Matched or Paired Samples T-Test - Hypothesis Testing

This Statistics video tutorial provides a basic introduction into matched or paired samples. It explains how to use the T-test and the student's t-distribution to determine whether or not if you should reject the null hypothesis in favor of the alternative hypothesis. It also explains how to construct a confidence interval and calculate the margin of error at a specified significance level.
Confidence Intervals & Margin of Error:
• How To Find The Z Scor...
Find The Z-Score Given Confidence Interval:
• How To Find The Z Scor...
Student's T-Distribution:
• Student's T Distributi...
Hypothesis Testing - Null & Alternative:
• Hypothesis Testing - N...
Type I and Type II Errors:
• How To Identify Type I...
One Tailed and Two Tailed Tests:
• One Tailed and Two Tai...
_______________________________________
Test Static For Means & Pop Proportions:
• Test Statistic For Mea...
Hypothesis Testing Problems:
• Hypothesis Testing Pro...
P-Value Method - Hypothesis Testing:
• P-Value Method For Hyp...
Hypothesis Testing - Proportions:
• Hypothesis Testing - S...
Difference of 2 Means - H Testing:
• Hypothesis Testing - D...
Cohen's D - Pooled Standard Deviation:
• How to find Cohen's D ...
Hypothesis Testing - 2 Proportions:
• Hypothesis Testing Wit...
__________________________________
Matched or Paired Samples t-Test:
• Matched or Paired Samp...
Chi Square Test:
• Chi Square Test
Chi Square Distribution Test:
• Chi Square Distributio...
Test of Independence - Chi Square Dist:
• Test of Independence U...
Final Exams and Video Playlists:
www.video-tutor.net/
Full-Length Videos and Worksheets:
/ collections

[@TheOrganicChemistryTutor]

Next Video: kzbin.info/www/bejne/fnynop-PiLGnbJo Final Exams and Video Playlists: www.video-tutor.net/

[@shuvashishsharma1299]

@TheOrganicChemistryTutor shouldn’t we use t(alpha/2) to find critical value for C.I?

[@riky2107]

for confidence interval shouldn't we use the two-tail t-score, in this case "2.26" 9df?

[@riky2107]

@@shuvashishsharma1299 I was thinking the same

[@maryamalhamdan1510]

The organic chem tutor honestly deserves a Nobel prize for his endless number of lessons ... this society needs to appreciate teachers more than artists and entertainers

[@enruiwang1537]

彳亍

[@alexb1312]

ur making teaching out to be way harder than it actually is

[@maryamalhamdan1510]

@@alexb1312 well bad math teachers have caused generation hate math for ever

[@ali_engin33r]

he may not have a nobel but having millions sure helps too

[@tomatrix7525]

I’m not joking when I say this man teaches me chem, ochem, physics, maths in general, calculus, physics

[@liamstevenson5350]

@Beckham Darren 15 minutes? You know what that reminds me of? Geico. 15 minutes can save you 15% or more on car insurance.

[@trevormasuku9606]

Thank you so much.You are talented .a good teacher.Its Annie in Zimbabwe.most lectures are failing

[@sabrinasaad8381]

He is your saviour haha

[@AbhishekVerma-kj9hd]

Very nice explanation sir love from India

[@tashinesempire6129]

I was literally just thinking of the same thing

[@laaori]

Just in time! You have gotten me through Chemistry, Physics, Calculus and now Stats. I’ve gotten really good grades in those past classes and I’d like to give you some credit. When I’m not a broke student I promise you I will remember the help you gave me, Mr Wahlberg.

[@mmdg1917]

is he the actor Mark Wahlberg??

[@thangtran145]

@@mmdg1917 yess, he is. He is mark walhberg.

[@DanishJaved]

are you doing good financially now / graduated?

[@DanishJaved]

been 4 years thought I'd ask lol

[@RomeoJ1996]

Entering this week of stats we’re covering hypothesis testing in MATH302, without even reviewing the material this week you just taught me exactly all I needed to know to solve paired samples

[@colleen1396]

Thank you so much for making this video! I wasn't able to attend my online class due to internet problems so I was really having hard time coping up with the activities. But this video helped me so much! Love from the Philippines! ❤️

[@Alchemist10241]

I've added videos of this channel to my review list and I review them once in a while but every time I learn some new points. 👍

[@dendeibrahimadekanmbi8022]

I thank you for these works you have been doing and also for sharing your knowledge. Some of your videos clearly explain what many textbooks can not do. Please, are you hoping to deal with topics such as: factor analysis and principal component analysis? Many Thanks

[@khaishir1968]

thank you so much for this i can use this now in my dissertation to calculate stuff !!Thanks a lot

[@choolwe5564]

Bro taught me math in high school now he’s teaching my stats in university he never fails

[@Cat-it7sl]

oh my god you're a lifesaver, i was so lost on all of this but i actually get it now. THANK YOU

[@amudh]

I am repeating what others have said here. This is exceptionally done!

