Email from Students
Jump to the end of this file to see the most recent message.
> I am in your STA220 class, and as you know there is a > quiz this coming Wednesday. We are allowed open book > but closed notes... what does this mean exactly. > Does this mean that notes which we take in class are > not allowed to be used to help answer questions on the > quiz? You may not look at class notes during the quiz. You may look at your textbook. > > If you could get back to me and give me any > suggestions as to how best prepare myself that would > be greatly appreciated! Okay! Here are some suggestions. Look at Assignment 1. Read the suggested sections in the text. The assignment lists concepts for which you are responsible. Make sure you understand them all. Don't worry about stuff in the text that is not in the homework and not in the list of concepts. Do the recommended homework problems. For odd-numbered problems, check the answers in the back of the book. If you have trouble with a problem, make a note of it and go to tutorial and ask Christine for help. Or come to Christine's or my office hours. Treat all in-class examples as homework problems with solutions. Trying to express each in-class example as a question is a useful thing to do; it will help you recognize the ones that are asked (and some will be). Take a look at last summer's Quiz One. This time, Quiz One will be somewhat harder, but it is still going to have something in common with last time. The most important thing is the homework problems, followed by the in-class examples. ---------------------------------------------------------------------------- > U said we could write what ever we wanted, as long as its in the text, > and we use those written notes in the text on the tests right? so is it > OK to make a tiny formula sheet, at the end of the book, as long as its > in the book? I really find it a waste of time flipping through pages, in > the book wording its harder to understand tooo. Its easier to write it my > way so that i can use it as recall cues and understand better. > Okay! I will post this and inform the class. ---------------------------------------------------------------------------- > > If I may ask, will you make sure that atleast one game and or gamling > question (toss, coin, cards) is on the test? My wordings imply that I > want them to be on the exam, but in all honestly, I suck at all sorts of > gambling problems, like a flush, who cares, is it really that important? > unless we all will be gamblers. > Well, honestly I have no interest in gambling games either. On the other hand, in this simple setting, lots of the basic principles are easier to illustrate. So, for example, if I say that having a girl or a boy in five births is like tossing a fair coin 5 times and observing heads or tails, I'm suggesting a model that's clear and simple. This is the justification for all those gambling examples, and there will be more than a few on the test. ---------------------------------------------------------------------------- > I do not understand how to get the answer when looking at table 2 for > questions where x is greater or equal to a number. The complement of "x is greater than or equal to 4" is "x is less than or equal to 3" . So P(X >= 4) = 1 - P(X <= 3). ---------------------------------------------------------------------------- > > For the midterm, do we have to know how the various formulas are constructed? > It is essential to be able to apply the formulas. If where they came from was explained in text or lecture, knowing that could be helpful too -- because a question could ask you to figure out something similar. ---------------------------------------------------------------------------- > I was going through my notes/hmwk a couple of times, and I have some > questions. > > 1. Notice that in all these examples, the random variables are > quantitative. But discrete random variables can be qualitative too. For > example, you could randomly sample a person from the population of > Toronto, and let x designate residence type: 1 = apartment, 2 = > single-family dwelling, 3 = homeless and 4 = other. Can you give an > example of a random variable that is both continuous and qualitative? > > I would think another example could be the weight in pounds of the food > time bough in a supermarket, b/c it is a) qualitative(types of food) and > b) continous. ( 0 -50, 50-100) etc. > > Am I right in saying this? The way I see it, weight of potatos is one variable, weight of shrimp is another variable. Both variables are quantitative and continuous. I cannot think of ANY useful examples of a variable that is both qualitative (unordered categories) and continuous. And even the useless examples I can think of are weird mathematial constructs. > > 2. Also > > Can you let m know if I am on the right track with these statements > > Dep events are always mutually exclusive - NO I agree. > > Mutually excl events are always depedent. Yes ( the rule stated in the > text) Yes, given the book's definition of independence. If you take as the definition P(A and B) = P(A) P(B), then there is a hairsplitting, trivial exception. If A and B are mututally exclusive so that P(A and B) = 0, and also P(A)=0, then my definition of independence (which is the traditional one) is satisfied. But I would never make a test or exam question based on this in STA220. > > /> I strongly disagree with this statement. ;-) > > Independent events are always mutually exclusive ( no, the rule is not > reversible) I agree. ---------------------------------------------------------------------------- > This question refers to yesterdays exam. It goes something like this: > > A