# ANOVA vs 2 sample t-test

N

#### noko1100

Hi, I hope someone can help me with my doubts.

I am confused regarding the use of these tools. I have read a lot info on internet ,

I know that:

2 sample T test help you to Compare two samples with each other

My doubt is that some websites says that you can use the ANOVA instead of 2 sample T test to compare two samples with each other

is this correct?

regards

Last edited by a moderator:

#### Bev D

##### Heretical Statistician
Super Moderator
Yes you can use ANOVA for comparing the means of two samples. You can also use ANOM (Analysis of Means), confidence intervals as well as Tukey Duckworth Endcount approach. There are many statistical tests for comparing the difference in means...each has its own strengths and weaknesses...

Last edited:
N

#### noko1100

Yes you can use ANOVA for comparing the means of two samples. You can also use ANOM (Analysis of Means), confidence intervals as well as Tukey Duckworth Endcount approach. There are many statistical tests for comparing the difference in means...each has its own strengths and weaknesses...

I would not want to take advantage of your help but which would be fundament for using two instead of three samples with the ANOVA

Minitab requires minimun 3 samples for the ANOVA

Last edited by a moderator:

#### Miner

##### Forum Moderator
No, Minitab does not require three. Attached is the Minitab Furnace.MTW data set analyzed both as a 2-sample t-test and as a 1-way ANOVA. Notice that the conclusion is the same for both methods.

#### Attachments

• ANOVA.jpg
44.3 KB · Views: 289

#### Bev D

##### Heretical Statistician
Super Moderator
Minitab requires minimun 3 samples for the ANOVA

Are you referring to '3' levels of a factor (example 3 different process settings for temperature or 3 different vendors for some material)
OR
a sample size of 3 (data points) for each level of a factor?

I am not familiar with Minitab - but Miner is and I trust his answer regarding the ability of Minitab to perform ANOVA on 2 levels of a factor (a "two-sample" test)

#### Statistical Steven

##### Statistician
Super Moderator
No, Minitab does not require three. Attached is the Minitab Furnace.MTW data set analyzed both as a 2-sample t-test and as a 1-way ANOVA. Notice that the conclusion is the same for both methods.

Just a point of clarification, the p-value will be IDENTICAL if you assumed equal variance. The difference in p-values in the assumption of unequal variance for the two-sample t-test.

#### Miner

##### Forum Moderator
Good catch. Got in a hurry.

#### Attachments

• ANOVA.jpg
47.9 KB · Views: 176