Question

You are assigned to compare call times between two different groups. What questions must you evaluate prior to selecting the appropriate statistical test? What are the risks of selecting the wrong test? Illustrate by providing specific examples.

Answer 1

The call time between the two different groups will depend on the following factors as per me:

• Are the people related by blood or close friends or spouse or acquaintances or totally unknown to each other.
• For what purpose the call has been done, if it is for an interview, it's supposed to go longer, if it's a casual talk it should be of a shorter duration.
• At what time the call has been done, I'll say that the calls done early morning and during afternoon are of shorter duration compared to the calls done during night and evening.
• At what time of the year the call has been made, normally during the time of festivals we get many calls but they aren't of longer durations, they are kind of very formal talks.

Drawing contingency tables with a proper chi-squared test can give an excellent estimate of association or dissociation. against or in favor of the null hypothesis. In R, drawing contingency tables with viewing p values can help one estimate if there is an association or not. If the p-value is less than say 0.05, then it's safe to assume there is an association rejecting the null hypothesis and if not, we can't reject the null hypothesis but at the same time won't blindly accept the null hypothesis either.