This man deserves the salary of my Professor. This is the best video to understand t-test, degrees of freedom and whether to reject or accept the null hypotheses. This excel method is even much easier than that of STATA which confuses me each time. I should be grateful if you could upload other methods using the excel sheet. Thank you very much Sir!
Yes, you can use the t test. You will notice I have a playlist on t test (see the video description for the link). You would use unrelated means. Also, you can google z scores for two sample proportions.
Just watched the first two minutes of the video and I gotta say that this graph explains the point of the ttest really well! It's become clear in my mind at last, thanks for your help!
The .5 is taking the square root in Microsoft Excel. It is the same thing as raising something to the 1/2 power. Make sure you like MyBookSucks on FaceBook (see link in video description). This will help others find the educational videos.
I think SPSS is one of the easier tools to use because it has a pretty simple interface, but has very limited graphs. SAS (JMP) has decent graphing capabilities too. Check out the the length of the "student" license to because they may expire after sometime. Over the years I have done more and more in Excel too. Excel can do most anything you can do any of these other tools.
Extremely well explained in a very simple way. Thank you very much sir.You just cleared some of my confusions. I have tested with the formula you mentioned here and tested along with Excel calculations and they are perfect match...Thanks a lot again....
Demetris Marshall Wonderful to hear and with sometime you may pick up a lot of stats. The world would be a different place if people understood basic statistics and economics.
Gregory Hadley Appreciate you taking the time to write me. I am always happy to hear my videos are helping. Hopefully you will like, share, subscribe, If you get a chance could you please like our FB page. www.FaceBook.Com/PartyMoreStudyLess It will help others find the videos.
THANK YOU SO MUCH! look. this was so clear. i was able to follow everything without a problem. I AM used to using spss to do these calculations but i havent done research in over 3 years and now i work from home so this was very helpful.
OMG, SO helpful! Thank you so much! I just discovered your channel, and I LOVE it! I always had problems to understand how to calculate things in Excel, but now I start enjoying it! Thanks! :)
Good question about t test and group size. In "theory" the group size should be the same. There could be all sorts of problems if the groups were not the same size and you tried to make them the same size (which observations would you drop?). Keep in mind the pretest and post test t tests need to be the same size. If you have an observation n the pretest and they do not participate in the post test their observation should be dropped from the pretest. Make sense?
+Zoran Lukic Who knows perhaps they just don't like statistics. If you get a chance, like and share some of the videos. Also like my facebook page (www.FaceBook.Com/PartyMoreStudyLess). This will help others find the videos.
I was going to call my business MyProfessorSucks but went with MyBookSucks instead. Good luck in your classes. Make sure you like MyBookSucks on Facebook (see link in video description).
Excellent explanation. Can I know how this video been prepared? any softwares? Presentation is too good.
11 років тому
Thanks for the reply :) I wanted to know how to find the p value. But doesn't matter I found out couple of days ago. Still, GREAT video, I downloaded it and used it for the last few days for my dissertation. I really appreciate your upload. Thanks so much.
Great to hear! Good luck in your classes as well. Make sure you like MyBookSucks on FaceBook (see the link in the video description). This will help me get the word about about the free videos. You can help me ease the pain of other students :)
Hey thanks this was a great video. Just had an off topic questions but what did you use to actually create this youtube video? How did you take the screen shot of what you were doing on your desktop as you made your lesson, and do things like highlight, erase, create a split screen. Is it a particular software? Thanks
Hi, great video thanks. I am still a little unsure specific to my own problem. I have data from 20 trading days on traders p&l. I am comparing to the S&P 500 returns over the same period. I want to know if the traders approach is producing returns due to chance or not. Can you say which t-test is appropriate, and what data is relevant? For example, do I just use the daily average % returns for the traders to see if their average is chance, or do I use both the traders and the S&P? Thanks
Andy Gibson Well thanks for that! Great to hear that you are losing your hated for statistics. I think the world would be a different place if everyone understood statistics and economics. Make sure you like our FB Page www.FaceBook.Com/PartyMoreStudyLess this will help others find the videos.
You are an amazing teacher. but how did you calculate those T values? i.e the table and 95% confidence with degrees of freedom? could you please explain more about them?
Can you help me with finding out how to populate the P-Value or the probability? Finding the significance against the benchmark. This presentation was so helpful.
