I am a bit new to multilevel and mixed-effects modeling. I have a dataset, where I have sensor measurements throughout the day for 4 weeks on N participants. I have some prompts on their smartphones 60 - 70 times a day. All I want to do is to see if the sensor measurement before the prompt differs from the measurement after the prompt. I think paired sample t-test is a good option, but with high intensity of data, I don't think it meets all the assumptions of the test. Any suggestions or leads will be really helpful.
Asked
Active
Viewed 391 times
0
-
Welcome to CV, RforResearch! This isn't an answer (and hence I make a comment), but when thinking longitudinally, there may be lots of ways to think about differences. See my answer to the question Which stats tests to use? Bee flight activity quantified during an eclipse – Alexis May 30 '18 at 22:03
-
Thanks Alexis. I am going through. trying to understand the nuances. – RforResearch May 30 '18 at 22:04
-
@Alexis, I think the key difference is that I have multiple participants, with multiple data points within a day for 4 weeks. – RforResearch May 30 '18 at 22:16
-
And what specific differences are you interested in? (My point was not that I answered your question, but that you might find the advice to think about specific differences meaningful to you worthwhile. :) – Alexis May 30 '18 at 22:28
-
Hey Alexis, I just want to check if the sensor measurement n seconds before the prompt is significantly different from the same measurement n sec after the prompt. Note that sensor is measuring all the time. – RforResearch May 31 '18 at 16:36