This is beyond great. I love how you are able to turn something this complicated, into something so simple. Great with references to WHAT you can actually use these things for, instead of merely knowing HOW you can calculate it. A big thumbs up from Denmark.
your lectures are only better source I found so far to understand these concepts. They are done in a very well detailed manner for understanding below par person like me. Thanks a lot Brandon.
Hello! Thank you for watching. Actually all of my videos are organized into sequential topics and then within each playlist also on my channel page. RU-vid does a terrible job letting you know what you are watching is part of an overall playlist. Click "Videos" under my name and then change "Uploads" to "Playlists" Let me know if you cannot locate them. Best, B.
Thank you for the positivity and the slow, deliberate explanations of each step. Your videos make me feel like I can actually get through my statistics class.
All your lectures are awesome. There is no way to describe how much I and others benefit from your approach. You maybe an outlier but we should learn from that too.
a friend linked me this thing. After watching these videos I feel much safer in the course Im following. These videos quite literally created a bridge between me failing and passing my course. Thank you so much!!
Thank you Brandon so much. You are an outstanding teacher. I appreciate the work you have done: guiding people along their journey in the world of statistics and data analyses. I also appreciate your having recognized the work being done by Josh over at Stat Quest. Both of you are my You Tube gurus for statistics. Your sincerity and conviction as a teacher are so wonderfully witnessed in these teaching courses. Again 'danyabad' (thank you)!
Excellent Sir, I am thankful for these videos, they really help out. I am a business student, although at PhD level but not very good at statistics. I am able to build these concepts through these videos. Thank you again sir for uploading those videos. Humble Regards
I never comment but your explanations are so helpful that I just have to say thank you and please keep these videos up! Your motivation very necessary in finals season thanks thanks thanks!!!
One of your best presentations. Not only understandable, but showing the app and especially the Excel formula was a blessing. Often, statistics ends up becoming more of an advance Excel course than a statistics course. Loved the extra help!
Thank you for making these videos so clear and precise. As a student learning Statistics for the first time, I found these very helpful and simple to understand. You my friend, are awesome. Ps: I love your encouraging messages at the start and end, it really helped motivate me to keep trying. :)
Thank you sir.. I was actually confused a lot before in my class with regard to the concept but you have made f-test my favourite topic. Thank you again...😊
thank you for your videos before I found them I was completely lost in my stats class could not follow along with discussion or class work at all the teachers explanations went right over my head but your videos saved me. thank you very much
Thanks so much for the pace and clarity. Pretty cool how global your fan club is! Names from around the world. This grateful grad student is saying "thanks" from Montana.
That was a great video Brandon. Very clear, well paced (for a beginner) and informative. I have liked, subscribed and passed your channel on to other students here in sunny Australia.
Thank you so much for your videos!!! They are helping me so much in my statistics class... Your videos really help me to have a clear understanding of all the concepts
Can u make an abridged video or a comprehensive video, where u can cover situation in which we can use a particular video. Also, plz make an video, emphasizing statistics in medical field - 1] How to read the result of study 2] How to determine that the study is plausible and not just manipulated via p value 3] Effects of power[ a dedicated video] 4] Cohen's d, this thing popped up on many sites 5] Biostatistical concept like odds ratio, relative risk 6] How to conduct studies and hence write thesis U are just awesome. The amount of Lynda certification u got is just awesome and ur videos speaks for them. I wish u a very beautiful career and happy life! Keep it coming.
Brandon, many thanks for putting this series on youtube. Can anyone tell why can't we find F-ratio for sample means similar to what we did for variances? why do we have to compare them using t-test or z-test?
I am blind so I appreciate your descriptive Ness… Thought even when I was sighted I never could recall the tail except for I believe it is either negative or positive and has some to do with the grid upper left upper right lower left or right?
In 6:30 it was stated that we are not comparing sample variance with hypothesized variance… BUT IN REALITY, WE DO SOMEHOW. We have population variance1 and population variance2 in the F statistic. because F statisitc is actually ((sample variance1)/(population variance1))/((sample variance2)/(population variance2)). So we may not be “comparing” sample variance wih population (hypothesized) variance but we are “comparing” sample variance RATIO with the population variance RATIO. That is similar to the test of 2 means where we compare the sample mean DIFF with the population mean DIFF.
