One-Way ANOVA: LSD confidence intervals

[@KiddSteely]

10 years later and youre still helping a ton of people out

[@jbstatistics]

Glad to be of help! Good luck on your exam!

[@trietleminh3929]

THANK YOU SO MUCH!!! WISH YOU ALL THE BEST. YOU JUST SAVED MY LAST BRAINCELL FOR STATISTICS.

[@navinreddy1087]

really got addicted to your teaching sir

[@dutonic]

This unit of stats has basically driven me to the brink of LSD anyway

[@isadorabordini9528]

You're an excellent teacher! Thank you so much for sharing your knowledge!

[@jbstatistics]

Thanks for the kind words!

[@melaniemerchant1949]

looking forward you post more advance materials!!! you are life saver for all!!!

[@isaacchewe1825]

JB you're the best 😊👍. You're really helping me prepare for my exams.

[@jbstatistics]

I'm glad to be of help!

[@n8chz]

This video has everything, LSD, SEX...

[@cococnk388]

Thanks, but I do not get the point.... Here is how I see the thing ...am I right? AR>P>AC>A but we refuse to do this conclusion because at a certain point "AR - P = 0" due to the range of the upper level and lower level (same reasoning for "AC - P") while the other pairwise differences falls within the negative half completely without crossing zero... More to that we can conclude because we are dealing with more than two groups using the same "alpha".... ? Increasing the probability of rejecting the null hypothesis?????

[@yiyingzhou5311]

I learned ANOVA from your nice videos. Thanks! Just to make sure the table that shows the confidence intervals should be with the 97.5% confidence, not 95%, right?

[@jbstatistics]

Those were 95% confidence intervals. We split the remaining 5% evenly into both tails, so there's .025 in each tail. I've defined t_.025 to be the t value with an area of .025 to the right.

[@sonsofaaron]

These are so helpful. Do you have videos where you discuss the Bonferroni method and the Turkey's method?

[@jbstatistics]

I'm glad to be of help. I don't have videos on the Bonferroni method of Tukey method yet, but I will at some point in the future.

[@Maha_s1999]

Yes, please Prof Balka

[@adrienneabu-assal755]

This helped so much!! Thank you!

[@leesweets4110]

Cant you run ANOVA on each three-group subset of the data to determine which group(s) is the odd one out?

[@savannaha5038]

I wish you were my stats teacher, the prof teaching my class is so bad at teaching that she somehow managed to overcomplicate something as simple as what p hat means to the point where I had no idea haha

[@leesweets4110]

Is this real data? Im no expert here and intuition is a bad way of doing math... but I feel like the confidence interval should NOT include 0 in the comparisons with the P group, and contain 0 in the comparisons between all other groups. The group that did not consume any alcohol should have performed significantly better in general and the difference with other groups should be large; the groups that did consume alcohol should have performed more similarly and therefore the difference in means should be nearer to zero. Am I misinterpreting?

[@jbstatistics]

The summary statistics are real, so the conclusions from the ANOVA and the LSD procedure as the same as for the real data, but the data used for the boxplots is simulated data based on the summary statistics. I'd agree that the conclusions are perhaps somewhat surprising as far as the nature of the problem goes, but it's not too surprising that there appears to be a reward effect (students in the alcohol + reward group did better than those who had alcohol with no reward). Students in the placebo group did perform better than those in the alcohol group, but the reward and caffeine seem to provide a boost over just alcohol.

[@boreiaproasteia7173]

If i have for group1-> n=12, group2-> n=10, group3 -> n=11 and group4 -> n=11 what can i do to calculate the LSD

[@panagiotisgoulas8539]

@jbstatistics Which program you use to visualise data? The data table on kzbin.info/www/bejne/oYCbpZKnYqaLfqcm17s how did you generate it R? When I run the aov command in T it won't generate sample means and standard deviations. Is there any trick to do so without calculating manually using the tapply function...? Sorry on the kinda off topic questions, but do you by any chance know as well on how can I possibly run a summary table in multiple regression that summarizes information (r sq, r sq adjusted, mallows cp and vif) for an dependent variable vs EVERY combination of independent ones. I seen this in minitab but I cannot find a way to run the vifs and mallows cp for every combination of dependent-independent in R. I asked in stack overflow but I haven't gotten a reply. Help!

[@rachelcyr4306]

?With the df isnt the df for MSE n(t)-k? wouldnt that be 40-6 = 34?

[@gordongoodwin6279]

It’s n-k, there are only four groups, so 40-4=36

[@fakhri7688]

Thank you sir

[@jackl593]

Very trippy

[@juandavidmunoz2781]

Thank you!!!!!!

[@AnilKumar-sm3ky]

Jb: Just because of you

[@StephtherefAmazing]

Who's here from psych 292 at UW?

[@dchangebegins]

Condition for pair of straight line