Introduction to Type I and Type II errors | AP Statistics | Khan Academy

Поделиться
HTML-код
  • Опубликовано: 27 дек 2024

Комментарии • 34

  • @greensky01
    @greensky01 4 года назад +154

    Something to help remember this:
    Benjamin Franklin stated it as: "it is better 100 guilty Persons should escape than that one innocent Person should suffer".
    Where:
    1) One innocent person suffer = type 1 error (alpha error) = more severe [false positive]
    2) 100 guilty persons escape = type 2 error (beta error) [false negative]
    3) You have more authority (increase Power) in a prison when you reduce type 2 error
    a) by having a bigger prison (larger sample size)
    and
    b)counting prisoners (increase precision of measurement)

    • @prinkasays
      @prinkasays 5 месяцев назад +2

      Best way to remember this

  • @sebster95
    @sebster95 2 года назад +92

    Type I error = Illusion (you are seeing an effect when there is not one)
    Type 2 error = 2 blind 2 see (you are failing to see an effect when there is one)

    • @Mkhl4Sure
      @Mkhl4Sure Год назад +3

      I love this explanation

  • @AlphaFoxDelta
    @AlphaFoxDelta 3 года назад +23

    We never try to prove that something is false, that isn't and shouldn't be your goal. If this weren't the case, you would spend eternity proving every possible case wrong. We simply collect evidence and use reason to prove that something may be true with a certain level of confidence, a level which is rarely 100%.

  • @ChristianneAngelica
    @ChristianneAngelica 3 года назад +9

    I just watched the entire stats series... thank you sir for making me pass this class :')

  • @rajvaswani7878
    @rajvaswani7878 4 года назад +17

    Easy way to remember type 1 and 2 errors:
    Type 1 includes explicit costs and is an error of commision (committing the wrongful)
    Type 2 includes implicit (opportunity) costs and is an error of ommision (not doing what was right)
    Type 1 errors are often more serious as they're explicit and thus more transparent. Often avoiding one error comes at the cost of committing another, weigh in which error matter the most in your scenario.

    • @muyin
      @muyin 4 года назад +24

      you call that easy??

  • @davidsweeney111
    @davidsweeney111 7 лет назад +11

    superb informative, exception clear presentation of fascinating concepts, appreciate it buddy!

  • @kidhuman1
    @kidhuman1 2 года назад +1

    2022 using this for poli sci stats thank you my man

  • @teacher.prashantdubeyforma5142
    @teacher.prashantdubeyforma5142 7 лет назад +6

    clear presentation

  • @sapnaratheesh2813
    @sapnaratheesh2813 Год назад

    Pl keep it precise and simple

  • @Carmadon100
    @Carmadon100 3 года назад +1

    better than my professor

  • @goryluv
    @goryluv 7 лет назад +11

    Looks like a confusion matrix

  • @Posby95
    @Posby95 2 года назад +4

    Why not just say «accept Ho» instead of «fail to reject Ho»? Seems like a confusing double negative for no reason.

    • @jonathansiegel2818
      @jonathansiegel2818 2 года назад

      The reason for this is that we NEVER know whether the null hypothesis is or is not true. This is because the null hypothesis is stating what the population parameter is (for example, the average score for all individuals between age 18 and 24 in the entire world. We don't know the answer to that. We assert a population value.. and then take a sample and compare the scores of that sample to our estimated population mean). If the sample mean is far off into the tails of our standardized normal distribution, we reject the null hypothesis (the statement about our population parameter).

  • @kwicke5636
    @kwicke5636 2 года назад

    superb explanation and visualisation! Thank you.

  • @aayangoverski3437
    @aayangoverski3437 10 месяцев назад +1

    1:44

  • @angelayang5183
    @angelayang5183 4 года назад +1

    thank you!

  • @nirmanonline
    @nirmanonline 6 лет назад

    Superb.. Easy to understand. Thanks

  • @simonreichmd
    @simonreichmd Год назад

    I believe p value is not equal to the probability that null hypothesis is true. To my knowledge it is a probability that we would observe to the same or more extreme results despite null hypothesis being true.

  • @MrWinter2
    @MrWinter2 3 года назад

    Thank you for posting! :)

  • @jagadeeshmandala4097
    @jagadeeshmandala4097 3 года назад

    Playlist link

  • @medicalexercisephysiologys2818

    Good

  • @rileyregan2900
    @rileyregan2900 3 года назад +2

    This video could have just started at 3:00 lol

  • @goddaniel9478
    @goddaniel9478 8 месяцев назад

    I didn't know VLAD TV made Khan Academy videos....

  • @dankormick7476
    @dankormick7476 8 месяцев назад

    If Ho is low NH must go, If Ho is high NH is your guy

  • @ishikahalder799
    @ishikahalder799 2 месяца назад

    🔥🔥🔥🔥🔥

  • @lucyhudson3921
    @lucyhudson3921 6 лет назад +2

    why is he talking about P values if there is an alternative hypothesis? This is a mix up of Fishers Significance testing and Neyman Pearsons Hypothesis testing which aren't compatible?

  • @amineamine3374
    @amineamine3374 7 лет назад

    C,est quoi ça

    • @siegfriedcrypto
      @siegfriedcrypto 7 лет назад +1

      Amine Amine Les types d'erreur des probabilités..

  • @zachaiedmonds6773
    @zachaiedmonds6773 7 лет назад

    first

  • @SuperYtc1
    @SuperYtc1 7 лет назад

    First