t Tests for One Mean: Introduction

An introduction to t tests for one population mean. I briefly discuss when we use the test, and when we would use a z test instead. I also briefly discuss the hypotheses of the test, and the p-value for different alternatives. I then work through an example. (If you are comfortable with the basics of hypothesis testing, and understand the difference between t and z procedures from confidence intervals, then much of this video may be review.) If you are just looking for an example, it starts at 7:00.

The reaction time data is simulated data with the same summary statistics as found in:

Armstrong et al. (2012). Mild Dehydration Affects Mood in Healthy
Young Women. Journal of Nutrition, 142: 382-388.

[LISA.WANG.]

Clear voice, good pace, simple and organized format, dark background easy on the eyes, great teaching with all the little details, it’s perfect— this is perfect, thank you very much.

[manchao]

best tutorial series on statistics on youtube

[fife]

I wish you'd continue to be enthusiastic in your videos like the earlier ones. Its so rare to find someone talking about statistics with actual enthusiasm in their voice. Its also rare that a person laughs when watching a video on two sample t-tests. But I think your videos are very concise and enlightening. Thanks for your work with these videos!

[TylerCJF]

I am in an applied statistics and research class and I am currently reviewing for our upcoming final. I enjoyed the video. It was straight forward and informative. One thing that I would like to see and what I am still trying to figure out is when you should apply the one sample t-test rather than another test such as the two sample t-test or the matched pairs test or even simple linear regression. For my final I will just be given a word problem and I will have to decide which test to use on my own without being told. There are 7 different tests that the professor may ask us about so differentiating between each is where I struggle. Anyway good video and I'm just giving my feedback. Thanks for reading.

[kayhankurul]

In next 3 hours, I will be examined for my class Statistical&Quality Control... I understood very clearly! :D

[erikhallberg]

Thank you. This video is a wonderful help for learning t-tests

[lensong]

When you explained the 3 situations whether u<u0, or u>u0 or u=/u0, I don't understand why the u=/u0 is the double area of u<u0? Since u=/u0 means either > or <, shouldn't it be the sum of both < and > (which is 1, arguably)?

[BrainNP]

Thank you so much! Very clear, easy to understand and it was fluid flowing! Thanks eh! :)

[stefanfarier]

All this time.... I was wondering "what exactly is a t-test"?

And I've been doing it for years

[sarawilliams]

so helpful, I've enjoyed all of the jbstatistics videos. thank you!

[RA-ijzx]

But if the t distribution has a large number of samples we can measure it as z, as my teacher said.

[Fabelot]

Thanks for a great explanation!
Don’t boxplots show the median where the thick line is, tho?

[Simeon]

"Drawing conclusions about young woman in general is dubious at best" haha thats true even when you are not talking about stats.

[omaromar-otih]

How you can recognize between testing hypothesis and chi square if your doctor give a question, because both have the same information...?!

[adamhall]

When I run the command pt(1.389, 24, .025, lower.tail=FALSE) in R I get .09280314, not .0888... What am I doing wrong?

[katherinedavis]

What do I choose in my calculator when using software?

[donnasmith]

Excellent videos. I teach statistics at the community college level. I have a Masters in Mathematics not in statistics. I notice you use the term zed for z, am I incorrect in saying z rather than zed? I am assuming zed is the Greek equivalent to z.

[richardl.]

In 12:10 if we were to use the critical value from the t table it would be 2.06 and since the value we got was 1.389 we fail to reject the Null hypothesis correct ?

[EvilSapphireR]

Aaaaand we never got the video for the two samples T test.

[aryamotegaonkar]

let the null hypothesis be that the given t-distribution has the required degrees of