[@tinkamugishavianney1784]

Its amazing how i sleep and dodge lectures keeping in mind organic chemistry has got my back.....God bless you your a hero to many

[@stephaniemoura9325]

This was soo hard to find! Exactly what I’ve been looking for! One sample for comparing the effectiveness of two variables!

[@turnabuse7035]

Thank you so much. I found question type, which i look for, eventually.

[@meenukrishna7684]

Exceptionally good....very clear and simple presentation

[@esthermuseba4278]

Many thanks for simplifying this topic.

[@lemmarie4091]

Thankyousomuch. I won't be able to answer my stats activities without your helpful tutorials. Godbless💙

[@fredessafigueroa1666]

Wow!! Super easy to follow, awesome tutorial.

[@kingrey405]

This man is a science-maths teacher. Any field he is very good 👍🏾

[@anthonyliu2264]

Appreciate your time to go over paired T test. Can you go over on SPSS because that is what my college is using instead of excel

[@Vintagetube310]

You’re the best, man!

[@nanjuldopsal260]

Many thanks for breaking this down.

[@KhaleesiZA]

YOU HELPED ME, SOOO MUCH FOR MY ASSIGNMENT WHICH I HAD NO UNDERTSANDING AND I LOVE YOU FOR THIS VIDEO!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! CAUse my lecturer was not explaining properly, the way you just did. so THANK YOU!!!!

[@ashutoshsingh7529]

Give this man a medal!

[@Richard-kz1ee]

My best online prof. Heart you Sir

[@Chikulamayembe57]

Brilliant presentation. I have learnt a lot from your videos Sir. On confidence interval construction, were you not supposed to use t_(9; 0.025)=2.262?

[@watermel]

Yes I thought the same thing!

[@salimalzadjali8387]

Yup

[@MonsieurGilbert]

i think so too

[@wubesemagn2021]

Thank you, It really is helpful.

[@andrewwashere]

Thank you so much! I’m a psychology student and this helps a lot

[@jorrpatel238]

Me too, let's connect? What's your Instagram?

[@anjumshaikh3270]

Thank you Sir.Greatful to u

[@jenkomatic]

Thank you for your video, it's really helpful! Do you know if the difference in mean values can also be used to quantify the effect size? Is there a separate test you could on the mean difference to quantify an effect size?

[@joserobertopacheco298]

Very Good explanation.

[@RamonTogonon-ec1ey]

God would certainly bless you with good health for your great job in educating a vast generation of academicians.

[@nageenajamrozbspsychologys7223]

this was helpful thanks sir

[@isaacwachira8803]

Shouldn't the null hypothesis be that the mean difference equals to zero?

[@Sawdust6666]

Thank you so much you are superb at what you do.

[@mufarachisi4236]

You're the best Mat God bless you

[@rafaella5]

Hello! Please I hope someone answer I really need help. How do I know if it should be before-after or after-before?

[@ethanbennink7680]

it works either way, just make sure that the tail is facing in the corresponding direction.

[@thibodeauxchris4171]

I am sure you get these comments a lot judging by the ones on this video, but you are educating a generation of STEM students. Your videos have gotten me through all of calc-calc 3, statistics, and DE. I owe you more than you will ever know. Your videos have allowed an old fart like me to return to school and change the lives of myself and my family. I hope you sleep well at night knowing that you have made a lasting impact on the entire world through your videos.

[@thibodeauxchris4171]

Forgot to add in linear algebra! I'd be toast without your hands-on videos.

[@signedturnip6442]

Shouldn't we use n-1 to do the standard deviation since we're talking about samples? or are the differences not considered samples in these cases?

[@xxv242]

Thank you very much!

[@cosmokramer7708]

I don't think we ever say there is sufficient evidence to support the alternative....we simply say we fail to reject the null. They are two different things and quite serious

[@theodoresweger4948]

Very well done thanks.

