Email from Students
> A question that I want to ask is that will you be going by the book, with > the textbook being our primary source of information or do we need to > concentrate more on your lecture notes. I am following the topics in the text book, mostly in order but skipping some things and adding a little bit of stuff. Your lecture notes will tell you what I am adding and what I am skipping. I am following the vocabulary and notation of the text like a slave. In this sense, the text is primary and lecture is secondary. Unless I explicitly say, "I'm not going to lecture this; read it yourself," you will not be responsible for anything in the textbook that I do not also say in class. In this sense, lecture is primary and the text is secondary. On the other hand, the homework problems are all going to be from the text, and I am also going to base some quiz and final exam problems directly on examples in the text. This is an incentive to read the text. What's REALLY primary are the homework problems. They tell you what you must be able to do. > > Secondly, Will the first quiz only have definitions on it and does it > cover chapter 1 and 2 entirely There will be some definitions, but also some calculations. Certain sections in Chapter 2 are skipped. Please see the course web page for details. > When i emailed you a long time ago, you mentioned that there was parking > at erindale park & it's free... how exactly do i get there? i usually take > the bus to school but i'm driving this wednesday (i'm coming from hamilton > QEW) Coming North on Mississauga Road, turn right on Dundas. The first left you can make is the park entrance. After a short drive the parking lot is off to your right, and parking is free. Walk across the foot bridge, and take the path uphill to the left. When you see a chain link fence a couple meters above you on the right side, that's the back of the parking lot. Climb up the rocks, go aound the fence and there you are. > What are 'Likelihood ratio bias', 'Stochastic bias' > and 'Parthnogenic hydrophobia'? Were they introduced > in class? Possibly in your travels you have observed that some people are attracted to fancy technical language that sounds impressive but means nothing. 'Likelihood ratio bias' and 'Stochastic bias' fall into this category. I made them up. Parthenogenic refers to reproduction by just the female, without any male contribution. I believe certain lizards and amphibians can do this if there are no males around. Hydrophobia refers to a pathological fear of water. It is another term for the disease rabies, and it is also used by psychiatrists sometimes. When you put the two words together, you get ... well, nonsense. But I thought it sounded good. > On the website, it says: > > Read the Probability handout, and do problems 3.1, 3.3, 3.4, 3.10, 3.11, > 3.13, 4.6, 4.7, 4.8, 4.9, 4.10, 4.11, 4.12, 4.13, 4.17. --> Does this > mean that we can do the problems with just the aid of the handout, thus > without having to read the entire chapter..? Or do we have to read ALL of > chaprter 3 and 4? You don't have to read anything in Ch3 and 4 except pages 174-176 in Chapter 4, as indicated below. The handout should be enough to let you do the problems on probability and discrete random variables. We save a lot of time and unnecessary effort this way. > > Now read the discussion of discrete and continuous random variables on > pages 174-176 in Chapter 4. Then read Chapter 5's Introduction and > Section 5.1. Skip Section 5.2, but read 5.3 carefully. Do problems 5.16, > 5.17, 5.18, 5.19, 5.21, 5.22, 5.23, 5.24, 5.28, 5.29, 5.32, 5.33, 5.34, > 5.35, 5.36, 5.37. --> Again, do we read the entire chapter 4, or just the > pages 174-176? And same for chapter 5... do have to read the entire > chapter or just the Intro, section 5.1, and 5.3? > Just the Intro, section 5.1, and 5.3. > I guess what Im asking is, do we read ONLY what the website says to read, > or do we still have to read the entire chapters for each of 3,4, and 5... > Because they're really long. Just read what the assignment says to read. > > Secondly, what is quiz two going to be like? Again all multiple > choice...? Definitions...? Formulas and calculations..? It will not be multiple choice this time. Most, but not all of the answers to Quiz 2 are probabilities -- that is, numbers between zero and one. > Just to confirm, for our second quiz we do not need to bring any computer > printouts of our homework right (MINITAB). I dont have the CD yet and am > waiting for it. > Confirmed. No Minitab until everyone has had a chance to borrow a CD. > Someone asked this question... > >Secondly, what is quiz two going to be like? Again all multiple > >choice...? Definitions...? Formulas and calculations..? > > In the answer it doesn't say explicitly that definition will not be on the > test. So, we assume that definitions could come in the quiz this time around > too? I am not trying to be tricky here. You will not be directly asked for any definitions. But if you don't know what the terminology and notation means, I would say you would be in serious trouble. Again, most of the answers are completely numerical this time, but not quite all of them. By notation, I mean things like P(x<2). > How would you find the probability for standard normal random > variable z if > P(-1.645< x <1.645). The entry for 1.64 (in Table 4) is .4495 and the > entry for 1.65 is .4505. Would we have to take a value that is between > .4495 and .4505 to answer the question? Thanks! > Yes. Since 1.645 is midway between 1.64 and 1.65, the probability you want is midway between .4495 and .4505. That's 0.45. So P(-1.645< x <1.645) = .90. This is the easiest case of interpolation in the table -- midpoint to midpoint. In general, if you go a certain proportion of the distance between two x values, you need to go that same proportion of the distance between the corresponding probabilities. I could give you a formula but it's not worth the trouble. Here's the rule. On quizzes and the final exam, the ONLY interpolation you