Hi
I would want to ask if for a longitudinal study with repeated measures (an outcome compared among three groups of patients with logistic regression, and measured at determinated timepoints: basal, weeks 8, 16 and 24, NOT basal vs week 8 or basal vs week 16,....), I am committing a type I error (or anyother).
Many thanks
Hello David,
There's nothing wrong with a multi-group study having multiple data collection points over some time frame. The real concern here is, what specific research question/s do you wish to be able to address from your study?
As an aside, you probably know that Type I error simply means that a hypothesis test yielded a statistically significant result (suggesting that the null hypothesis is false), when in fact the null hypothesis is true for the target population. Anyone who has ever conducted hypothesis tests in their lifetime might well have (unknowingly) committed one or more type I errors.
This is one reason why replication of studies, especially ones where the consequences of hypothesis testing errors being mistakenly taken as "truth" are substantive, is vital.
Good luck with your work.
What is it exactly that you want to test? We might be able to help you better with some more information...
Hi Koen
The aim is to compare the predictors of clinical response to a treatment in 3 groups of patients (naive, one prior treatment or two prior treatments) using logistic regression at timepoints weeks 8, 16 and 24. The comparison is carried out at each timepoint, that is, I do not compare week 8 with week 16, etc. (I guess in the latter case I should use a repeated measures ANOVA or similar to avoid Type I error).
Many thanks
📷
David Olivares Don't you have some more measurements? at least also the starting situation? 3 is not enough for a logistic regression.
Hi Koen
For the logistic regression I use as DV response yes/no and as IVs, sex, age, smoking habit and previous treatment (naive, one or two previous), and this for each timepoint. Of course the patients at last follow up are a fraction of week 16, and so on. There is not basal measure (it would be necessary to wait at least 8 weeks to get the first response measure).
Many thanks
David
David Olivares Why no measurement at starting time?
Could you share some data? I could try with my software:
www.lerenisplezant.be/fitting.htm
- Badges
- Science topic
- Similar topics
- Correction