At about 7:17, you calculate the first component of the first factor sum of squares. You do this by repeating nine times, the boys average (7.7) minus the grand mean (9) and squaring the result, (7.7-9)^2, then summing all of this to get 16 as the boys sum of squares. It seems that you could get a more granular and accurate measure of the sum of squares if you subtracted the grand mean (9) from the individual scores. Why is this not done? I realize that it would increase the sum of squares and affect the F tests later on, but it seems more accurate. Why is it that using the boys mean in place of each score is acceptable when it seems less accurate?
I think they are two different things. One is called sum of squares between the group, one is called sum of squares within the group. So basically our goal is to see whether different group are the same. So we compare differences between group mean and ground mean, and differences within the group. If they are very different, we reject hypothesis. Think about this way. You probably have three class , which all have average 81 points, 82 points and 83 points. If you calculate average of (81+81+82) to get your group mean, then the differences between each group mean and your ground mean would be very small. However, it is possible that most of students in group one are around 81, while in group two there are some students get 90, while some get 72, which also give average 81. However, obviously these two groups will be very different even if they have same average.Then you calculate differences between each data point in a group and the mean of the group it is in.(called sum of square within) Then, you compare sum of square between and sum of square within. If the ratio is very large, that means your different group perform very different even if they have similar mean
great and easy to follow, however, there are formulas for doing this quicker which unless you have all the room in the world or you have unlimited time you have to use, I wish your provided those
Thanks! I use a variety of different software including photoshop, illustrator, HTML5, Final Cut, GarageBand and a few other products. Each video takes me a long time to create. My rule of thumb is each minute of video takes about 12 hours to create, so a 7 minute video takes me about 84 hours of development time. Good luck on your presentation.
Amazing Tutorials!! Many thanks I have only one comment (kind of silly), but the square of 1.3 = 1.69 and not 1.8 Thanks again and they are really very helpful tutorials with great explanation... Regards
Thank you for video! When you calculate Sum of squares for Gender you do this separetly for boys and girls. I expected that for second factor age it would be separate calculation for each category 10YO, 11YO, 12YO. Could you please clarify why it is used different approach for caluculation sum of squares for factor 1 and factor 2?
Many thanks for the comment. Make sure you like MyBookSucks on FaceBook (see link in video description). This will help others find the educational videos.
Thank you for you compliments. It is much appreciated. Make sure you like MyBookSucks on FaceBook (see link in video description). This will help others find the educational videos.
Appreciate that! Always good to know my videos are helping. Make sure you like MyBookSucks on FaceBook (see link in video description). This will help others find the educational videos.
Still useful 10 years later. Best use of video to explain a topic in a lecture style way, no useless animation but perfectly understandable. That must take a lot of work. Thank you for making these video's. Best video's on youtube for stats
I liked your video on 2 way ANOVAs, it helped me understand it better, but in no galaxy, universe, world, star system, dimension is statistics fun. Sorry, just not.
Ben Liu I am doing a lot of rounding in the video. If you get a chance like my FB page (www.FaceBook.Com/PartyMoreStudyLess), share videos, like and subscribe. It will help others find the videos.
at 7:08, how did you get 16? I calculated (7.7-9)^2*9=(-1.3)^2*9=15.21, and by my calculation (-1.3)^2=1.69. So, either I'm not able to understand your calculation or you made a mistake. Please, explain it to me! for what did you round a few your calculation?
I watched 2 x 2 hour lectures from my uni about factorial analysis and had no idea wtf he was talking about. I watch this youtube video which is 18 minutes and understand it... you are a life saver! Thank you!
I believe that is a rounding error. Keep in mind that in all of statistics we are estimating, so rounding should not make a lot of difference. It is hard to carry more than one decimal when you are doing calculations by hand.
Good catch. I hope (I don't think) it makes a difference, but I will correct that when I revise the video. Thanks again for the catch, you have an eve for detail.
Good luck on your exam!!!!! Make sure you like MyBookSucks on FaceBook (see link in video description). This will help others find the educational videos.
