1
$\begingroup$

I've been doing some analysis projects at work and I've been supplied with some dummy data regarding whether an applicant has applied on a weekday or weekend (I have set this as 0 for weekend and 1 for weekday. The other variable I have is the total time it took for an applicant to apply and get recruited.

Now from a statistical stand-point if I use a regression analysis against this dummy data and I receive a P value < 0.05 - can I say safely say that there exists a linear regression relationship and that the day the individual applies has an impact on overall recruitment? Of course the upper and lower ranges are also +ve as well as an adjusted R squared of 0.70.

To verify the relevance of this variable I was thinking of using a multiple regression analysis to test for the P-values and see whether there is an associated increase with any other variables (to test for mutlicollinearity).

$\endgroup$

1 Answer 1

1
$\begingroup$

First off, you never mentioned what the purpose of your analysis is. It would be helpful if you updated your question with that information. What is it that you are trying to determine? What is your objective?

The regression analysis you are using in this case is equivalent to the independent samples $t$-test (assuming the same applicant doesn't appear in the dataset multiple times). You can verify this by running the regression analysis and then a $t$-test between the two groups and examining the resulting $p$-values. They will be identical. Your analysis says that there is a statistically significant difference in the total time it took for an applicant to apply and get recruited between those who apply on the weekend and those who apply on weekdays. In other words there is as association between weekend/weekday applicants and an increase/decrease in total recruitment time.

Whether or not it is actually the weekend/weekday that is driving/causing the difference in total time is an entirely different question. You will likely need to control for a whole host of other factors to gain a better understanding of what may actually be $causing$ the differences.

$\endgroup$
3
  • $\begingroup$ Sorry about that - I had done a multi-regression analysis looking at other variables such as each stage of the recruitment process (i.e. time spent with recruiters, time spent interviewing applicants, short-listing etc.) I was then given some more data (The dummy data provided). So I guess it would make more sense if I re-ran my multiple regression analysis and considered the weekend/weekday data to see if it improves the adjusted R squared or impacts the P-values of any other variables? $\endgroup$
    – IronKirby
    Jun 23, 2015 at 9:42
  • $\begingroup$ You haven't answered the main question that needs answering. What is the objective of your study? What do you hope to gain by looking at this data and running your analyses? In your explanation, do not use any statistical terminology. $\endgroup$ Jun 23, 2015 at 10:17
  • 1
    $\begingroup$ Oh, apologies. The purpose of this analysis is to identify what are the leading contributors to the recruitment process and identifying whether there exists any 'wasted' steps that could be removed from the process resulting in a quicker recruitment time. This is the objective. By running my analysis I hope to identify which variables weigh more heavily in the recruitment process than the others so I can focus on these variables and see if I can identify any changes that are required. $\endgroup$
    – IronKirby
    Jun 24, 2015 at 3:28

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

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