[Limdep Nlogit List] Calculation of simulated uptake in NLOGIT/Limdep

[Jason Ong]

I recently received an interesting comment from a reviewer after I used
NLOGIT to calculate my simulated uptake.
https://url.au.m.mimecastprotect.com/s/grJsCK1DvKT8pyqkRhMfKu5RluU?domain=limdep.com

"To calculate uptake, one first calculates the utility of each alternative
and then includes this in a logit probability formula. But the parameter
estimates from effects coded or dummy variables do not represent the
utility of a particular attribute but rather the relative utility. With
effects coded variables this is relative to the mean of the levels of that
attribute. With dummy coded variables this is relative to the omitted
attribute level. So for the uptake calculation, if the uptakes are
estimated using the parameter estimates these should be interpreted as
uptake relative to the mean of each of the attributes, not uptake compared
to no app (see Hu et al, 2022). Therefore, the effect of 'image only' is
the effect on the utility of the app of providing 'image only' data
compared to the average utility of the different ways you provide data, not
compared to providing no data at all.


Any advice on the interpretation of simulated uptake results from NLOGIT?


Best regards,

*Jason Ong*
PhD, MMed(Hons), MBBS, FAChSHM, FRACGP
*X/Twitter:* @DrJasonJOng
*Link Tree:* https://url.au.m.mimecastprotect.com/s/bTKiCL7EwMfN8DP2zSqhBuyR5uQ?domain=linktr.ee

*Professor*
Monash University
London School of Hygiene and Tropical Medicine
NHMRC Emerging Leadership Fellow

*Sexual Health Physician*
Melbourne Sexual Health Centre (MSHC)
Head, HIV/STI Economics and Health Preference Research (MSHC)
co-Head, MSHC HIV/STI Clinical Trials Unit

*co-Editor-in-Chief*
Sexual Health https://url.au.m.mimecastprotect.com/s/A6z0CMwGxOtxmy5YnIJiMu8qCoB?domain=publish.csiro.au