00:07:39 Kelsey MacCuish: don’t worry about joining a breakout room! ill explain that in a bit :-) 00:12:48 Matthew Choy: Should review take up the entire lab today? 00:13:29 Kelsey MacCuish: yes it will take up the entire time 00:13:37 Matthew Choy: Got it got it 00:13:54 Christopher Patterson: good ! 00:13:56 Hiruni Jayasekera (she/her): tired!! 00:13:59 Ekua-Yaaba Monkah: great how are you!? 00:14:08 Annette Mercedes: Stressed from coordinating flu clinics 00:14:39 Annalisa Watson (she/her): Just this one lol 00:14:47 Phoenix Ding: I have two next week 00:14:49 Andriana Marijic Buljubasic (she/her): Yes so many 00:14:51 Annette Mercedes: Just this one, rest are early october 00:14:59 Matthew Choy: I’ve just been getting mauled by my chem class :( 00:15:07 Matthew Choy: 3a 00:15:09 Annette Mercedes: Yikes Chem!! 00:15:13 Matthew Choy: No robak 00:15:19 Matthew Choy: She is good but lots of material 00:16:34 Annalisa Watson (she/her): Will we get the Q3 answers before the midterm? 00:19:39 Ijeoma Uche: Where can we find this presentation 00:19:54 Annalisa Watson (she/her): https://docs.google.com/presentation/d/1UPk_V41-Lg9vLJbtnLkJvbriLdqqwfsKm5av-j3T-Zo/edit#slide=id.g40f7b8e44c_0_15 00:19:58 Maddy Griffith: @ijeoma it’s on the course schedule as a lab :) 00:20:06 Nikki Marucut: ^ it’s on the course schedule page! 00:20:06 Ijeoma Uche: Thank you! 00:25:58 Sam Holland: Line regression (especially interpreting slope!!!) 00:26:01 Hallie Roth (she/her): Simpson’s Paradox 00:26:03 Phoenix Ding: PPDAC and sampling 00:26:04 Annalisa Watson (she/her): Correlation and r squared 00:26:04 Maddy Griffith: Linear regression/coefficient 00:26:06 Hiruni Jayasekera (she/her): calculating conditional/marginal 00:26:07 Ala Koreitem: sampling 00:26:28 Andriana Marijic Buljubasic (she/her): The recent concepts — regression, sampling, designing 00:26:29 Annalisa Watson (she/her): Variables too 00:26:34 Nikki Marucut: What kinds of data do you use (and the code) for histograms and bar plots 00:33:13 Matthew Choy: Walking = x heart = y 00:33:17 Ekua-Yaaba Monkah: x= walking speed, y- heart rate 00:33:17 Julia Hankin: Walking speed = X 00:33:19 Nikki Marucut: Yes^ 00:33:34 Nikki Marucut: Because when you walk your heart rate increases 00:33:53 Ekua-Yaaba Monkah: this is very helpful, I always get tripped up with these problems 00:33:58 Nikki Marucut: Walking has to happen first to indicate what the heart rate is and that’s my way of thinking about tit 00:34:44 Lupita Ambriz: x= temp y=enzyme activity 00:34:45 Ekua-Yaaba Monkah: x= temp, y= enzyme activity 00:34:47 Annette Mercedes: Temp = x, enzyme = y 00:34:52 Phoenix Ding: X= temp, y = activity 00:35:07 Annalisa Watson (she/her): Enzyme function has an optimal temperature range 00:35:12 Nikki Marucut: Yes^ 00:35:58 Cristal Escamilla: x=child's age, y=later score on test 00:35:59 Lupita Ambriz: x=age of first word y=later score 00:35:59 Ekua-Yaaba Monkah: x= child's age at first word, y= later score on test of mental ability 00:36:03 Matthew Choy: ^ 00:36:22 Nikki Marucut: I think the word “later” indicates that it should go on the y axis 00:36:26 Annalisa Watson (she/her): ^ 00:36:40 Nikki Marucut: **Later score** 00:37:12 Nikki Marucut: It helps to talk through it! Thank you 00:37:19 Matthew Choy: ^^ defo helps talking it out 00:39:45 Annalisa Watson (she/her): We wouldn’t say that correlation measures spread? 00:41:10 Hiruni Jayasekera (she/her): how do we estimate r by looking at a graph with a line of best fit? 00:41:27 Annalisa Watson (she/her): Thank u 00:42:21 Stacy (Seohyun) Ahn: sorry did you say r squared can be a measure of spread unlike r 00:42:49 Annalisa Watson (she/her): There’s a good graphic in the lecture notes for r and line of best fit 00:43:34 Hallie Roth (she/her): How is r2 different from variance? 00:43:44 Hallie Roth (she/her): They seem similar but different 00:43:53 Annalisa Watson (she/her): / What do we need to know about variance? 