Statistical Significance and p-Values Explained Intuitively

[@danielalexanderyt3232]

Took me 50 minutes to thoroughly go through this video and take notes for my AP psych test. Speaking of it, HOW IS THIS AP PSYCH

[@TheReef70]

Okay so I understood that the bigger the number the weaker the evidence and truer to the null hypothesis. The pace, clarity and preserving the material to basics is SUPER FREAKING supportive of a growth mindset as an adult learner.

[@DataDemystified]

Happy to help how I can!

[@soldotnamoose]

Wow, I wish my professor would have shown us this video in the beginning of our statistics class instead of the end. This is simple enough for me to understand a little more about statistics. Thank you

[@smitasingh8764]

This is the best explanation I got so far. I have watched so many videos but I didn't really get the intuition of what this p-value is. You made it all pretty easy to understand!

[@DataDemystified]

Glad it was helpful!

[@arunavamaulik19]

couldn't agree more!

[@K_Fogs]

I have watched so many KZbin videos just trying to get my mind around 'statistically significant'. The way that you phrased this helped me to FINALLY understand. That lightbulb went on. Thankyou. :)

[@yubuwang8823]

This video is amazing!!!!!! I'm a Chinese student, I do a lot of research about this topic, no one explains this as clear as you are! I will pay to see your lecture!

[@DataDemystified]

Thank you! No need to pay :). Just tell your friends if they’re also trying to learn!

[@DHDH_DH]

I am also from China and am currently taking an introductory statistics course in the U.S. No one on KZbin gives a better explanation than you on the significance level and P-value.

[@jeffgalak9402]

@@DHDH_DH thank you!

[@Shanlii]

This was the best explanation regarding this topic! Dr. Galak explained it so well in a manner that makes sense and sticks; his intuitive explanation allows me to use this logic and apply it any question on the USMLE step 1 exam regarding this concept. Thank you so much!

[@AGWalp]

Ph.D. candidate (second doctorate): Outstanding, sir! The easiest (best) description that I have ever read or watched. I feel more confident in my comments about p and H0. Thank you for not getting into the "weeds!" That is where most of us get lost, and the jargon only makes it worse.

[@triptajaswal8276]

Thankyou 😭 it is bestest video for understanding p value ... God bless you

[@DataDemystified]

Glad it was helpful!

[@jeffreysayed8621]

Thanks for your explanation. You clarified the concept in plain English. I saw in another video how in Statistics significance has nothing to do with big and important. Something can be small or unimportant but still be "statistically significant."

[@Mabunt3]

I scoured the internet and my textbook and this is the best explanation I've come across. Thank you!

[@zohaibakhter3049]

The intuition that you provided in the beginning was exactly the piece I was missing to solve this cryptic p value puzzle! Thank you so much, please never stop teaching.

[@DataDemystified]

Thank you for the incredibly kind comment!

[@emmanuelessien9527]

This is the best explanation to statistical significant studies. Even a grade student will understand this. Thanks Dr.

[@vupham7361]

OH MY GOD! This video really helps me understand the meaning of p-values not just some statistical jargon.

[@whyisgamora4191]

OH MY GOD THANK YOU. I've watched soo many videos. My professor never explained that rejecting Null was something desirable in a study & always used abstract definitions. I finally understand it now. Subbed

[@DataDemystified]

Happy to help!

[@HKC6924]

God sent.. very good explanation.. after a lot of head scratching, I now finally understood the whole idea behind the p values and rejecting the null hypothesis. Thank you 👍

[@DataDemystified]

Glad I could help!

[@asma7918]

Finally I understood p value.. Best explanation ..bless you..

[@a1x45h]

This is the only video that made sense to me. Watched like 10 videos about p value and null hypothesis. This one was surprisingly the most intuitive. Thank you so much!!! :)

[@DataDemystified]

I'm so glad this helped you understand the concept!

