Hello
I am testing different groups of T-cells (1,2,3 and 4) for the level of CD69 upregulation after stimulation with a molecule at different concentrations 1,2,3 and 4, see the attached plot. If I want to see if there is a significant difference in CD69 expression between the different T-cells should I do one-way anova at each concentration? or two-way anova would be enough?
Many thanks
Neither.
You should model the (logit!) of %CD69 depending on concentration and on the group of T cells. Thecnically, this is a two-factorial model with a continuous predictor (concentration), a categorical predictor (group of T cells) and their interaction.
It might be ok to use the logit-transformed %CD69 values, however, slightl more appropriate would be a quasi-binomial model or a beta-model (they use the %CD69 values as they are but a logit-link to model the expectation).
In addition to Jochen Wilhelm 's response I note that your graph has no label on the y-axis. This does not bode well for publication.
D. Booth
Good morning David Eugene Booth. I too thought there was no Y-axis label, but it became visible when I clicked on the graph. ;-)
Cheers,
Bruce
i think that you have two variables one is %CD69 and the other one is Types of T Cells concentration in all 4 are same so, you prefer the One way anova with assumptions.
my new video on ANOVA in efficacy Studies using SPSS
https://youtu.be/76L9zS0uJoA
- Badges
- Science topic
- Similar topics
- Biology