Thanks for the question! If you need to build a confidence interval by hand when you don't have actual data (for instance on a stats exam), you should use the formula μ=xbar ± t* (SE), where t* is calculated using the qt() function and SE=s/√n. I get into this idea at around 1:55 of the vid.
Hi, thanks for the video! It is super helpful in rounding out the resources I have as a student in R. I was wondering, for the function t.test, is there a way to change the null hypothesis? For example, if instead the null hypothesis is true mean is greater than x, is there a way to do so using the t.test function? Thanks a ton!
Hello, many thanks for your video. I have a question and the question is calculate mean, median, mode and standard deviation of the given dataset and the 95% confidence of the interval. It is a large dataset with 8 variables and each variable has 1640 objects. I understand everything except the 94% confidence interval part. I know how to find out mean, mode, median etc. But what should i do with this 95% interval question. Should i do it like you did? Please answer!
Hey! It sounds like you just need to do what I did in the vid. Something like t.test(dataset$variable, conf.level = .94) or t.test(dataset$variable, conf.level = .95), depending. Hope this helps!
Hello thank you so much for your reply. I am doing masters at Stockhom University in 'Environmental Science and biogeochemistry'. I have some courses of data science. Is that ok for you if i ask your precious suggestion?
Hi! A 95% confidence level leaves out 2.5% on the left and on the right, so the critical t-value must capture 97.5% of the probability to its left. I get into this at around 2:35.
@@EquitableEquations one more question while getting lower or upper bound should we use one sided t value or two sided t value to -/+ from sample mean?
Hi! Violation of normality isn't a problem unless the sample is very small. Many texts give n=30 as a rule-of-thumb threshold, but in practice this is very conservative.
@@EquitableEquations Ah sorry, I meant to edit this sooner, because per group my data doesn't meet that criteria (observations per group are between 11-72). Would this work? Or is it only building C.I.'s for the pseudomedian and not the mean? wilcox.test(dat, conf.int = T, correct = T, exact = F, conf.level = .95)
@@EquitableEquations I think that's the way to go. I don't have symmetry either because most values = 0. I also have heterogeneous variance, which I've heard also causes problems... Thank you!
Thank you for this! I was having trouble setting a different level than the default .95! This cleared that up for me!
Thanks!!
how to perform t.test in Rstudio if I do not have data, I only have a standard error(se)
, and mean sample(xbar)
Thanks for the question! If you need to build a confidence interval by hand when you don't have actual data (for instance on a stats exam), you should use the formula μ=xbar ± t* (SE), where t* is calculated using the qt() function and SE=s/√n. I get into this idea at around 1:55 of the vid.
Thank you so much!
You're welcome ;)
Thank you!
It is simply great!
Thanks!
Great video thank you so much!
You're so welcome!
Can you tell me how to perform simultaneous vonfidence intervals in R or RStudio??
Hi! There are several ways to go about this. I expect to have a vid on the subject later in the year.
Hi, thanks for the video! It is super helpful in rounding out the resources I have as a student in R. I was wondering, for the function t.test, is there a way to change the null hypothesis? For example, if instead the null hypothesis is true mean is greater than x, is there a way to do so using the t.test function? Thanks a ton!
Hi! Check out this vid on significance testing in R:
ua-cam.com/video/iMAE47S_Xq4/v-deo.html
It's also done with the t.test() command.
Hello, many thanks for your video. I have a question and the question is calculate mean, median, mode and standard deviation of the given dataset and the 95% confidence of the interval. It is a large dataset with 8 variables and each variable has 1640 objects. I understand everything except the 94% confidence interval part. I know how to find out mean, mode, median etc. But what should i do with this 95% interval question. Should i do it like you did? Please answer!
Hey! It sounds like you just need to do what I did in the vid. Something like t.test(dataset$variable, conf.level = .94) or t.test(dataset$variable, conf.level = .95), depending. Hope this helps!
Hello thank you so much for your reply. I am doing masters at Stockhom University in 'Environmental Science and biogeochemistry'. I have some courses of data science. Is that ok for you if i ask your precious suggestion?
Sounds like an exciting program! Good luck!
Why is it .975 ? How to get lower and upper bound from 95%
Hi! A 95% confidence level leaves out 2.5% on the left and on the right, so the critical t-value must capture 97.5% of the probability to its left. I get into this at around 2:35.
@@EquitableEquations one more question while getting lower or upper bound should we use one sided t value or two sided t value to -/+ from sample mean?
Does this work for non-normal data? (I assume the CO2 data is normal since you used the t-test on it)
Hi! Violation of normality isn't a problem unless the sample is very small. Many texts give n=30 as a rule-of-thumb threshold, but in practice this is very conservative.
@@EquitableEquations Ah sorry, I meant to edit this sooner, because per group my data doesn't meet that criteria (observations per group are between 11-72). Would this work? Or is it only building C.I.'s for the pseudomedian and not the mean? wilcox.test(dat, conf.int = T, correct = T, exact = F, conf.level = .95)
@@brazilfootball Even n=11 is fine if the distribution is relatively symmetric. You can also consider bootstrapping.
@@EquitableEquations I think that's the way to go. I don't have symmetry either because most values = 0. I also have heterogeneous variance, which I've heard also causes problems...
Thank you!
Why did he choose to put 90 as the confidence level at the end?
Hi! I just wanted to show how to do a different confidence level. There's nothing special about 90%.