[@kathrynayoob3522]

Thank you. Super helpful. I have been studying for the National Counselor Exam. Now I understand why the numbers seem arbitrary. Because they kind of are!

[@aliasgari5152]

After watching 3 videos on p-value got to this one and I'm happy now. Keep doing.

[@DataDemystified]

So glad that this was helpful!

[@huyarz61]

Your explanation is really helpful! I spent about an hour to watch others videos and try to understand what level of significance means, but most of them merely just mention it in a sentence without explanation. I was a bit frustrated until I watched your video !

[@ceevaaaaa]

OMG! This was like the best, most intuitive candid explanation ever. After struggling with so many videos, finally this one made sense. Thank you very much Galak sir.

[@DataDemystified]

Really glad you found this useful!

[@xxerggher2667]

After watching so many videos on p-values, this is the clearest explanation by far. Liked and subscribed!

[@DataDemystified]

I'm so glad you found it useful! If there are other topics you think could use some intuitive clarification, please let me know!

[@xxerggher2667]

@@DataDemystified Thanks for the reply. Maybe some people would appreciate a clear explanation of p-hacking.

[@Sluo1947]

Awesome explaination. I never understood why want to find evidence against H0, now it's clear.

[@DataDemystified]

Thanks! I'm so glad this helped you understand the concept better!

[@Ficccs]

Thank you, well explained! Been having a tough time understanding this

[@DataDemystified]

Happy to help!

[@exploring197]

I would love to see the continuation video on "How to calculate P-value?". Thanks for your efforts. Keep teaching.

[@joannameringoff6870]

Loving this - a little slower talking would be wonderful but can always go back a few seconds and re-listen.

[@DataDemystified]

Thank you! Yes, I'm working on finding just the right speaking speed. Feedback taken!

[@yunusskom1743]

Mesej yang jelas, struktur yang jelas, mudah difahami, terima kasih

[@jaypalmer2463]

this was the best explanation I have ever come across - thank you!!!!

[@BeyondNoise9]

Amazing !!! Please ,please, keep uploading more such videos. You have no idea how much they help us !!

[@DataDemystified]

Thanks! Have a look at the other videos on the channel as there is already quite a bit of similar content.

[@unnatigupta940]

OMG Thank you so much i was having hard time understanding the concept of P-value, but you made it very simple, even the people from non data scientist field can understand it.

[@DataDemystified]

Very happy to help!

[@mrazafarooq]

That is the best explanation so far for statically significant. Please introduce a course for statistics used in science to arrange the data and understand what statistical measures can be taken for the data?

[@martasobralgarcia3996]

This is the best video I ever saw explaining something on statistics! The explanations are great to understand the logic of it.

[@DataDemystified]

So glad it was helpful!

[@rockmamabaker8370]

Wow! I love this video! I've been trying to get my head around this for months. Really clear video. Great information. I feel slightly better about the topic- but statistics is hard and it takes a while to stay in my brain! Thank you!

[@TheKothay1]

Thank you for using the legal analogy to explain null hypothesis. Had been so confused so far. Also much thanks for such a simple explanation of the concepts of significance and P value. Looking fwd to some more videos.

[@DataDemystified]

I’m glad that this video helped clear things up for you!

[@Do_Worrk]

This is an amazing explaination. Subbed! Can you go into Confidnece Intervals, Bayesian Estimates, and Effect Sizes?

[@platinum11110]

I discovered the channel today, went to see what other videos there were but saw that it's been a year since his last video 😪

[@glowingmemories]

I've watched a few videos on P values/statistical significance and this is the one that made it click. Thanks!

[@DataDemystified]

Glad I could help you make the concept click!

[@Daizyonceagain]

This is the first spss lecture I enjoyed watching

[@DataDemystified]

Thanks! Glad you found it useful!

