Timeline for Repeated Measures & Standar Error of Measurement (SEM)
Current License: CC BY-SA 4.0
20 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Mar 16 at 22:49 | comment | added | mdscience | I think that the question if there is some variant of repeated measures for a single subect... | |
| Mar 15 at 20:03 | comment | added | Jeremy Miles | Post (example) data in the question. I don't understand this description. | |
| Mar 15 at 18:27 | comment | added | mdscience | In this case I have 10 rows where put 2 devices (A and B), how I have 5 different settings and 2 devices I have 10 rows, 5 for each device. An in columns I have the combination of S1_Y, S1_Z, S2_Y, S2_Z.... Do you mean that it´s could be wrong? | |
| Mar 15 at 16:30 | comment | added | Jeremy Miles | Those aren't your subjects. Those are your predictors. Subjects are the things (typically people) that your measures are repeated over. | |
| Mar 15 at 12:10 | history | edited | mdscience | CC BY-SA 4.0 |
added 27 characters in body
|
| Mar 15 at 11:37 | history | edited | mdscience | CC BY-SA 4.0 |
edited title
|
| Mar 15 at 11:35 | history | edited | mdscience | CC BY-SA 4.0 |
added 470 characters in body
|
| Mar 15 at 9:26 | comment | added | mdscience | Positions, positionsXrooms, rooms, roomsXdevices, positionXrooms, positionXroomXdevices. But really I´m interested in positionXroomXdevices. | |
| Mar 15 at 9:20 | comment | added | Jeremy Miles | Where are your subjects that it is within? How many do you have? | |
| Mar 15 at 7:49 | comment | added | mdscience | Thinking in this, could be interesenting compare results with Anova post-hocs and differences explain by the range of SEM? How can relation both conceptually? SEM is based on reliability index and standard deviation. | |
| Mar 15 at 7:45 | comment | added | mdscience | Repeated measures factors: 1x5 (position) + 1x2 rooms. Between subjects (Devices 2). Why is not a repeated measures Anova in this case? | |
| Mar 15 at 3:29 | comment | added | Jeremy Miles | If the question ended before the sentence "I want to know:" and just said "How do I analyze this" it would be relatively straightforward to answer. | |
| Mar 15 at 3:28 | comment | added | Jeremy Miles | This is not a repeated measures anova. If it is repeated measures, your N is 1, and you cannot analyze anything. Remove that or no one can answer the question. | |
| Mar 15 at 3:27 | comment | added | Jeremy Miles | There are a lot of questions here, and a lot to unpack, so it's hard to write an answer. One thing to consider is that reliability and anova are very closely linked. Reliability is estimated using the ICC - the intra-class correlation, which is based on the sums of squares, anova is also based on the sums of squares. If your means are the same in the different conditions, your reliability (ICC) will be zero. | |
| Mar 14 at 18:07 | comment | added | mdscience | But from another, more practical perspective, I would like to analyze the behavior of the 2 devices in rooms Y and Z. How? Going back to the beginning, I would like to know if the change between positions is real, that is, if this change is larger than the changes that are observed for records made with the same position. This is why I had thought about the SEM and the MDC. However, I don't know if it could actually fit this approach. | |
| Mar 14 at 18:07 | comment | added | mdscience | On the one hand, I would like to report a value that characterizes the differences between the 6 test-retest measures for each of the 5 positions. Right now I have thought about the standard deviation, but I don't know if with this design I could opt for some other measure of reliability. Right now I have done a repeated measures anova in which I am only interested in the posthoc to compare the same position of the same method between different rooms. | |
| Mar 14 at 16:26 | comment | added | Jeremy Miles | I've tried to make the question clearer, please check I haven't got anything wrong. I don't think this is repeated measures anova. I don't quite understand your question: You have three categorical predictors, and one continuous outcome. You can therefore test 3 main effects (plus post-hoc analysis), three two-way interactions, and 1 three-way interaction. What do you want a confidence interval of? Finally,I"m not sure that anyone without substantive knowledge of the area can say whether the difference is relevant on a practical level. | |
| Mar 14 at 16:23 | history | edited | Jeremy Miles | CC BY-SA 4.0 |
Tried to clarify.
|
| S Mar 14 at 15:49 | review | First questions | |||
| Mar 14 at 16:27 | |||||
| S Mar 14 at 15:49 | history | asked | mdscience | CC BY-SA 4.0 |