Why do we teach everyone the t-test?

The same reason we give kindergartners a pair of safety scissors instead of a paring knife: it's a tool that can do a lot of jobs without the risk of someone misusing it and getting hurt.

I think it's mostly just looking under the lamp light. We explore it because it's explorable. Great video though I love your content :)

Broo, we only study this on the sixth semester of statistics, i consider this black magic.l

i am a senior double majoring in applied mathematics and statistics. I was considering graduate studies in statistics but people are discouraging me. What is your experience so far with bio stat graduate studies?

Content of my third book... The Welch test eliminates the t-tests, but t-tests are still more manageable for some purposes. Tables for the Welch test with real degrees of freedom are rarely available .

Great video, loved the general approach of looking at the t-test! Now, we need a video on non-parametric tests and bayesian hypothesis testing 🙂

Now this is one level of abstraction deeper than my prior

I feel like what we dont teach enough are its limitations. Unequal variances in two groups? Correlation between two groups? Unequal sample sizes? There are so many things that can lead to inflated type-I error rates and low power of the t-test. Especially practitioners should know this.

Haha Great video! Definitely the proofs proof that we don't need it 😂

Why not

Wish i had this video 10 years ago

I don't understand

Lol the description goes crazy. We are the grateful swine :P

I am not that impressed with t-tests. T-tests play a role in statistical analysis, but I believe their practical usefulness is limited, despite their solid theoretical foundation. In many real-world scenarios, the assumptions underlying t-tests are frequently violated, making their results less reliable. We should shift focus away from t-tests in introductory statistics courses. I have already started doing so in my own teaching.

From halfway of the video I lost ...

nice!

hi @very normal, Can i Ask material, or course, video i learn all about stat up to now, pls!

I'm not sure if this is a dumb question.
In my data analysis 101 class I wanted to compare the t-test to conjugate priors. It's been years since I've taken that class, but I recall being surprised that the t-test was better. (I think I compared the size of the confidence interval to the size of the credible interval while containing the true mean. A smaller interval was better). My question is whether my approach makes any sense in general.

I find the t-distribution so interesting, because it's such a useful , simple, and important distribution in usage, but it is extremely recent and relatively complex (I gave up my attempt to understand its derivation when I saw the gamma function). I think it really shows how mathematics is constantly evolving and has plenty of uncharted avenues for unimaginable growth.

The title made me think this would be about how t-test are bad and how everyone should just do bootstrap hypothesis testing instead or something. Still enjoyed the video though! When do you think the t-test is superior to a bootstrap test?

These videos would be more enjoyable if you didn't use this low vocal register, great stuff though

To annoy students?

Oooh! Great video, I saw this on my inference class, you should do one showing how the ratio of two scaled Chi^2 distributions is distributed F(df1, df2)

Your explanations were pretty damn good, but I think it would be better if you could record some series of videos and put each separate topic on a single video instead of mixing everything in a single video.

Another great video 😂 I love the π and p joke! Keep up the great quality videos! They’ve truly helped me grasp statistics so much and it’s been very helpful to have this knowledge now that I’m working in multiple psychology labs!

Baba is you theme on the background)

Its funny how profs protect you from the actual math behind it all when you’re starting out haha

Hey, very insightful video. I'm curious, what about other tests and their power? For example, Mann-Whitney test? Or different versions ot t-test, for when variance is known vs when variance is unknown or when samples are different.

excellent video as i prepare to teach my students hypothesis testing & confidence intervals for quantitative data. thank you

Hope your channel reach 100k subscribers this year!❤❤❤

Anyone that turns their nose up at a T-test is an unserious person.

This video came at a perfect time. I'm taking a statistics class that will go through UMP tests, it's nice to get a glimpse of what I'll be seeing soon.

Because it’s the Student’s t-test, duh

Hey thank you for the great videos.

I am planning to pursue a PhD in statistics, and I found your video truly inspiring. However, I am also a bit concerned about the career prospects after completing a PhD in this field. If possible, could you provide an analysis of the potential career paths and opportunities available for statistics PhD graduates? I would greatly appreciate your insights.

great video, a nice reminder about one of the most basic tools in statistics

THANK YOU FOR THE VIDEO!!! I've been struggling to understand t-tests (and hypothesis testing in general) for a while now.... This came at a PERFECT time!!

First comment hehe