Hi,
I am doing a state wise comparison of nutrition related data of 2 different surveys and I am using their state wise estimates. I have raw data for only 1 data set and I understand that I might have to transform variables to get estimates from this one as per the other.
While I know that I could do geographical mapping for this, what sort of analysis can I undertake while using summary (percentages of different indicators) of 2 different surveys within a broadly same time period (findings presented in 2015, 2016 for other).
I want to compare statistically significant differences of these estimates in states..
Secondly what tests/checks used to check the comparability/ robustness of these 2 surveys findings?
Please help.
Hello Marian,
Since the purpose and development of the two surveys may not be the same, and since the target groups of respondents and how they were sampled (and rates of non-response) may not be the same, any comparisons are likely not to be a strict, "apples to apples" assessment.
You can, of course, run comparisons within a state (or region or nationally) of the proportion of respondents in what you consider to be comparable categories. For example, obesity rate within a state. A simple binomial test will help indicate whether a difference is non-chance (for example, see this on-line calculator: https://www.medcalc.org/calc/comparison_of_proportions.php). The caveat from the first paragraph still applies!
Good luck with your work.
First you decide your hypothesis, then i will able to tell you which test is applicable.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics