I have Likert scale data of farmer perception as a dependent with 27 items with five scales. However, my data set contains one outlier. One of the ways that I know to remove them is by using Boxplot to identify outliers and then delete them manually. However, It is impossible to remove that case. I would like to know a simple way to remove such outliers in SPSS and Stata without deleting the case. The problem is attached herewith.
Thanks.
Helllo Mequannt,
I'm not clear as to how it is that you find it impossible to remove a case that you wish to remove from a data set. However, you can always set the observed value of the unwanted case as a missing value (go to Variable View on the data worksheet and for the relevant variable, just enter that case's value under the "Missing" column. It will then be ignored in any subsequent analyses involving that variable.
The bigger question, though, is why you think it's important to ignore or jettison that case from the analysis in the first place. If the value has been incorrectly recorded, then either fix it or drop the case. If the value is an impossible value, then fix it or drop the case. If the value can't be shown to be incorrect, then it's probably best to keep it. Sometimes, we learn a lot more about a variable and cases in the data set by looking at those with distinct values rather than by deciding to treat them as inappropriate.
The danger of trimming data is, you'll tend to bias estimates of parameters (if the data that are dropped are legitimate), you'll alter relationships across variables, and these effects are more pronounced with smaller sample sizes.
Good luck with your work.
- Badges
- Science method
- Similar topics
- Biological Science
- Ecology