I do have an example of doing Two Way Analysis by hand. I have an entire playlist on Two ANOVA. You can find the link to the playlist in this videos description. The first video in the playlist shows how to calculate Two ANOVA by hand. Good luck on your thesis as well. Btw, What is your topic of your thesis?
For main effects, it would be clearer if the whole thing were framed in terms of a ratio between between groups/factor variance and within-group variance. But for the age factor, you organize the data by gender. This makes the overall functioning of the f-ratio for the age effect unclear for beginners
I think it may have helped to have mention or emphasise that the within-group variance is calculated for each 'group', 'treatment', however you call it, rather than combining the factors together completely
Excellent! Very nice presentation, and very clear. Thank you for taking the time to put this together. I don't suppose you might do a video on the estimation of parameters in linear regression models?
I was going 2 say you made a mistake on SSe but I see it was just wrong on screen. Thank you for doing it step-by-step bc I've been looking everywhere for someone to break it down 4 me. Math is not my best subject...at all!!
Can someone please help me? I have followed all of the steps however when I tried to calculate the sum of squares both factors I got a negative value. Does that mean I've gone wrong somewhere?
Unfortunately, the link for related videos does not seem to be visible. The graphic appears directing to related links, but the link is not visible. Please advise.
thanks again for clarifying that SSW is "error." Not getting that concept can completely wreck a calculation---writing a good interpretation is impossible without it.
at 10:40... your number rounding is very confusing... you round to the firs decimal place for the first age group.. (-2,5)^2=6,25 ->6,3 but then for the secon you round to the second decimal ->0,25 and then when you try to add them you get a different result to what you show in the video (6,3+0,25+9)=46,65 and NOT 46,5. 46,5 would be the answer if you didn't round the first number and left is 6,25.. also that would be a lot less confusing.
thanks for the video. also, you might not care, but i do, so, for the future, when you're talking about biologically determined characteristics it's "sex," and when you reference identification, it's "gender."
Thanks for your comment. I do like challenging ideas like the one you put forth because labeling does bias us. For example, separating children's educational achievement by race (black v. white) instead of by socioeconomic class (poor v. wealthy). The first does not give us insight into the cause of low educational achievement while the second does give us insight. Even though most government reports use "sex" instead of "gender" it does not necessarily make it correct. A larger question, especially when doing statistics, does using "gender" give us more or less insight into our analysis than "sex." Nonetheless, good luck in your studies.
Hi, I have trouble with the Sum of Squares Total. I used SPSS to calculate, SSFactor1, SSFactor2 and the SSWithin are correct. But the SSTotal is not Its have 2 values: "Total" and "Corrected total". The "Corrected Total" is the same values as you calculated here, but the "Total" one, what is it and how can i calculate it? I used the tool "Univariate..." in Analyze> General Linear Model. Many thanks!
This is an excellent teaching tool. Thank you! BUT! There was a couple times where you structured the data so it was confusing. Specifically, when you were adding the sum of squares for age alone. you still put them in a boys and girls column. It made me think that the gender information was affecting the result for that measure. I had to watch it three times just to figure out that you really were only determining the summer squares for age alone. There wasn’t a reason to split it up by gender in that instance. It made it a lot harder for me to understand what was going on
Hi there, thank you for a very nice layout. I wonder about 11:00 min in the video, where you have the BOYS mean instead of Girls mean (7 green, 10 Brown, 14 Purple) substracted from the grand mean. I thought that the sum of Squares of the Girls should be 90, and the Sum of Squares of the second factor (age) should be 136,5. That leads to a negative Sum of Squares in both factors of -36,5. Am I right or have I misunderstood something?
@statisticsfun maybe a dumb question, but can I calculate the SST as well if I don’t have any individual test scores but only group means? Thanks a lot for the video btw
searched the website for couple days to try to figure out how to solve this by hand and EVERYWHERE i looked... people did them by excel or other programs and just wrote out the answers without showing any work... thank you SO much for this video. helped a lot.
Hilal özkeçeci Glad that you are no longer having statistics nightmare, best to keep your dreams to happy dreams. 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.
