Sampling from a Distribution, Clearly Explained!!!

แชร์
ฝัง
  • เผยแพร่เมื่อ 7 พ.ค. 2017
  • What does it mean to sample from a distribution and why would anyone ever do it? Find out by watching.
    For a complete index of all the StatQuest videos, check out:
    statquest.org/video-index/
    If you'd like to support StatQuest, please consider...
    Buying The StatQuest Illustrated Guide to Machine Learning!!!
    PDF - statquest.gumroad.com/l/wvtmc
    Paperback - www.amazon.com/dp/B09ZCKR4H6
    Kindle eBook - www.amazon.com/dp/B09ZG79HXC
    Patreon: / statquest
    ...or...
    TH-cam Membership: / @statquest
    ...a cool StatQuest t-shirt or sweatshirt:
    shop.spreadshirt.com/statques...
    ...buying one or two of my songs (or go large and get a whole album!)
    joshuastarmer.bandcamp.com/
    ...or just donating to StatQuest!
    www.paypal.me/statquest
    Lastly, if you want to keep up with me as I research and create new StatQuests, follow me on twitter:
    / joshuastarmer
    #statquest #statistics

ความคิดเห็น • 105

  • @statquest
    @statquest  2 ปีที่แล้ว +2

    Support StatQuest by buying my book The StatQuest Illustrated Guide to Machine Learning or a Study Guide or Merch!!! statquest.org/statquest-store/

  • @noorghadban2238
    @noorghadban2238 ปีที่แล้ว +18

    I am a grad student taking Biostatistics this semester and this channel is saving my life. Truly CLEARLY explained. No click bait whatsoever

    • @statquest
      @statquest  ปีที่แล้ว

      Thank you very much! :)

  • @Dataforliving
    @Dataforliving 5 ปีที่แล้ว +56

    I appreciate the effort to simplify and will always give thumbs up when the concept is crystal-clearly explained. Otherwise, I'd rather request to explain what it means to "do a test on a sample" and why is a large p-value expected. There seems to be a disconnect between one condition and its effect as stated here. I appreciate all your other videos though! May I suggest a real-world example (like the ones to which you resort in most of your other videos)

  • @joaoostrowski
    @joaoostrowski 5 ปีที่แล้ว +16

    I recently found about Stat Quest channel and I fell in love with the videos, watching everything from the basics. For this video, though I wanted to give feedback, too many new concepts were introduced for trying to give a broader explanation of sampling distributions. I'm confident with my AP studies but I got a bit confused with the video, Maybe if the focus were on sampling distributions for estimating a statistic that we don't have from the population it would've been more clear. This is at least the strategy followed by Khan Academy when introducing this part for their AP.

  • @DanielTysonM
    @DanielTysonM 3 ปีที่แล้ว +11

    I think you could clearly explain this particular topic even better, so that it is on par with your other videos 😊 otherwise, I completely treasure your channel!

  • @sarrac3682
    @sarrac3682 4 หลายเดือนก่อน +1

    Thank You!!!! I love your videos, and how they explain things in simple terms. I am a social sciences student but have statistics for social sciences in uni. Your videos wake up my brain and help me understand.

    • @statquest
      @statquest  4 หลายเดือนก่อน

      Wow, thank you!

  • @shahjahanbd2000
    @shahjahanbd2000 ปีที่แล้ว +1

    I like your examples and explantions! Easy to understand complex concepts!

  • @geralddevera
    @geralddevera 6 ปีที่แล้ว +5

    Thank you very much Sir for this recommendation! Thank you for always singing prior starting the discussion! It helps students like me to be motivated! :D

  • @ThomasHaberkorn
    @ThomasHaberkorn ปีที่แล้ว +3

    I just bought 'The StatQuest Illustrated Guide to Machine Learning' and I'm loving it. Wonder if you would also do a book about the
    'Statistics Fundamentals'.. would love it too

    • @statquest
      @statquest  ปีที่แล้ว +1

      Great suggestion!

