Hi altruist, I have count data which is already aggregated for male and female for multiple categories, i.e. physical activity level. So data are in columns in a way like male_walking, female_walking, male_vigorous, female_vigorous. And under each colum, count ( and non-normal) data is there for about 30 cases/parks (rows). Can you suggest which test I should use to compare whether activities vary between gender or not. Really stuck with this issue for last two days. Thanks.
Attached is a specimen data.
I think you can use paired two samples mean difference test if your sample size for each park is same. For instance, if you surveyed 50 person in each park, then count can be used as variable. Then you can find the mean difference between make and female for each category. Details of the test can be find here https://stattrek.com/hypothesis-test/paired-means.aspx
Hello Shahi, do you mean paired t-test ? For count data its not suggested in literatur. Moreover my data is not normal. And my subject is not person, a park. At each park, population data is aggregated by count in each variable.
Hi Shuvo Vaia, if it is not yet solved this, here is an option you may try. If your are not particular interested for every park or is there any significant differences among parks, you can reorganise the count data in one column, but give a group number (i.e. 1 for female, 2 for male) as you are interested in the difference between these two groups. Then run a non-parametric test (most likely Mann Whitney U test), this will bypass your issue of non-normal distribution. I tried with your sample file, there is no significant difference between male and female count in the parks you observed (U = 59.5, p = 0.478).
Hi, Faysal,
Here are two articles which address your question: the first is univariable and involves an ordered attribute, the second is multivariable and involves categorical attributes--either metric is easily handled by either method):
https://odajournal.com/2014/03/27/unioda-vs-students-t-test-comparing-two-migraine-treatments-2
https://odajournal.com/2015/12/09/globally-optimal-cta-model-of-world-war-ii-recruit-training-camp-location-preference
A brief review of maximum-accuracy methods:
https://odajournal.com/2017/04/18/what-is-optimal-data-analysis
- Badges
- Science topic
- Similar topics
- Analytical Chemistry
- Column