Cross tabulation and Hypothesis Testing

Hypothesis testing is self-explanatory. You did this so many times. You should know.

There is a null and there is an alternative hypothesis. You are given a significant level at which you check the hypothesis at. It is most likely 0.05 unless stated otherwise. You either accept or reject the null hypothesis. Null is never accepted.There is a 1 tailed and 2 tailed test. If > or < sign is used then it's a 1 tailed test. If =/= sign is used then it's a 2 tailed test. You use the normalised significance value. At 0.05 it's 1.645. At 0.01 it's 2.33. Get the p-value or the probability of obtaining a z-value. If the obtained p-value lies BEYOND the specified significance level then null is rejected. If not null is failed to be rejected.

Cross tabulation reflects a joint distribution of two or more variables. We use a chi-square test for this. Can include a third or more variables.

Chi square is fucking bullshit. So much shit. God fucking dammit.

why chi square is used
 * ease of comprehension - can be easily interpreted and understood by people with little statistical knowledge
 * versatility - may provide better insight into complex phenomenon than a standard multivariate analysis
 * clarity - clarity of interpretation allows better link between research result and managerial action
 * simplicity - used for finding relationship between 2/3 variables only as it becomes too complicated with 3+ variables and it becomes hard to have enough respondents for each variable.