Timeline for How to analyze nested data of repeated measurements? ANCOVA or linear mixed model?
Current License: CC BY-SA 4.0
12 events
when toggle format | what | by | license | comment | |
---|---|---|---|---|---|
Sep 13, 2023 at 7:23 | vote | accept | Steven C | ||
Sep 13, 2023 at 7:02 | answer | added | dipetkov | timeline score: 1 | |
Sep 13, 2023 at 3:22 | comment | added | Steven C | There's nothing wrong with what you understand :). First stage perform the task three times without music, and the second perform the task three times with music. The reason for this is that not all music is strictly synchronized with the task in the second stage. In fact, only two stimuli (e.g, A, B) were performed at the same time as the task, while the other two (e.g. C, D) were performed between the two phases. I tried to simplify the design in the problem description, but it seemed to cause some misunderstanding 😂. | |
Sep 12, 2023 at 21:48 | comment | added | dipetkov | Last clarification: Were the two stages (with and without music) always in the same order? Since you called them "first" and "second", I assumed this means: first perform the task three times without music; then perform the task three times with music. But now I'm wondering if the ordering of with/without music was randomized within a session? | |
Sep 12, 2023 at 14:14 | comment | added | Steven C | Well, participants were asked to do the same task with the same difficulty for three times in the same stage. And in a different stage, the task is absolutely the same. The intention to separate a 'whole' block into three small blocks is to avoid the influence of fatigue. Therefore, it seems unnecessary to consider the block factor (just in my opinion). An another question is factor Session seems to be nested in Subject (ID). I'm wondering that Session should be considered as a random effect or still fixed Effect. | |
Sep 12, 2023 at 12:59 | comment | added | dipetkov | I'm not 100% certain I understand the design, sorry. I guessed that blocks of tasks means different levels of difficulty but now it seems it means performed one after another without a break? If it's really the same task performed repeatedly, there is the question of whether subjects are getting better at it. So now I'm also thinking whether you don't need interactions in the model as well. | |
Sep 12, 2023 at 11:35 | comment | added | Steven C | Actually, participants did the same task for each block (All of the tasks are same). Hence, I don't think Block is an important factor in the design and I just take the performance of each Block as a sample. So that for each stage, three samples of performance were considered. Am I doing the right thing? Or should I average the three blocks of each Stage to one sample? | |
Sep 10, 2023 at 14:52 | comment | added | dipetkov |
I think I would combine Stage and Music into one factor, Stimulus with levels {0, A, B, C, D}. Obviously, 0 means "without music". And also I'd introduce another random factor Session. So something like: Performance ~ Task + Stimulus + Session + (Session | Id) . I've added a Task factor because I think you should account for "the three blocks of tasks" in some way.
|
|
Sep 10, 2023 at 14:52 | comment | added | dipetkov | This is an interesting experiment. What about the blocks of tasks? Why isn't Block or Task a factor in the models? Also, it's not clear whether the same tasks were performed on different days. (So there are 4 repetitions of the same tasks performed without music.) | |
Sep 10, 2023 at 13:46 | history | edited | kjetil b halvorsen♦ |
edited tags
|
|
S Sep 1, 2023 at 12:36 | review | First questions | |||
Sep 1, 2023 at 12:38 | |||||
S Sep 1, 2023 at 12:36 | history | asked | Steven C | CC BY-SA 4.0 |