Hi many thanks for the video. I would like to ask how could I analyse one independent variable with two different levels’ influence on one dependent variable. For example, how do high level of organisational support and low level of organisational support affect on employee satisfaction? Could you please suggest how to do this? And how to classify the organisational support into high and low level of this collect via seven likert scale. I am confused about using multi group analysis and ANOVA. Many thanks for your help
extremely high quality, very intelligent, and elegant. The speed is great. not slow at all. I can watch it while thinking. I would absolutely ignore any not constructive comments.
fabulous video....do u have something such descriptive about how to do 2 way anova in minitab??? Or do you know if some one else have uploaded such video? can u just give me the like if it is there
7 years later, and your still helping me in 2020. Passing my final exam now.
Felt in 2021
It is my first comment on RUclips. I can not wait to share it with my friends. It is thousands of help. Thanks for the pretty vision and explanation.
Thank you for making statistics so much simpler to understand. Your videos have helped me and thousands more.
At about 7:17, you calculate the first component of the first factor sum of squares. You do this by repeating nine times, the boys average (7.7) minus the grand mean (9) and squaring the result, (7.7-9)^2, then summing all of this to get 16 as the boys sum of squares. It seems that you could get a more granular and accurate measure of the sum of squares if you subtracted the grand mean (9) from the individual scores. Why is this not done? I realize that it would increase the sum of squares and affect the F tests later on, but it seems more accurate. Why is it that using the boys mean in place of each score is acceptable when it seems less accurate?
I think they are two different things. One is called sum of squares between the group, one is called sum of squares within the group. So basically our goal is to see whether different group are the same. So we compare differences between group mean and ground mean, and differences within the group. If they are very different, we reject hypothesis. Think about this way. You probably have three class , which all have average 81 points, 82 points and 83 points. If you calculate average of (81+81+82) to get your group mean, then the differences between each group mean and your ground mean would be very small. However, it is possible that most of students in group one are around 81, while in group two there are some students get 90, while some get 72, which also give average 81. However, obviously these two groups will be very different even if they have same average.Then you calculate differences between each data point in a group and the mean of the group it is in.(called sum of square within) Then, you compare sum of square between and sum of square within. If the ratio is very large, that means your different group perform very different even if they have similar mean
Great video!
Just wanted to say at 11:20, the calculations in blue should be 14- 9 not 12 - 9.
Nevertheless, the process is the same :) thank you
great and easy to follow, however, there are formulas for doing this quicker which unless you have all the room in the world or you have unlimited time you have to use, I wish your provided those
Thanks! I use a variety of different software including photoshop, illustrator, HTML5, Final Cut, GarageBand and a few other products. Each video takes me a long time to create. My rule of thumb is each minute of video takes about 12 hours to create, so a 7 minute video takes me about 84 hours of development time.
Good luck on your presentation.
and yet, priceless when it comes to value!
Can I use two way anova in solving 3 factor experiments?
Thank you so much, man!
Woow!
Thank you again!
Thank you for you very nice job!!!
Amazing Tutorials!! Many thanks
I have only one comment (kind of silly), but the square of 1.3 = 1.69 and not 1.8
Thanks again and they are really very helpful tutorials with great explanation...
Regards
Same question,
You are welcome, welcome, welcome! I really appreciate hearing from students and it does make my day to hear that my videos are helping.
Thank you very much for your compliments. I believe the 1.3 = 1.8 is a rounding error. I don't carry all the digits through the entire problem.
Advice: Play in x1.25. Great vid!
at 12:49 in the video you show that the boys score of 6-7= (-2.0)2 Wouldn't it be (-1)2? Plus isn't the sum of squares equal to 31?
28 is the correct sum of squares if the correct value is considered for (6-7)² = 1 instead of 4.
Thank you for video! When you calculate Sum of squares for Gender you do this separetly for boys and girls. I expected that for second factor age it would be separate calculation for each category 10YO, 11YO, 12YO. Could you please clarify why it is used different approach for caluculation sum of squares for factor 1 and factor 2?
