I have a small dataset collected from SMEs and I need to compare the demographic characteristics between SMEs with access to external finance and SMEs who do not have access to finance on SPSS. The two groups are unequal in the sample size.
Is there a test on SPSS that compares categorical data between two groups despite the difference in sample size?
Regards,
Marouan
Hello Marouan Sghair,
Yes. There is no restriction about having same sample sizes when you compare two or more independent groups. However, some disadvantages exists when you use unequal sample sizes. It is important to declare these flaws in the discussion of your paper if your analysis is used to research purposes.
About the second question, you could use contingency tables (crosstabs) and chi-square test in SPSS to evaluate the association between two categorical variables. But, when running that test, you could also compare the column proportions between groups (z-test for proportions). Don't forget to adjust the p-values if your categories have more than two groups (e.g. Bonferroni correction). Also, keep in mind the assumptions of the chi-square test for independence.
Best regards.
Dear Juan Hernández-Lalinde,
Thank you very much for your detailed response.
According to my previous experience running Contingency tables (crosstabs) or Bivariate correlation (Chi-Square) analyses, neither these tests provide the option to compare two groups. As far as I know, these tests are used to measure the association between two or more categorical variables. Yet, I am looking at a measurement test that can perform comparison analyses between categorical variables from two groups.
In my thesis, I need to look at the differences of "legal structure", "Ownership structure" and "industry sector" between SMEs with access to finance and their counterparts that are financially excluded.
Kind regards,
Hello Marouan,
You are very welcome.
You are right when you say that these procedures does not compare (directly) two or more groups. Indeed, the principal function of Chi-Square test for independence is to evaluate the association of two categorical variables.
Having said that, you can do a z-test for proportions at the same time you are running the Chi-Square test in SPSS. As far as I understand from your post, you have two categorical variables, one with three levels (legal, ownership and industry) and the other with two levels (with and without access). Therefore, you could run a 3x2 crosstabs a perform the Chi-Square test but focusing your analysis in the proportion comparison activating this option in the "cells buttom". Here is a reference from the IBM SPSS help center:
www.ibm.com › sig_tests_colprop_ex
Resultados de la Web
Comparing Column Proportions - IBM
Best regards.
Sorry, I think the previous link does not work. Here is:
https://www.ibm.com/support/knowledgecenter/SSLVMB_23.0.0/spss/tables/sig_tests_colprop_ex.html
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS