Two-level multilevel model using SPSS (chapter 3 v5); cross-level interactions in HLM

This is the fifth of several videos illustrating how to carry out multilevel modeling involving two levels. Specifically, it illustrates the testing of cross-level interactions in HLM.

[joseamaldonadomartinez]

I watched your entire series (chapters 3 v1 - v5) and also have Heck et al. (2014) book. Thanks Mike Crowson for helping me learn about HLM

[Spiriiiit]

Greetings, Mike Crowson. First of all thank you very much for your insightful Videos about HLM, they are a great help. However, i have some difficulty in translating the analyses you did there onto the datastructure i aquired in my study. Basically, i have a Panel repeated measures design wherein i conducted a trust game. There were 66 subjects who participated and 10 timepoints per subject, where they had to decide on how many gamepoints they want to send. Now my hypothesis is that due to the egoistic behaviour of the receiver in the first round (which i programmed in advance), there should be a decline of points from round 1 to round 2 (which can already be seen in the descriptive statistics of my dataset). In the remaining 9 rounds however, there should be a slow growth in amount of points spent due to consistently cooperative behaviour by the receiver (who could decide to keep all points to himself or else equally divide them). So basically i have a two level multilevel model, where the individuals who participated are on level 2 and the 10 measurement units are nested within those on level 1. Did i understand that correctly? Now i tried to compute a nullmodel and afterwards models with fixed intercept, fixed slope and also fixed intercept and random slopes, i also tried to incorporate predictor-variables that i got on level 2 to see whether the variance can be explained by that but i have great problems in doing so in SPSS. Basically i got the same error message which you warn of in one of your videos: that the iteration has been stopped and convergence could not be achieved. Thus SPSS did not give me any estimates for the random Slope UN (2, 2). Also i have written two new variables into my dataset which model the 2 Slopes: 0 1 2... (for the Decline from Round 1 to Round 2) and then 0 0 1 2 3... (for the Increase after Round 2 towards the End). Could you maybe tell me, at which point i can utilize these in the mixed model anlyses i have to conduct? Do you have any Idea how i could fix the problems i have described and also how to conduct the analyses?

[amarzya]

Hello Mike! Could you explain how you would add a covariate (for example a baseline measurement) in the equation?

[tahseenbhutto]

I have a question, I have 115 leaders who filled a survey for 307 employees' creative behavior and those 307 employees filled up a survey of these 115 leaders about leadership survey. Do we need to use simple SEM or multilevel SEM analysis here?

[leoniehofmann]

Maybe you can explain, how it would work in SPSS if theres a variable that we want to control for, a simple covariate like "gender"?

[khadeejamunawar]

Could you also elaborate how you made this graph on excel? Thanks in advance