  • @ywang4781
    @ywang4781 5 ปีที่แล้ว +2

    These are such awesome and helpful videos! I feel so much more interview ready. Thank you so much! :D

    • @statquest
      @statquest  5 ปีที่แล้ว

      Awesome!!! Good luck with the interview. :)

    • @Dupamine
      @Dupamine 4 ปีที่แล้ว +1

      how was the interview bro

  • @saratkumarp2906
    @saratkumarp2906 4 ปีที่แล้ว +1

    Do we have to increase the sample size each time and calculate mean for each sample size ? In tat case mean of the sample sizes will form a normal curve irrespective of the type of curve of the original population
    Am i right?

  • @rrrprogram8667
    @rrrprogram8667 6 ปีที่แล้ว +2

    Nice... Thanks foe the video

  • @PuneetMehra
    @PuneetMehra 20 วันที่ผ่านมา

    This is your 9th video after Intro and you directly jumped to "large p-value", without explaining what is p-value and what is "large p-value", "t-test", etc in any of the previous 8 videos!

    • @statquest
      @statquest  20 วันที่ผ่านมา

      Sorry about that. You'll start learning about those in the next video.

  • @mkilptrick
    @mkilptrick 5 ปีที่แล้ว +48

    What is meant by p-value? t-test?

    • @Dataforliving
      @Dataforliving 5 ปีที่แล้ว +2

      Same question here.

    • @-long-
      @-long- 4 ปีที่แล้ว +4

      they were explained in different videos, check this channel's playlist

    • @kadenhunter805
      @kadenhunter805 3 ปีที่แล้ว

      You probably dont care but if you are stoned like me during the covid times you can stream pretty much all of the latest movies on instaflixxer. Have been binge watching with my brother for the last months =)

    • @jacobyjeremias9585
      @jacobyjeremias9585 3 ปีที่แล้ว

      @Kaden Hunter Yup, have been watching on Instaflixxer for months myself :)

  • @sukumsr1212
    @sukumsr1212 3 ปีที่แล้ว

    Hello @statquest, these are wonderful videos and a great contribution to us all. thank you. Is there a way we can get the slide decks as PDFs. It could help a direct import to OneNote so that we can then make notes on the slides.. thanks in advance for considering

    • @statquest
      @statquest  3 ปีที่แล้ว +1

      I've made study guides for some of my StatQuest videos here: statquest.org/studyguides/

  • @zuhdibadwan8742
    @zuhdibadwan8742 2 ปีที่แล้ว +1

    Give this guy a medal

  • @letsjoinhands
    @letsjoinhands ปีที่แล้ว +1

    So in summary random samples are taken from statistical distributions to perform statistical tests on the sample data in order to compare our expectations (of the results of those tests) with the actual results of those tests that could be different from our expectations (based on some already known facts about the original distribution from which those samples were taken). Is that right?

    • @statquest
      @statquest  ปีที่แล้ว +2

      Yep, that's it. However, there are also other uses. For example, over the weekend I wanted to learn how Partial Least Squares worked, so I generated some data from a normal distribution and plugged it in. This was helpful because I knew a lot about this data and this made it easier to interpret the results.

  • @daviddunmer3889
    @daviddunmer3889 3 ปีที่แล้ว +2

    Can you please point out some problem sets for practicing these concepts? otherwise retention is quite low.

    • @statquest
      @statquest  3 ปีที่แล้ว +1

      I have a few "quizes" for topics, including this one, on this page: statquest.org/video-index/ Just look for "StatTest!"

  • @jhanvilotwala1234
    @jhanvilotwala1234 2 ปีที่แล้ว +2

    Hey Josh! Can you please explain what are t-tests?

    • @statquest
      @statquest  2 ปีที่แล้ว +1

      I have some videos that explain t-tests in a very unconventional way. If you are curious, check out: th-cam.com/play/PLblh5JKOoLUIzaEkCLIUxQFjPIlapw8nU.html

  • @blakef.8566
    @blakef.8566 2 ปีที่แล้ว +1

    How does a computer make the selection based on probabilities described by the histogram. Do you have any papers or texts that talk about the algorithm?

