conditional logit model_help with data interpretation

[senzu]

0
down vote
favorite
I have performed a clogit following the instruction of Aizaki et al.2012 but I am not sure on how to interpret some results that seems quite strange. A brief explanation of what I am trying to do:I am looking at the tourist preference on different mountain habitat in which the environment change from natural to more anthropic one . I run this code

clogit(RES ~ ASC + People_15 + People_25 + Forest_50. + Forest_0. + Lake_50. + 
Lake_0. + Price + strata(STR), data = mont)

enter code here

[HTML][HTML]Call:
clogit(RES ~ ASC + People_15 + People_25 + Forest_50. + Forest_0. + Lake_50. + 
Lake_0. + Price + strata(STR), data = mont)

              coef exp(coef)  se(coef)    z p
ASC            2.22e+01  4.17e+09  3.87e+03 0.01 1
People_15     -1.78e-08  1.00e+00  7.07e+03 0.00 1
Pleople_25    -8.00e-08  1.00e+00  1.08e+04 0.00 1
Forest_50.     1.26e-07  1.00e+00  1.04e+04 0.00 1
Forest_0.     -2.15e-09  1.00e+00  7.40e+03 0.00 1
Lake_50.      -1.28e-07  1.00e+00  7.15e+03 0.00 1
Lake_0.       -8.59e-08  1.00e+00  6.75e+03 0.00 1
Price          1.24e-09  1.00e+00  9.81e+01 0.00 1

Likelihood ratio test=363  on 8 df, p=0
n= 495, number of events= 165
I have this warning message: Warning message: In fitter(X, Y, strats, offset, init, control, weights = weights, : Ran out of iterations and did nerge

What does it mean? Moreover I don't understand why the coef are so small and why the p is 1. Any help would be very helpful!!!Thank you

 

[hlsmith]

Not a regular R user, but you pretty much have null results. Are your categorical variable coded wrong. What happens when you run a summary or look at your top and bottom rows.


Side note I have never ran conditional logistic in R.

 

[senzu]

This the data.frame. It should be similar to the one presented by Aizaki http://eprints.lib.hokudai.ac.jp/dspace/bitstream/2115/57786/1/v50c02.pdf

 head(mont)
  ID   BLOCK QES ALT   RES ASC Peo_25 Peo_15 Peo_5 Forest_100. Forest_50. Forest_0. Lake_100. Lake_50. Lake_0. Price STR
1  1 block 1   1   1 FALSE   0      0      1     0          1         0        0             0            0           1    40 101
2  1 block 1   1   2  TRUE   1      0      1     0          0         0        1             1            0           0    40 101
3  1 block 1   1   3 FALSE   0      0      0     0          0         0        0             0            0           0     0 101
4  1 block 1   2   4 FALSE   0      0      1     0          1         0        0             1            0           0    50 102
5  1 block 1   2   5  TRUE   1      0      0     1          1         0        0             0            1           0    90 102
6  1 block 1   2   6 FALSE   0      0      0     0          0         0        0             0            0           0     0 102

>



and this the summary

 summary(mont)
       ID         BLOCK     QES         ALT         RES          ASC    Peo_25 Peo_15 Peo_5 Forest_100. Forest_50. Forest_0. Lake_100. Lake_50. Lake_0.      Price            STR      
 Min.   : 1   block 1:375   1:99   1      :  1   Mode :logical   0:330   0:463   0:396   0:296   0:329      0:438     0:388    0:363         0:404        0:388       Min.   : 0.00   Min.   : 101  
 1st Qu.: 9   block 2:120   2:99   10     :  1   FALSE:330       1:165   1: 32   1: 99   1:199   1:166      1: 57     1:107    1:132         1: 91        1:107       1st Qu.: 0.00   1st Qu.: 902  
 Median :17                 3:99   100    :  1   TRUE :165                                                                                                            Median :40.00   Median :1703  
 Mean   :17                 4:99   101    :  1   NA's :0                                                                                                              Mean   :37.56   Mean   :1703  
 3rd Qu.:25                 5:99   102    :  1                                                                                                                        3rd Qu.:50.00   3rd Qu.:2504  
 Max.   :33                        103    :  1                                                                                                                        Max.   :90.00   Max.   :3305  
                                   (Other):489

Can you notice something strange?

 

[hlsmith]

So you are using conditional because data are matched? Matched on what and how many per match? I am guessing Block is the matching variable, but I don't see it in the model, though not familiar with the procedure. Well the warning says it converged, so you got estimate, but perhaps the ran out of iterations means it was not able to maximize the algorithm so the estimates are crap. I wonder if there is a way to request more iterations or set-up a burn-in rate (it goes through say 10,000 iterations before trying to converge.

 

[hlsmith]

Sorry for not knowing your answer, but also wondering how good of predictors these are, is it possible these are valid estimates? What type of results do you get if you ran bivariate or non-conditional logistic regression?