I have question. If the samples are different from each other can we divede them. For example in contro grou 10 and research group 12. How we can calculate them. Thanks!
Just wanted to point out a couple of things. In this video, he does square root as the cell^0.5, another way you can do this is =SQRT(cell). Also, in the formula, the (n+n-2) for the purposes of this example is also the degrees of freedom, so you can just take your n+n-2 value if you are working with the 2 samples and the value you get is the df you should be looking for on the table of critical values.
And thank you for this video, this has helped me loads in the recent report I've had to write, before watching this video, I was still intimidated by t-tests, this video however broke it down step by step to the point where it's almost as straightforward to me as chi-squared tests, so again, thank you.
Hi, Thank you for the video, this was cristal clear. However, I wonder how we can get the final table of the "t values needed for rejection of null hypothesis". For example, I wish to take a 95% degree of confidence with a degree of confidence equal to 80. Do we have to calculate it or can we find it somewhere else?
Victor Cadel You should have a t test table in the back of your statistics book. It is very similar to z scores. As your sample gets larger and larger the t score approaches the z score. Here is a link to a program that calculates t scores for you stattrek.com/online-calculator/t-distribution.aspx Good luck!
Interesting. As the sample size gets larger the n-1 becomes less and less important. As n approaches the population the n-1 adjustment becomes irrelevant.
if i have 2 group of patient control(n=12) and cancer (n=78) and i want to analyse the data of protein biomarker expression level so in that condition what type of Ttest i should use, please help
I want to do a t-test between a population and sub-group from within that population to see if the sub-group varies significantly from the population. Should I use the test for unrelated or related groups?
Min 12.41, how did you calculate those T values ? You did not say anything about it. You simply put the table and say 95% confidence with degrees of freedom goes like this. Where do i get that info ? I'm using minitab, it does says it anywhere ?
t values are pre-calculated based on sample sizes and the decision of the null hypothesis being one or two questions and the confidence level (0.05 is 95% confidence and is the common standard in scientific research) you would like to impose upon the null. In the example the null stated there would be no difference between groups (higher or lower); that sets this null as a 2 tailed test which is most common. If the null was re-worded to say the control group will score lower, that would be a one tailed test. So all you need is your "n" and then - 1 or 2 for the number of tails in the test, then you can go to a standard chart (online or in your text book) to determine the t values as they are all preset based on the standard bell curve. Hope this helps.
very well explained tutorial, but when can you say that there is a (highly) significance between the means in this example, because the p-Value is 5%? Thanks a lot for an answer
aabb aabbcc Yes that is correct, to elaborate -- I reject the null hypothesis because the calculated t score (3.14) is greater than the t score critical value (2.101). Make sense?
imperialman1988 I am using excel to calculate my t statistics. At about 11:59 in the video I discuss and show which p values to use (or your critical value). My critical values help me set up the rejection region. Make sense?
Jole'ne Gettridge I am going to tell you a statistics secret. The 95% is completely arbitrary. You confidence can be anything you like it to be. It depends on your research and how confident you want to be.
not quite the 'using excel' i was expecting, more accurately 'using excel as a calculator.' good explanation of the t test, but not the explanation of using excel's ttest function i was hoping for.
I'm just wondering why do you need to times the pooled variance denominator with the [(n+n)/(n*n)] (i.e. what is the significance). Other than that, GOOD JOB and half of my lab report owe it to your videos :)
Jing Bian Great to hear that my videos are helping you. I have an entire playlist on t statistics and I would encourage you to watch some of the videos on how to do the same calculations by hand (ua-cam.com/play/PL8B759A5C1C5C12AF.html). The first video in the playlist is probably the most helpful.
Poor explanation on my part, I am doing a correlation of 8 CEO's, half divorced 9 9.5 9.75 10 11 11.5 10 10.25 81 10.125 being the tenure on the job. I decided to use this test and the pared means both by comparing the married and divorced tenures. So, 9 9.5 9.75 10 against these 11 11.5 10 10.25 does this make more sense?
section A of the equation that you write as n-1*s^2+n-1*s^2 my own professor is expressing as s^2=Ex^2/N-Xbar^2. and no subtraction of 1 from N, just N*S^2+N*S^2. your version is working perfectly, but i CANNOT get my own professors to work... what is going on here?