I had the same question. Although the final answer is correct, calculating means based on Age would have been more intuitive.
how did u get the square of -1.3 to be 1.8? even if u round figure it, it must be 1.7
Awesome!!!!! Thank you for spreading the word and congratulations on your A+. You deserve all the credit.
Many thanks! Make sure you like MyBookSucks on FaceBook (see link in video description). This will help others find the educational videos.
Many thanks for the comment. Make sure you like MyBookSucks on FaceBook (see link in video description). This will help others find the educational videos.
Thank you for you compliments. It is much appreciated. Make sure you like MyBookSucks on FaceBook (see link in video description). This will help others find the educational videos.
Appreciate that! Make sure you like MyBookSucks on FaceBook (see link in video description). This will help others find the educational videos.
Appreciate that! Always good to know my videos are helping.
Make sure you like MyBookSucks on FaceBook (see link in video description). This will help others find the educational videos.
Still useful 10 years later. Best use of video to explain a topic in a lecture style way, no useless animation but perfectly understandable. That must take a lot of work. Thank you for making these video's. Best video's on youtube for stats
I liked your video on 2 way ANOVAs, it helped me understand it better, but in no galaxy, universe, world, star system, dimension is statistics fun. Sorry, just not.
Around 7:05, you square -1.3, the answer is 1.69 or about 1.7, not 1.8
FUCK ME IVE BEEN SPENDING HOURS FIGURING OUT WHAT I WAS DOING WRONG
He used 1.3 to write it down, the actual value is 1.3333333 or 4/3. squaring that makes it 1.8
Nathaniel Peter ,Thanks a lot!
how did it became 1.3333333?😭😭😭
Saved Me and my final in 2024, insanely helpful video
Hi this is helpful. but (-1.3)^2=1.8? It should be 1.69.
Ben Liu I am doing a lot of rounding in the video. If you get a chance like my FB page (www.FaceBook.Com/PartyMoreStudyLess), share videos, like and subscribe. It will help others find the videos.
at 7:08, how did you get 16? I calculated (7.7-9)^2*9=(-1.3)^2*9=15.21, and by my calculation (-1.3)^2=1.69. So, either I'm not able to understand your calculation or you made a mistake. Please, explain it to me! for what did you round a few your calculation?
I had the same result :(
I watched 2 x 2 hour lectures from my uni about factorial analysis and had no idea wtf he was talking about. I watch this youtube video which is 18 minutes and understand it... you are a life saver! Thank you!
Hi, can I have more than two independent variables for 2-way Anova :)?
I believe that is a rounding error. Keep in mind that in all of statistics we are estimating, so rounding should not make a lot of difference. It is hard to carry more than one decimal when you are doing calculations by hand.
Thank you so much! I have an exam this week and my teacher goes through things super fast and textbooks just don't do it for me.
1.8 x 9 is not equal to 16 right?!
Thank you! Make sure you like MyBookSucks on FaceBook (see link in video description). This will help others find the educational videos.
Hi Hannah,
Not sure why the link does not work. I did add the links in the video description of this video. Hope that helps.
David
Good catch. I hope (I don't think) it makes a difference, but I will correct that when I revise the video. Thanks again for the catch, you have an eve for detail.
Good luck on your exam!!!!!
Make sure you like MyBookSucks on FaceBook (see link in video description). This will help others find the educational videos.
my prof has us do this a completely different way and now i am so confused :/
Campion. Did he teach you how do calculate the SS for unbalanced data?
You are very welcome!
Make sure you like MyBookSucks on FaceBook (see link in video description). This will help others find the educational videos.
I do have an example of doing Two Way Analysis by hand. I have an entire playlist on Two ANOVA. You can find the link to the playlist in this videos description. The first video in the playlist shows how to calculate Two ANOVA by hand.
Good luck on your thesis as well. Btw, What is your topic of your thesis?
isnt the square of -1.3, 1.69 at 7:22??
He used 1.3 to write it down, the actual value is 1.3333333 or 4/3. squaring that makes it 1.8
I also struggled for the same question that -1.3 is 1.69. Thanks Nathan for explaining it. 1.3333333 is indeed 1.8 rounded up.