[@nansubugaangella9934]

❤thank you, you actually save lives

[@cjk9518]

Thank you so much,God bless

[@D4rkVenge]

Hey there, hopefully you see this. It's ok if you don't, I LOVE your videos, thank you so much. I was on the part when you find the Standard Deviation of the Mean of the Differences. and I found my Standard Deviation a different way; I used Sum of Squares to find the Standard Deviation - S (Standard Deviation of the Mean) = √Σ(x-Mean of Differences)^2 / N-1 (Degrees of Freedom). When I found the Standard Deviation with the variance formula, instead of your Standard Deviation, 13.025, I instead calculated 12.76. Do you also get that SD when you calculate the SD of the Mean with the Variance Formula?

[@adityendrapratapsingh7999]

When calculating 95% confidence interval, shouldn't you use t-value for 0.025, df=9( t table at 08:14). You used the t-value for 0.05 which will make it 90% confidence interval (10:53).

[@surihapsari6595]

Remember, it's a one-tailed test! :-)

[@itsrishika7022]

yea I had the same concern

[@angeleatitia1664]

@@surihapsari6595 CI is always 2-tailed

[@flameon8185]

t value must be 2.262 right?

[@chrisnedelchew]

You are right, CI is always 2-tailed. He just made a mistake.

[@dizzyboy92]

Excelent video. Just commenting to boost it.

[@analuciaaraujoraurau197]

Hey guys! In first place, thank you so much for the video. One question: Is it possible to use another parameter than the population mean in the hypothesis? Let's say I want to measure if the sample mean of a sports team's height is greater than a certain limit. This limit is not the population mean, but just a value established by rules. Could I run the t test?

[@eulastamayo328]

thank u for this

[@sinaabedini7487]

You are the greatest

[@mabelting963]

It helped a lot🙏🙏

[@malkim201]

Very valuable thank you so much

[@hallaallam7091]

Thanks you saved my statistic exam

[@harrybarden6053]

got on the wrong bus, on my way to uni and watching your videos before class. love your work

[@rsadsultanov3071]

Don't you practice online lessons?

[@peruskins96]

When do you take the absolute value of T and critical values?

[@faisalsadiq7698]

Thank you so much....God bless you

[@mohammedakhrib499]

Thank you so much

[@commonwang1457]

I have a question. what conditions we need to confirm in order to constract confidence interval?

[@Brian_LeFevre]

You’re a legend

[@michelleacab6024]

Hello. It is possible to use paired t-test if I get the mean scores of before and after, not its total scores?

[@nimoosman7716]

Thanks man

[@AN-ts4rf]

Man, thank you one more time!! for the second year in a row your saving my ass!!

[@bertram1724]

how do I get the P-value from this?

[@amirhosseinnouri6821]

Thanks

[@joanchong6622]

hi, may I ask if the sample size data of before and after is not the same, how can I solve this?

[@zarazafar4886]

Is it always after - before?

[@ghaniyd6292]

Hi, do you know the source book/journal for paired t-test?

[@tjpradeep9217]

Bro how to choose one tail or two tail for sample distribution

[@akhilanvijayan]

when I calculate SD I am getting a different value

[@user-pq7nx4kc4t]

On calculating the Confidence interval, why did you get the t critical at alpha not alpha/2?

[@yajatchoksi2652]

how do we/you know that the grap is left or right-sided and how to know that the grap is one tail or two tail

[@CamiloSanchez1979]

Why am I getting -21.173, -5.03 on my calculation for the confidence interval when I use the equation given by professor?

[@riky2107]

for confidence interval shouldn't we use the two-tail t-score, in this case "2.26" 9df?

[@Beingzuuu]

what if the n is more than thirty, do we still use this formula and just replace the t with z?

[@itzboyon6983]

This teacher is not a human, trust me! He is a God, a life-saving God! 😢😢😢

[@Strikerfm1]

I did this on my own before watching the video. What I did was get the avg weight of before the program, the avg weight of after the program, and the SD from the after the weight loss program. I then used the formula and still ended up rejecting the null hypothesis. Was this approach correct? here is the data . AVG before (206.5) AVG after (193.4) SD 18.1945. I got a calcutalted t value of -2.277 which is way past the -1.833 .

[@hyerijeong9332]

Shouldn't we use .025 instead of .05 for t when we do the confidence interval?

[@leahhall7504]

No, I don't think so. I also am working on stats atm and struggled with that question. From what I have found, even when you construct a confidence interval for two-tailed tests, you still use the critical t value that corresponds to the given alpha. I think you divide alpha by two when you are finding the p-value of a two-tailed test.

[@darknessdash1853]

the example is a one tail t test, so you dont divide alpha by 2, you can understand that from the null hypothesis, since it uses greater than or smaller than < > it should be one sided.