+alvin dulin You are very welcome. If you get a chance, like and share some of the videos. Also like my facebook page (www.FaceBook.Com/PartyMoreStudyLess). This will help others find the videos.
This man deserves the salary of my Professor. This is the best video to understand t-test, degrees of freedom and whether to reject or accept the null hypotheses. This excel method is even much easier than that of STATA which confuses me each time. I should be grateful if you could upload other methods using the excel sheet. Thank you very much Sir!
I am always posting new videos. I am going to add a video on how to do t test using SPSS too.
Yes, you can use the t test. You will notice I have a playlist on t test (see the video description for the link). You would use unrelated means. Also, you can google z scores for two sample proportions.
You saved me then I can submit my research now. Please, try to continue sharing your knowledge. Thank you so much!!!
I like the way you parse the complex formulas into smaller pieces, then solve by “reassembling” the smaller pieces.
Just watched the first two minutes of the video and I gotta say that this graph explains the point of the ttest really well! It's become clear in my mind at last, thanks for your help!
omg .. it is so overwhelming to see a teacher like you !! god bless you !!
Thank you! I'm taking an online course in which my professor is less than willing to help actually teach! Your video made it easy! Thank you again!
The .5 is taking the square root in Microsoft Excel. It is the same thing as raising something to the 1/2 power. Make sure you like MyBookSucks on FaceBook (see link in video description). This will help others find the educational videos.
I think SPSS is one of the easier tools to use because it has a pretty simple interface, but has very limited graphs. SAS (JMP) has decent graphing capabilities too. Check out the the length of the "student" license to because they may expire after sometime.
Over the years I have done more and more in Excel too. Excel can do most anything you can do any of these other tools.
Extremely well explained in a very simple way. Thank you very much sir.You just cleared some of my confusions. I have tested with the formula you mentioned here and tested along with Excel calculations and they are perfect match...Thanks a lot again....
This is fantastic!! I was completely lost before and my book sucks but this explained everything step by step.
You are an amazing teacher. I've never been too good with stats. but I've learned so much just from watching your vids
Demetris Marshall Wonderful to hear and with sometime you may pick up a lot of stats. The world would be a different place if people understood basic statistics and economics.
Thank you so much for this tutorial. I not only learned a lot, I was impressed with your skill as a teacher.
Gregory Hadley Appreciate you taking the time to write me. I am always happy to hear my videos are helping. Hopefully you will like, share, subscribe, If you get a chance could you please like our FB page. www.FaceBook.Com/PartyMoreStudyLess
It will help others find the videos.
THANK YOU SO MUCH! look. this was so clear. i was able to follow everything without a problem.
I AM used to using spss to do these calculations but i havent done research in over 3 years and now i work from home so this was very helpful.
Do you mean to samples or two observations? If you have two samples, then yes a t test will work fine.
OMG, SO helpful! Thank you so much! I just discovered your channel, and I LOVE it! I always had problems to understand how to calculate things in Excel, but now I start enjoying it! Thanks! :)
Good question about t test and group size. In "theory" the group size should be the same. There could be all sorts of problems if the groups were not the same size and you tried to make them the same size (which observations would you drop?).
Keep in mind the pretest and post test t tests need to be the same size. If you have an observation n the pretest and they do not participate in the post test their observation should be dropped from the pretest.
Make sense?
I think so any one can understand easily because your have great skill of explaining
very well explained...now i'm going thru all your videos on statistics..thanks a lot..
why do they hit the unlike? the explanation is so beautiful
+Zoran Lukic Who knows perhaps they just don't like statistics. If you get a chance, like and share some of the videos. Also like my facebook page (www.FaceBook.Com/PartyMoreStudyLess). This will help others find the videos.
I was going to call my business MyProfessorSucks but went with MyBookSucks instead. Good luck in your classes.
Make sure you like MyBookSucks on Facebook (see link in video description).
Excellent explanation. Can I know how this video been prepared? any softwares? Presentation is too good.
Thanks for the reply :)
I wanted to know how to find the p value. But doesn't matter I found out couple of days ago. Still, GREAT video, I downloaded it and used it for the last few days for my dissertation. I really appreciate your upload. Thanks so much.
Great to hear! Good luck in your classes as well. Make sure you like MyBookSucks on FaceBook (see the link in the video description). This will help me get the word about about the free videos.
