# How to calculate significant differences

#### Richard.Chambula

##### New Member
Hi Everyone..

I am stuck on one of my assignment questions. I have two samples, let's say List of employees in one Department A each with an average salary, as well as for Department B each with an average salary. I have been ask to calculate if there is a significant (p=0.05) difference between the average salaries of employees in Department A and the average salaries of Employees in Department B. Can anyone please help me on this? How do i do it?

#### trinker

##### ggplot2orBust
what are you thinking? You have class notes and a text book I assume.

#### Richard.Chambula

##### New Member
what are you thinking? You have class notes and a text book I assume.

I just want an example, or a brief explanation so I can apply it on my 15,000 records of data.. I have read notes and listened to youtube videos, but the more I read, the more I get confused.. I have tried the Excel "t-Test: Two-Sample Assuming Unequal Variances", Now I don't know how to interpret the results.. Maybe you can tell me what is the best test for this kind of statistical problem..

#### trinker

##### ggplot2orBust
"t-Test: Two-Sample Assuming Unequal Variances"
Based on what you've said above this approach sounds reasonable.

Now I don't know how to interpret the results
If you post the results we can help you interpret them. The p.value and/or test statistic is really important to interpreting the results.

#### Karabiner

##### TS Contributor
Performing a statistical test on a mean difference with a sample of
15'000 subjects sounds rather absurd. There is no need for a t-test,
one can just interpret the descriptive statistics.

With kind regards

K.

#### Richard.Chambula

##### New Member
Performing a statistical test on a mean difference with a sample of
15'000 subjects sounds rather absurd. There is no need for a t-test,
one can just interpret the descriptive statistics.

With kind regards

K.
Thanks, how do I then determine the significance difference based on the descriptive statistics..

#### Richard.Chambula

##### New Member
Based on what you've said above this approach sounds reasonable.

If you post the results we can help you interpret them. The p.value and/or test statistic is really important to interpreting the results.
Thank you, below are my results. I have given the results different colors, blue for lets say Department A and red for department B.

t-Test: Two-Sample Assuming Unequal Variances

Average salary - associate professors_typeIIA| Average compensation - associate professors_TypeIIB
Mean 554.6814404 | 477.6027397
Variance 6227.756571| 7325.62579
Observations 361| 584
Hypothesized Mean Difference 0|
df 810|
t Stat 14.12084311|
P(T<=t) one-tail 6.1329E-41|
t Critical one-tail 1.64673699 |
P(T<=t) two-tail 1.22658E-40|
t Critical two-tail 1.962897021|

Thank you again..

#### trinker

##### ggplot2orBust
Well you have to know if this is a one or two tailed test (is it directional?). Then look at the appropriate p values. If it's less than your preset alpha level there's a sig. diff. Or you could look at the critical value. If the t stat surpasses that value (larger than on positive end or smaller than on negative end then you have significant diff.