Explanatory variables in percentages

[Adamu Idris]

Dear respected Members

Pls, I am using xtlogit in my model. Due to large coefficients, a famous Professor in this forum advised me to multiply the affected variables by 100. I followed the thread titled "Percentages versus proportions as explanatory variables" that was posted on 23 Apr 2014, 16:19 by Tellvictor.

Pls, could anyone help me on how to solve the following issues with regards to explanatory variables in percentages?
1. How do I calculate odd ratio interpret them ratios when my explanatory variables in percentages (for example 55.32 instead of 0.55.32)? Am aware on how to calculate and interpret them when they are in proportion.

2. How do I interpret marginal effects when my explanatory variables in percentages (or example 55.32 instead of 0.55.32)? Am aware on how to calculate and interpret them when they are in proportion.

Your suggestions are highly welcome and appreciated

Adamu

[Adamu Idris]

Pls in addition to my earlier post on Explanatory variables in percentages, How do I calculate odd ratio and interpret them when the variable is in log transformation and lastly, how to calculate and interprete marginal effects when the variable is in log transformation.

Thank you all

[Phil Bromiley]

You didn't get a quick answer. You'll increase your chances of a useful answer by following the FAQ on asking questions - provide Stata code in code delimiters, readable Stata output, and sample data using dataex. Also, don't ask too many questions in one posting.

I'm afraid I don't work in odds ratios. Others may know better than I, but I suspect you need to do margins for different values of your x's.

[Dacosta Xavier]

Hi, I am very new to this group and at the early stage using Stata. I found the above question similar to what I am facing at the moment with a slightly different. I have no clue to interpreting the result from listcoef, help percent because all predictor variables (Gender, Education, Marital Status, and Age) are negative. This is related to Logistics Regression (binary). Appreciate if any of you could help. Regards.

[Marcos Almeida]

With regards to the first query, I suggest to have the variable written as a proportion. Then, it will be up to you to describe the results as a percentage modification, or not.