How to choose an appropriate statistical test

Поделиться
HTML-код
  • Опубликовано: 26 янв 2025

Комментарии • 84

  • @tilestats
    @tilestats  Год назад +3

    If you like this video, you might also like my new video about how to select a multivariate analysis or ML method
    studio.ruclips.net/user/videoXMZ_0aloVAc/edit

  • @we-are-electric1445
    @we-are-electric1445 2 года назад +76

    A really good summary. There are not enough videos like this. There are too many videos out there telling you about statistics but not how to use them and which test should be used where.

  • @Travelll_World
    @Travelll_World 2 месяца назад +3

    Can't find - "simple, soft spoken, very easy to understand, proper pause in speech, perfect diagrammatic with examples explanation" on internet. Wonderful teacher 🙏🙏🙏

  • @clma75
    @clma75 Год назад +9

    I found this very helpful for choosing how to analyze my data for my project. This is the first video out of a gazillion watched that effectively summarized this information in one video,.

  • @sebinpark1638
    @sebinpark1638 2 года назад +17

    I learned 1 year of statistics in 18 minutes Thank you

    • @wenmockan
      @wenmockan Год назад

      The sign of a good teacher

  • @riyaaa.sharmaa
    @riyaaa.sharmaa Год назад +26

    This is literally the first time I’ve understood the tests. Very crisp and a very very good video explaining all concepts. Thank you sooo so sooo much!!!!!!!

  • @alirasteqar243
    @alirasteqar243 2 года назад +6

    More Than 10 years I have tried to find something like this. Great, Superrrrrrrrrrrrrrrr. All other except this video are not complete or totally garbage. Thankssssssssssssssssssss.

  • @calyinedu
    @calyinedu 6 месяцев назад +3

    This has to be the best video I have seen in my years of trying to understand this.

  • @AMADIGIFT-ui8ud
    @AMADIGIFT-ui8ud 7 месяцев назад +2

    This is so great. You're a good teacher, you explained what I have been trying to understand for more than a year in less than 30 minutes.
    You just got a new subscriber.

  • @user-jq6lu
    @user-jq6lu Год назад +3

    Thank you so much... was at my wit's end as I flunked stats last semester and had to resit the module. Chanced upon your video playlists which helped greatly! Am never good at maths. Viewed almost all 47 videos where you explained in such good layman terms until even I can understand. Today's exam results day and I finally cleared stats!!! Have subscribed...

  • @SUMITKUMAR-dy8ic
    @SUMITKUMAR-dy8ic 6 месяцев назад +1

    This video explains in detail about the concepts and the same time takes away apprehension ,skepticism whether we will ever be able to understand the basics of it..
    Hats off to the content creator.. i have been trying for months to recollect regarding statistical test but could not find means to do it.
    Thanks a lot

  • @BJ_design
    @BJ_design Год назад +1

    Great video, simple to the point and kept a consistent graphical layout which helped my understanding even more :)

  • @aganmariadonnafem.7915
    @aganmariadonnafem.7915 2 года назад +7

    Very well-explained! I love your approach in teaching using the algorithm and examples. Thank you so much! 🙏

  • @anubhutisingh8863
    @anubhutisingh8863 Год назад +2

    Very well presented with easy to understand examples. Thanks a Lot.

  • @dirkaloha
    @dirkaloha Год назад +2

    I’m so grateful for this explanation. It will help me on my DNP research paper. I saved it for future reference.

  • @fionaxacur1283
    @fionaxacur1283 Год назад

    Thanks!

  • @radhabane7941
    @radhabane7941 Год назад +3

    Thank you so much...
    Very grateful for this informative presentation with video.. The examples covered are simple and easy to understand.. I m pursuing my phd and was really stuck with data analysis chapter

  • @aubreyhoward2673
    @aubreyhoward2673 2 года назад +4

    SUPER helpful summary video. Thanks!

