When looking at a statistics hypotheses such as, Suicide will be supported when faced with a terminal illness just at letting one die? When looking at surveys it was a yes/no question. Does this make suicide and letting one die both nominal since there is no order to the question? I am tying to figure out which formula to use in a cross tabulation that will give the most accurate level of association lambda, gamma, and Person's r. But need to know if I am looking at nominal or ordinal?
1 Answer
$\begingroup$
$\endgroup$
2
To clarify, does the outcome have exactly two alternative values? If not, what are the values and do they take a meaningful order? If it's just two values, it doesn't matter whether they have an order, since you'll only be comparing one to the other somehow.
Second, with what kind of variable(s) are you cross-tabulating the outcome? Nominal? Ordinal (with three or more values)? Continuous?
-
$\begingroup$ What I am trying to discover is the strength of association between chosen variables as well as how to interpret the findings using SPSS functions and tests such as lambda, gamma, and Person's r along with other possible tests. I have the let die as my IV and suicide as my DV when I complete the cross tabulations. But I can not figure out which test would give me the best interpretation of the findings. $\endgroup$– maryMay 3, 2016 at 23:12
-
$\begingroup$ I really don't understand the data from your description. It sounds like you asked people whether, given (someone's terminal illness), would they let the person die naturally or support suicide? Is that true? In this case, the outcome is a variable with two values: die naturally or support suicide. When you say cross-tabulations, what are the other variables? Like, maybe, age of respondent, sex of respondent. What are they? Forget the tests for the moment and describe the data. $\endgroup$– RicoMay 3, 2016 at 23:31