I have conducted a choice bioassay：two plants (plants with same goowing days, control and treatment, respectively) were cultivated in a pot apart from each other with 5 cm，then 10 insects were released in the middle of the two plants. The number of insetcts on each plant was calculated at 1, 2, 4, 8, 24 h after the release of the insects. During the bioassay, inescts with non-choice or dead are excluded from analyses. The experiment was repeated 10 times.
I want to test the preference of the insects on the tested plants at different time-point, so which statistical method is suitable? Independent samples t-test or paired samples t-test ? Thanks in advance!
Your data is binomial, so use a binomial model.
You can analyze the numbers of insects on one of the plants. A proportion of more than 0.5 would indicate a preference, less than 0.5 would indicate a disfavor.
Did you use the same two plants in each repetition? If so, the results would refer to exactly these two plans you used. The data would not give you any hint how this might look if you took two other plants (i.e. there is no information how this experiment would replicate!).
Hi Jochen, thanks for your reply.
Yes, the same two plants were used in each repetition. Thus ten replicates were set for each time-point, and I just compare the preference at different time-points. I wonder whether the results at different time-points were independent or not, for the same 10 insects were used along the 24 h. That's important to choose independent samples t-test or pared samples t-test. Is that right?
t-test is still a bad choice. Your data remains being counts. The dependence between time-points may be modelled using a random effect (so would get a generalized mixed linear model of the binomial family).
- Badges
- Science topic
- Similar topics
- Biological Science
- Zoology
- Invertebrate Zoology
- Entomology