Hi Im currently on my 2nd year of studying BSc Psychology with Open Uni in the UK. I have looked at literary 100's of website and loads of RU-vid stuff plus reading my study books for DAYS until I finally came across your posting. Thats just the kind of example I wanted. No blah blah or complicated scribblings, just plain and simple 'this is how you do it'. Thank you
I just want to say thank you, statisticsfun ! Every aspect of your work, from technology side of it, to statistics, practical examples, timing, clarity of instructions and calculations, was excellent.
I am an english teacher. I have been trying my best to understand statistics for couple of days as I want to be a researcher. This explanation is godly great!!! and very easy to understand!!!
You have *no idea* how amazingly useful this is. Thank you muchly. The use of the arrows and moving the figures like you do, helps this visual learner understand where the numbers come from.
You rock with your explanations..I"m a grad student at UC Berkeley working on my Master of Public Health and Master of Public Policy. I wish we had someone like you at Cal to explain things clearly...You have a true gift!
Thank you for making this clip!! You made it so clear and easy to understanding!! I wished that I came across this earlier!!! Glad to catch this before my Com Exam!!!!~
@statisticsfun It's funny because I never saw statistics as boring but only the way it was taught as boring. Your lectures on the other hand is much more interesting than my books - which is why your site MyBookSucks make sense. I'm looking very much forward to your future videos. Sometimes things are boring if you don't know the use of it. How about applied statistics - how to design experiments?
You are very welcome, always glad to help. Make sure you like MyBookSucks on FaceBook (see link in video description). This will help others find the educational videos.
What a nice thing to say, much appreciated. Happy to know you can understand my American accent too. Make sure you like MyBookSucks on FaceBook (see link in video description). This will help me spread the word about my educational videos all over the UK!
Wow! Thank you and I appreciate your feedback. Make sure you like MyBookSucks on Facebook (see link in video description). This will help others find the educational videos.
@Norfeldt I use a variety of different programs to create my videos -- not just one product. The idea behind MyBookSucks and the tagline "party more study less" is to attract those students who struggle and eventually learn to hate math subjects (stats, econ, , maths). Most students just want to pass these subjects and move on.
I have a questions: is it cohen's d for the groups with the same smaple size ? or we could also use this way of calculations for different group sizes? I think for different group sizes std should be calculated with weights? If so, can you shortly tell how? Otherwise, it was a very clear explanation, and thank you a lof for this video. Always when I watch your videos they are so helpful, and I feel I really learn here! :) Thank you!
Good question. It should be the same. Meta analysis is comparing and contrasting results from different studies and you should be able to use Cohen d to compare two studies. Make sense?
Another question is that according to other sources, the standard deviation in the denominator should be the pooled standard deviation: s = sqrt[((n1-1)s1^2 + (n2-1)s2^2)/(n1+n2-2)], why is this case different?
That is almost subjective depending on what you are looking at. Often the actual Cohen D number is reported and not include "small/medium/large" wording.
I have seen different levels too. In the end the level of small, medium or large is going to depend on what you are studying. As long as the actual Cohen d is provided it really does not matter.
Great vid. Could you explain why a statistical significant results could yield a very minimal effect size and how the distributions plots would look like and how they would overlap graphically? Thank you. This would be very helpful.
A very very good video. What program did you use to make this presentation with all the effects? (I don't like the last comment "Study less party more" - how about "Learning that makes you understand"
He didn't calculate the standard deviation, he just used the standard dev scores for each distribution (which may have been arbitrarily chosen) in the effect size calculations
Great video (thanks), but please make it clear that there is a bug in your formula regarding the calculation of S pooled (I was alerted by people here in the comments). Thanks again, and keep the great work.
Thank you for simple explanations, I am a visual learner and the visuals help so much. If I could make one request it would be to provide more information on how to apply this information in the real world. I am presented with problems of interpretation. :-(
That was well fucking done, sir. Note how I am much more inclined to use swearing as I'm not in class and you are not actually my lecturer, yet you anticipated any of my questions before I ever even asked, the way any lecturer worth his position should. thankyouthankyouthankyou.
The denominator should be square root of the pooled variance. It is not the "average" standard deviation. If there are equal sample sizes, one can calculate the average variance (called pooled variance) in an arithmetic fashion. If you want a stable estimate of standard deviation, base it on the variances. In the example at 3:44 the appropriate denominator is 1.92
Always great to hear and I appreciate the feedback. Make sure you like MyBookSucks on FaceBook (see link in video description). This will help others find the educational videos.
Susan King Great to hear that my videos are helping you. I accepted your linkedIn request and gave you a bit of linked in love (endorsed you for a couple of skills). Good luck in your studies too.
@@morpheus7144 Not an expert, here, but yes. If the means of two sample sets differ only slightly and each sample set is highly variable (a large spread), the observed effect is small. In other words the investigated effect on the non-control group would be more difficult to illustrate experimentally (because it is so small). Larger samples sizes would be required to offset the difficulty in reliably demonstrating such a small effect, that or an experimental paradigm with more precise and relevant measures to better tease out the effect of interest.
I use a variety of different software to create the videos including Flash, HTML5, Illustrator, Photoshop and Final Cut Pro. As yo can imagine it takes a long time to create the videos.
Very cooool What if we have 3 experimental and 1 group; and we want to find effect size between the EXPERIMENTAL groups what to write in the numerator (group 2-group3) or vice?
why would you assume the standard deviations between the two groups are equal? Is that usually true? I'm imagining that those bell curves represent the posttest measures of the effect of the independent variable, and wouldn't those have different means and standard deviations for the two groups? How is this calculated in the SD's are different?
Hello,this video helped us a lot in our thesis paper :) I;d just like to asked what source you used for the verbal interpretation for cohen's d? Thank you sir.
Hi, may I ask for the Mean of Control group/Experimental group, is it the mean of post-test? Or the mean of change from post- to pre-? Thanks, this question bothers me for a long time.
I've read different benchmarks by Cohen ( 1988). I've read that >0.2 is small, >0.5 is medium, and >0.8 is large. Can you verify your benchmarks? (from The Importance of Effect Sizes in the Interpretation of Research by Schuele, C Melanie; Justice, Laura M, 2006)
I am doing a study where three treatment groups are involved compared to one control group..So will it be okay if calculate effect size using single treatment grp and control group, will there be any difference?