[@tonynikolaos3527]

Clear explanation, Indeed! Thank you! An area where I feel a “black box magic” happening, that would hope you can explain clearly is: in random permutation tests, data is shuffled randomly from both groups, some (less then factorial for practical purposes, yet sufficient ) number of times, and the subtracted means result makes a normal disturb plot. Now, we know: garbage in-garbage out. The randomization and permutations is possible “garbage”, over exposed and shafted with the original data from the 2 groups. Now, what I feel is magic, I’m supposed to believe, and come to conclusion for validation or rejection of the Ho, based on that original data, from Groups 1 and 2, when mixed with “garbage” to actually retain some of its qualities(or dilute them if it doesn’t) , through the random shuffled permutations, and if it has retained sufficiently( or diluted sufficiently) show me a P value that will disclose strong or weak tendency toward rejecting or accepting the Ho. So , basically, we use the random permutations, as if, analogizing, as if we could use some substance, and say it has those features, and then we mix it with “all kinds of “garbage tests” and eventually are able to say: see, it kept its features, through all those substantive permutations, must be strong, or, you see, it got diluted so much that lost those features, must be not enough strong, what you saw as feature was random feature case, so we keep the Ho. Something like that? Thanks a million- if you can touch upon this theme!

[@wotjake]

THANK YOU! This made so much more sense than other explanations. You Rock!

[@DataDemystified]

Glad you found it helpful!

[@bens1mo637]

Great explanation! Great, easy-to-understand, and clearly enthusiastic about psychology!

[@maliamercer140]

So well explained! Thank you for putting out great, simple content that is very intuitive. After watching other videos yours sunk in the best. Nice production quality too. Cheers!

[@PeterJunkerPaperPunker]

General suggestion: I wish you would speak a bit more slowly. Firstly because I I could better absorb the information and, secondly, because your voice naturally has a whispery fluctuation that obscures some syllables. In one place in this video you end a sentence with “zero” and the last syllable is barely there, before rushing on to the next sentence. I wouldn’t ordinarily respond with criticism, but I love your content so much I want to absorb it all as easily as possible.

[@DataDemystified]

Thanks Peter for e feedback! I know I speak to fast when I get excited...and I get very excited about making these videos! That said I will def try and slow down. As a stop-gap you can play videos at x.75 speed to help out. Thanks again!

[@cleverhype]

Omg, this is so much easier to understand in the way you explain it. Thank you!

[@hasiburbijoy007]

Finally, You give me a strong evidence to understand Statistically Significant! Thanks and Keep it up!!❤❤❤

[@DataDemystified]

Glad you found it useful!

[@hasiburbijoy007]

@@DataDemystified hope we'll learn more and more useful thing like this!!!❤

[@arunavamaulik19]

I find your videos extremely intuitive and would love to see you make a whole series on probability and statistics. There are quite a few on the internet but not many are intuitive.

[@DataDemystified]

Thank you! Are there any specific topics in probability and statistics you find confusing? It's always nice to know where to focus my efforts.

[@arunavamaulik19]

@@DataDemystified I would love to see a series on Bayesian inference! To be honest it could be any topic but a complete mini series would be helpful. Eg: I loved your video 'what's a distribution?', you could follow that up with a few more videos on similar topics. In any case, I hope you continue making these videos! pure gold

[@jeffgalak9402]

@@arunavamaulik19 Thank you for the suggestions!

[@rootenshi]

After hours of looking around and searching this topic including in my university book, KZbin videos, and scientific websites.. I finally understood the Statistical Significance from your explanation! So intuitive and clear. Thank a lot :) Also... Definitely subscribing to this great channel :D

[@DataDemystified]

I'm so glad you found this useful! If there are other topics in statistics that you find challenging to internalize, let me know. I'm always looking to make content that will actually help people!

[@rootenshi]

@@DataDemystified I really appreciate it. You will definitely be the first place I go to for demystifying data :)

[@SethWieder]

This was super helpful in my early understanding of theoretical statistics and what the null hypothesis means. Thank you!

