There is a phase 2 trial of a novel weight loss drug, comprising a 50mg dose, 100mg dose and placebo, with n=30 for each group.
The primary endpoint is absolute weight loss at 3 months.
I am new to the subject of statistical analysis, however i have proposed to conduct an ANCOVA with weight loss as the dependent variable, treatment group as independent variable and baseline weight as a co-variate.
I have then proposed to conduct a post hoc analysis to see if there are any differences for gender.
I was wondering whether this would be appropriate, or whether an ANOVA, with gender and treatment groups as fixed factors would be more suitable.
many thanks
I would include gender as a fixed factor in the model, what gives you the possibility to look for an interaction of dose and gender.
Using weight as continuous covariable is fine, here I would also suspect some possible interaction with dose.
I suggest NOT to think in terms of "ANOVA", "ANCOVA" and "post-hoc tests". Instead you should focus on the linear model (all these procedures mentiones above are just addressing some special aspects of linear models!). Use the models to estimate (not to test) the effects, and interpret the estimated effect sizes.
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Modeling
- Analyses of Variance
- ANOVA