I am grateful for finding your videos. I have been having great difficulties in finding the difference between Parametric and Non-Parametric Tests. I am a Statistics major student at the University of Sierra Leone (USL) and I believe I am in the right place for learning more about my course. Thank you @DATAtab
Remember that these are only assumptions, It is not required normally distributions in order to use parametric tests. Usually these tests are very robust (e.g., ANOVA), and will probably have the same results.
This is very critical subject in statistical techniques applicable in research. Mostly I was confused in such a way that they are talking about qualitative and quantitative research.
Hey @DATAtab, I think parametric test like the t-test/Anova do not care about the distribution of the population data or the sample data -> it is important that the sample means should follow a normal distribution. Also parametric tests like simple regression /multiple regression only care about the distribution of the error-term following a normal distribution, so again not that the population or samples follow a normal distribution
Hi, many thanks for your feedback! hmmm, if the sample data is normally distributed, isn't the mean also normally distributed? Since one draws usually only one sample, and thus there is only one mean value, one cannot examine in practice in any case whether the mean values are normally distributed, but only whether the data are normally distributed! Please correct me if I am wrong! Regards, mathias
@@datatab Hi Mathias, it is not my place to correct you or your team, but many thanks for reacting to my concerns!! I know that you often find the assumption in statistic-books that normality (or near normality) of the population data is a precondition for t-test´s and ANOVA´s. But when I was younger I learned that with a big enough sample size the CLT gives us a normal distribution of the sample means regardless of the sample data distribution or of the population data distribution. And that seemed to be the fantastic magic of the CLT: since we often do not know the complete population (and its distribution) we can rely on the CLT to give us a normal distribution of the sample means to perform test´s and give Confidence Intervals for the means etc. Also there exist t-test´s for binomial data and the distribution shape of binomial data is certainly not normally distributed. This is a rather naive question from my side but because of what I wrote above: why than bother with non-parametric tests at all? The only reason that comes to my mind is "whenever I have veeery small samples and no knowledge of the population distribution"
@@datatab did you mayhap have had time to talk to collegues about question (why to bother with nonparametric tests at all when the CLT (+ robust sample size) seems work "magic" regardless of the underlying population or sample distribution). Also, If I have been unclear in what I meant, I found a video with Dr. Todd Daniel that more visually describes what I meant about the CLT and sample mean distribution (or any other sample parameter-of-interest distribution). th-cam.com/video/GcfYy27nBgY/w-d-xo.html
Hello! How to apply the non-parametric u-mann-whitney test for two samples EE:5,7,6,9,8,7,10,4,3,6,7,8 and EC:4,8,5,7,10 ,10,3,4,7,9 ? (values represent marks obtained) Thank you!
Hi, the Central Limit Theorem suggests that with a sufficiently large sample size, the sampling distribution of the mean approaches a normal distribution, regardless of the distribution of the population. Typically, sample sizes of 30 or more are often considered adequate for the central limit theorem to hold, which can justify the use of parametric tests. However, this is a general guideline and might not apply in all situations, especially if the data are heavily skewed or have outliers. Regards Hannah
This was a really nice and informative video. But I do have a question. Under what category (Parametric or Non Parametric) will tests like Wald test and Likelihood Test come??
Hello, thank you for your feedback! I'm not 100% sure and would have to research it myself first, but I think both are not parametric! But as I said, better check again yourself!
the data set was non parametric and I did spearman correlation for that...and I want to know what test should I do for regression analysis...is there separate rate test for non parametric data
I think that's because EVEN though it's only a small difference and a small sample size with doing a parametric test it'd enough to reject the null hypothesis???
![[LIVE] : ONE ลุมพินี 86 | คู่เอก "คมเพชร vs ชาติพยัคฆ์"](http://i.ytimg.com/vi/m3q_dRYDuU8/mqdefault.jpg)
If you like, please find our e-Book here: datatab.net/statistics-book 😎
I am grateful for finding your videos.
I have been having great difficulties in finding the difference between Parametric and Non-Parametric Tests. I am a Statistics major student at the University of Sierra Leone (USL) and I believe I am in the right place for learning more about my course. Thank you @DATAtab
Great to hear! Many thanks for your feedback!!! Regards Hannah
@@datatabI have test at wednesday, good luck for me.😅😅😅😅
@@Chillersnowman You got this!!
Thank you. The table showing the parametric tests and their non-parametric counterparts is very useful.
Glad it was helpful!
Thank you SOOOOO much for this fab video. Beautifully presented and clear diction throughout... Perfect for my revision.
Super easy to follow! Thank you for helping me to find my way in the jungle of statistics :)
Best thing I came across today .... Awesome explanation
Why am I interested in this topic? Because I have an exam I don't want to fail
Then all the best : )
That's me right now
Same here
Funny
😢 same we are
wow, what kind of presentation this is, I'm satisfied
Nicely explained!!
Lots of love from India
Glad you liked it!
Such a useful dear ma'am.. May God Bless You!
Thanks for these videos.
I am glad to found your videos'...Thank you some much
So helpful videos!
Glad you like them! Regards Hannah
Remember that these are only assumptions, It is not required normally distributions in order to use parametric tests. Usually these tests are very robust (e.g., ANOVA), and will probably have the same results.
