Binary logistic regression

[laurenhosking]

I want to create a logistic model to predict the body fat group i.e. lean or bulky. Select the significant variables, check all the criteria to analyse your model and make an sensitivity-specificity analysis. However I’m unsure what my classification or continuous varible would be or if I’d have any

[PaigeMiller]

You have to use independent variables to do any prediction. Do you have independent variables? What are they?

[laurenhosking]

I m given variables including sport (basketball, cross-country, track, softball, volleyball, swim & tennis), height, weight, percentage body fat, BMI (body mass index), sit ups per minute, 40 yard dash, vertical jump, sit reach test, bench press (60lbs and max), leg press (200lbs and max), medicine ball throw and shuttle run.

[PaigeMiller]

So that list contains the independent variables you could put into your model. Sport seems to be the only categorical variable in the list.

[laurenhosking]

So are these variables classed as continuous variables or classification variables?

[PaigeMiller]

Categorical variables (like sport) are also considered to be classification variables (and you would use the CLASS statement when you do the model). The rest seem to be continuous, but since I don't have your data, you have to look at the data and confirm that.

[laurenhosking]

They are independent however I am very confused how to carry out this test

[PaigeMiller]

Here's an example using PROC LOGISTIC

https://documentation.sas.com/?cdcId=pgmsascdc&cdcVersion=9.4_3.4&docsetId=statug&docsetTarget=statu...

[laurenhosking]

Thank you! Would you say to include every one of the independent variables? 

[PaigeMiller]

The answer depends on the data.

[laurenhosking]

Tgis is my data it goes down to 57 entries but 

[PaigeMiller]

No, I didn't mean you should show me the data, I meant you have to analyze the data and see what variables belong in the model.

[laurenhosking]

I also made a variable based on body fat which I know is the response 

[laurenhosking]

@Rick_SAS