    • @statquest
      @statquest  2 ปีที่แล้ว +1

      Say like we have a histogram with three options along the x-axis, "a", "b" and "c". The probability associated with a = 0.5, the probability for b=0.3 and the probability for c = 0.2. (NOTE: if we add up those probabilities we will get 1). Now, have a computer generate a random value between 0 and 1. Since there is a probability of 0.5 that that number is between 0 and 0.5, select 'a', we will select 'a' if the value is between 0. and 0.5. Since there is 0.3 probability that that number is between 0.5 and 0.8, we will select select 'b' in that case. Lastly, since there is 0.2 probability that that number is > 0.8, we will select select 'c' in that case.

    • @blakef.8566
      @blakef.8566 2 ปีที่แล้ว +1

      @@statquest Oh that's very simple, thank you

  • @nitin.cherian
    @nitin.cherian 5 ปีที่แล้ว +7

    Josh, what is p value as mentioned in this video ? Is there a video on that ?

    • @statquest
      @statquest  5 ปีที่แล้ว +4

      Yes, th-cam.com/video/5Z9OIYA8He8/w-d-xo.html
      One thing I might not have put enough emphasis on in that video is that p-values are how statisticians decide if something they see is special or not.

    • @nitin.cherian
      @nitin.cherian 5 ปีที่แล้ว +2

      @@statquest Josh, I understood the calculating the p value. Could you please upload a video on how p values are used by statisticians, how p values are significant?

    • @statquest
      @statquest  5 ปีที่แล้ว +6

      @@nitin.cherian Yes! I will.

  • @Oscar-jg9gg
    @Oscar-jg9gg ปีที่แล้ว

    @ 2:27, what are statistical tests? Also, is the reality distribution sampling and our expectation the statistical distribution?

    • @statquest
      @statquest  ปีที่แล้ว

      I guess this video assumes that you are already familiar with statistical tests. :(

  • @santoshgujar5237
    @santoshgujar5237 2 ปีที่แล้ว +1

    Thank you, Sir

  • @lilyha2470
    @lilyha2470 4 ปีที่แล้ว

    Hi, any videos on X2, T-test, ANOVA and ANCOVA?

    • @statquest
      @statquest  4 ปีที่แล้ว +2

      Interestingly, t-tests and ANOVA are just types of linear models (i.e. a form of linear regression). I would recommend watching my Linear Models play list (you can skip the videos where I explain how to do the work in R, which is a programming language): th-cam.com/play/PLblh5JKOoLUIzaEkCLIUxQFjPIlapw8nU.html Alternatively, you can just pick and choose the videos you want to see from my index: statquest.org/video-index/

  • @InDepthPrep
    @InDepthPrep ปีที่แล้ว +2

    you introduced new concept likr t test p value

    • @statquest
      @statquest  ปีที่แล้ว

      True. I made this video a long time ago for my colleagues at work who already knew about those concepts.

  • @dorindragos
    @dorindragos 4 ปีที่แล้ว

    May one use a t-test on such small samples (3-strong each)? Shouldn't one use smth like a Mann-Whitney test?

    • @statquest
      @statquest  4 ปีที่แล้ว

      It is very common to use t-tests on small samples (n=3). How good an idea that is depends on how similar the underlying distribution is to a normal distribution. If the underlying distribution is very different from being normal, then n=3 is probably too small for a t-test. However, the underlying distribution is vaguely normal, then n=3 is fine.

  • @Hossein_am98
    @Hossein_am98 ปีที่แล้ว

    thank you for your video
    in fact what i didn't understand and was ambigue for me what that i didn't know what is P-test and t-test is and i did'nt really understand what is N

    • @statquest
      @statquest  ปีที่แล้ว

      OK. Thanks for the feedback! I'll modify my playlist so that this topic comes after those. (By the way, all of my videos are here: statquest.org/video-index/ )

  • @amnont8724
    @amnont8724 ปีที่แล้ว

    2:49 - Hey Josh, given that the distribution is the same, why is the P-Value high if we take a few different samples? Aren't the samples sort of related because they belong to the same distribution?