[@noahchoi1213]

if it was a two-tail test you would use .025 on each tail, but since its left-tailed its just .05 on the one side.

[@arlenemaesalera949]

is it a general rule to subtract after to before (after - before)??? really need the help.

[@philipcho231]

No. But in this case they wanted weight reduction so before-after would be easier to explain

[@Thevisibleman844]

How can I calculate it by using the calculator

[@nimoosman7716]

Wilcoxon rank sum test ?

[@dthakur9056]

What if t value is in negative

[@Ventixo]

I dont understand why for the weight loss, the null hypothesis isn't mewd is less than 0. Isnt the status quo that the weight loss program IS effective??

[@samueldarenskiy6893]

shouldnt we have t alpha/2 since we are dealing with confidence intervals? So, ~2.26

[@jadhavjadhav4480]

v good

[@ratnanirupama]

Isnt the t value formula meant for an unpaired t test?? this video is on paired t test right

[@mdichathuranga1]

hi thanks for the video, just need to get a point clarified when getting the confidence interval shouldn't the (Alpha /2) be used which is 0.025, But 0.05 .... because for confidence interval we usually use t(Alpha/2) right ? .... in your previous video, you mention we should be using t(Alpha/2) : kzbin.info/www/bejne/g4anZGxmn9mah9U would you be able to advise on this ?..... Thanks, best reagards.

[@bffchannel9860]

Ishan Chathuranga nah cuz it’s 1 tail

[@mdichathuranga1]

@@bffchannel9860 thanks 👍

[@surihapsari6595]

Can anyone tell me why my upper and lower values of 95% CI of the difference in SPSS are different to those in the video (-22.418 & -3.782)? All other numbers are the same, the mean, std. deviation, std. error mean, calculated t-value, and df.

[@surihapsari6595]

Using the CI formula in this video kzbin.info/www/bejne/in61c3p8fs9jiK8 gives me the same values as SPSS.

[@itsrishika7022]

when calculating the CI you have to use t at alpha/2 so t0.025 at df = 9 which would give you these values

[@peace58489]

Im confused. I thought when the Statistic > critical value you reject the Null Hypothesis but you did the opposite. Please explain why because i thought this rule was consistent throughout statistics.

[@woodchuk1]

joseph smith In this case, the alternate hypothesis is directional, meaning that the test is one-tailed. It happens to be left-tailed because the alternate hypothesis involves a mean difference less than zero. Since it’s left tailed, the critical value is negative. The rejection region is always to the more slender side of the curve; therefore, if the critical value boundary is negative (on the left side of the curve), the rejection region is farther to the left of the critical value boundary, that is, where the curve is skinnier. This amounts to saying that the null is rejected if the test statistic is more negative than the critical value (which is itself negative in this case).

[@sheilafajardo240]

no it depends on the test used

[@devotion_tws]

wait why md

[@zacharyh8253]

Why is it -3.18 for Tc instead of -0.318 ? Rounding?

[@theodoresweger4948]

Just because I have nothing better to do I used my HP35s for a trend line and I got a r factor of +.914 which means it has a decent probability, and the m factor was .624 and b factor was 64.4 giving a formula final weight equals aprox .624*initial weight+64.4 comes close have fun...Not responsible for minor errors this is statistics!!!!

[@philipcho231]

Anyone know why null is >=0 not =0? I thought null was about no change.

[@KamranKhan-pm9sj]

hello i need your help

[@__evanhart5320]

the standard deviation i get on my calculator ( because i can input it there) is 12.3. idk why im always a few points off when it i do it that way

[@justinmulyagananda1057]

Did you use the Sample Stdev?

[@amaanahmed8936]

@@justinmulyagananda1057 yes i got 12.3 too, i used the general standard deviation formula, is it wrong?

[@ermutanda3802]

I know this is super late but why did he use the t values for 1-a and not 1-a/2 when computing the confidence interval? The way he did it he found it for a 90% confidence level no for a 95% confidence level like requested

[@KamranKhan-pm9sj]

Sometimes we do (after - before) and sometimes (before - after) ... what tge reason?

[@gw7403]

Null hypothesis should be u(d)=0 instead of u(d)>=0. It is also more informative to calculate and compare probability value to significance level instead of conceptualizing t values.

[@lavishangel4069]

How do I pay my tuition fee?

[@marlanarsi2590]

Qusetion ? Ha : is greater than 0 because before was heavy large number minus after smaller weight gives numbers above zero