[@DataDemystified]

Happy to help

[@emmanueldarkwa8311]

Excellent teachings. Please I need videos on sample size calculations and z score calculation

[@me951129]

Unfortunately giving more than one thumbs up isn‘t possible. This was brilliant!!

[@apoorvadk9635]

Thank you! Kudos for such a simplistic way of presentation!

[@DataDemystified]

Thank you!

[@bobbiejoohagan2639]

AGREE! This was SO helpful. And so clearly and simply communicated. Thank you!

[@DataDemystified]

Thank you!

[@hosseinhayer5618]

Best video about null hypothesis. thanks

[@al__karita]

Thank you for sharing your knowledge. Keep up the good work.

[@DataDemystified]

Thank you!

[@sachinchoudhary9043]

i don't think anyone can explain in a simpler way than this.

[@DataDemystified]

Thank you!

[@yvetteelizalde7396]

Great work! Going to share with all my CFA study friends!

[@DataDemystified]

Glad you found it useful! Just out of curiosity, is CFA: certified financial analyst? I ask because at CMU we have the College of Fine Arts (CFA) which is very different!

[@jbiswajit]

Thanks Professor. This was super easy to understand. Please keep posting videos.

[@DataDemystified]

Glad you found it useful!

[@festnotfirst]

This is a very easy to understand explanation, thank you.

[@iot3136]

Hi Prof. Jeff. Thank you very much. This is the first time in my life I learned why we need a Null Hypothesis. I am from a biology background. Using t, χ2, F tests day-to-day, following like a recipe, for years. However, I feel I must understand the background of this. Thank you for sharing your knowledge selflessly. I tried many times to get help from StackExchange forum to understand these concepts without many mathematical notations. However, their moderators with EGO, think either many should not understand these concepts in plain language or using their platform to show how elite they are. Thank you very much for bringing knowledge to every level, for different disciplines. If you do not mind I have few questions that I'd like to have your answers to. 1. So as I understood from the video, setting a null hypothesis is a mathematical requirement? 2. What p-value really tells in hypothesis testing? because some say it'd not about probability. 3. why having a low p-value is strong? (if we have to see this is a significant thing that occurs every time - e.g. support a vaccine is effective, don't we need a bigger probability to accept this?) 4. Is normal distribution a hypothetical thing? because why we have to use different t, χ2, F distributions for different data types to calculate a p-value. I am not sure why I fail every time to understand these; Is it because of mathematics or is it the language (wording) use to explain these concepts. Thank you very much for your great effort taken to teach students around the world.

[@DataDemystified]

So glad I could help! 1) Not a mathematical requirement, but rather a logical one. Statistical tests start with some kind of null hypothesis (which varies based on statistical test). What you are testing is whether you can reject that or not. 2) A p-value computes the likelihood that you would observe whatever data you got GIVEN that the null hypothesis is true. For example, let's say a vaccine really doesn't work at all (something you can't ever know for sure), but you observe that the difference between a treatment and placebo group in terms of infections is something like 1 out of 100 infected for the treatment group and 10 out of 100 for the placebo. The p-value would say; how likely is it that we'd see something like that if, in reality, there is no effect of the vaccine. If you got a p-value = .01, that would say: there is a 1% chance that the vaccine is ineffective in reality, but we'd observe such a result (or a strong result). 3) the lower the p-value, the less likely it is that you'd observe the results you got AND the reality is that there is no actual effect. 4) all parametric statistical tests assume that the data are distributed in some way. A normal distribution is just one example. F distributions and chi-sq are just others. Depending on the nature of your data, some distributional assumptions are more appropriate then others. There are also non-parametric tests that make no distributional assumptions, but that's a longer conversation! Good luck to you!

