I have a bioassay arena (Petri dish) with two wells in opposite quadrants. I can put different stimulus source (s) of interest into two wells. 5 larvae were released at center of Petri dish and count the number of larvae present in both wells after an hour. I replicated each bioassay 20 times. Would t-test be the good for analysis?
Jhalendra,
your response variable is a count, viz. number of larvae in each well out of five. Perhaps, logistic regression would be appropriate, although some assumption on data distribution should be considered. Please read the following topic in RG:
https://www.researchgate.net/post/Which_is_the_most_appropriate_method_to_analyze_counts?ev=prf_cont_person_add_comment_question_target
I think that, Mann Whitney Test is better than T test because you have a non parametric variable
You are running dual choice experiment. Here the number of subjects (your larvae) should be in even number (say 6 instead of 5), yet i would recommend 10 larvae per replicate. and reduce the number of replicate to 10.
You can now go for both T-test (independent) if you use separate batches of larvae for each replicate. Also u can go for Chi-square analysis for 50:50 distribution.
If you are using same larvae for all the following replicate use repeated measures analysis.
Good luck.
Thank you very much to all for your inputs/ideas/recommendations.
I am using fresh and actively crawling larvae only (
If you test the normality of th data and it conforms to the requirements of a parametric test, then I see no reason why t-test should not be used
Distributions of counts data tends to be normal with the increase of numbers (events), but this is not the case (to my opinion), thus data could be a binomial, poisson or negative binomial distribution. Parametric tests might be inappropriate very likely.
T-test suits your hypothesis and experimental setup!! Why not use it ??
Thank you all for your suggestions/ideas. How about using Wilcoxon Signed- Rank Test? I reviewed some papers and statistical base of this analysis and realized that this would be the best option to analyse these data.
Dear Dr. Jhalendra Rijal ,
You can rely on the Ki-square test or multiple ANOVA analysis or ANCOVA.
This question will soon be for your question, it may help you:
https://www.researchgate.net/post/Which_statistical_test_should_I_use_to_analyze_data_from_several_dual_choice_tests_with_insects
- Badges
- Science topic
- Similar topics
- Biological Science
- Biota
- Eukaryota
- Fungi