Message/Author |
|
|
Dear Madam or Sir, I am trying to do a three-level analysis with diary data where L1= occasion (3 times a day) L2= day (5 days) , and L3 = Person (44). I got the following error message: *** ERROR One or more between-level variables have variation within a cluster for one or more DAY clusters. Check your data and format statement. Between Cluster ID with variation in this variable Variable (only one cluster ID will be listed) DAYFLOW 5 DFLOW_SD 5 I suspect this is because I have 5 days and day-level variables can vary between those days. Do you have any idea how I can resolve this? My syntax was this: USEVARIABLES ARE OccFlow DayFlow DayCP DFlow_sd Day Person x1mean x2mean; CLUSTER IS Person Day; BETWEEN = (Day)DayFlow DFlow_sd (Person) x1mean x2mean; Missing are all (999); DEFINE: x1mean = Cluster_mean(DayFlow); x2mean = Cluster_mean(DFlow_sd); CENTER DayFlow DFlow_sd (GROUPMEAN Person); ANALYSIS: TYPE = THREELEVEL; ESTIMATOR = ML; MODEL: %WITHIN% OccFlow; %BETWEEN Day% OccFlow; DayCP ON DayFlow; DayCP ON DFlow_sd; %BETWEEN Person% DayCP DayFlow DFlow_sd x1mean x2mean; OUTPUT: SAMPSTAT Tech1 Tech8 CINTERVAL; Best wishes, Jakob |
|
|
You can't put variables on the BETWEEN list if the values vary for cluster members. Be definition, all cluster members must have the same value on a variable. |
|
|
Thank you very much for your reply. Could you then please explain to me how one could model a three-level longitudinal multilevel model in Mplus where you have occasions nested in days nested in persons? Could I have two "within" lists where both are allowed to vary but are still treated as distinct levels? Best wishes, Jakob |
|
|
How the BETWEEN and WITHIN options work for TYPE = THREELEVEL are described on pages 571 and 568-69 of the user's guide. |
|
Back to top |