Nice videos! By the way, do you have a pdf about the topics of your videos? I would like to skip the first 10 playlists but reading those topics. Thanks in advance 😀
Hi Brandon, I have basics understanding of Stats concepts. Can you please help me with the complete curricula or syllabus to get a complete statistical knowledge? I didn't find any such resources useful with the complete flow of where to start but your videos has helped me a lot. Thanks Keep it up.. have a great day ahead.
Very clear video. Love it. But in some contexts they still use 2 tailed test. Its not like u said in 13:07 and this does change the result. So what should I do. upper right tailed or 2 tailed. Thank you
First of all, thank you very much for the great videos!!! I have some questions: At 10:10, why this is one tail test? The purpose is comparing two can machines to see if their variances are equal, this should be two tail test, I think. And the Excel way, by watching your video, it seems no place to choose one-tail or two-tail. Sorry I could not try Excel on my computer because my Excel version is too old and 'F.INV.RT' is not available.
Dear Sir: I am very sorry. You have explained it very clearly one minute afterwards. I am checking with my own textbook. It is funny that the book on my desk treats this kind of case as two tail test. I think yours makes more sense. Thank you for your insight!
7:54 do I have to check whether the population is normally distributed before conducting this test? How do I go about doing this? I understood everything, but this slide. Please clarify
Hi @Brandon, thanks for the video. Perfect explanation. My question is - at 7:50 (ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-UWQO4gX7-lE.htmlm50s) you mention that two normal populations with equal variances. Here the "equal variance" of two populations is an expectation, and that's what we are testing for. Currently, the structure of that sentence feels like the equal variance of two populations is a requirement to use F-distribution.
Nicely put! But the critical value for the two-tailed F-test is not correct. If 0.05 is chosen as the significance level, then 2.368 (using excel function FINV(0.025,24,21) to get 2.368) instead of 2.054 should be used as critical value. Otherwise, the type I error is inflated.
Hello! Thanks for your comment. The video is correct as shown. As seen, this is problem is set up as a one-tailed (always upper-tail) test at .05 (=F.INV.RT(0.05,24,21), F-crit of 2.054. If it were done as a two-tailed test the F-crit would be 2.368 (=F.INV.RT(0.025,24,21) but it is not in this case. Since the larger variance is always placed on top of the F-ratio, we only look at the upper tail regardless. With the larger variance in the numerator, a F-value in the lower tail will not occur since F >= 1.
@@BrandonFoltz Hi Brandon. Thanks for your prompt response! Nice to have the discussion. If the alternative hypothesis is about "variance1 not equal to variance2", then, per hypothesis testing logic, null hypothesis would be rejected if the ratio of variance1 to variance2 is either significantly close to 0 (one direction) or significantly larger than 1 (the other direction). Now, since we "strategically" always put the larger variance as numerator and smaller variance as denominator, the calculated F-test statistic value will always be >=1. In other words, all "significantly close to 0" values will be flipped to "significantly larger than 1" values. But that does not mean the critical value on the right tail should be based on 0.05. It should still be based on 0.025 (several textbooks have already explained this in details. for example, Biostatistical Analysis book by Jerrold Zar, 5th edition, Example 8.7 in Chapter 8.). It is tricky to calculate the P-value though for such scenario because it should be 2*prob(F >= F_observed). But if your alternative hypothesis is about "variance1 greater than variance2", then I agree that 0.05 should be used to calculate the F-crit.
I was following perfectly until you said that the confidence level was .95 and then you said "remember?" well I rewatch the whole thing and it was never mentioned how you got to that conclusion so I got lost. how do you determine the number of the probability? I'm working on my MS of I-O PSychology and as a part of the proposal and other CLC assignments we have to do this and I can't figure out my part, this was partially helpful like 95% since I have never been exposed to this subject before.
From 4:40, what if the info. was presented such that X had n=22, x bar=4.9808 oz, s squared=0.0137 and s=0.1171 and Y had n=25, x bar=5.0592 oz, s squared=0.1130 and s=0.3361. Would you still have to make the bigger variance(from Y) the top number seeing that X comes before why? So that it would be in fact, s quared of Y / s squared of X?
