1
$\begingroup$

I work for a bank in south america and we do something called elasticity price test to determine what is the best "Price" (fee for a product), here is how it goes:

We select two similar groups and then we give, for example, 20% fee for the first group and 15% fee for the second one, and then we wait to see if the group with smaller fee is going to borrow more money ( so more that we will make more money from this loan ).

So we use a varible called "unitary margin" that represents how much money we make from one person in each group.

So for example:

*Group 1 (20% fee) -> with unitary margin of 22.00 dollars

*Group 2 (15% fee) -> with unitary margin of 22.40 dollars

I want to be able to explain if it is significant or if i just happened by chance and that is a bad idea to give a 15% fee for everyone.

I have all the data, what test should i study to make better decisions ?

$\endgroup$
  • 1
    $\begingroup$ You need to have some information about the variability of the responses in the two groups. $\endgroup$ – BruceET Dec 12 '18 at 9:21
  • $\begingroup$ I have that, as a said I have all the data possibly needed , I have everything. Is it a t test case? $\endgroup$ – Victor Luiz Dec 12 '18 at 12:06
  • $\begingroup$ A two-sample t test uses sample sizes, sample means, and sample standard deviations for independent samples from two populations to test whether population mean sare equal. It is not clear to me if 'unitary margin' is a sample mean or based on a sample mean. If so, you should be able to use a t test. Data must be at least nearly normal for small samples. For large samples a t test should be OK unless data are far from normal (e.g., markedly skewed or with many far outliers). For a 'pooled' t test one assumes the two population SDs (or variances) are equal. If in any doubt use Welch t test. $\endgroup$ – BruceET Dec 12 '18 at 17:35

Your Answer

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

Browse other questions tagged or ask your own question.