Yes that's right of course! Many thanks for the feddback! Regads Hannah
This is very critical subject in statistical techniques applicable in research. Mostly I was confused in such a way that they are talking about qualitative and quantitative research.
Qualitative data is in quantitative research. Qualitative research is totally different where coding is done that does not deal with mean medians
Thank you for the video........... But you did not talk about the chi-square test and fisher's exact test. Where do they fall in the chart?
Hey @DATAtab,
I think parametric test like the t-test/Anova do not care about the distribution of the population data or the sample data -> it is important that the sample means should follow a normal distribution.
Also parametric tests like simple regression /multiple regression only care about the distribution of the error-term following a normal distribution, so again not that the population or samples follow a normal distribution
Hi, many thanks for your feedback! hmmm, if the sample data is normally distributed, isn't the mean also normally distributed? Since one draws usually only one sample, and thus there is only one mean value, one cannot examine in practice in any case whether the mean values are normally distributed, but only whether the data are normally distributed! Please correct me if I am wrong! Regards, mathias
@@datatab Hi Mathias, it is not my place to correct you or your team, but many thanks for reacting to my concerns!!
I know that you often find the assumption in statistic-books that normality (or near normality) of the population data is a precondition for t-test´s and ANOVA´s.
But when I was younger I learned that with a big enough sample size the CLT gives us a normal distribution of the sample means regardless of the sample data distribution or of the population data distribution. And that seemed to be the fantastic magic of the CLT: since we often do not know the complete population (and its distribution) we can rely on the CLT to give us a normal distribution of the sample means to perform test´s and give Confidence Intervals for the means etc.
Also there exist t-test´s for binomial data and the distribution shape of binomial data is certainly not normally distributed.
This is a rather naive question from my side but because of what I wrote above: why than bother with non-parametric tests at all? The only reason that comes to my mind is "whenever I have veeery small samples and no knowledge of the population distribution"
@@datatab did you mayhap have had time to talk to collegues about question (why to bother with nonparametric tests at all when the CLT (+ robust sample size) seems work "magic" regardless of the underlying population or sample distribution).
Also, If I have been unclear in what I meant, I found a video with Dr. Todd Daniel that more visually describes what I meant about the CLT and sample mean distribution (or any other sample parameter-of-interest distribution).
th-cam.com/video/GcfYy27nBgY/w-d-xo.html
Simply explained 👍
Thanks : ) Hannah
@@datatab My name's Namrata not Hannah. It's okay though
@@namratajain921 : ) But my name is Hannah : ) it was like "Thanks, Regards Hannah"
@@datatab 😂hahahaha, I'm sorry bro ❤️
Thank you 🎉
You’re welcome 😊
Very nice explanation
How do we come to know that data is normally distributed
Please have a look here: datatab.net/tutorial/test-of-normality
Normality test need to be done
Link for your video on dependent and independent variables???
Why is the t-test, chi2-test, anova function no longer available? Just looking today, and it's no longer a test that you can use.
Thank you very much
Many thanks!!! 😊 Regards, Hannah
MANY THANKS IT WAS SO HELPFUL
Thank you
Thanks...Very helpful.
Thanks for your feedback! Regards Hannah
If one sample is normally distributed and another not normally distributed which correlation i have to use?
Spearman, because one sample is not normal and thus are the assumptions for pearson not met.
Chi square can be used for non parametric tests right?
Yes : ) Regards Hannah
Hello! How to apply the non-parametric u-mann-whitney test for two samples EE:5,7,6,9,8,7,10,4,3,6,7,8 and EC:4,8,5,7,10 ,10,3,4,7,9 ? (values represent marks obtained) Thank you!
so helpful
Glad it was helpful! Regards Hannah
Are 30 respondents enough for parametric tests?
Hi, the Central Limit Theorem suggests that with a sufficiently large sample size, the sampling distribution of the mean approaches a normal distribution, regardless of the distribution of the population. Typically, sample sizes of 30 or more are often considered adequate for the central limit theorem to hold, which can justify the use of parametric tests. However, this is a general guideline and might not apply in all situations, especially if the data are heavily skewed or have outliers. Regards Hannah
This was a really nice and informative video. But I do have a question. Under what category (Parametric or Non Parametric) will tests like Wald test and Likelihood Test come??
Hello, thank you for your feedback! I'm not 100% sure and would have to research it myself first, but I think both are not parametric! But as I said, better check again yourself!
Thank you 👍
Thanks for your Feedback!!! Regards Hannah
Can you explain why nominal measurements use nonparametric?
Are non parametric data guantitative?
Yes!!! The distribution of the data is just not normal!
the data set was non parametric and I did spearman correlation for that...and I want to know what test should I do for regression analysis...is there separate rate test for non parametric data
What did you end up doing?
Could you please clarify 1:55 through 2:17? Why would a small sample size and/or a small salary difference lead one to reject the null hypothesis?
I'm also struggling to understand this too. It sounds contradicting, so maybe I'm thinking about it wrong 🤔 😕
I think that's because EVEN though it's only a small difference and a small sample size with doing a parametric test it'd enough to reject the null hypothesis???
3:00 table of tests
Watch your wideo about it?
1:30 parametric tests are more powerful
For my todays exam
How can I contact to you
How can I help you? Regarding advice I have currently unfortunately no time!
She really sounds like killjoy from valorant, I think I'm in love.
😃
😂
udemy itseems