T-test, ANOVA and Chi Squared test made easy.

Рет қаралды 274,364

2 жыл бұрын

Statistics doesn't need to be difficult. Using the t-test, ANOVA or Chi Squared test as part of your statistical analysis is straight forward. You do need to understanding the underlying principles of hypothesis testing and p-values of course. You need to understand when to reject the null hypothesis and accept the alternative hypothesis if you are going to make any inference about the population from your sample data. If you are doing any data analysis or statistical analysis then this video is a must. If you are an aspiring data scientists or doing quantitative research, then this is a good place to start.
SUBSCRIBE:
--------------------
Click here: kzfaq.info_...
BECOME A MEMBER
---------------------------------
/ @gregmartin
This channel posts global health and public health teaching videos and videos about how to find the right job in global health. If you haven't already, please consider subscribing to this channel and becoming part of this community. The channel also provides teaching on research methods, statistical analysis and how to write and publish a scientific paper.
LETS CONNECT:
---------------------------
Twitter: @drgregmartin
Linkedin: / drgregmartin
Facebook: / thisweekinglobalhealth

Пікірлер: 130

@gregmartin 11 ай бұрын

Get my FREE cheat sheets for Public Health, Epidemiology, Research Methods and Statistics (including transcripts of these lessons) here: www.learnmore365.com/courses/public-health-epidemiology-research-methods-and-statistics-resource-library

@gilangignasraharjo6138 Жыл бұрын

It's not "Super Duper Easy" but you made it easier. Thank you sir

@gregmartin Жыл бұрын

So nice of you! Always happy to help. Thank you for the amazing feedback!

@hisnibs1121 2 жыл бұрын

Thank you Greg. This is the second of your videos on quantitative analysis I've watched, and I really like both your style and the pace at which you proceed through this stuff. It's ideal for me and for my current purposes - revising for work purposes, against a tight deadline, stuff I haven't done for many decades, and probably only half-understood at the time, .

@gregmartin 2 жыл бұрын

Glad you like them! Thank you for your awesome feedback!

@gunamrit 4 ай бұрын

Ryan Reynolds teaching me stats is the best combo I could ask for ! Thank You so much!

@gregmartin 3 ай бұрын

😄 Absolutely thrilled you enjoyed it!

@bagm27drano27 6 ай бұрын

i had to research the appropriate stat tests with no guidance for my dissertation (not a biostat student). this short video put my mind at ease. thank you!

@mirandacox3751 2 ай бұрын

Love your videos, still! I always find your explanations the best ever. Thanks for the work!

@md.masumomarjashim 2 жыл бұрын

The code with statistical videos would make our day.

@manuelagauthier3932 8 ай бұрын

Thank you so much Greg. Great work.

@LifeWithJonG 11 ай бұрын

Greg! Spectacular resource. Thanks!

@rauldhruva2372 Жыл бұрын

I had to watch the video couple of times and a playback speed of 0.5... Concept is crystal clear now Thanks a lot!

@gregmartin 11 ай бұрын

Glad it helped! Thank you

@Biostatistics 2 жыл бұрын

Hi Greg, this is another wonderful video. Please may I know what software you use for video editing as well as animations? Thank you in advance

@taffy6886 5 ай бұрын

Thank You. Your teaching style is excellent. I have had so many light bulb moments while watching your videos.

@user-fk8hi3qn5o 10 ай бұрын

Awesome stuff, thank you!

@maryellenelizabethhart7370 Жыл бұрын

Thank you. These are perfect. Thank you for sharing your passion. I appreciate you.

@gregmartin 11 ай бұрын

Glad you like them! You are so welcome

@MoumitaHanra Жыл бұрын

Amazing! Not everyone teaches us say all possible scenarios of implementing each hypothesis test..very helpful!

@gregmartin 11 ай бұрын

Thank you

@haraldurkarlsson1147 2 жыл бұрын

Greg, Another solid video. One thing that I was taught in statistics is that we never accept the NULL hypothesis. We either reject it or fail to reject it. I guess it is because there is always a chance - even at p=0.05 that the hypothesis might be correct even when we reject it etc. (false positives and such). Also does the variance not have to be the same in these samples or populations that we compare?

@practicemail3227 11 ай бұрын

you are correct, it it either we fail to reject or fail to reject.

