Our World Statistics Day conversations have been a great reminder of how much statistics can inform our lives. Do you have an example of how statistics has made a difference in your life? Share your story with the Community!
Choose Language Hide Translation Bar
Highlighted
Level V

## Generalized Linear Model - Question for binomial distribution

Hello everyone,

I have a question regarding the Generalized Linear Model platform when we use a binomial distribution.

The response variable can be specify using two continuous columns as Y in this order: the count of the number of successes, and the count of the number of trials.

What if, for some reasons, the number of successes is greater than the number of trials for a few rows (e.g. successes= 60, trials  = 50)? Is it automatically corrected during the analysis (i.e. the number of trials is updated -> 60)? Or should it be absolutely corrected before launching the analysis because it can lead to mistakes during the calculations?

1 ACCEPTED SOLUTION

Accepted Solutions
Highlighted
Staff

## Re: Generalized Linear Model - Question for binomial distribution

JMP accepts the data in only one way: first the number of event, then the total trials. There is no correction. JMP assumes that your data is correct.

Learn it once, use it forever!
3 REPLIES 3
Highlighted
Level VII

## Re: Generalized Linear Model - Question for binomial distribution

I can't answer whether JMP will automatically correct the number of trials for all rows, but obviously you have an issue with data collection or entry.  If you have more successes than trials, what other data collection/entry errors are in the data set?  I would correct these before proceeding with any analysis.

Highlighted
Staff

## Re: Generalized Linear Model - Question for binomial distribution

JMP accepts the data in only one way: first the number of event, then the total trials. There is no correction. JMP assumes that your data is correct.

Learn it once, use it forever!
Highlighted
Level V

## Re: Generalized Linear Model - Question for binomial distribution

Thanks for your answers  @statman and @markbailey . Of course I totally agree with you: the cleaning step should be the very first step before doing any analysis. In some very particular cases the total number of trials can be slightly updated in comparison to the initial number provided. I just wanted to know, by curiosity, how it was handled from a technical point of view and if the correction was automatically done or not. But definitely I agree that it is better to clean first

Article Labels