@@nathanielpeter5411 but 7.7-9 is -1.3 why did he take 1.3333333?
For main effects, it would be clearer if the whole thing were framed in terms of a ratio between between groups/factor variance and within-group variance. But for the age factor, you organize the data by gender. This makes the overall functioning of the f-ratio for the age effect unclear for beginners
I think it may have helped to have mention or emphasise that the within-group variance is calculated for each 'group', 'treatment', however you call it, rather than combining the factors together completely
Jenn, nice to hear my videos are helping you towards your PhD. Btw, what is the subject are of your PhD?
I am not sure what Bit string means relative statistics.
Thanks for that. I will double check the video. I appreciate the help.
one of the only few genuinely decent and complete explanations of two-factor ANOVA
As always you are very welcome and good luck this week too.
Oh, and thank you very much for the videos, you explain it very clearly!
very slow...I wish there was a short and fast interview of what the whole concept and operation is and then get into details of operation....
u could always fast forward buddy he is slow because the concept is a bit confusing
I fell asleep twice .. third time's a charm
I checked! Holy Cow Batman. You have great ears and eyes and a keen perspective on detail. You are correct.
Sir will u help me out to how to analyse the data for crd design if the data is based on germination test done in lab.
great vid!! very comprehensive and easy to understand!!
Amazing videos! I´m so glad i found these. Thank you for taking so much time in making these :D
Excellent! Very nice presentation, and very clear. Thank you for taking the time to put this together. I don't suppose you might do a video on the estimation of parameters in linear regression models?
I was going 2 say you made a mistake on SSe but I see it was just wrong on screen. Thank you for doing it step-by-step bc I've been looking everywhere for someone to break it down 4 me. Math is not my best subject...at all!!
Very helpful. There are minor mistakes still such a long process anyway gets benefit of doubt. Thanks for the easy explanation
hi great videos , in 7:13 (-1.3)^2 is 1.7 not 1.8, pls note this calculation error everyone who's watching
is there any mistake at 8:43? 1.3^2 is actually equal 1.69, and if the result is 1.8 anyway, the sum of squares should be 16.2
(-1.3)^2= 1.8?
Shouldnt it be 1.7? Since (-1.3)^2= 1.69? Round it off we'll get 1.7
SO helpful..
Linda Adeyemo Great to hear and good luck in your studies too.
Maybe it's me but I think you may need to go back and check your math in a few spots (e.g. 6-7=-2?????
Can someone please help me??? I dont see how 1.3 squared equals 1.8. I keep getting 1.69. someone please help
Please give an example of three-way ANOVA, if you do not mind. Your videos are great!
1.3 squared is 1.69, so a rounded number of 1.8 is really high! So, why 1.8?
Which calculators do cross predictions so we found out that uncertainty with certainty and exclusive principles become nill forever
Can someone please help me? I have followed all of the steps however when I tried to calculate the sum of squares both factors I got a negative value. Does that mean I've gone wrong somewhere?
Basically the second half on MCEN 3047 on this video
Is this a type 1, 2 or 3 sum of squares? I know it doesn't matter because it's balanced but I am curious?
Uhm, do we round off the grand mean or the average mean in the mean table if I got 7.75 to 7.8? Yay or nay?
I cannot express my thanks enough. These videos are great!
What about Sum of squares between?
thanks for the video but 7:09 (-1.3)^2 = 1.69 not 1.8
Unfortunately, the link for related videos does not seem to be visible. The graphic appears directing to related links, but the link is not visible. Please advise.
thanks again for clarifying that SSW is "error." Not getting that concept can completely wreck a calculation---writing a good interpretation is impossible without it.
Can I use two way anova for solving 3 factor experiments?
Thank you so much, I'm finally getting this.
Such an awesome video. Thank you so much for taking the time to show so simply and transparently how this all works.
Thank you very much Sir! I am so grateful!
Thank you very much, this video is really useful !
