zeros, you can calculate the double summation based simply on the number of 1's in the observation. You can use egen with rowtotal. The number of pairs from a given number of entities has a standard equation solution- I think it is N(N-1)/2. [But check this with some examples] That is, with three 1's, you want the number of distinct pairs that can be formed. Likewise for four 1's etc. Then, you use egen by panel to get panel totals.

Thanks, Phil! I think rowtotal will miss some information when capture experts' experience. For example, expert 1 has management and RD experience while expert 2 has service and reporter experience, the sum of their experience both equal to 2, but actually their experience is quite different. Could you please explain more?

[Fred Lee]

One thought is that calculating the "distance" or "diversity" of each pair of expert in the same panel and then average it to measure the diversity of panels. Does it work?