@razzle_gapl9639 Ай бұрын

Exactly what I was going to say when I heard his explanation of samples too unlikely to get when the null hypothesis is true. I would just add that we never ACCEPT the alternate hypothesis either, because that would lead to a lot of legal issues when there’s an error

@haraldurkarlsson1147 Ай бұрын

@@razzle_gapl9639 I don't know if you have heard of a book called Fictitious Science which was written by a psychologist working in academia. It is a real eye-opener in regards to p-hacking etc.

@thomasabbott1757 7 ай бұрын

Great video, thank you I had a quick question, can you explain the difference between a two independent sample with dichotomous outcome test and a chi-square test of independence that uses two independent samples? I feel like you can use both for the same data set. For example, prevalence of CVD among smokers and nonsmokers. Thank you

@yingma3228 Жыл бұрын

Hi Greg, I am curious about what tool you used for analysis in this video?

@benpetro436 8 ай бұрын

Very good video .... i really like the way u deliver in as much layman as u can where can i get the dataset that you used in this video?

@tombayes9706 2 жыл бұрын

Great informative video Greg, thanks! One question - if your ANOVA result is significant, leading you to follow up with post-hoc tests, such as a t-Test with some kind of family wise error rate adjustment such as bonferroni, why bother with the ANOVA in the first place? Couldn't you just run the t-Tests and adjust the p-values for the number of tests (as per bonferroni), making the ANOVA redundant. I appreciate for lots of groups this would be time consuming but is that the only benefit of the ANOVA in that case, efficiency? I can't see how it offers any extra protection against type 1 error if the end result is to run t-Tests and report corrected p-values anyway? Appreciate any thoughts! Thanks and keep up the great vids.

@tomworrall190 6 ай бұрын

I think this is a great question. I don’t know but perhaps skipping the ANOVA step and just doing many between-two-group comparisons would increase the risk of a type 1 error, and would be akin to data dredging? E.g., comparing samples from all Western European countries to see if life expectancy differs. Is the ANOVA step somehow ‘protective’?

@kelseysbookrecs 4 ай бұрын

you're single-handedly getting me through school

@gregmartin 4 ай бұрын

I'm so happy to be able to help! Keep up the great work

@adelalsaeedi 2 жыл бұрын

How I wish there would be a translation of a video to support the channel. The channel is beautiful in science. Raising the number of views is suggested to you Cc

@gregmartin 2 жыл бұрын

Thanks. Will see what I can do.

@federicogarland272 2 жыл бұрын

your videos are amazing!

@gregmartin 2 жыл бұрын

Glad you like them! Thank you so much!

@anggipermanaharianja6122 2 жыл бұрын

This vid is legit.... please keep the good work!

@gregmartin 2 жыл бұрын

That's kind of you to say - Thanks!!

@mrutyu_s 2 жыл бұрын

Helpful , thank you for this video 👍🏻

@gregmartin 2 жыл бұрын

Glad it was helpful!

@dr024 2 жыл бұрын

is it wise to do a t-test in those life expectancies when those seems to be population density plots? we only perform tests if data are from samples right?

@maloukemallouke9735 Жыл бұрын

hi thank s for videos i am wondering how minimum numbers of items i should take in my data to make test significant?

@kpamorjames566 Жыл бұрын

Good day. Was wondering how I could carry out a T-test and anova using data gotten from bacteria culture

@lynnn5614 2 жыл бұрын

you’re an angel thank you so much

@gregmartin 2 жыл бұрын

You’re welcome 😊

@thangamaninmamit-civil6860 Жыл бұрын

For water quality analysis of different season’s which test will be more suitable? Thank you

@jamesleleji9470 2 жыл бұрын

Thanks for providing this tutorial, Greg. Suppose we have a baseline data and post intervention data. How can the following test of significance be performed in R or what kind of test is appropriate to perform? T-Test to Determine the Statistical Significance of the Baseline Food Security T-Test to Determine the Statistical Significance of the Post Intervention Food Security T-Test to Determine the Statistical Significance of the Difference of Differences of Food Security

@abderrahimzilali3293 2 жыл бұрын

Hi Leleji, First you need to determine the response Y you're intersted in for which you want to compare the difference in means between the two conditions (Baseline and Post intervention). To apply the t-test which is a parametric test, you data need to comply with certain assumptions namely independance and variance homogeneity and normality of the residuals. And no need to remind that the response Y should be Quantitative. # R code > t.test(Y_Baseline, Y_PostIntervention, alternative = "two.sided", var.equal = TRUE)

@chiellinilee5152 Жыл бұрын

Thank you very much. Excellent teaching video.