[@iot3136]

Thank you very much 🙏🏽 This is really helpful for me. I’m really greatful to you. Wishing you all the very best for you academic and personal life! Take care 🍀

[@jakedrums1469]

He just explained it 300 times better than my teacher

[@shreemahalakshmitextile3319]

😂😂😂😂😂😂

[@shreemahalakshmitextile3319]

Our teachers are working for the money only they do not have any problems if we do not understand 😢😢😢😢

[@shreemahalakshmitextile3319]

That’s right 😂😂😂😂

[@reuvenmeirherrera9580]

wow. very good at simplifying and explaining this VERY back and forth/ hard to grasp ideas ty

[@nutrimum2904]

THANK YOU!!!! What a confusion you've dissolved. Wonderfully explained! Thank you :D

[@DataDemystified]

Happy to help!

[@jamessujitmalo3910]

Great job. Thank you. Got a lot of valuable insights from the video.

[@venkateshgopalarathnam1933]

Brilliant explanation of intuition behind p-value!

[@thekalyangk]

What an AMAZING explanation !! Super clear. Thank you !!

[@DataDemystified]

Very glad you found it useful!

[@robertue1]

Thank you for sharing this video, you made it really easy to understand and digest. Keep up the good work.

[@DataDemystified]

Glad it was helpful!

[@edangenetfeleke1053]

Really amazing with good and simple example keep it up!!!!!!!!!!

[@DataDemystified]

Glad you found the video helpful!

[@stephenaloia6695]

Very straightforward - thank you!

[@DataDemystified]

Glad you found the video useful!

[@mjunk3241]

I really like your aim to convert the concept into commonsense. One thing that has been boggling my mind is the 'degree of freedom'. Can you demystify it? Thanks..

[@DataDemystified]

Thanks! I LOVE the idea of a video on degrees of freedom. That’s going high on my list of ones to make. Thanks for suggestion.

[@bobby7354]

This is such a great video & explanation, thank you so much for bringing p-values & any other kind of data into context, it's so helpful!!

[@DataDemystified]

Glad it was helpful!

[@robinvandepoele2117]

Hi! Just wanted to thank you for the video and to let you know that it’s so good, it and the video about confidence intervals are part of the obligatory curriculum in med school here in Belgium! Great work and keep it up

[@DataDemystified]

Thank you! Really glad it helps!

[@vikashprasad557]

Thank you, it was very helpful. Just one feedback, if you control your pitch it will be very helpful

[@minhhiennguyen6798]

really well explain video ! Hope to see more content from you

[@nistarasamyatva195]

Please make one on standard deviation. Thank you so much for the video.

[@DataDemystified]

I don't have one right now on that exact topic, but I do have a video about how to think about distributions, which is pretty close to what you're asking about. Have a look: kzbin.info/www/bejne/eGS5YY2lftB9pc0 ....that said, I do think a specific video on SDs makes a lot of sense and I'm adding it to my list of videos to make. Thank you for the suggestion!

[@ryangarbe8396]

Would love for you to make a video about calculating p-values!

[@DataDemystified]

Great suggestion. I don’t think I’ll do a video on the math since that’s not the point of this channel, but I like idea of discussing where the heck a p value comes from in the first place. Thanks!

[@MrGodulo7]

Thank you. Seek to pursue research even more based on where I work and the need to produce evidence to back up the work that we do. Hope to interact some more on this platform

[@danielsmith874]

Hi, is .352 a possible p value?

[@rainbowsandunicornsandmath7950]

Thanks! This is a great overview for my AP Research students!

[@DataDemystified]

So happy to help out! Though I have to admit I don’t recall an AP research class when I was in high school. What do you cover in it?

[@rainbowsandunicornsandmath7950]

@@DataDemystified This is from the AP site's course description (they say it much better than would I). I believe it's a fairly new offering (last 8ish years or less?). The students conduct their own research project and submit a paper on it. It's an awesome experience! My students are working on their methods, but I want them to learn a little about statistical analysis so they can gather useful data. AP Research, the second course in the AP Capstone experience, allows students to deeply explore an academic topic, problem, issue, or idea of individual interest. Students design, plan, and implement a yearlong investigation to address a research question. Through this inquiry, they further the skills they acquired in the AP Seminar course by learning research methodology, employing ethical research practices, and accessing, analyzing, and synthesizing information. Students reflect on their skill development, document their processes, and curate the artifacts of their scholarly work through a process and reflection portfolio. The course culminates in an academic paper of 4,000-5,000 words (accompanied by a performance, exhibit, or product where applicable) and a presentation with an oral defense.

