1
$\begingroup$

I'mI'm trying to make an analysis of increase in sales vs promotion (kinda correlation). And want to see the results segregated by Males and Females. I can segreate data by males and females, find correlation independently and then can compare. But I want to know if there is any standard way to test this at a single test.

$\endgroup$
1
$\begingroup$

As I understand, you want to check, whether promotion can be predicted by some continuous variable, while additionally accounting for gender for any possible differences. This sounds like a task for (multinomial) logistic regression, depending on whether your promotion variable is binary or has multiple outcomes.

You can use your continuous variable and gender as your predictors, perhaps adding an interaction term.

| cite | improve this answer | |
$\endgroup$
  • $\begingroup$ Thanks.. But lets say, I have gender as independent variable along with multipe other continuous variables like promotion money spent, R&D Cost etc... to find out the impact on Sales, what test can be used?? $\endgroup$ – sriharsha KB Apr 13 '15 at 11:25
  • $\begingroup$ @sriharshaKB Again, you might be able to formulate your problem as regression analysis. Could you maybe elucidate your hypothesis or make your goal more explicit? What are you trying to find out? What do you mean by "impact"? $\endgroup$ – Fato39 Apr 13 '15 at 12:02
  • 2
    $\begingroup$ If sales are increasing with increase in promotion money spent and R&D costs and IF it's better with Male salesman or a female salesman. If this does not form a valid hypothesis, Kindly help me formulate one. Im a little new to this field. Thanks in Advance :) $\endgroup$ – sriharsha KB Apr 13 '15 at 19:26
  • $\begingroup$ @sriharshaKB That sounds like a legitimately formulated hypothesis which might be tested using linear regression - of which a correlation coefficient is a small part when testing the relationship between two variables. It might be a good idea to read through a basic text on the subject (and worry about the logistic/multinomial part later), such as this one. For studying the effect of gender, see interaction. $\endgroup$ – Fato39 Apr 14 '15 at 9:39
1
$\begingroup$

I have successfully used the Mini-tab's General Regression Model. In newer versions, it is just Regression Analysis' Linear Fit Model. I can add both categorical and numerical variables here. The result equation came individually for each categorical variable like

a. Regression line for Male b. Regression line for Female

For noobs like me, best article I found to decide which test you wanna run is here

| cite | improve this answer | |
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.