@gregmartin 11 ай бұрын

Thank you so much for the amazing feedback

@sergiochavezlazo5362 2 жыл бұрын

Finally, thanks !

@gregmartin 2 жыл бұрын

:) Glad you liked it Sergio

@lxtlkingbastard 8 ай бұрын

Thank you for making this !!

@jokovik5492 2 жыл бұрын

Greg, is there an online resource where I can practice questions about exactly which test to use in different circumstances? Watching your videos has given me confidence, but I still want to test myself out..

@gregmartin 2 жыл бұрын

I'll see what I can find and get back to you

@jokovik5492 2 жыл бұрын

@@gregmartin Thanks!

@Chenault68 Жыл бұрын

Thanks!

@ishangoinyambo6523 11 ай бұрын

Is anyone else here because of the Harvard sandal? While am still here thank you for this good explanation👏

@ryanwilliams6413 Жыл бұрын

Super video- thank you

@gregmartin Жыл бұрын

So nice of you - thanks, Ryan.

@emmanuellesaintmerant9657 7 ай бұрын

how are you getting these numbers. its a lot information.

@madhukumar6106 2 жыл бұрын

Thank you!

@gregmartin 2 жыл бұрын

You're welcome!

@Remifentanil 7 ай бұрын

Great Video. You should put time stamps in the video.

@amylawsongill 7 сағат бұрын

I like this guy and was going to subscribe but the MUSIC has stopped me

@KATERINAMARGETI 2 жыл бұрын

Thank you so much

@gregmartin 2 жыл бұрын

You're most welcome

@naveenarora5310 4 ай бұрын

Teaching is very good

@gregmartin 4 ай бұрын

Thank you! Glad to hear that.

@nurmuhammad2072 2 жыл бұрын

Thanks a lot Sir.

@gregmartin 2 жыл бұрын

You’re most welcome.

@MohdFarhan-ye9kc 2 жыл бұрын

Thank You.

@gregmartin 2 жыл бұрын

You’re welcome.

@nipun7609 Жыл бұрын

What are those decimals on the y axis?

@karinahulka1486 Жыл бұрын

Thanks a lot!

@gregmartin Жыл бұрын

You're welcome!

@felixlucanus7922 Жыл бұрын

Nice video. It may be good to clarify that the assignment of the null hypothesis to a statement is arbitrary. For example if the "life expectancy" example was worded differently, the null hypothesis, H0, could have been "Mean life expectancy IS NOT the same" and H1 could have been "Life expectancy IS the same." The meaning of the null hypothesis in this case would have been flipped from the example the video.

@gregmartin 11 ай бұрын

Great point! Thanks for the feedback

@notdave5081 Жыл бұрын

Thanks for the vid! Tho could you maybe not use the scratching noice when putting new things in? It really is painful haha

@sarrasouissi4542 Жыл бұрын

First, thanks Greg. I have 32 patients, each with three changing coordinates over the course of 40 days. I mean, for each patient, I have 40 values on an axe x, 40 values on an axe y, and 40 values on an axe z. I have no idea what kind of test to use.

@Nikolay76Gogol Жыл бұрын

At 06:30 why are you talking about paired test? These were different groups of people. Why do you think they are paired or linked?

@alexfrank1831 2 жыл бұрын

TIL when to use which test to check significance of hypotheses: - If two means, then use paired t-test - If three means then use ANOVA - If more use Chi-squared

@alexfrank1831 2 жыл бұрын

Let's see if my understanding needs to be corrected ; )

@gregmartin 2 жыл бұрын

you got it but you don't always do a paired t-test for two means. I normal t-test is fine unless the two means are frame samples where the observations can be matched. So "before and after" for example. For each person in the "before group" he or she is matched with themselves in the "after" group.

@NurzaBoo 2 жыл бұрын

Hi - this is so much easier to understand than my MSc lessons! Just to check if i've understood...There's actually very little difference between a t-test and a chi-squared test and how you would check if you can reject the null hypothesis. If your p-value is less than your threshold of 0.05 then its significant and you can reject. It's just the type of graph due to the type of data used that changes? But if you're given the data in a table, the results you need (CI, p-value, SD, etc.) are the same?

@gregmartin 2 жыл бұрын

I think you’ve got it! I have a video about hypothesis testing on my global health channel. You might like it.

