Thanks for the compliment Andrea! I haven't gone anywhere -- just a little pressed for time over the past year or two. I fully intend to get back to video production in the not-too-distant future. All the best.
Could I make a request/suggestsion for content exploring Logistic/Poisson Regression & GLM in general & more far reachingly Time Series Analysis? (no pun indended). Fantastic videos by the way. I enjoyed and learned a lot from them!
Can you reformulate this question '' Is it unlikely to see a difference of this size due to chance alone?" ? I find negation harder to grasp than sentences said in the affirmative.
This is really fascinating videos for stats....... I am wondering >> When we would use inference for variance or inference for mean? .... in other words, when we need to detect the statistical significance between two samples, which method we have to use? Thanks,
Hey, I'd like to ask a question here, does the population has to follow normal distribution? cause based on central limit theorem, the sampling distribution of population mean will follow normal distribution regardless of population distribution.
Sample standard deviation does not result in an exact t distribution. Pooled variation results in exact t distribution Mann-whitney U (wilcoxon rank-sum test) and bootstrap do not assume normality and permutation test
The procedures in this video assume we have two independent samples (e.g. a random sample of 50 students from Stanford, and a separate random sample of 50 students from Harvard). A paired-difference t procedure is appropriate when there are pairs of measurements that are dependent (e.g. before and after measurements on the same individual, tire wear on the left and right wheels of a car, IQ of identical twins, health status on individuals matched according to age, sex, smoking status, etc.)
Hi. I need to ask suppose I have 2 independent samples from independent populations. Their variances are known but unequal. Which test should I use to test difference in their means? Welch t test is when variance is unknown right?
If you (somehow) happen to know the population variances, then it would be a z test. (Same idea as for one-sample problems.) The formula would be the same as the Welch t formula, but with the population variances replacing the sample variances. You'd get a z test statistic, and use the standard normal distribution to find the p-value.
@@statisticsonsteroids How do you know the population variances? Yes, if you happen to know the population variances, and are sampling independently from two normally distributed populations, then a z test is appropriate. That comes up in practice pretty much 0% of the time, but you might see it in some text or assignment questions.
literally sobbing because this channel saved my college life
I learned more with this video alone than with two weeks worth of lecture. Thank you so much for this video! You rock! :)
How is it that you can explain these concepts in under 10 minutes and so well?? Yet my lecturers can't even do it in a whole semester? 😭🙏🏾
When you say you derive equations elsewhere, would be super helpful to provide a link to that video so I can watch more of your videos
Thank you bruv, I've been trying to get this concept for days
I'm glad to be of help!
Would have liked to see an example.
Yes
Your videos are amazing. It's a shame you haven't been publishing for more than a year. Come back, please!
Thanks for the compliment Andrea! I haven't gone anywhere -- just a little pressed for time over the past year or two. I fully intend to get back to video production in the not-too-distant future. All the best.
Could I make a request/suggestsion for content exploring Logistic/Poisson Regression & GLM in general & more far reachingly Time Series Analysis? (no pun indended).
Fantastic videos by the way. I enjoyed and learned a lot from them!
u r awesome!
Can you reformulate this question '' Is it unlikely to see a difference of this size due to chance alone?" ? I find negation harder to grasp than sentences said in the affirmative.
Learn more from you then my professor
best prof! sad that covid-19 made our lectures get cut short but these videos are great
This is really fascinating videos for stats.......
I am wondering >> When we would use inference for variance or inference for mean? .... in other words, when we need to detect the statistical significance between two samples, which method we have to use?
Thanks,
Hey, I'd like to ask a question here, does the population has to follow normal distribution? cause based on central limit theorem, the sampling distribution of population mean will follow normal distribution regardless of population distribution.
Sample standard deviation does not result in an exact t distribution.
Pooled variation results in exact t distribution
Mann-whitney U (wilcoxon rank-sum test) and bootstrap do not assume normality and permutation test
I'm not sure whether you're simply stating things, or implying I said something different and wrong. If it's the latter, I didn't.
@@jbstatistics no, your video is excellent. I'm just taking notes
Random question, what font is used for the video?
What's the difference between a two sample test and a one sample test with matched pairs?
The procedures in this video assume we have two independent samples (e.g. a random sample of 50 students from Stanford, and a separate random sample of 50 students from Harvard). A paired-difference t procedure is appropriate when there are pairs of measurements that are dependent (e.g. before and after measurements on the same individual, tire wear on the left and right wheels of a car, IQ of identical twins, health status on individuals matched according to age, sex, smoking status, etc.)
Is the inference for two means calculated the same way as inference for two variances?
No, the methods are quite different.
jbstatistics Dear jbstatistics
thank you for your reply.
Are you interested in making a video of interferences for two variances?
thank you, very very helpful.
+Rabah Ferjani You are very welcome!
why are all teaching videos having such small volumns, you guys need to speak it up.
Thanks...
You are very welcome.
Hi. I need to ask suppose I have 2 independent samples from independent populations. Their variances are known but unequal. Which test should I use to test difference in their means?
Welch t test is when variance is unknown right?
If you (somehow) happen to know the population variances, then it would be a z test. (Same idea as for one-sample problems.) The formula would be the same as the Welch t formula, but with the population variances replacing the sample variances. You'd get a z test statistic, and use the standard normal distribution to find the p-value.
@@jbstatistics Thank you. So I can safely use z test anytime I know the variances regardless of whether they are equal or not, right?
@@statisticsonsteroids How do you know the population variances? Yes, if you happen to know the population variances, and are sampling independently from two normally distributed populations, then a z test is appropriate. That comes up in practice pretty much 0% of the time, but you might see it in some text or assignment questions.
@@jbstatistics yes yes ofc, I'm just speaking theoretically. Thank you for answering on a 9 year old video to help me out🙏