Syntax error when indicating "equalslopes" in proc logistic model equation

87 Views Asked by At

I've been trying to run a proc logistic stepwise regression model using an ordinal outcome. Because I am trying to account for the assumption of proportional odds, several of my variables have uneven slopes. So, in my code, I am indicating both equalslopes and unequalslopes, however I continue to get this syntax error.

ERROR 22-322: Syntax error, expecting one of the following: ;, ABSFCONV, AGGREGATE, ALPHA, BACKWARD, BEST, BINWIDTH, BUILDRULE, CL, 
              CLODDS, CLPARM, CODING, CONVERGE, CORRB, COVB, CT, CTABLE, DETAILS, DSCALE, EXPB, FAST, FCONV, FIRTH, GCONV, 
              HIERARCHY, INCLUDE, INFLUENCE, IPLOTS, ITPRINT, L, LACKFIT, LINK, MAXITER, MAXSTEP, NOCHECK, NODUMMYPRINT, NOFIT, 
              NOINT, NOLOGSCALE, OFFSET, OUTROC, PARMLABEL, PCORR, PEVENT, PL, PLCL, PLCONV, PLRL, PPROB, PSCALE, RIDGING, 
              RISKLIMITS, ROCEPS, RSQUARE, SCALE, SELECTION, SEQUENTIAL, SINGULAR, SLE, SLENTRY, SLS, SLSTAY, START, STB, STEPWISE, 
              STOP, STOPRES, TECHNIQUE, UNEQUALSLOPES, WALDCL, WALDRL, XCONV.  
ERROR 76-322: Syntax error, statement will be ignored.

I am sure I am putting in my code correct, even when I do not state any specific variables for either the equalslopes option and the unequalslopes option, I continue to receive the same message. Here is my code below.

proc logistic data=upper_limit;
class Profitrank dealer_state_id dlr_zip RUCA2 area_description/ param =  ref;
model profitrank =  num_booked num_approved num_apps bk_conversion_rt appr_rt num_new num_used Percentage_of_New Average_Vehicle_Age average_dti 
average_pti average_revolving_balance average_revolving_balance_rate average_public_records average_inquiries average_inquiry_6_months 
average_open_rev_trades average_term average_major_derog average_minor_derog average_open_install average_scorecard average_fico_score average_app_risk_score
average_LTV average_consumer_rate average_truist_rate Dealer_state_ID RUCA2/link=clogit selection=stepwise SLE=0.10 SLE=0.10 EQUALSLOPES=(appr_rt average_pti average_public_records
 average_open_rev_trades average_major_derog average_minor_derog average_open_install average_fico_score average_app_risk_score RUCA2) unequalslopes details maxiter=500;
run;``` 

Any help with this would be greatly appreciated!

1

There are 1 best solutions below

0
On BEST ANSWER

You need to use options that work with the version of SAS you are using. The EQUALSLOPES option was added in SAS/STAT 14.1 which is SAS 9.4m3 released in 2015.

Note that SAS is a subscription license. Which means you have already paid for the right to use the newest version. Get someone at your company to install the newest version of SAS. You might also need to use a newer version of Enterprise Guide to take advantage of the new features.