I am looking for a way to analyze repeated measurements.
I have data of subjects who have varying number of measurements (from 2 to 10) over fixed time periods (Week 1, week 2, etc)
The subjects are divided into two groups (A and B).
What I want to check in my analysis is:
1. For all subjects together, does Week1 differ from Week 2, Week 2 from Week 3, etc
2. Are the changes over time in group A different from the changes in group B? I expect from my data for the group A to have higher deltas between timepoints in comparison with group B.
Some issues with my data:
1. some values are missing for most individuals;
2. plotting the data over time reveals non-linear trend: There is a trend of increasing values during the first 3 weeks, and from weeks 3 to 10 - a decrease of the values;
3. measurements of different patients are quite variable (think body weight - from 50kg to 120 kg) - a wide variation
- Repeated measures ANOVA is not a good option as I understand because it cannot deal with missing data.
- Linear mixed models (LMM) seems to be a good fit, as it allows for missing data, and allows entering the subjects as a random factor (so each subject has their own intercept).
The problem I see is the slope - it is not linear.
I know SPSS does not have a non-linear mixed effects model at all and I am not skilled in any of the other statistical programs. Is there any other solution for my data or workaround to use LMM?
Thank you, that does sound good.
Could you recommend me where could I read about *how* to achieve this?
Both the transformation of time variable as well as fitting a parabolic model to LMM?
Hello Erika Salciute-Simene. Like Jochen Wilhelm
, I was also thinking about a quadratic fit as one possibility. But I also wondered if a piecewise fit with a knot at 3 weeks might make sense. When you plot the data, do the fits look fairly linear in each of the segments (weeks 1-3 and 3-10)? Or do they look a bit curvilinear? Thanks for clarifying. If it looks very linear in both segments, I would expect a piecewise model to give a better fit. (I guess it would be more defensible, though, if there was some a priori reason to expect a change in the slope at week 3 specifically.)PS- SPSS does not have a built in (native code) way of generating the variables needed for piecewise regression, but IIRC, there is an extension command for this (probably programmed in Python).
ERIKA
Patient Program Durations (2 weeks etc ) may be confounded with underlying patient pathologies. Eating disorder patients (an example that fits your time scale and body mass interest, though I have related examples with bariatric interventions as well as hospital post surgical stays and Covid durations) may involve a group that responds well to a short term more educational Intervention and a group with more complex psychiatric issues). incidentally, the. reversal in the latter weeks of your data Would not be surprising with eating disorder populations where there are complex Dx -related issues).
In any case, With the data at hand, you might — over the common 2weeks— be advised to first conduct a RANOVA with Subjects divided into Duration Groups (Certainly 2Weeks vs >2Weeks possible, but ideally more than 2 Groups) — this would provide some initial look at interactions with durations and the AvsB intervention.
Subsequent analyses might be best then chosen given results of this first wave analysis.
Regards,
ALVAH
PS You may select a transformation for mass via classical plot of cell SD vs Means (eg, Bittner AC (2022). Analysis of variance ANOVA) Assumptions Review:… The XXIVth Ann. ISOES Conf.. 28-33.
Likewise extended analysis of error terms (useful for understanding, interactions) may be found at Bramwell et al.(1992). Repeated measures Analysis… InterJ. Indus Eng., 10, 185-197
- Badges
- Science topic
- Similar topics
- Mathematics