Hello everybody,
For my master theses I'm doing research on high-growth firms. I have a sample of approximately 1100 firms from 6 countries. These companies are assigned to 1 of the 19 industries they operate in (services, manufacturing, retail etc.).
Now I want to test the hypotheses: "that high-growth firms are overrepresented in the service industry" (so more firms in the service sector compared to other industries)
Firstly, I want to do it for the whole sample and afterwards for every country seperatly.
Does somebody know how to test this in SPSS?
Thanks in advance
David
Hello David,
One simple way to evaluate this question would be via the chi-square goodness of fit test. In SPSS, that would be: Analyze/Nonparametric/One-sample. The null hypothesis is that frequencies are equal across service sectors.
Good luck with your work!
Hi, David,
The Chi square looks relevant to me as well, but with 1,100 firms grouped in 19 industrial categories (nominal) and ranked into low-, medium-, and high-growth categories (ordinal) or even just into low- and high-growth categories, you will have to be careful to avoid cells with less than 5 companies. This is likely to be an even bigger problem when you test each country individually. You may need to recode some of the industry classifications to avoid this problem, and to do so you will need a good justification from the literature.
Also keep in mind that the Chi square default (frequencies equal across industries) will need to be weighted by the number of companies in each industry. You might avoid this by converting frequencies to percentages in your Chi square test.
I also agree with the chi square position. Alternatively, univariate analysis using odds ratio will determine effect size. With the use of service industry as the reference variable, it will be possible to effect sizes with 95% confidence interval. The result with which ones have higher or lower odds than the reference variable and decision on the statistical significance of these ratios.
- Badges
- Science topic
- Similar topics
- Econ
- Manufacturing