Hello all! For the second half of Day 1, we have covered the syntax and result interpretation of the Intercept-only (null) and Level-2 (between-individual) predictor model. Please post any questions you have for today's content here!
Hi Dr. Dimotakis!
Thank you so much for an amazing course. I am already learning a lot and really excited about the course.
I think my questions may go beyond the Day 1 content, but I still wanted to bring them up in case there’s an opportunity to touch on these topics as the class progresses.
1. Are there any differences in analyzing ESM interventions vs. standard ESM surveys? For example, if a participant misses a treatment in the morning, should we remove their evening response from the analysis? Otherwise, their response might resemble a passive control condition. I’m also curious about any methodological considerations that are unique to analyzing intervention-based ESM data.
2. What are some points we need to consider when analyzing dyadic ESM data? (e.g., CFA, main hypothesis testing)
3. In class, we briefly discussed Monte Carlo integration. I was wondering about common practices for cleaning ESM data before analysis. I’ve seen that researchers often include only individuals with at least three days of data—are there other typical criteria? For instance, is it common to remove day-level observations that are missing independent variables?
4. What are some important considerations when the dependent variable is binary?
Looking forward to today's class!
Best,
Sophia
Dear Professor Dimotakis,
Thank you for yesterday's class. Your introduction to MLM/ESM was great.
I have three questions regarding what we saw yesterday. They are pretty basic, but I want to make sure that I get things right:
1. For the intercept-only model, I understood the following:
- γ00 is at the between level and represents the mean of the dependent variable for all individuals in the sample (grand mean; fixed effect).
- U0i is at the between level and represents how much a specific individual departs from γ00 (random effect). This gives us the mean of a specific individual.
- Rti is at the within level and represents how much a specific individual departs from their respective mean on a given day (random effect). This gives us the mean of a specific individual on a given day.
Is my understanding correct?
2. I have a hard time understanding the model with between-level predictors, specifically how a between-level variable (e.g., neuroticism) can explain the (I think) between-level variance of a within-level dependent variable (e.g., anxiety). Could you please re-explain this notion?
3. Could you please re-explain what the added-value of grand-mean centering is?
Thank you very much.
Best,
Constantin
Great questions, all. We'll discuss interventions and data cleaning later in the course, and review the material so far in the beginning of our session today.