    • @statquest
      @statquest  ปีที่แล้ว +1

      When the samples are taken from the same distribution (mean that there is not a fundamental difference in the samples) the p-value should be high. To learn more about this, see: th-cam.com/video/vemZtEM63GY/w-d-xo.html

    • @amnont8724
      @amnont8724 ปีที่แล้ว

      @@statquest Ohh, because the closer the P-Value comes to 1, the more confident we are that the samples are NOT from a different distribution?

    • @statquest
      @statquest  ปีที่แล้ว +1

      @@amnont8724 The wording associated with p-values is tricky and subtle. The closer the p-value is to 1, the less confidence we have that the data come from different distributions.

    • @amnont8724
      @amnont8724 ปีที่แล้ว

      @@statquest Yes the wording regarding P-Values feels very tricky, especially when every other TH-camr and lecturer besides you uses a slightly different explanation. Thank you :D

  • @Keke-xp1ee
    @Keke-xp1ee 2 ปีที่แล้ว +1

    Why an I seeing about p test before learning about p test in the playlist. Please make an ordered playlist

    • @statquest
      @statquest  2 ปีที่แล้ว

      Which playlist are you looking at?

  • @lion87563
    @lion87563 ปีที่แล้ว

    As far as I understand, sampling from a distribution versus sampling in signal processing is this: Instead of sampling a y based on x, we now sample x based on y.

  • @shekharspeaks129
    @shekharspeaks129 ปีที่แล้ว +1

    AT 2:51, you said "since the distribution is same, T-test should be give a large p-value." I do not understand this point.

    • @statquest
      @statquest  ปีที่แล้ว

      To understand p-values, see: th-cam.com/video/vemZtEM63GY/w-d-xo.html

  • @ShashankSingh-yi5vi
    @ShashankSingh-yi5vi 2 ปีที่แล้ว

    Sir please cover spss in the same way...dying for it

    • @statquest
      @statquest  2 ปีที่แล้ว +1

      I actually asked SPSS for a copy of their software so I could make videos about it and they said "no".

    • @ShashankSingh-yi5vi
      @ShashankSingh-yi5vi 2 ปีที่แล้ว +1

      Hey Josh first of all thanks for such a wonderful job. I always recommend you channel to my colleagues. Also you can try trial version of IBM SPSS. Please do this. Thanks

  • @pdkovacs
    @pdkovacs 4 ปีที่แล้ว +2

    Does "taking a sample from a distribution" as opposed to "taking a sample from the population" mean that you take a sample from a sample?

    • @statquest
      @statquest  4 ปีที่แล้ว +1

      This is a really good question. The answer is, "it depends". If the distribution is based on a small sample of data (for example, if we measure how tall 20 people are and calculate the mean and standard deviation of the heights, and then create a distribution based on the estimated mean and standard deviation), then, in a way, we are using the distribution to give us new measurements that are related to the original sample. However, typically when we "sample from a distribution", the distribution isn't based on an existing sample - instead we just pick values for mean and standard deviation that suit our needs. For example, at 2:59 in this video, I could set the standard deviation to 1 for both distributions and the number of measurements from each distribution to 3, and then plug in different values for the means to determine how far apart they must be in order to consistently get a small p-value. Does that make sense?

    • @pdkovacs
      @pdkovacs 4 ปีที่แล้ว +2

      @@statquest Yes, I think I understand -- and now I've realised that you actually give a definition-like description of "taking samples from a distribution" at the very end of the video. (I've posted my question as soon as I started to watch this video.) Thanks a lot for the explanation and thanks a lot for all your excellent tutorials on TH-cam. They give insights into aspects of some of the most important stats concepts I cannot easily find in other materials on the Internet.

    • @statquest
      @statquest  4 ปีที่แล้ว

      @@pdkovacs Hooray! I'm glad I can help.

  • @haorantao612
    @haorantao612 ปีที่แล้ว +1

    These videos really saved my life!! I hereby announce that Josh Starmer is the hottest guy alive!!!

  • @carzetonao
    @carzetonao ปีที่แล้ว +1

    Why my high school teachers and university professor can not teach it in this way.i think probably they both are confusing for what they are saying, so they chose to read all concepts from textbook.

