Inference for Two Means: Introduction

Поділитися
Вставка
  • Опубліковано 25 січ 2025

КОМЕНТАРІ • 36

  • @silviaying2934
    @silviaying2934 3 роки тому +3

    literally sobbing because this channel saved my college life

  • @ody3382
    @ody3382 5 років тому +6

    I learned more with this video alone than with two weeks worth of lecture. Thank you so much for this video! You rock! :)

  • @thilishamoodley8883
    @thilishamoodley8883 6 років тому +16

    How is it that you can explain these concepts in under 10 minutes and so well?? Yet my lecturers can't even do it in a whole semester? 😭🙏🏾

  • @johnbarneswood
    @johnbarneswood 5 років тому +9

    When you say you derive equations elsewhere, would be super helpful to provide a link to that video so I can watch more of your videos

  • @tunjiadewoye448
    @tunjiadewoye448 3 місяці тому +1

    Thank you bruv, I've been trying to get this concept for days

  • @jenna.f.8882
    @jenna.f.8882 8 років тому +25

    Would have liked to see an example.

  • @MrAndreaCaso
    @MrAndreaCaso 8 років тому +3

    Your videos are amazing. It's a shame you haven't been publishing for more than a year. Come back, please!

    • @jbstatistics
      @jbstatistics  8 років тому +4

      Thanks for the compliment Andrea! I haven't gone anywhere -- just a little pressed for time over the past year or two. I fully intend to get back to video production in the not-too-distant future. All the best.

    • @OscarSouth
      @OscarSouth 6 років тому

      Could I make a request/suggestsion for content exploring Logistic/Poisson Regression & GLM in general & more far reachingly Time Series Analysis? (no pun indended).
      Fantastic videos by the way. I enjoyed and learned a lot from them!

    • @chojojo5323
      @chojojo5323 6 років тому

      u r awesome!

  • @Ahmed-jl7uh
    @Ahmed-jl7uh 3 роки тому

    Can you reformulate this question '' Is it unlikely to see a difference of this size due to chance alone?" ? I find negation harder to grasp than sentences said in the affirmative.

  • @oscarwong5226
    @oscarwong5226 2 роки тому

    Learn more from you then my professor

  • @Ashley-pf7hz
    @Ashley-pf7hz 4 роки тому

    best prof! sad that covid-19 made our lectures get cut short but these videos are great

  • @nourelislam8565
    @nourelislam8565 5 років тому

    This is really fascinating videos for stats.......
    I am wondering >> When we would use inference for variance or inference for mean? .... in other words, when we need to detect the statistical significance between two samples, which method we have to use?
    Thanks,

  • @ruis2345
    @ruis2345 5 років тому

    Hey, I'd like to ask a question here, does the population has to follow normal distribution? cause based on central limit theorem, the sampling distribution of population mean will follow normal distribution regardless of population distribution.

  • @scuti7073
    @scuti7073 2 роки тому

    Sample standard deviation does not result in an exact t distribution.
    Pooled variation results in exact t distribution
    Mann-whitney U (wilcoxon rank-sum test) and bootstrap do not assume normality and permutation test

    • @jbstatistics
      @jbstatistics  2 роки тому

      I'm not sure whether you're simply stating things, or implying I said something different and wrong. If it's the latter, I didn't.

    • @scuti7073
      @scuti7073 2 роки тому

      @@jbstatistics no, your video is excellent. I'm just taking notes

  • @42boombox
    @42boombox 4 роки тому

    Random question, what font is used for the video?

  • @randallkim5903
    @randallkim5903 7 років тому +1

    What's the difference between a two sample test and a one sample test with matched pairs?

    • @jbstatistics
      @jbstatistics  7 років тому +2

      The procedures in this video assume we have two independent samples (e.g. a random sample of 50 students from Stanford, and a separate random sample of 50 students from Harvard). A paired-difference t procedure is appropriate when there are pairs of measurements that are dependent (e.g. before and after measurements on the same individual, tire wear on the left and right wheels of a car, IQ of identical twins, health status on individuals matched according to age, sex, smoking status, etc.)

  • @naadde
    @naadde 10 років тому +1

    Is the inference for two means calculated the same way as inference for two variances?

    • @jbstatistics
      @jbstatistics  10 років тому

      No, the methods are quite different.

    • @naadde
      @naadde 10 років тому

      jbstatistics Dear jbstatistics
      thank you for your reply.
      Are you interested in making a video of interferences for two variances?

  • @rabahferjani6663
    @rabahferjani6663 9 років тому +2

    thank you, very very helpful.

    • @jbstatistics
      @jbstatistics  9 років тому

      +Rabah Ferjani You are very welcome!

  • @hatelovebowel4571
    @hatelovebowel4571 2 роки тому

    why are all teaching videos having such small volumns, you guys need to speak it up.

  • @edwardraywer4198
    @edwardraywer4198 6 років тому +1

    Thanks...

  • @statisticsonsteroids
    @statisticsonsteroids Рік тому

    Hi. I need to ask suppose I have 2 independent samples from independent populations. Their variances are known but unequal. Which test should I use to test difference in their means?
    Welch t test is when variance is unknown right?

    • @jbstatistics
      @jbstatistics  Рік тому +1

      If you (somehow) happen to know the population variances, then it would be a z test. (Same idea as for one-sample problems.) The formula would be the same as the Welch t formula, but with the population variances replacing the sample variances. You'd get a z test statistic, and use the standard normal distribution to find the p-value.

    • @statisticsonsteroids
      @statisticsonsteroids Рік тому

      @@jbstatistics Thank you. So I can safely use z test anytime I know the variances regardless of whether they are equal or not, right?

    • @jbstatistics
      @jbstatistics  Рік тому

      @@statisticsonsteroids How do you know the population variances? Yes, if you happen to know the population variances, and are sampling independently from two normally distributed populations, then a z test is appropriate. That comes up in practice pretty much 0% of the time, but you might see it in some text or assignment questions.

    • @statisticsonsteroids
      @statisticsonsteroids Рік тому +1

      @@jbstatistics yes yes ofc, I'm just speaking theoretically. Thank you for answering on a 9 year old video to help me out🙏