You made a calculation error. 12:58. you said 6-7 = -2 = 4
Did you just assume children's gender?
so much more complicated than it needs to be...
at 10:40... your number rounding is very confusing... you round to the firs decimal place for the first age group.. (-2,5)^2=6,25 ->6,3 but then for the secon you round to the second decimal ->0,25 and then when you try to add them you get a different result to what you show in the video (6,3+0,25+9)=46,65 and NOT 46,5. 46,5 would be the answer if you didn't round the first number and left is 6,25.. also that would be a lot less confusing.
Why does 1.3^2 is 1.8? Isn't it 1.69?
it is 28 but what is wrong is that in the 4 row 6-7=-1 not -2
thanks for the video. also, you might not care, but i do, so, for the future, when you're talking about biologically determined characteristics it's "sex," and when you reference identification, it's "gender."
Thanks for your comment. I do like challenging ideas like the one you put forth because labeling does bias us. For example, separating children's educational achievement by race (black v. white) instead of by socioeconomic class (poor v. wealthy). The first does not give us insight into the cause of low educational achievement while the second does give us insight.
Even though most government reports use "sex" instead of "gender" it does not necessarily make it correct. A larger question, especially when doing statistics, does using "gender" give us more or less insight into our analysis than "sex."
Nonetheless, good luck in your studies.
Hi, I have trouble with the Sum of Squares Total. I used SPSS to calculate, SSFactor1, SSFactor2 and the SSWithin are correct. But the SSTotal is not
Its have 2 values: "Total" and "Corrected total". The "Corrected Total" is the same values as you calculated here, but the "Total" one, what is it and how can i calculate it?
I used the tool "Univariate..." in Analyze> General Linear Model.
Many thanks!
statisticsfun ama hiç fun gibi değil
This is an excellent teaching tool. Thank you!
BUT!
There was a couple times where you structured the data so it was confusing.
Specifically, when you were adding the sum of squares for age alone. you still put them in a boys and girls column. It made me think that the gender information was affecting the result for that measure. I had to watch it three times just to figure out that you really were only determining the summer squares for age alone. There wasn’t a reason to split it up by gender in that instance. It made it a lot harder for me to understand what was going on
Hi there, thank you for a very nice layout. I wonder about 11:00 min in the video, where you have the BOYS mean instead of Girls mean (7 green, 10 Brown, 14 Purple) substracted from the grand mean. I thought that the sum of Squares of the Girls should be 90, and the Sum of Squares of the second factor (age) should be 136,5. That leads to a negative Sum of Squares in both factors of -36,5. Am I right or have I misunderstood something?
@statisticsfun maybe a dumb question, but can I calculate the SST as well if I don’t have any individual test scores but only group means? Thanks a lot for the video btw
searched the website for couple days to try to figure out how to solve this by hand and EVERYWHERE i looked... people did them by excel or other programs and just wrote out the answers without showing any work... thank you SO much for this video. helped a lot.
thanks a lot for the effort, statistics isn't a nightmare anymore!
Hilal özkeçeci Glad that you are no longer having statistics nightmare, best to keep your dreams to happy dreams. 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 for this lecture...
Hi many thanks for the video. I would like to ask how could I analyse one independent variable with two different levels’ influence on one dependent variable. For example, how do high level of organisational support and low level of organisational support affect on employee satisfaction? Could you please suggest how to do this? And how to classify the organisational support into high and low level of this collect via seven likert scale. I am confused about using multi group analysis and ANOVA. Many thanks for your help
2024 , and helps me ❤ thank you
Thanks a Lot! helped me understand Two way ANOVA much better from your video than from my professor. Keep making more !!
Thank you!!! so much!!! you can't understand how much this video help me!! really! thank you!
Good tutuorial, except for a minor the algebric error at 12:51 [(6-7) = -1].
extremely high quality, very intelligent, and elegant. The speed is great. not slow at all. I can watch it while thinking. I would absolutely ignore any not constructive comments.
Just a comment: When you start calculating SSE's, I'm pretty sure you're subtracting gender/age means not age means right?
fabulous video....do u have something such descriptive about how to do 2 way anova in minitab??? Or do you know if some one else have uploaded such video? can u just give me the like if it is there
thanks , your video really helpfull