Message/Author 


I am doing an LGC analysis in which I have 18 respondents. I have taken their responses to the same questionnaire (with 10 questions, 5 loading on one factor, 5 loading on another) over 9 weeks. Thus I have 90 columns and just 18 rows in my data file. When i try to run the model to see how the two underlying factors vary over time, it gives me the fatal error of "number of observations not greater than the number of variables". Is the number of respondents (18) an unsolvable problem? Or is there some way in which I can make this work? I need to know because I cant go and get more responses that easily. 


The only thing I can think of is changing your data from a wide to a long format which is shown in Example 9.16 for a growth model. There are limitations to this approach for a CFA model in that all residual variances are assumed to be the same and also factor loadings. You probably need more observations or a simpler model. 

HeSook Kim posted on Tuesday, April 12, 2011  3:08 am



My raw data's cases are 2844. But i don't know why the following result told me that the number of observations are 1422. INPUT INSTRUCTIONS title: growth model data: file is C:\Users\Public\Documents/stress study.dat; variable: names are stsum1 stsum2 stsum3 stsum4 ax_d4 stpar1 stgr2 stper3 stape4 stmat5 strpa2 strgr2 strper2 strape2 strmt2 strpa3 strgr3 strper3 strape3 strmt3 strpa4 strgr4 strper4 strape4 strmt4 axde1 axde2 axd3 sexw1; usevariables are stpar1 strpa2 strpa3 strpa4 sexw1; missing are all(999); model: i s  stpar1@0 strpa2@1 strpa3@2 strpa4@3; i on sexw1; s on sexw1; I S ; [I S ]; I with S; OUTPUT: SAMP; STAND; MOD(5); INPUT READING TERMINATED NORMALLY growth model SUMMARY OF ANALYSIS Number of groups 1 Number of observations 1422 Number of dependent variables 4 Number of independent variables 1 Number of continuous latent variables 2 Observed dependent variables Continuous STPAR1 STRPA2 STRPA3 STRPA4 Observed independent variables SEXW1 Continuous latent variables I S 


It seems that you may have more variables in the NAMES list than you have columns in your data set causing two records to be read for each observation. If you can't resolve this, please send the input, data, and your license number to support@statmodel.com. 


I am about to run an Autoregressive Latent Trajectory model. since my unit of analysis is Brand, that's not easy to get too many observations. I was wondering what is the minimum sample size or the number of observations to run the model? 


I would say between 3050 as a minimum. A Monte Carlo study could be used to investigate this further. 

Back to top 