Quantcast
Channel: Statalist
Viewing all articles
Browse latest Browse all 72764

Include categorical variable to a logit model with market shares (percentages) as dependent variable

$
0
0
Dear Statalist members,

in my research I try to come up with an airline market share model. This model should be implemented as multinomial logit model and as nested logit model and the results should be compared. I am using Stata 13.1 for Windows computer.
The dependent variable is market share (market_share) between 0-1. My independent variables are ticket prices one day prior (TP_day_prior) departure and two month (TP_2m) prior departure and the frequency (freq). The data is collected for every day between April-June 2015 and for three different airlines (two legacy and one low-cost carrier) that are given the numbers 1-3 in the variable airline_code.
Since my dependent variable is continuous between 0-1 I tried to use .glm for the mulitnomial logit model:
. glm market_share TP_day_prior TP_2m freq, family(binomial 1) link(logit) vce(robust)
However this does not take into account the categorical variable airline_code nor the different days that can be seen as different cases. Does someone know how to include the categorical variable and the variable case in the model? I think this is necessary because the passenger have to choose between 3 airlines with different attributes each day and the market share for each day adds up to 1.
Furthermore does someone know if Stata is able to create a nested logit model (nest 1: legacy carriers; nest 2: low-cost carrier) with this data?

I would be really grateful if someone is able to answer my question.

Kind regards and thanks for the effort,

Britta Wilken

Viewing all articles
Browse latest Browse all 72764

Trending Articles



<script src="https://jsc.adskeeper.com/r/s/rssing.com.1596347.js" async> </script>