00:15:59 Elizabeth Allen: Hi Kelly! 00:16:58 Barbara Pierpont: Will we receive a link to the recording? 00:17:35 Jennifer Ward: @Kate Kozak Hi! :) 00:18:26 Kathryn Kozak President: @ Jennifer Ward, Hi! :) and hello to everyone. 00:19:07 kim pham: hahaha..... 00:19:19 John Smith: Hello All. 00:19:39 Jonathan Hagy: https://fivethirtyeight.com/features/not-even-scientists-can-easily-explain-p-values/ 00:19:49 Sasha Friedman: "The probability that our results occurred simply by random chance" is informal definition, but I have found very helpful to students. 00:20:24 Sasha Friedman: And yes, simulations are absolutely the way to go. 00:20:33 Kelly Fitzpatrick: 4/20 00:20:43 Kelly Fitzpatrick: proportion 00:20:44 Kathryn Kozak President: Qualitative. 00:20:44 George Soliman: Test of proportion of lefties 00:20:50 Michael Heeren: left handed/ qualitative 00:20:52 Collin Byrnes: left hand or not 00:22:01 Sasha Friedman: .12 of the population 00:22:06 Kelly Fitzpatrick: Repeat question? 00:22:06 Michael Heeren: 14400 00:22:06 Collin Byrnes: 1440 00:22:10 George Soliman: 14,400 00:22:15 Mark Bellavia: 14400 00:23:11 Leo Fitzsimons: what about the ambidextrous? :) 00:23:24 George Soliman: But if the alternative hypothesis is true, then it would actually be 24,000 lefties 00:23:43 Paul McCombs: yes that is why not left handed I believe 00:23:47 Giorgio Turri: A sample of 20 out of a 120,000 members: doesn’t this mean we can expect a large error margine? 00:23:54 Collin Byrnes: wouldn't it just be something other than 14400? 00:24:07 Giorgio Turri: Due to pure statistical significance? 00:25:08 Sasha Friedman: yes with sample of 20 margin of error would be large 00:25:57 Sasha Friedman: The 120000 is irrelevant to margin of error 00:27:42 Kelly Fitzpatrick: Yes R thank you! 00:27:56 Kathryn Kozak President: Thank you Micheal. 00:28:00 Elizabeth Allen: Will this presentation be available? 00:28:41 Kathryn Kozak President: Yes, all material should be posted in myAMATYC (my.amatyc.org). 00:28:44 Anne Dudley: Yes, recordings will be at my.AMATYC.org 00:28:51 Elizabeth Allen: Thank you! 00:30:57 Paul McCombs: what was the website for R code again 00:32:03 Kathryn Kozak President: The website will be in his presentation on myAMATYC. 00:32:13 Kelly Fitzpatrick: https://www.sullystats.com/statistics-6e/r-guidebook/ 00:32:20 Paul McCombs: thank you 00:32:58 Kelly Fitzpatrick: I use R for every topic students enjoy it and now that everyone is home on a computer they are using it more! 00:33:27 qingshou kong: Is the simulation really necessary? 00:33:29 George Soliman: Once he mentioned R I immediately thought of you Kelly haha 00:33:49 Kathryn Kozak President: I use R also. My students really have figured it out, even the students with less computer skills. 00:33:49 Kelly Fitzpatrick: Thanks George:) 00:34:21 Sasha Friedman: The simulation is the best way for students to learn about sampling distribution models and pvalues. 00:34:43 Kathryn Kozak President: That is what I use to teach sampling distributions. 00:35:24 George Soliman: When you say "all stadiums", you mean all stadiums EXCLUDING Coor's Field, correct? 00:35:44 George Soliman: So in other words, "all OTHER stadiums" 00:36:00 Kathryn Kozak President: No, all stadiums would include Coor’s Field. 00:36:05 Kelly Fitzpatrick: Good point George 00:36:19 Kelly Fitzpatrick: ? 00:36:39 Kathryn Kozak President: I would think you would include Coors field. 00:36:49 George Soliman: Because if you include Coor's Field, the mean would be inflated 00:36:55 Kathryn Kozak President: True. 00:36:57 Sasha Friedman: Samples vary due to randomness 00:37:00 Michael Heeren: different averages of the samples 00:37:02 Kelly Fitzpatrick: I would agree with George.... 00:37:06 Heidi Lindsey: sampling error 00:37:31 Kathryn Kozak President: George, I agree, but I don’t know what Michael does. 00:37:41 Michael Heeren: I think it would be interesting to graph the p-value distribution. :) 00:37:44 Michael Kaye: Only looking at home run data can be misleading. Coors Field is bigger than most stadiums. So one must hit a ball farther in Coors Field in order for it to be a homerun. 00:38:16 George Soliman: Great point Michael! Not all HRs are created equal, since ballparks are different sizes throughout the MLB. 00:38:21 Giorgio Turri: I was going to say: how does one control the other, even hidden, variables? 