I am looking for a statistical method, which would give a value, that represents a fit (in one way or another) between my results compared to previous works' results of the very same kind. To illustrate what I mean:
+------------+------+--------+
| | Male | Female |
+------------+------+--------+
| Country 1 | 38 | 22 |
| Country 2 | 59 | 47 |
| Country 3 | 55 | 35 |
| Country 4 | 62 | 39 |
| Country X* | 47 | 26 |
+------------+------+--------+
The first four rows are the previous results from 4 different countries. The numbers represent a % of all correct answers among a certain category of population.
The fifth row represents results, that I got for an identical questionnaire (in a different country).
Any idea which statistical method would suit in this case?
This is not such a straight forward problem. Dr. Google came up with some interesting links when using the phrase "comparing multiple proportions". One such that got mentioned multiple times is the Marascuillo procedure, see here: https://www.itl.nist.gov/div898/handbook/prc/section4/prc474.htm
You might check the Meta Analysis literature. This is an approach to finding general results from many studies. It is based on getting the effect size from each study and analyzing these. It might be useful for you to compare the results of your studies using their effect sizes. One of the best references I have found on this is:
Glass G. V (1976). Primary, secondary, and meta-analysis of research. Educational Researcher. 5 (10): 3–8.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Applied Statistics
- Survey Design
- Questionnaire