I have a dataset where the same survey was sent out to a community of individuals over the course of seven waves. I will note that not all participants started at the same wave, nor did all participants complete all waves.
One question in the dataset asked participants if a particular event had occurred in their life with a binary answer format (yes/no). I want to look at if there are significant differences within participants on certain dependent outcome variables before vs. after this event occurred. I have used Excel coding to identify ~90 participants from this dataset who stated the event had not occurred in their life in their first wave, but stated the event had taken place in at least one subsequent wave. This means not all participants stated the event took place at the same wave.
Therefore, I'm wondering what would be the best course of analysis? I originally was suggested to do a repeated measures ANOVA, but I cannot neatly select two within subject variables when participants have different pre/post timepoints. It was also suggested I dummy code each case for whether the case occurred before or after the event for each participant, which makes sense, but I am unsure how to use this grouping variable in a repeated measures ANOVA (this task is also tedious by hand, but doable).
You need a longitudinal model in a logistic regression. Start with a spaghetti plot to see what is going on.. logistic regression and programs can be found through Google search. Most packages contain this method.
Full details on the model can be found in the Davidian course notes shown in the attached screenshot. The Harrell course notes give you a worked example in R.. Best wishes, David Booth
Multilevel regression (HLM, random coefficient) modeling may be a practical option for you to deal with the time-varying nature of the event IV. In the HLM, your event variable could be treated as a time-varying (Level-1) predictor (IV; dummy-coded) along with Wave (Time) as a continuous Level 1 predictor. Additional time-invariant predictors (if any) could be considered at (Level 2), for example, gender or age at Wave 1.
Multilevel modeling also neatly addresses the issue of unequal numbers of time points between individuals and / or unequal spacing of time within and between individuals.
Christian Geiser Thank you for these helpful suggestions! For anyone's reference I found the following tutorial for running HLM models in R:
https://towardsdatascience.com/hierarchical-linear-modeling-a-step-by-step-guide-424b486ac6a3
Also, would the attached example data structure make sense for this sort of analysis at level 1? Practically, I would use dummy coded event occurrence in column 4 as the "time-varying (Level-1) predictor (IV; dummy-coded)"; wave in column 2 as my "continuous Level 1 predictor"; continuous dv in column 5 as outcome
It looks good except that I would code "event" as 0 and 1 and "wave" as 0, 1, 2, 3, .... Doing so will facilitate your interpretation of the regression coefficients (in particular, the intercept) because then both independent variables have a meaningful zero point.
- Badges
- Science topic
- Similar topics
- Statistics
- Survey