[@DataDemystified]

That sounds amazing. I wish I had that! Thank you for being a teacher and caring about your students!

[@marcbeattie4484]

Such a great definition. Thank you!

[@DataDemystified]

Glad you found this useful!

[@e505mp95]

You're doing a very good video! Best explanations and examples quoted! It helps me a lots!! Biggest Thanks!!! Would you consider make video on how to use excel or google sheet to solve problem like Compute: P(-1.5

[@DataDemystified]

Thank you! I'm not sure that excel tutorials are in my future, but I'll definitely think about it!

[@wannabeevangelist8481]

I have taken moment to like the video and subscribe to your channel, thank you berry berry much, it feels good to finally understand

[@DataDemystified]

Thank you and happy to help!

[@aleksmusic94]

That was very helpful and clearly put. Thank you so much!

[@DataDemystified]

I’m really glad you found the video informative!

[@zinnijha7556]

Thanks for simplifying for us. I wondered why .05 ...:)

[@DataDemystified]

You're welcome!

[@yellowsniper1995]

Amazing video! Just subscribed! 👍🏽

[@JaneChirwa-db8ij]

terrific explanation. Thank you.

[@tanushreebhuimali2289]

Thank you making me understand the concept!

[@ankitakamath3120]

Thank you, your explanation was very effective. I finally understood the essence of p-value from your example. Subscribed! :)

[@jsklair]

Great video as always. Can you do similar primer videos on those alternative methods you mentioned ("Confidence intervals", "Bayesian estimates", and "Effect sizes")?

[@DataDemystified]

Absolutely. I already have one on Confidence Intervals (see here: kzbin.info/www/bejne/i4Wro6ltjqatnq8). I also have one that intuitively explains how to think about Bayes' theorem without really calling it that (see here: kzbin.info/www/bejne/b56yomyaYsqLrqs). I have one about Effect Sizes in the queue!

[@AshishKumar-xw3xk]

huge help man ! thanks a ton , god bless you for making this video.

[@frankl1]

liked the video within the first 3 minutes, owesome

[@ВікторФіщук-б7я]

Thank you for this great example!

[@AndreaEsmeraldaHernandez-h9k]

Thank you very much for your explanation! I finally understood!! Great Job

[@towobolafunmilayo3224]

Thank you, you real made it better to understand

[@timhowell6929]

Very nice explanation, thank you!

[@Jason-o5s]

Cheer~~~the quality of being worthy of attention--- importance.😊

[@AlvinCheong]

These videos are super helpful for me! I'm a mechanical engineer at a large tech company in the Bay Area and this helps me understand my data. Can you discuss Cpk vs. Ppk and all the topics surrounding such?

[@DataDemystified]

Thanks. Glad you're finding them useful! Cpk vs Ppl is a pretty technical topic. What in particular would you like to understand about it?

[@debla123]

Really liked the way you explained

[@sahareeshjabeen]

excellent explanation really helped me!

[@DataDemystified]

Glad to hear it!

[@aragzan9]

Simply amazing explanation

[@DataDemystified]

Thank you!

[@mehnazmannan1386]

How should we think about how to choose the null hypothesis? Can the null hypothesis be “a dropped ball will always fall”? Or should it be “a dropped ball will fall”? Does the distinction of “always” matter in crafting a null hypothesis? Or is the null hypothesis something that has a chance of being rejected, but then how would we even necessarily know so in advance? Tldr is what are the assumptions, if any, in crafting a null hypothesis?

[@jamirie61046]

Great information! Just subscribed.