You can help me ease the pain of other students :)
Hey thanks this was a great video. Just had an off topic questions but what did you use to actually create this youtube video? How did you take the screen shot of what you were doing on your desktop as you made your lesson, and do things like highlight, erase, create a split screen. Is it a particular software? Thanks
Hi, great video thanks. I am still a little unsure specific to my own problem. I have data from 20 trading days on traders p&l. I am comparing to the S&P 500 returns over the same period. I want to know if the traders approach is producing returns due to chance or not. Can you say which t-test is appropriate, and what data is relevant? For example, do I just use the daily average % returns for the traders to see if their average is chance, or do I use both the traders and the S&P? Thanks
Just awesome. Hated stats before seeing your videos. Love your work Sir!
Andy Gibson Well thanks for that! Great to hear that you are losing your hated for statistics. I think the world would be a different place if everyone understood statistics and economics. Make sure you like our FB Page www.FaceBook.Com/PartyMoreStudyLess this will help others find the videos.
You are very clear and easy to understand! I will be looking for more from you :-)
Not sure what you mean? In this example there are two choice (two columns).
You are an amazing teacher. but how did you calculate those T values? i.e the table and 95% confidence with degrees of freedom? could you please explain more about them?
Wonderful to hear! Good luck on your research too.
Hi sir! Can I ask how do you know which confidence level will you use? is it you just randomly choose between 90%, 95% and 99%?
+hany rosal
95% is great to accept the experiment, if you got 99% confidence that would be much better
90% would be good but not great
You are amazing! Great video, literally saved my life! Keep it up!
Can you help me with finding out how to populate the P-Value or the probability? Finding the significance against the benchmark. This presentation was so helpful.
I have question. If the samples are different from each other can we divede them. For example in contro grou 10 and research group 12. How we can calculate them. Thanks!
Just wanted to point out a couple of things. In this video, he does square root as the cell^0.5, another way you can do this is =SQRT(cell). Also, in the formula, the (n+n-2) for the purposes of this example is also the degrees of freedom, so you can just take your n+n-2 value if you are working with the 2 samples and the value you get is the df you should be looking for on the table of critical values.
Gary thanks for chiming in and adding some clarity. There are a lot of ways to to the same thing in Excel.
And thank you for this video, this has helped me loads in the recent report I've had to write, before watching this video, I was still intimidated by t-tests, this video however broke it down step by step to the point where it's almost as straightforward to me as chi-squared tests, so again, thank you.
I just read your comment.About p -value I just need to know how to calculate in excel. Thanks again
The best I've seen so far!
Hi, statisticsfun
So the t-statistic for the test of the difference of means is 3.14
May I ask what the p value is?
Thank you in advance
Hi,
Thank you for the video, this was cristal clear. However, I wonder how we can get the final table of the "t values needed for rejection of null hypothesis". For example, I wish to take a 95% degree of confidence with a degree of confidence equal to 80. Do we have to calculate it or can we find it somewhere else?
Victor Cadel You should have a t test table in the back of your statistics book. It is very similar to z scores. As your sample gets larger and larger the t score approaches the z score. Here is a link to a program that calculates t scores for you stattrek.com/online-calculator/t-distribution.aspx
Good luck!
statisticsfun Perfect, thank you again for you help
Interesting. As the sample size gets larger the n-1 becomes less and less important. As n approaches the population the n-1 adjustment becomes irrelevant.
That is interesting and that makes sense to me. I am curious to see what you come up with.
can you tell me when to use n-1 or n-2 in getting the df? because I'm really confused...
Best video on T Test, thanks a ton
Pankaj Kaushik Well thank you for such a nice complement
You are an amazing teacher
I just checked and there seems to be sound. Perhaps it is something on your end.
if i have 2 group of patient control(n=12) and cancer (n=78) and i want to analyse the data of protein biomarker expression level so in that condition what type of Ttest i should use, please help
Yes you should be able to use this test for each of your experimental groups.
