i have the data of 10 companies profits and revenues for last 10 years.
I want to test whether there is any significant difference between those 10 companeis in terms of revenues and profits.
I am thinking of using ANOVA. can i use repeated measures ANOVA or simple ANOVA or any other appropriate inferential statsitical tool.
You have time series data so should make sure that you use relevant techniques. Time series data typically have an internal structure, in your case hidden factors such as economic recessions, government policies, import/export climates, societal trends etc will mean that the data from each company is not independent.
ANOVA requires that the groups are all independent which will not be the case.
NIST have good resources as a starting point:
http://www.itl.nist.gov/div898/handbook/pmc/section4/pmc4.htm
To use ANOVA, there are conditions must be met:
1. The DV should be measured on a continuous level.
2. The DV should be normally distributed
3. You should have adequate sample size in each group {you said you have 10 groups}
4. The sample size calculation depends on the effect size that you will use
5. After you run ANOVA, if there are significant differences, you can run post hoc test to identify the difference between which groups
Good luck
In addition to my answer above, if you want to test the differences among the 10 companies {as a total over 10 years}, then ANOVA is appropriate.
However, if you want to see the differences in every single year {10 times} then you need Repeated measure MANOVA.
considering that Repeated measure ANOVA will be suitable for one company over several times only.
These links may help you
https://statistics.laerd.com/statistical-guides/one-way-anova-statistical-guide.php
https://statistics.laerd.com/spss-tutorials/one-way-anova-using-spss-statistics.php
https://www.researchgate.net/post/What_is_different_between_RM_or_Ordinary_one-way_ANOVA_What_kind_of_experiment_is_suitable_for_each_of_them
ANOVA is not a good technique for solve your problem. Try to model the behavior of the proffits with a mathematical model (polynomial, ARIMA, etc) and use chi square goodness of fit test. Good luck.
Yes, you can do it... You can do difference analysis for revenues and profits among the ten companies by using ANOVA, and difference analysis between two groups by using LSD methods.
- Badges
- Science topic