00:38:30 Prof. Kiley (she/her/hers): Aah, the lurking variables 00:38:30 Kate Sims-Drew: @Michael Kaye I believe they intentionally made it larger because the ball carries farther in the thinner air. 00:38:32 George Soliman: Especially with a shorter porch at Yankee Stadium, for example 00:38:34 Kelly Fitzpatrick: either way P-Value can be difficult to explain and teach 00:39:06 George Soliman: p-values are rocket science to some stats students 00:39:18 Kathryn Kozak President: What about the fact that ASA says to not talk about statistical significance. 00:39:28 Michael Heeren: Is Coors field longer to account for the fact baseballs travel farther? 00:39:39 Patrick Staten: bicycle science lol 00:39:58 George Soliman: Anyway if Michael is accepting live questions, I'll ask him about whether or not he's including Coor's Field in his null hypothesis 00:40:34 Melissa Quigley: Sample size increases power 00:40:43 Kathryn Kozak President: Larger samples increase the power as Melissa says. 00:40:45 Sasha Friedman: decreased margin of error due to increased sample size 00:40:57 Sasha Friedman: increases power, given a fixed level of alpha 00:40:59 Heidi Lindsey: by increasing sample size, variability is reduced, so the results are more reliable. 00:41:02 Catherine Moushon: larger samples decrease margin of error 00:41:17 Catherine Moushon: Well said Heidi! 00:41:21 Giorgio Turri: Is it the sample size or its relative size (respect to the total population)? 00:42:06 Sasha Friedman: No. Sample size relative to population size is a common mental mistake. Sample size is what matters. Not population size. Unless population is really small 00:42:25 Kelly Fitzpatrick: time and money problem with large n too 00:42:37 Sasha Friedman: Although not intuitive, it's not about having a sample size that is a certain proportion of population size. 00:42:45 Sasha Friedman: it's strictly the sample size 00:42:55 Giorgio Turri: Well, 10,000 is a big number, but if I am dealing with 10^23 atoms, a 10,000 sample of atoms is negligible. 00:43:15 Giorgio Turri: (I am an atomic physicist) 00:43:24 Sasha Friedman: if it is gathered randomly, 10000 is plenty 00:43:29 Brian DeSantis: Heart attack 00:44:00 George Soliman: Typical Type I error 00:44:10 Michael Minic: The formula for sample size actually does not depend on population size (unless the population size is very small) 00:44:14 Sasha Friedman: again, it's not about what proportion of the population size the sample size is. It's strictly the sample size. 00:44:34 Giorgio Turri: But over a huge population, you can not assume complete uniformity. Thus a relatively small sample has a significant chance of be ‘peculiar’ 00:45:06 Kathryn Kozak President: I am ashamed to admit that I am not real confident of what an effect size is. Does anyone have a good definition? 00:45:15 Sasha Friedman: If the sample is randomly gathered from population (thus representative) then it does NOT need to be a certain proportion of population. 00:45:39 Steven Diaz: LOL 00:45:52 Giorgio Turri: Uhm, we should continue the conversation later. I am not convinced. 00:46:09 Susan McCourt: @Sasha: random samples are not necessarily representative. 00:46:10 George Soliman: So in other words, effects of a Type I error? 00:46:23 George Soliman: As in negative* effects 00:46:44 Judy Williams: I was just going to suggest creating a conversation about this is the Stat community on myAMATYC. Julie Hanson?? 00:46:48 Kelly Fitzpatrick: I should start teaching p-hacking, very interesting 00:46:49 Sasha Friedman: OK, then if the sample is representative. 00:48:11 George Soliman: Good stuff Michael, thank you! :) 00:48:11 Sasha Friedman: that third link is disturbing.... but true! 00:48:21 Sasha Friedman: Thanks Michael! 00:48:40 Elizabeth Allen: Thanks Michael. I enjoy your talks. 00:49:39 Kelly Fitzpatrick: Yes 00:49:45 George Soliman: Yes 00:49:58 Heidi Kiley: Isn't the data from Coor's Field part of the population data? 00:50:17 Kathryn Kozak President: I agree with Heidi. 00:50:22 Giorgio Turri: I don’t find the raise the hand tab. 00:50:22 Jackie Johnston: Thank you! 00:50:26 Catherine Moushon: Thank you Michael! Good ideas to give a little zest to my Intro Stats classes. : ) 00:50:28 Sasha Friedman: Thanks for the presentation! 