Timeline for Why does standardizing the DV in multilevel modeling change the pattern of results?
Current License: CC BY-SA 4.0
6 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Oct 18, 2020 at 17:59 | vote | accept | Brigadeiro | ||
| Oct 17, 2020 at 15:56 | comment | added | Erik Ruzek | There are many good articles on this topic. See researchgate.net/publication/… and researchgate.net/publication/… | |
| Oct 17, 2020 at 15:54 | comment | added | Erik Ruzek | It depends on whether you do any centering of the IV. Imagine that you person mean center the IV, then at the occasion level, the coefficient is telling you the average within-person DV change associated with a 1 unit increase in the IV. That could be a 1-unit deviation from the person mean, but recognize that it is a linear coefficient so it is a 1-unit change at any point in the range of the IV. If you do not person-center the IV (and do not include the person mean at level 2), then the coefficient is an uninterpretable blend of the within- and between-person association. | |
| Oct 16, 2020 at 4:02 | comment | added | Brigadeiro | thanks for your response - very helpful. Quick follow-up question: given that the model implicitly person-mean centers the DV (with the intercept term), effects then are interpreted as change in the IV associated with change in the DV related to each person's mean, correct? | |
| Oct 14, 2020 at 21:18 | history | edited | Erik Ruzek | CC BY-SA 4.0 |
deleted 53 characters in body
|
| Oct 14, 2020 at 21:01 | history | answered | Erik Ruzek | CC BY-SA 4.0 |