  • @amanytaahmed3583
    @amanytaahmed3583 Месяц назад +1

    Amazingly illustrated! Thank you so much! ❤️

  • @Alka_Kavyanshi
    @Alka_Kavyanshi 2 года назад +3

    Thankyou so much for your help ❤️❤️❤️🙏🙏🙏🙏 very informative yet brief description 👍

  • @sarambe
    @sarambe 2 года назад +1

    a beautiful video it is saving my life and. I will share this with my lecturer to use as an easy example for students

  • @tilestats
    @tilestats  3 года назад +5

    I got this comment "Great summary video! Do parametric assumptions for the Pearson test mean that all the independent and dependent variables have a normal distribution? So if just one of the independent or dependent does not fit normal distribution in our dataset, Pearson should not be used?"
    Answer: Note that the variables are not defined as independent or dependent in correlation. It does not matter, which variable you put on the x- or y-axis (in contrast to linear regression). Each pair of variables should show a bivariate normal distribution but more importantly, the data should not contain outliers as I describe in the video about the Spearman correlation.

  • @abhijitmondal4850
    @abhijitmondal4850 9 месяцев назад +4

    Each and every ward and sentences are wisely chosen by the author.
    ONE YEAR OF BIOSTST IN A NUTSHELL. ❤❤

  • @mdmahmudulhasanmiddya9632
    @mdmahmudulhasanmiddya9632 2 года назад +2

    U r very knowledgeable person sir.

  • @danorts9577
    @danorts9577 Год назад +2

    Awesome summary of the topic!
    Thank you, it helped me a lot!

  • @smr0018
    @smr0018 Год назад +1

    Awesome
    Thank you so much for the comprehensive explanation.

  • @elenapopova4569
    @elenapopova4569 2 года назад +3

    Awesome video! I am completely new to stats and this video obviously will safe my live! Thank you very much! I wish you all the best!

  • @inna_b_26
    @inna_b_26 Год назад

    the best explanation I've heard

  • @ragulajay7197
    @ragulajay7197 2 месяца назад +1

    Amazing explanation , “The best”

  • @tedransom8087
    @tedransom8087 3 года назад +1

    Thank you! I wish I had this guide 10 years ago.

    • @tilestats
      @tilestats  3 года назад

      Thank you!

    • @tilestats
      @tilestats  2 года назад +2

      It is now possible to buy this video as a pdf.
      payhip.com/TileStats

  • @JulieZhou-1017
    @JulieZhou-1017 5 месяцев назад +1

    So good, I will spend more time watching all your videos!!!

  • @anubratadas1
    @anubratadas1 Год назад +1

    Thank you so much for this fantastic summary!

  • @GauravK-y2d
    @GauravK-y2d Год назад +1

    Hey, thanks a lot for making this video. It is really helpful.

  • @digitalsustainability
    @digitalsustainability 3 года назад +3

    This is very succinct and useful! Thank you!

  • @sananalungu6425
    @sananalungu6425 Год назад +1

    Really good explanation ❤

  • @chensoki
    @chensoki 3 дня назад

    when continuous data did not satisfy parametric test requirement, what do you think about transforming the data (exceptions - really out outlier) and still proceeding with parametric test?

  • @felekeassefa1320
    @felekeassefa1320 2 года назад +2

    Great video !

  • @sridharr2025
    @sridharr2025 7 месяцев назад +1

    Very useful video, thank you

  • @sahilkathuria9850
    @sahilkathuria9850 Год назад +1

    Best summarisation video I have come across. Just one doubt, in case of Paired study for a continuous variable , and if data distribution is not normal, say higly skewed cant we use Wilkcoxon signed rank test as you mentioned in the video that we need to have symmetric distribution in order to apply Wilcoxon test

    • @tilestats
      @tilestats  Год назад +1

      If the data (differences between the pairs) is skewed, you can use the Sign test.

    • @sumanamaggy1093
      @sumanamaggy1093 9 месяцев назад

      Data between the pairs is skewed means, their means are so different? How to find it without doing t test or should we simply do histogram and find wthere they are skewed or not?

  • @SzehangChoi
    @SzehangChoi 6 месяцев назад +1

    And you my friend, you are the real hero

  • @RevisionNotes_hcleelars
    @RevisionNotes_hcleelars 7 месяцев назад +1

    Very good summary

  • @rafaurendez3683
    @rafaurendez3683 2 года назад +2

    Great video! Already subscribed!

  • @dr.anasababneh284
    @dr.anasababneh284 3 года назад +1

    wow, this is amazing. Keep the great work. I've subscribed to support you

  • @pedrohenriquecarneiro904
    @pedrohenriquecarneiro904 Год назад +1

    Nice video!

  • @carlospazymino7405
    @carlospazymino7405 Год назад +1

    Amazing thank you so much.

