I have read some articles using animal models where researchers just eliminate the death subject, but what if those subjects have been measure sometimes and could provide important data to my statistic analysis.
I would like to know which is the best statistic test I can perform in this case.
Hi Sergio
You may consider to use linear mixed effects models or generalized estimating equations (GEE). Observations will contribute to the results up to the point where measurements stopped.
Do not remove dead animals from the analysis. You are correct- they provide information, and it may be valuable. Mixed effects/ GEE is a good approach. If appropriate, consider time to death as a response variable.
Dear Sergio
1. The t-test for comparing two related samples need to be developed if you have two samples. Please refer to Chapter 10, Dennis Howitt and Duncan Cramer (2008).
2. Or ANOVA for repeated measures if you have 3 or more than 3 groups. Please, refer to Chapter 18 of the above references.
Regards,
Zuhair
I may have been wrong in the manner I wrote my question, I know about the t-Test and ANOVA tests, BUT in order to perform those, I need to have the same number of observations in my groups. That is my problem here, in the first measure the article says there were 10 subjects, in the next timestamp there were only 7 and finally there were only 3. The variable of measurement is numeric and continuous. I want to compare the means for this measurements, BUT I don't want to eliminate any subject, I want to take into account their death because it seems to be a benefit in the variable measured prior to their death.
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Modeling
- Analyses of Variance
- ANOVA