Timeline for What to set "intercept factor" and "slope factor" as in model?
Current License: CC BY-SA 4.0
8 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Oct 18, 2023 at 11:12 | answer | added | Christian Geiser | timeline score: 1 | |
| Oct 18, 2023 at 7:39 | comment | added | Terrence |
I think the simsem package provides the flexibility Jeremy prefers. You can find vignettes on the web site about power analysis for model fit/comparison: github.com/simsem/simsem/wiki/Vignette However, if you are interested in power for testing specific parameters, the summaryParam() function provides that for sim() output (see the ?summaryParam help-page example).
|
|
| Oct 18, 2023 at 3:33 | comment | added | Jeremy Miles | You don't need power to plot a slope. You need power to test a null hypothesis. I'm not familiar with semPower but from reading the help file, it appears that you need to use the betaST or betaIT options. I'm not a huge fan of this sort of wrapper function, I prefer to code a simulation myself so that I know that I understand what it is doing. | |
| Oct 17, 2023 at 20:46 | history | edited | Matt W | CC BY-SA 4.0 |
added 21 characters in body
|
| Oct 17, 2023 at 20:21 | comment | added | Matt W | I am planning to estimate a latent growth model to analyse changes in depression, as indicated by PHQ-9 scores collected at 7 intervals. I also plan to use predictors to analyse whether curves are different for different populations within the sample (e.g male vs female, high scores on adhd questionnaire vs low scores), although I have not included this in the power analysis as I am initially just trying to calculate the sample required to plot a slope. I added my current R code to the main body. | |
| Oct 17, 2023 at 20:20 | history | edited | Matt W | CC BY-SA 4.0 |
added 385 characters in body
|
| Oct 17, 2023 at 20:06 | comment | added | Jeremy Miles | This rather depends on the model you are planning to estimate. Can you provide more details? (But essentially it's the same as the slope and intercept in regression. You just have each person measured mutiple times, instead of once, as you do in regression). | |
| Oct 17, 2023 at 19:26 | history | asked | Matt W | CC BY-SA 4.0 |