t-test for two independent samples in Stata®

Learn how to compute Student's t-test for two independent samples in Stata using the *ttest* command.

Note that you can type *db ttest* into the Command Window to open its dialog box.

Copyright 2011-2019 StataCorp LLC. All rights reserved.

[zuLess]

These videos say what you are doing, but not why you are doing it. "Click on this and click on this and abrakadabra you have done it". What have I done exactly? What is the difference between all these variations of t-tests and when are we supposed to use them?

[Comrade-wvlu]

thabk you sir! Your lectures are hidden gems. But interpreting data briefly would help us a bit more. BTW you have done an excellent work. Wish you all the very best!

[fjmenendez]

I feel confused. As I view it, the right analysis to be done is a t test two independent groups (not samples, groups). You did it. But the title is "t test for two independent SAMPLES in Stata". And you said this in some moments.

In summary: I think there is a confusion in the title and at some moments in the video. Am I wrong? I look forward for your clarification.

And thanks for the playlist. I am learning a lot from it!

[statacorp]

That depends on your data. It sounds like they are independent samples but if your dataset contains people who moved from one region to another, it could be paired. You will want to check with the person who collected the data to be certain.

[ronaldmulema]

Hi All; It is a wonderful tool for hypothesis testing.

[abdullahrabaya]

Thank you for nice explanation and easy way to carry on. How you can interpret the results ?

[LiveOnThursdays]

When we say the difference between the two groups is 0, that means no difference between the two groups correct? Yup and i just answered my question when i typed it out lol

[salomeerekaha]

Many thanks. Well explained video. How would you interprete this result in a written scientific paper?

[gauravmohan]

How to we explain why the confidence intervals are overlapping though. I see your p<0.05, then how is the CI overlapping?

[respiro]

Hi, the logic sequence should be, first variance analysis (if are equal or no, sdtest maybe) then t-test?

[asmasultana]

How to decide equal/unequal in the command

[BengalsfanLife]

 I was wondering if you could help explain to me why the absolute value of t value for the coefficient in command reg is the same as the absolute value of two sample t test in command ttest.

[TheTimiiee]

hi, if u have values for regions (north, west, east and south) and u had to test if north = south for example, are they paired samples or independent samples?

[msanatanrao]

T-test is used when the sample size is less than equal to or less than 30 but here the sample size is 60 . Anybody please clarify me that??