00:50:32 Heidi Kiley: That would be like leaving out the Mensa members in the previous left versus right handed example, no? 00:51:14 Gayle Gosch: How will we gain access to this Powerpoint? 00:51:26 Kathryn Kozak President: It will be posted in myAMATYC. 00:51:48 Sasha Friedman: George thinks his sample should go beyond a threshold of a certain proportion of the population 00:51:55 Kelly Fitzpatrick: You need to compare two different groups. The groups need to be independent not dependent. My opinion. 00:52:08 Susan McCourt: This is inspiring? Any suggestions for motivating learning about Type I and Type 2 errors (beyond what's typically in textbooks)? 00:52:31 Susan McCourt: *inspiring!! 00:52:42 Kathryn Kozak President: Maybe we should not do one sample tests and just do two sample tests. 00:53:15 Kelly Fitzpatrick: good point 00:53:25 mike panahi: how do we get the power points? 00:53:36 Kathryn Kozak President: They will be posted in myAMATYC 00:53:43 Kathryn Kozak President: myAMATYC is my.amatyc.org 00:54:36 Roxy Peck: @Giorgio The sample size is the critical thing as it is what has the biggest effect on the standard error. So just because the population is huge doesn't mean that you need a huge sample in order to detect an effect.. 00:54:38 Susan McCourt: thanks 00:54:41 James Beedie: My Question is as a math tutor is how do I explain P-values? 00:54:56 Roxy Peck: Thanks Michael. Great talk. 00:55:01 Elizabeth Allen: I don't know about others, but I find that teaching the p-value concept right now very frustrating. I want my course to be relevant and I want to be able to give my students the best information possible. But I don't really know where to go from here 00:55:02 Holly Wendel: Thank you! 00:55:04 Mary DeHart: Thanks, Michael! 00:55:04 Mary Guzman: Thanks :) 00:55:04 Cindy Box: THANKS MICHAEL! 00:55:05 Daisy Lam: Thank you!! 00:55:06 Sylvia Brown: Thanks 00:55:06 Kate Sims-Drew: Thank you! 00:55:07 Stephen Burrus: Thank you, Michael. 00:55:09 Patrick Staten: Thanks 00:55:09 Giorgio Turri: Thanks great talk :-) 00:55:09 Laura Egner: Thanks! 00:55:09 Julie Hanson: Thank you Michael. 00:55:09 Wade Wells: Thanks, Michael! 00:55:09 Jim Anderson: Thank You!! 00:55:10 Heidi Lindsey: Thanks Michael!! 00:55:10 Patrick Torres: Thank you, Michael! Great presentation! 00:55:11 Lisa Zyga: This was great, thank you!! 00:55:12 Terry Varvil: THANK YOU MICHEAL FROM TAMPA!! 00:55:12 Melissa Quigley: Thank you! 00:55:13 Dan Mullins: Thank you!!! 00:55:13 Kathryn Kozak President: Thank you. 00:55:14 Barbara Pierpont: Thank you! 00:55:17 Sudha Kolathu Parambil: Thank you. 00:55:17 Cindy Scofield: Thank you! 00:55:18 Heidi Kiley: Thank you thank you!!! 00:55:19 Tony Paz: Thank you! 00:55:19 James Morgan: Thank you Michael! 00:55:20 Kristel Ehrhardt: Thank you! :) 00:55:22 kim pham: thanks so much! 00:55:24 Faye Dang: Thanks Mike! 00:55:28 Mario Morin: Thank you - great presentation! :) 00:55:29 Mitra navab: Thank you very much 00:55:30 Pauline Pham: thank you 00:55:30 Daphney Hill: Thank you! 00:55:31 Jeffrey Grell: Thanks Michael. Great talk! 00:55:31 FELICIA PETRENCO: Thank you. 00:55:31 Barbara Leitherer: Thank you Michael - that was great. 00:55:32 Faun Maddux: Thank nyou! 00:55:33 Susan McCourt: Look for mike Sullivan and George Woodbury's videos too! 00:55:36 Anne Dudley: If you have already registered for a session, please use the link in your email, Otherwise, follow one of the links below to register for the session of your choice. A Better Future for Online Discussion Boards Built From the Ground Up: College Algebra with Corequisite Support Non-STEM Corequisite Content and Resource Development Enhancing Students Experience of Learning Statistics If you have already registered for a session, please use the link in your email, Otherwise, follow one of the links below to register for the session of your choice. A Better Future for Online Discussion Boards Built From the Ground Up: College Algebra with Corequisite Support Non-STEM Corequisite Content and Resource Development Enhancing Students Experience of Learning Statistics 00:55:38 Ryan Petitfils: There's a cute example about Type I and Type II errors related to asking a girl out on a date: google 365 Data Science Type I vs Type II error 00:55:39 Farrah Chmilnitzky: Thank you! 00:55:57 mike panahi: Thank you from dallas your friend Mike Panahi 00:55:58 Collin Byrnes: Thanks a lot for your time--very thought provoking!