would ever need to do is this one, since 90% and 10% come up so often. I guess this happens because we have 10 fingers. Anyway, on quizzes and the exam, if any other number you need is between two numbers in a table, just take the closer one and you will get full marks. This happens once on Quiz 3, I believe. > Is it possible for you to post the answers for the even-numbered > questions so that we can check our homework as we are progressing. I'm > not asking for all the calculations but the bottom line final answers > for checking purposes. I find it helpful to check my answers to see > if I am doing the questions properly before I continue on with the > rest of the homework. > Thanks! I'm sorry, I don't have the answers to the even numbered questions. When I assign them, I just think about how it would be solved, and pick the ones that would be good practice. Then, if someone asks in office hours, we work out the numerical answer together. Generally, you can check your progress with the odd numbered questions, and ask about the others in tutorial or office hours. But there must be a solution set for the even numbered questions. Maybe one of the other professors has a copy. I'll ask around, and put it on reserve if I can get my hands on it. > In section 8.1 in the textbook, pg 328 it gives an example about > building specifications in a city require that the average strength of > sewer pipe must be more than 2,400 pounds. > >From the way we discussed it in class I would have said that the null > hypothesis is Ho: u >or = 2,400. Meaning that they DO meet the > specification. The alternate hypothesis would be that Ha: u<2,400. > Meaning that they DO NOT meet the specifications. The book gives the > answer from the point of view of the city conducting the test. They say > the null is Ho:u2,400. > > Can you try to clarify this in an email? I know I'm asking for a lot. > It's a good question. My reply is too late for Quiz 4, but it may be useful for Quiz 5 and the final. Your question illustrates that in choosing null and alternative hypotheses for a one-tailed test, the direction you choose will sometimes depend on the way you are thinking about the data. One clear rule is that the research hypothesis is the alternative hypothesis. In the example, after a bit of talk, we find the following statement in italics: "we want to decide whether the mean breaking stength of the pipe exceeds 2,400 pounds per linear foot." This implies Ha: mu>2400, and consequently H0: mu <= 2400. This approach is an efficient way to get the right answer on a quiz or exam. But also it is consistent with something mentioned in lecture -- that is, deciding the alternative implies ACTION. Here, the action is to order some of that pipe, and the only way a responsible administrator would do so is if she is convinced it exceeds specifications. On the other hand, if the pipe were already in the ground, ACTION would be to dig it up, and this would be only justified if you could reasonably conclude that it did not meet specifications. Then, we would have Ha: mu < 2400. I hope this makes sense and you agree. But if it seems like hair splitting, remember that you can always tell the alternative hypothesis from the phrasing of the question. > I just had a quick question about the computer assignment. When you are > asking asking for the rejection region in question 4a and ask to answer > all the standard questions. Would you like us to answer only up to > question 3 or would you like us to continue. Because the next question > (4b) is with respect to question 4 of the standard questions. > Yes, they usually overlap. Don't bother with the book's questions unless they are asking something in addition to my "standard questions." > > In the text, page 350, example 8.5... I cant seem to find the rejection > region, and the t-value of -3.00 using minitab. Also, I don't know how to > set up my H0 and Ha to appear in the form on minitab: > Minitab is for calculations. And, you can decide whether to reject H0 based on the p-value. > H0 <= 20 > > Ha > 20 > > By looking at the SAS Analysis of 10 Emission Levels on page 350, I get > the N, Mean, Std. Dev, and Std. Error... all the upper part, but I cant > get my sheet to look like any of the bottom stuff.. the stuff underneath > the "Hypothesis Test" heading on pg. 350. You don't need to. Don't worry. All you need from Minitab is numbers. Format does not matter. > In question 5, we are having problems > distinguishing what the null and alternative hypotheses are. We thought the > null hypothesis was p = .5, but the question is worded so that p=.5 seems to $ > the research hypthesis (alternative). We both know that the equality needs t$ > be in the null hypothesis, but we are confused. We are also second guessing > our answers to question 4 now. > Ouch. You have a good point here. I did not phrase this well. My intention was for you to test H0: p = 0.50. > Hi Jerry! I'm a bit confused about what to put on my minitab output > which is to be handed in. I know that we are to do questions 1-9 on > minitab however, for question 4 and 5, we are to answer all standard > questions. Does this mean we have to state the null, alternative > hypotheses (in words) etc., on our output? No, sorry about this confusion. Just use Minitab to get numbers. But think about the standard questions for any word problem; you could write the answers on a separate sheet, not to be handed in. > Also, is question 6 telling > us to use minitab to answer questions 4 and 5 (a,b,c)? Thank-you! > 4a 4b 5a 5b > I was just wondering if it would be alright if I organize my minitab output > into a word document? Thank you. > Fine, good idea. > I have a question about the minitab homework for stats220. For question 6, > you said we are being asked for 4 items and I am not quite sure what the > fourth one is. I know you ask for confidence intervals, p-values and the > test statistic. Are you asking for both the t and z-test statistic making > four, or something else? > Two tests and two confidence intervals.