1
$\begingroup$

I have run correlation on SPSS for the variables of number of employees in a bank and Net profit margin for the data of 10 years. Results show the correlation coefficient '-0.1649' and p-value '0.649'.

I am unable to interpret the result and need help.

$\endgroup$
1
  • 1
    $\begingroup$ It just indicates that there is not strong evidence that the correlation is different from 0. To determine if there is a slight negative correlation as the estimate suggests you would need a much larger sample size. $\endgroup$ Feb 3, 2019 at 19:48

2 Answers 2

1
$\begingroup$

It appears you have data for one bank for 10 years, so N = 10. The low, negative correlation only indicates that for the bank you chose, for the 10 years you have, the linear relationship was small and negative.

Unless you are willing to posit that either a) These 10 years are a random sample of all years from your bank or b) These 10 years are a random sample from those 10 years for some larger population of banks then you won't be doing any inference and the p value isn't important. (And positing either of those things seems unjustified, at least to me).

What I would do is plot the ratio of profit to employees over time and see if anything interesting appears.

$\endgroup$
2
  • $\begingroup$ Yes, I have taken data for one bank earlier. Later on, I have run correlation for the 10 years data of 6 banks (60 observations) and found the following results. R = 0.231, Sig (2-tailed) = 0.076. I am actually confused whether p-value will matter over here as per which the above results show that the same are not statistically significant because p-value > 0.05 whereas correlation (R) = 0.231 that means slight positive correlation. $\endgroup$ Feb 4, 2019 at 20:13
  • $\begingroup$ If you try to do this for 6 banks then your data is not independent and correlation really isn't appropriate. You would want a multilevel model. $\endgroup$
    – Peter Flom
    Feb 5, 2019 at 11:09
0
$\begingroup$

Your data shows a slight negative correlation between number of employees and net profit. (correlation coefficient of -0.17).

Your data is consistent with a world in which there was no true correlation between number of employees and net profit, in that it is not unlikely that the data shows a slight negative correlation due to random chance (p-value of 0.65).

In short, almost all rational people would not take this data as proof that there is a negative association between number of employees and net profit.

$\endgroup$

Your Answer

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

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