Dear List Serve Members
I have some questions regarding crosstabulation tables.
With CrossTabulations missing data becomes important does it not. I have
800 cases and there will be some missing data. What percentage of missing
data is acceptable. I know the number of cells that have less than 5
responses should be kept to a minimum.
How many are we allowed before our conclusions are dismissed?
When should we start collapsing categories to increase our cell numbers?
Yates correlation is used when there are low cell counts. I hear this is not
a popular procedure to massage the data.
I have not tried SPSS Exact Tests yet but I imagine this should help with
small sample sizes. I will do this also.
Franco Giuliano Bassoriello