00:44:45 Hallie Roth (she/her): Ah got it, thank you! 00:45:40 Ekua-Yaaba Monkah: What would the scatter plot look like if r=0.6 compared to r=1? 00:46:11 Matthew Choy: I think r = 1 is where all the points are on the line but r = 0.6 has more points clustered around the line 00:46:16 Matthew Choy: Kind of hard to word 00:46:22 Nikki Marucut: Yes ^ 00:46:41 Nikki Marucut: So would that be considered a strong correlation? 00:46:48 Nikki Marucut: How would you interpret that? 00:46:56 Nikki Marucut: Got it 00:46:58 Nikki Marucut: thanks 00:48:40 Phoenix Ding: Linear? 00:48:41 Nikki Marucut: Linear? 00:48:41 Maddy Griffith: Linear shape 00:48:42 Mariah Jiles (she/her): moderate 00:48:46 Ekua-Yaaba Monkah: linear 00:48:47 Annalisa Watson (she/her): Medium strength 00:49:17 Annette Mercedes: How are the first two questions different? 00:49:28 Matthew Choy: Would curved be ok too? 00:49:58 Nikki Marucut: Moderate, r = .6 00:49:59 Phoenix Ding: Medium strong around r=0.7 00:50:42 Nikki Marucut: positive 00:50:43 Lupita Ambriz: Positive 00:50:43 Phoenix Ding: positive 00:50:45 Hiruni Jayasekera (she/her): positive 00:50:50 Annette Mercedes: Positive 00:51:26 Nikki Marucut: Not really 00:51:26 Hiruni Jayasekera (she/her): not really 00:51:26 Lupita Ambriz: Yes 00:51:26 Ekua-Yaaba Monkah: no 00:51:30 Matthew Choy: I think its subjective 00:51:31 Annette Mercedes: slight 00:53:22 Mariah Jiles (she/her): does this only apply to scatter plots? 00:56:01 Nikki Marucut: Not always 00:56:18 Nikki Marucut: Isn’t it arbitrary? 00:56:24 Nikki Marucut: It’s just a starting point 00:56:29 Matthew Choy: So basically talking about extrapolation? 00:59:07 Hallie Roth (she/her): When we interpret slope, can we just say “one-unit” change or should we actually provide the unit if it’s given? E.g. if unit was person in a population, or years of age 00:59:33 Hallie Roth (she/her): Actually I think you just answered this :) 01:03:17 Nikki Marucut: Great okay thanks 01:05:20 Annette Mercedes: I have to leave for a training session. Will the recording be available today so I can watch the last part? 01:06:12 Annette Mercedes: Thank you so much!! This has been super helpful 01:06:17 Annalisa Watson (she/her): A one dollar increase in GDP per capita is associated with a .1166 dollar increase in health expenditure per capita 01:12:07 Lupita Ambriz: Yes 01:12:08 Nikki Marucut: Yes 01:13:23 Nikki Marucut: Fit <- lm(OFC2 ~ OFC1, OFCdata) 01:13:31 Nikki Marucut: summary(fit) 01:13:38 Nikki Marucut: ? 01:13:45 Annalisa Watson (she/her): I think it is tidy(fit) ? 01:14:12 Nikki Marucut: ^ ahhh yah 01:14:15 Nikki Marucut: That’s right thank you 01:16:08 Ekua-Yaaba Monkah: Do we need to type "formula" and "data" as well? 01:20:49 Hiruni Jayasekera (she/her): Moderate, positive, linear, relationship with no major outliers 01:20:53 Matthew Choy: Non-linear, positive, moderate, no outliers 01:21:08 Nikki Marucut: I would say it is very roughly linear, moderate, positive no outliers 01:22:02 Annalisa Watson (she/her): What are the options for form? 01:23:01 Ekua-Yaaba Monkah: Could we say curved for shape? 01:23:21 Annalisa Watson (she/her): Does strength refer to how strongly linear? 01:23:36 Matthew Choy: So basically we could say either linear or non-linear as long as we back it up appropriately 01:23:49 Andriana Marijic Buljubasic (she/her): How far does a point have to be to be considered an outlier? 01:30:03 Matthew Choy: Transformed has a stronger r^2 01:30:43 Matthew Choy: Ohh okok 01:31:08 Chitra Nambiar: X is transformed 01:31:12 Matthew Choy: ^ 01:31:15 Lupita Ambriz: Ofc1 transformed 01:31:18 Matthew Choy: X axis 01:31:48 Olufunke Fasawe: Why is only one axis? 01:31:58 Chitra Nambiar: log ? 01:32:04 Olufunke Fasawe: I thought both axes should have been transformed 01:32:06 Matthew Choy: It sort of matches up with the roughly linear part of the first graph 01:32:42 Stacy (Seohyun) Ahn: yes 01:35:11 Hiruni Jayasekera (she/her): ln 5 = 1.6 01:41:53 Olufunke Fasawe: Yes, it does 01:41:55 Olufunke Fasawe: Thank you 01:42:06 Nikki Marucut: Are we going over Two way tables 01:42:12 Nikki Marucut: Sorry ^ off topic 01:42:12 Phoenix Ding: 0.9 01:43:55 Matthew Choy: Isn’t it 2.09? 