  • @AkshayrajKore
    @AkshayrajKore 7 หลายเดือนก่อน

    What is p-value ?

    • @statquest
      @statquest  7 หลายเดือนก่อน

      Great question! To learn about p-values, check out this video: th-cam.com/video/vemZtEM63GY/w-d-xo.html and th-cam.com/video/JQc3yx0-Q9E/w-d-xo.html

  • @mahdiamrollahi8456
    @mahdiamrollahi8456 3 ปีที่แล้ว

    So, when we have the population, why we need to sample data?

    • @statquest
      @statquest  3 ปีที่แล้ว

      If you have the population, there is no need for sample data. However, you rarely have the population. For more details, see: th-cam.com/video/vikkiwjQqfU/w-d-xo.html

  • @ehtishamullah1510
    @ehtishamullah1510 2 ปีที่แล้ว

    Now youtube shows PD curve for average watch time against video time

  • @bearbearbear7098
    @bearbearbear7098 2 ปีที่แล้ว +1

    Not clearly explained.....what is t-test and p-value ?

    • @statquest
      @statquest  2 ปีที่แล้ว +1

      Sorry. This is an old video - one of my first - and back then I thought that only my coworkers would watch my videos, and since we had been talking about p-values and t-tests, I thought they would understand. However, if you want to learn more about statistics, see: th-cam.com/video/oI3hZJqXJuc/w-d-xo.html th-cam.com/video/vikkiwjQqfU/w-d-xo.html th-cam.com/video/SzZ6GpcfoQY/w-d-xo.html th-cam.com/video/yQhTtdq_y9M/w-d-xo.html

  • @theRealtensigh
    @theRealtensigh 2 ปีที่แล้ว

    What is the p-value? I watched it twice and I couldn't figure it out. Sorry if it's obvious to everyone, I'm a rookie.

    • @statquest
      @statquest  2 ปีที่แล้ว +1

      To learn more about p-values, see: th-cam.com/video/vemZtEM63GY/w-d-xo.html and th-cam.com/video/JQc3yx0-Q9E/w-d-xo.html

    • @theRealtensigh
      @theRealtensigh 2 ปีที่แล้ว +1

      @@statquest thank you, I will check that out

  • @ChadMc74
    @ChadMc74 3 ปีที่แล้ว +1

    I took a p test once

  • @Ilemaurice687
    @Ilemaurice687 2 ปีที่แล้ว

    Hey plzzze i need help if possible can you help me plz

    • @statquest
      @statquest  2 ปีที่แล้ว

      It really depends on what you need.

    • @Ilemaurice687
      @Ilemaurice687 2 ปีที่แล้ว

      @@statquest wait i will send you the question

    • @Ilemaurice687
      @Ilemaurice687 2 ปีที่แล้ว

      @@statquest thx u

    • @Ilemaurice687
      @Ilemaurice687 2 ปีที่แล้ว

      @@statquest Question:
      To examine the bone mineral density of women with ankle fractures, the investigators recruited 10 postmenopausal women with ankle fractures and 12 healthy postmenopausal women to serve as controls. The stiffness index of the lunar Achilles in each woman was obtained. The mean stiffness index for the ankle fracture group was 76.4 with a standard deviation of 5.83. In the control group, the mean was 82.3 with a standard deviation of 6.34. Assume that both samples are drawn from normal populations.
      (i) Test at 5% level of significance, whether the variances of the stiffness
      indices for the two groups are equal.
      (ii) Using p-value approach, examine whether these data provide sufficient
      evidence to conclude that, in general, the mean stiffness index is higher
      in healthy postmenopausal women than in postmenopausal women
      with ankle fractures? Take a=0.05
      (iii) Obtain a 95% confidence interval for the difference of two population
      mean stiffness indices. Does this interval confirm the conclusion derived
      in part (ii).

    • @Ilemaurice687
      @Ilemaurice687 2 ปีที่แล้ว

      @@statquest plz help me

  • @billmichae
    @billmichae 3 ปีที่แล้ว

    One remark Josh: Please control terminology used. You are too loose when "speaking" probability or statistics language.