  • @abm_kamrul_hasan
    @abm_kamrul_hasan 2 года назад +1

    Well explained. Thank you.

  • @fiddlesticks6146
    @fiddlesticks6146 22 дня назад

    i'm seeing quantitative continuous data and qualitative ordinal vs nominal data on the flow chart, but what about quantitative discrete??

    • @tilestats
      @tilestats  21 день назад +1

      Those kinds of data are usually count data and I would suggest you have a look at Poisson regression in that case.

  • @abdul2651
    @abdul2651 4 месяца назад +1

    Amazing vid. Got sub!!

  • @ema60024
    @ema60024 5 дней назад +1

    Thank you so so much

  • @supa.scoopa
    @supa.scoopa 7 месяцев назад +1

    Wow. Thank you so much

  • @애라-y6e
    @애라-y6e 29 дней назад

    good video but what if I have to compare two group of data, and one of it appeared to be normally distributed and other is skewed, so should I proceed with parametric or non parametric test

    • @tilestats
      @tilestats  29 дней назад

      Both groups should be normally distributed to use, for exemple, a t-test. So use the non parametric test in such a case.

    • @애라-y6e
      @애라-y6e 29 дней назад

      @@tilestats even though the p-value for the skewed data is normally distributed?

    • @tilestats
      @tilestats  29 дней назад

      p-value from what test? Shapiro-Wilk?

    • @애라-y6e
      @애라-y6e 29 дней назад

      @@tilestats yupp

    • @tilestats
      @tilestats  29 дней назад +1

      If the p-value is larger than 0.05 from Shapiro-Wilk in both groups, you can use a t-test. Watch this video to understand how to check normality
      ruclips.net/video/mCo1wSLhIvI/видео.htmlsi=xsW_J-x9fuunXRmP

  • @nabeelakhtar7528
    @nabeelakhtar7528 Год назад

    The video assumes that we have just one variable. In that case we can easily follow the flow chart and get the test. But, what if there are different types of variables? like two continuous variables? or one categorial variable and one numerical variable? how do we select a test?

    • @tilestats
      @tilestats  Год назад +1

      For two continuous variables you can use, for example, correlation or linear regression. If you have one categorical variable and one numerical variable and like to see if there is a difference between the categories, you should follow the tree from "Difference". For multivariate methods, you can have a look at this video:
      ruclips.net/video/XMZ_0aloVAc/видео.html

    • @nabeelakhtar7528
      @nabeelakhtar7528 Год назад +1

      @@tilestatsi have seen the video you recommended, however, it is quite complicated. Can you make one similar to this one, but which takes two variables into account? It would be good if you make a chart so we can follow easily. Thanks!

  • @sumanamaggy1093
    @sumanamaggy1093 9 месяцев назад

    I have done case control matching so i have 110 data points out of which 55 are matched with other 55, now i have continuous variable, which test to use

    • @tilestats
      @tilestats  9 месяцев назад

      Sounds like a paired t-test. Check this video and see if it fits your design.
      ruclips.net/video/dqK0wlWMrD8/видео.html

  • @malishavanderberg4228
    @malishavanderberg4228 6 месяцев назад +1

    Thank you for this

  • @Mayuri-f7i
    @Mayuri-f7i 9 месяцев назад +1

    Thank you

  • @NEPALAYA185
    @NEPALAYA185 Год назад

    Hy, which analysis is usually followed to compare the trends of census data? For example migration. Thanks in advance.

    • @tilestats
      @tilestats  Год назад +1

      trend analysis is usually performed by some regression method. If the data seems to be linear, try linear regression.

    • @NEPALAYA185
      @NEPALAYA185 Год назад

      @@tilestats Thank you for your response. :)

  • @rachelmarieaclao4032
    @rachelmarieaclao4032 Месяц назад +1

    thank you sir

  • @niloofarkh4779
    @niloofarkh4779 11 месяцев назад +1

    perfect , thank u so much

  • @백경민-j1l
    @백경민-j1l 3 года назад +1

    Very nice, thanks!

  • @imammaobi
    @imammaobi 3 месяца назад

    @Emmanuel Onuoha .Interesting .

  • @ela6865
    @ela6865 Год назад

    thanks

  • @ryeedzaman380
    @ryeedzaman380 Год назад +1

    stat test tmr anyone??