Hi All, I am building an SEM model using PROC CALIS PATH analysis. My dependent variable is HIV Status i.e either positive or negative. Some of my other variables are categorical/binary e.g condom use (0/1), pregnant_ever (0/1) In the model I put together, it seems to run but with the following error that suggests a poor fit "WARNING: Standard errors and t values might not be accurate with the use of the Moore-Penrose inverse." I was wondering if someone can advise me on the best way to fix this error...or point me to some relevant literature fitting SEM PATH with a categorical dependent variable. 26 ods results on; 27 ods graphics on; 28 proc calis data=new_sib method=wls plots=pathdiagram; 29 path 36 37 IPV <--- violence, 38 client <--- violence, 39 police <--- violence, 40 childhood <--- violence, 41 42 ptsd <--- mental, 43 depression <--- mental, 44 49 binge_drinking <--- mental, 50 hiv <--- binge_drinking, 51 hiv <--- age_firstsw, 52 hiv <--- education, 53 hiv <--- born, 54 hiv <--- condom_use, 56 hiv <--- pregnant_ever; 57 2 The SAS System 17:13 Saturday, February 15, 2020 65 66 pathdiagram 67 diagram=standard; 68 fitindex on(only) = [chisq df probchi rmsea cn srmsr bentlercfi agfi] noindextype; 69 70 run; WARNING: 74 of 3005 observations in data set WORK.NEW_SIB omitted due to missing values. NOTE: Convergence criterion (GCONV=1E-8) satisfied. NOTE: The Moore-Penrose inverse is used in computing the covariance matrix for parameter estimates. WARNING: Standard errors and t values might not be accurate with the use of the Moore-Penrose inverse. NOTE: PROCEDURE CALIS used (Total process time): real time 2.01 seconds cpu time 1.14 seconds 71 ods graphics on; 72 ods results off; Regards Kennedy Otwombe ------ This email has been scanned for spam and malware by The Email Laundry.
... View more