I tried to conduct Chi-Squared test in R but I came upon an obstacle. In my data I have some observations which I would like to delete, because they aren't significant enough, specifically for this table I want to get rid of values such as "Don't know", "Refusal" and "No answer":
> TAB <- table(gender, health)
> TAB
health
gender Bad Don't know Fair Good No answer Refusal Very bad Very good
Female 974 4 4021 6563 0 1 203 3587
Male 688 8 3319 6407 1 1 146 3691
I want to delete observations which take these values from the whole data frame. I tried doing that by
> remove <- c("Refusal", "Don't know", "No answer")
> ess_data1 = ess_data[! health %in% remove, all()]
But in the result I still get responses such as before but with count equal to 0:
health
gender Bad Don't know Fair Good No answer Refusal Very bad Very good
Female 974 0 4021 6563 0 0 203 3587
Male 688 0 3319 6407 0 0 146 3691
Is there any simple way to filter this data frame to obtain table without these redundant observations and values?