In two way (4*4) fish post harvest research, some of my final output data is skewed. Couldn't find much how to transform data using an arcsine transformation in SPSS. Please do help me by sharing related article or related SPSS help guide or even directing me to a useful youtube tutorial video would be great. Thank you.
1. Does SPSS have an arcsine transform function?
2a If not calc with some other program then go to 2b otherwise
2b if yes use same steps as any other transformation e.g log
3. if 2b doesn't work try a Box-Cox or log approach
You can get all of these in most stat packages. For full details see:
https://www.google.com/search?q=How+do+I+do+regression+transformations%3F&rlz=1C1CHBF_enUS874US874&oq=How+do+I+do+regression+transformations%3F&aqs=chrome..69i57.24756j0j1&sourceid=chrome&ie=UTF-8
Best, D. Booth
You say in the title make the data "normal" and then seem to just be interested in skewness in the description. Anyway, the arcsin way not do what you want (it may reduce skew). As David Eugene Booth says, the Box-Cox family of transformations is more flexible so worth considering.
Fooling around with data transformations loses the natural measurement scale of the variable.
Skewed data often come from identifiable processes such as counted data. You are better finding and appropriate model than trying to warp your data to fit some preconception. Normalising your data may well not normalise your errors, which is more important than the distribution of the data.
Worst case scenario, I'd go with ordered logistic regression.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS