Help - Logistc regression in SAS Enterprise

#1
Dear all, i would like to ask your help because i need to be able to reproduce in SAS enterprise the same results as these ones from the book "An introduction to statistical learning", Springer, P.136:

Coefficient Std. error Z-statistic P-value
Intercept −10.8690 0.4923 −22.08 <0.0001
balance 0.0057 0.0002 24.74 <0.0001
income 0.0030 0.0082 0.37 0.7115
student[Yes] −0.6468 0.2362 −2.74 0.0062

However, no matter the combinations i tried my results are always different.
Could someone please let me know what can be done?
Note: data set attached

Thanks you very much,
David
 

hlsmith

Not a robit
#2
Your attachment did not open and people typically have reservations about opening mystery files. If it is the Gareth book, it appears the datasets are available online, which dataset is it?


Also, what syntax does Enterprise generate for you? Please post it.
 

hlsmith

Not a robit
#3
Alright, it appears to be the Credit file. Pg 136 is not available in the Google book version, what is the dependent variable? Gender or married?
 
#5
This is what SAS enterprise generates:


Analysis of Maximum Likelihood Estimates
Parameter DF Estimate Standard Wald Pr*>*ChiSq
Error Chi-Square
Intercept 1 -11.3572 1.3941 66.3645 <.0001
balance 1 0.00565 0.000701 64.8208 <.0001
income 1 0.000022 0.000026 0.7055 0.4009
student No 1 0.159 0.3911 0.1654 0.6842
 

hlsmith

Not a robit
#6
You still aren't telling my which variable it is predicting.


Two things, you won't get the exact same results because they are using R correct? Also, if you stare at you results long enough, you can see an error you made. It becomes a skill, staring at inanimate objects until you get tire. Give it a try and let me know what the full output looks like, DV and sample size - so I can be sure I downloaded the right dataset.
 
#7
I can see that one of the mistakes might be at the student independent variable, however i do not know how to fix it.

Response Profile
Ordered default Total
Value Frequency
1 0 33
2 1 967

Number of Observations Read 1000
Number of Observations Used 1000

R-Square 0.1195 Max-rescaled R-Square 0.4746

Testing Global Null Hypothesis: BETA=0
Test Chi-Square DF Pr*>*ChiSq
Likelihood Ratio 127.2646 3 <.0001
Score 116.1234 3 <.0001
Wald 65.436 3 <.0001
 

hlsmith

Not a robit
#8
Independent variable predict the dependent variable. So balance, income, and student are your independent variables predicting??


I am not going to download the file, since the version online with the book appears to be truncated. I don't use Enterprise, so I can't help there. what does the book present for R code? Yes, I would switch reference groups for student and see what happens.