01:43:56 Nikki Marucut: For every 1 unit increase in OFC1 there is an increase of 2.09 units in OFC2? 01:44:01 Stacy (Seohyun) Ahn: for every increase of log(OFC1) value, there is 2.09 units increase in OFC2 01:44:02 Hiruni Jayasekera (she/her): slope is 2.09 where every one unit increase of ofc1 is associated with a 2.09 unit increase in log ofc1 01:44:05 Hallie Roth (she/her): A one unit change in OFC1, is associated with a 2.09 unit increase in OFC2 01:44:16 Maddy Griffith: For every log unit increase in OFC1, we can expect to see a 2.09 unit increase in OFC2 01:44:24 Nikki Marucut: Ah LOG 01:44:35 Hallie Roth (she/her): Logggg 01:44:37 Nikki Marucut: Okay noted 01:44:38 Nikki Marucut: LOL 01:46:07 Matthew Choy: Could we word it like “every 1 unit increase in log(OFC1)” or does it have to be every 1 log unit increase in log(OFC1)? 01:46:23 Matthew Choy: Okok tyty 01:47:24 Phoenix Ding: Y = 2.09*log(40.14)+6.03 01:47:26 Hallie Roth (she/her): OFC2 = 6.03 + 2.09 * log(40.14) 01:49:33 Mariah Jiles (she/her): No? are we are extrapolating data 01:49:34 Hiruni Jayasekera (she/her): we would want to look at 40.14 within the range of the original graph 01:49:37 Jessica Fields (she/her/hers): ln(40.14)=3.69, which is outside the range of data we have, so I think it would be considered extrapolation? 01:49:43 Phoenix Ding: I think no, because the upper limit we have is around 15 01:50:35 Matthew Choy: To clarify log refers to ln() for our use case right? 01:50:52 Hiruni Jayasekera (she/her): ^i believe yes matthew 01:50:56 Hallie Roth (she/her): So in this case we don’t need to exponentiate, right? 01:51:03 Hallie Roth (she/her): To find the value (even if out of range) 01:51:14 Matthew Choy: tyty Hiruni 01:51:56 Hallie Roth (she/her): Got it 01:52:17 Nikki Marucut: This was great 01:52:29 Nikki Marucut: I’m glad we spent time on this 01:55:08 Phoenix Ding: (A+C)/(a+b+c+d) 01:55:11 Hiruni Jayasekera (she/her): a+c/a+b+c+d and b+d/A+b+c+d 01:55:11 Maddy Griffith: A+C/A+B+C+D and B+D/A+B+C+D 01:55:15 Olufunke Fasawe: A+c/A+B+C+D 02:01:12 Olufunke Fasawe: So, are we to always report the 4 different conditions for conditional distribution questions? 02:03:21 Silvana Larrea: Thanks Kelsey! Good luck everyone in their midterm! 02:03:34 Nikki Marucut: ^ same to you silvana 02:03:59 Hallie Roth (she/her): Thank you! 02:04:57 Maddy Griffith: Thank you! 02:05:00 Stacy (Seohyun) Ahn: thanksss 02:05:03 Phoenix Ding: Thank you 02:05:12 Julia Hankin: Thank you! :) 02:05:12 Cristal Escamilla: thank you! 02:13:36 Olufunke Fasawe: Thanks Kelsey 02:14:07 Ekua-Yaaba Monkah: yes please 02:14:14 Edythe Glazer: Where are the questions? 02:19:54 Annalisa Watson (she/her): There’s a lab slide deck on the course schedule! 02:24:44 Annalisa Watson (she/her): I have a couple questions from the practice midterm 02:24:56 Hiruni Jayasekera (she/her): ^me too! 02:31:51 Ekua-Yaaba Monkah: zip code would be nominal correct? 02:32:33 Annalisa Watson (she/her): The primary outcomes were back-related dysfunction (Roland Disability score, range: 0 to 23) and symptom bothersomeness (0 to 10 scale). Outcomes were assessed at baseline and after 8, 26 and 52 weeks. 02:36:22 Christine Won: Is this a question from a practice midterm? 02:36:27 Hiruni Jayasekera (she/her): yes 02:36:34 Christine Won: Which midterm? 02:41:19 Annalisa Watson (she/her): Yeah I think the answer is wrong 02:45:20 Annalisa Watson (she/her): There’s a data sheet in another link on the class website 02:45:40 Hiruni Jayasekera (she/her): https://ph142-ucb.github.io/fa21/src/resources/fa18-mt1-supp.pdf 02:49:49 Edythe Glazer: select(- variable)? 02:52:01 Jessica Wright: case_rate? 02:54:27 Annalisa Watson (she/her): Thank you Kelsey!