I want to do a t-test between a population and sub-group from within that population to see if the sub-group varies significantly from the population. Should I use the test for unrelated or related groups?
hey sir. we want to know how did you get the degrees of freedom? please sir we are really in need of your immediate answer
Min 12.41, how did you calculate those T values ? You did not say anything about it. You simply put the table and say 95% confidence with degrees of freedom goes like this. Where do i get that info ? I'm using minitab, it does says it anywhere ?
t values are pre-calculated based on sample sizes and the decision of the null hypothesis being one or two questions and the confidence level (0.05 is 95% confidence and is the common standard in scientific research) you would like to impose upon the null. In the example the null stated there would be no difference between groups (higher or lower); that sets this null as a 2 tailed test which is most common. If the null was re-worded to say the control group will score lower, that would be a one tailed test. So all you need is your "n" and then - 1 or 2 for the number of tails in the test, then you can go to a standard chart (online or in your text book) to determine the t values as they are all preset based on the standard bell curve. Hope this helps.
You are very welcome!
very well explained tutorial, but when can you say that there is a (highly) significance between the means in this example, because the p-Value is 5%? Thanks a lot for an answer
aabb aabbcc Yes that is correct, to elaborate -- I reject the null hypothesis because the calculated t score (3.14) is greater than the t score critical value (2.101). Make sense?
Thank you so much. It was really helpful and I learned a lot.
Very nice video... well explained
Could you share some short method using the T Test function in Excel?
The t.test function will do the trick.
The t.test function will do the trick.
I am sorry, can I ask how we can have the P value calculated for this test? Is that the P value for the two tailed calculated by the excel?
imperialman1988 I am using excel to calculate my t statistics. At about 11:59 in the video I discuss and show which p values to use (or your critical value). My critical values help me set up the rejection region.
Make sense?
quick question. How do you determine the confidence? Is it always 95%?
Jole'ne Gettridge I am going to tell you a statistics secret. The 95% is completely arbitrary. You confidence can be anything you like it to be. It depends on your research and how confident you want to be.
I used the same chart for confidence levels; however, my t test value was -0.27912. Does that mean I would reject the null hypothesis?
not quite the 'using excel' i was expecting, more accurately 'using excel as a calculator.' good explanation of the t test, but not the explanation of using excel's ttest function i was hoping for.
I'm just wondering why do you need to times the pooled variance denominator with the [(n+n)/(n*n)] (i.e. what is the significance).
Other than that, GOOD JOB and half of my lab report owe it to your videos :)
Jing Bian Great to hear that my videos are helping you. I have an entire playlist on t statistics and I would encourage you to watch some of the videos on how to do the same calculations by hand (ua-cam.com/play/PL8B759A5C1C5C12AF.html). The first video in the playlist is probably the most helpful.
hello, I have downloaded this video. and many thanks for that!
Excellent teaching I understood how to calculate t test
partha sarathi Great to hear and good luck in your classes.
Can I calculate t-test even if my # of participants for experimental group is different than the # of participants for control group?
Thank you! This video was super helpful!
Awesome explanation.
Thanks for intersting presentation. Thanks1
The p value would be very small and less than .025.
You're the best.
what if i have no control, can i use t test for two experimental group? please anwer :(
How do you calculate the p value in this case?
Thanks!
Poor explanation on my part, I am doing a correlation of 8 CEO's, half divorced 9
9.5
9.75
10
11
11.5
10
10.25
81
10.125
being the tenure on the job. I decided to use this test and the pared means both by comparing the married and divorced tenures. So, 9
9.5
9.75
10
against these 11
11.5
10
10.25
does this make more sense?
section A of the equation that you write as n-1*s^2+n-1*s^2 my own professor is expressing as s^2=Ex^2/N-Xbar^2. and no subtraction of 1 from N, just N*S^2+N*S^2.
your version is working perfectly, but i CANNOT get my own professors to work... what is going on here?
What if one of the groups has only 2 choices, how would I set that up?
thank you for video. very well explained...:D
This is great! Thank you!!!
thanx you really helped
Excellent!
skip to 4:14 for actual useful and relevant information u were looking for
excellent! thank you!!!
Hi I got everything until you got to E25 ^ .5 Why .5? That would be around 10:56 on the video.
I was hoping to see how this is actually done using t.test() function
Great!
why 18?
THERE'S NO SOUND!
This is really helpful! Thanks!
+alvin dulin You are very welcome. If you get a chance, like and share some of the videos. Also like my facebook page (www.FaceBook.Com/PartyMoreStudyLess). This will help others find the videos.