@saradhiananth1613 2 жыл бұрын

@@gregmartin link to the other channel, Prof. Martin 🙏🏾

@praneethkrishna7218 Жыл бұрын

thanks for making it more complicated , or may be I'm too stupid for this

@thatomofolo452 2 ай бұрын

Got it 💯

@gregmartin 2 ай бұрын

Keep it up!

@cmcchannel2808 2 жыл бұрын

Thanks Mr Greg Please I need this strategy of preparing lessons. I mean to say the vedeo editing

@gregmartin 2 жыл бұрын

Noted

@wilhelmhesse1348 3 ай бұрын

One thing missed is how to calculate the P values

@cloudycandy9011 Жыл бұрын

To someone reading this, pls helpp I'll be doing two-sample t-test, but I don't know what variable i should use. Our alternative hypothesis is: there is a significant difference between using manual method and using GCHM (our own grain-collecting machine) in collecting rice grains.

@cloudycandy9011 Жыл бұрын

Thankyou, someone pls reply😢

@flake8382 Жыл бұрын

Clear as mud champ! :(

@venkatk1414 Жыл бұрын

Struggling to understand 😢

@fiercelion8626 4 ай бұрын

धन्यवाद जी। 👏👏

@gregmartin 4 ай бұрын

Thanks

@ariffaiz313 Жыл бұрын

I can't help but seeing a smarter version of ryan reynold teaching me

@gregmartin Жыл бұрын

Thank you!

@bradleysande9832 7 ай бұрын

Sir did you watch Deadpool

@rhysstevomusic1338 11 күн бұрын

How is a relationship based on a difference of proportions between samples. I think this is wrong

@Ali.Sadiq406 5 ай бұрын

والله والنعم

@rajxabc Ай бұрын

low freq boost on voice processing enhances fumbling .

@gregmartin 19 күн бұрын

Thanks for the feedback

@ZeptejSeFilipa 6 ай бұрын

coool

@averysaurus 2 жыл бұрын

Nice but we can’t ever accept hypotheses, only fail to reject.

@asierlarreategui 5 ай бұрын

Great video. Pitty that your audio and sound effects are not great. Almost annoying.

@gregmartin 4 ай бұрын

I appreciate the feedback and will make sure to adjust the audio effects for a better listening experience. Thanks for pointing it out!

@abinaya7900 Жыл бұрын

Yes i am reading the positive comments and yes i am grateful for this video but when you keep stuttering and repeating the same words 4 times a row, or using large connecting words, it is very difficult to understand what you are trying to convey, or I am just really dumb xD

@abdullahahmadi5021 Жыл бұрын

How you can speak that fast, thinking fast

@lavas3l677 3 ай бұрын

the explanation was good but the music and the sound effect are really distracting i wish the video was without it

@gregmartin 2 ай бұрын

I appreciate the feedback, thank you!

@user-gd5tm3es1i 3 ай бұрын

❤️‍🔥

@gregmartin 2 ай бұрын

Thanks

@LISA.WANG. 2 жыл бұрын

11:10

@Sergiuss555 11 ай бұрын

very difficult to listen with all the stutter, uneven sound volume, and added sound effects

@timothybennett1765 Жыл бұрын

Poor sound quality

@Invergence_x Жыл бұрын

You call this Explantation easy 😅

@Tyokok 8 ай бұрын

maybe just me, this video is not very well explained for beginners, language description wise, maybe for people who has some basics. and 2nd half of video is more about coding instead of explain or summary from math perspective.

@larshmallows 3 ай бұрын

Not absolute beginner no, need some basic knowledge on statistical data analysis. Wouldn't say it's coding, the numbers may make it appear so but they're just outcomes of the test

@cardiyansane1414 7 ай бұрын

Nice but those statistics…. Not accurate

@szocsdaniel Ай бұрын

I wonder why he used European vs African life expectancy and not IQ as an example... ;)

@withlovestephaniedenise7024 Жыл бұрын

Dear God, please help me! SMH

@LimanskiBerberin Жыл бұрын

Not needed to speed up the talk. I would rather watch 2 minutes longer video.

@juliaelia8794 Жыл бұрын

hard to follow way too fast...dissapointing

@cphone8093 Жыл бұрын

Anyone who says stats are uncomplicated, 'super-dooper easy' (!) has spectrally low levels of empathy. He then follows with a rapid-fire explanation and throws-in complex digressions...this is no teacher. If you are a teacher you know some things you should never say to learners, and you know how to modify your pace of delivery. This guy's an amateur.