Multilevel regression with 2 levels in SPSS: Review of examples from Chapter 3 of Heck et al. (2014)

[@turkmura81]

perfect combination of all Heck et al. (2014) demos. Thank you, sir!

[@mikecrowson2462]

Glad you liked it, Murat! Cheers :)

[@bilalasghar9616]

Thank you so much for such a wonderful explanation of Multi- level analysis or Multi-level Modelling...i was for searching for such videos since a long time!!

[@mikecrowson2462]

Hi Bilal, thank you for visiting! I'm glad you found this helpful. Cheers!

[@bilalasghar9616]

@@mikecrowson2462 Yes Sir!! It is quite helpful...

[@philippegoldin5049]

Thank you so much for such a clear and concise elucidation of LMM.

[@tazdingo4524]

Sir. You are a great teacher. thank you for everything.

[@helenacarvalho4031]

Thank you for your very clear explanations

[@DB-in2mr]

Hi Mike , thanks again for this piece of statistical art and explanations. On simple question: in your null model (checking for random intercepts only and no predictors) I can easily calculate the L1 variance estimate (66.550655), but I do not find the right computation procedure to compute L2 variance estimate (10.6422). I tried some procedures but I do not understand how to calculate the variance between the school intercepts. I know that the table "Estimates of Covariance Parameters" contains it. Thanks a lot!

[@tegegntadesse9685]

thank you so much for your clear explanation.

[@yangwang4574]

Thank you very much. This is very helpful! A follow-up question about the significant interaction - what analyses should be the next step after observing a significant level 2 by level 1 interaction? Thank you very much!

[@learn5081]

Model 4: how do we interpret the random slopes of SES? In reporting the results, do we just focus on the fixed effect of SES?

[@pataloise-young4215]

Great video. Why would one include BOTH a student-level SES variable and a school-level SES variable (which appears to only be a mean of the individual data) in the model?

[@janna8864]

Thanks Mike, your videos are so clear and helpful! I see you have a video about generating robust standard errors in SPSS - is it possible to estimate those for multilevel models?

[@lukuseula]

Hi Mike, Thanks for the video. I was wondering, how to calculate effect size in mixed models linear analysis. So if there is statistical significance in the test, how to calculate, for example, Cohens d. In your video you say that students who are higher with respect to ses, are prediced to score higher with the respect of math achievment score. Estimate is 3.47. So how to calculate, what is the effect size? I found this kind of solution, but could you tell me how to do it in your example analysis? Cohen's d = (difference in means) / (pooled standard deviation)

[@ikikennogeensik]

Nice info for analysis of multilevel data but how to get the right frequencies for multilevel data descriptives?

[@md.mahiuddinsabbir6125]

Hello Mike, Many thanks for your tutorials. I am very new to HLM and I have a question. My results indicate that ICC based on Model 1 is 0.010150305 and Intercept [subject = Industry] Variance is insignificant (p=0.545 which is 0.273 even if considered as one-tailed), while residual is significant. Now, should I continue further with that particular dataset to perform HLM? Or should I just infer that the response (dataset) does not show evidence to perform HLM as this concern is raised by my manuscript's reviewer? Your response is much appreciated. Thanks!

[@bringmesunshinetrio6913]

I'm also interested in this. My results are ICC is 0.03 on Model 1 and the intercept for school variance is not significant. So does this mean I should not do multi-level modelling for this outcome variable?

[@footballsailors2328]

Hi Mike, Thanks. I thing the excel lead files that create the plots are unavailable.

[@MarckMamaniFloresxD]

Thanks Teacher..!!!

[@mikecrowson2462]

You are very welcome, Marck!

[@eboamuah6811]

Hello Mike, Why do I get a lower ICC when the level 2 variables are added to the model? How do I explain that?

[@mikecrowson2462]

The ICC captures the amount of variation in the dv accounted for by the clustering variable. [see www.theanalysisfactor.com/the-intraclass-correlation-coefficient-in-mixed-models/]. When you add level 2 predictors that account for between-cluster differences in the dv, it necessarily will go down. [Of course, adding level 2 predictors that do not account for between cluster variation will result in the ICC remaining the same.] Since the variance of the level 2 residuals is a component of the computation of the ICC, if this number decreases, then so does the ICC.

[@yiyuz3733]

Hi Mike, first and foremost many thanks for your video which is very very helpful! I followed the step, but I did face a problem that stated: For my intercept covariance parameter I get this message, "This covariance parameter is redundant. The test statistic and confidence interval cannot be computed." Any help would be greatly appreciated!

[@yeeunchoi3770]

Thanks a lot!

[@akramhossain9576]

very good

[@MarckMamaniFloresxD]

RSTUDIO FOR STATISTICS? :D

[@mikecrowson2462]

Hi Marck, let me see what I can do on this. Cheers! :)

[@MarckMamaniFloresxD]

@@mikecrowson2462 Yes, Rstudio is the software currently used in statistical analysis and other applications. : D

[@shivanithakur8938]

Sir, Can you pl provide your email ID ?? Pl ... I want to ask something related to my analysis.

[@maymann477]

Thanks a lot!!