lottery ticket is sold for a dollar. > 1000 tickets are sold. > you buy ten tickets > what is the E(x) of this game. > > So I followed the logic > 1. P(winning) = 1/1000. Therefore P(losing) = 999/1000 > > x p(x) > -10 999/1000 > +990 1/1000 > > > > E(x) = (-10*999/1000) + (990*1/1000) > = -9.00 > > This answer was not in the multiple choice, Am I wrong? is the P(winning) > 10/1000? so the P(losing) should be (1-10/1000).....hmmmmmmmmmmmmmmmm....... > > I am lost, oh god I am so lost > I am lost too, but that is beside the point. If you buy 10 tickets, your chance of winning is 10/1000, not 1/1000. The event of interest consists of 10 sample points out of 1,000. All 1,000 sample points are equally likely. ---------------------------------------------------------------------------- > Hi, I went from a 92 on quiz one to a 64 on the midterm. I wasn't expecting > the midterm to be so hard based on last years quiz marks. I said the quiz was a cream puff, and it was. I said the midterm was going to be harder, and it was. Last year I skipped probability entirely, and it was a mistake. Average marks on last year's quizzes, over different material, have no bearing on the midterm this year. At the same time, it is certainly natural for you to expect the future to be like the past, and I did not intend to mislead you. The website from last year had the students' quiz marks so they could check and make sure our records were correct. I provided this year's class with the link to last year's website because I thought the quiz questions and solutions might be helpful. I was not thinking that you'd be checking out their quiz averages, because the information was irrelevant. But of course you did. I was thinking like a professor, not like a student. My goal is to do both, so I would say I made a mistake here. > Why did you make it > so much harder this year? The average this year is 64, while it seemed to be > in the 80's last year. 80% and up is an A. At the Universty of Toronto, it is not customary for the average mark to be A. In fact, the university has explicit guidelines for the maximum and minimum percentage of As and the maximum percentage of Fs. I do not have to follow these guidelines, but I try to. Basically, I blew it last year. This year I am doing much better (by this citerion), at least so far. If the marks in STA220 this summer were based ONLY on the midterm, I would be slightly exceeding the guidelines for both the maximum percentage of A marks and the maximum percentage of F marks. To me, this means that the difficulty midterm was PERFECT, by U of T standards. > Furthermore, the weight of this test is going to kill > me no matter how well I do on the final. I don't understand why 2 chapters > should carry so much weight in the course. Probability is important; that's a good enough reason. And originally, the midterm was supposed to cover chapters 3-6 -- that's all of probability instead of just the core part. But I needed to make it earlier so people who were shocked by the midterm would have a chance to drop. If the test had covered more material it would have been harder, and your score would probably have been lower, though of course not necessarily. So as I see it, the fact that the test only covered 2 chapters was not optimal, but from the perspective of a student interested in getting the highest possible mark, it was a gift. > Despite the difficulty of the > test, I DO understand the material, but I feel I just "choked" on the > midterm. I'm sorry about this, but measurement error is always possible. It's part of the whole system where we give tests that imperfectly reflect people's knowledge. I believe that the midterm did a pretty good job of diagnosing most students' understanding of the material, though maybe not yours. > In light of the low class average (vs. last yr.) and the > nonproportional weight of chapters 3&4, is there anyway you could provide a > make-up test for those who didn't do so well. Or could you provide some > future quizes on future matterial to offset this midterm?? There is no way. The course is perfectly on target. > I really need to > do well in this course, because my average is above 80 and this course will > bring it down heavily. Thanks for you time!! You are carrying a 69.6 into the final exam. This limits your highest possible mark in the course to 84.8, which would be rounded to an 85. This is an A. Things definitely could be worse. Good luck (and good skill) on the final exam, because it's a mixture of both! ----------------------------------------------------------------------------- On Wed, 15 Jun 2005, Christine Lim wrote: > The book says: > n large -> z test > n small -> normal population -> sigma known -> z > n small -> normal population -> sigma unknown -> t > > You said (sigma is always unknown): > n large -> population normal -> t (or z) > n large -> population unknown or not normal -> t > n small -> population normal -> t > > Am I right? Yes. However, after I talked with you I changed my mind, because the textbook is very well organized. So now I am following their pattern, except that I have promised the students that on the final exam, sigma will never be known. And because t -> z as df -> infinity, for large samples you can use either one. So what I said in class was n large -> z test or t-test; it does not matter normal population -> t is preferable but n is large and if you use z, okay n small, non-normal -> stop > > Say n=11680, and they should be using t. Should they > use df=120 or df=infinity? I know you said you will > avoid quetions like this in the final, but they appear > in the homework problems. What do you want me to tell > them? I would say use df = infinity -- that is, z > > What if n=100. df=60 or df=120? > > n=200, df=120 or df=infinity? > > How do you want them to decide? The book's intent is clearly that they should be using z (infinity). To me, it seems that the book in general promotes rigid thinking. I assume this is so at least everything will be clear. I guess the most convenient course for the students to follow is to just use the bottom row. And this way they will get the answer in the back of the book, which is always reassuring. But on the final, the wrong answers will be far enough away that they will be okay using either t or z for large samples, interpolating in the t-table or not. ----------------------------------------------------------------------------- > With 4 assignments and a cumulative 2hr exam, up to what chapter will we be > responsible for with only two classes left? > Well, in the summer each class is equivalent to one week, so in that sense we have 2 weeks left. We will definitely do Chapter 9, not including confidence intervals for differences between means, and not including anything for differences between variances. The question is, how much, if any, of Chapter 10 will we cover? This should be clear by the end of the class Monday. In spite of what the syllabus says, we likely will not do any of Chapter 13. The final will be comprehensive, but with HEAVY emphasis on material after the midterm. More details later. ----------------------------------------------------------------------------- > My comment is about the final exam. I am concerned that with all the material > covered many people may not be able to decipher what formula to use for what > question and what chapter the question is based on. Normally I would expect > this sort of dilemma in a final exam, but due to the format of the textbook > and assignments I don't think we have had enough practice doing review-type > questions where the general topic of the question is not given. All the > questions that we have been assigned are located in subsections of the > textbook, making them easier because the general topic of the question is > known. This is why I think it is unfair to put us on the spot in a test > situation asking us review-type questions when we have had no practice with > them in the past. What I would suggest is if you would include the section > number in the exam question to keep everyone on track. Thanks for your time! > Well, I agree that selecting the right technique is a significant part of what you will need to do on the final exam. It is also a very significant part of what you will need to do if you ever need to use this stuff in another course or after you graduate. There are no Chapter and section numbers on the job -- just data and questions about the data. A student who has been listening in lecture and/or reading the text with attention, and trying to UNDERSTAND the methods should be able to select the right method most of the time or all the time. In fact to me, ability to select the appropriate method is the strongest indication that a student knows what is going on. So I'm definitely not going to label the questions as you suggest. But also it's very good that you have noticed that selection is an issue. Between now and the final you should look over the homework problems, and scan them for cues about what you should do -- cues other than what section they're in. ----------------------------------------------------------------------------- > > I noticed question 9.9 (pg.448) contains material other than what we > were told to focus on for section 9.1 and 9.2 such as pooled estimates > and confidence interval stuff. Should we ignor this question. Are > there any other questions we should ignor? > Sorry about that. I intended to just assign parts a and b of Question 9.9. The assignment is now corrected. Thanks! ----------------------------------------------------------------------------- > I'm having difficulty getting the correct answer for some of the homework > problems in assignment 8. There as follows: > -9.11--> the answer in the back of the book says fail to 'reject Ho at > alpha .10'. I dont understand how you know to use .10? Why not .05 or .01? Well, .10, .05 and .01 are the most common significance levels. Since you can't reject at 0.10, you can't reject at .05 or .01 either, so I guess that they were trying to tell you as much as possible with one statement. On the other hand, the .10 level was NOT indicated in the problem, as you point out, so there is no way you (or I) would know to use that particular value of alpha. But, if "Reject H0 at 0.10" were one of the alternatives in a multiple choice, we'd both pick it. > > -9.27--> I found out the correct test statistic which is t=-.46, but where in > the question does it tell you the direction? Is it <,> or not equal? Based > on the answer in the text it says, 'fail to reject Ho at t=-.46',..which it > seems to imply that the rejection region would be on the right (meaning >)in > order to fail to reject Ho, but how do you know that? Am i missing something? The question says to "compare the mean DIQ" of the two groups. This is non-directional. The answer in the text does NOT say "fail to reject Ho at t=-.46." It says "t=-.46,fail to reject Ho." It gave you the value of the test statistic, and correctly stated that Ho was not rejected. It said nothing about the alternative. > > -9.63-->In order to do this question you need a n1,n2 x1 and x2 right? > So, I assume that my two n's are both 80 right? But are my x's .10 and > .05?, because I tried doing that and im not getting the correct > answer. Can you please tell me which values are what? > n1 = n2 = 80; p-hat1 = .10; p-hat2 = .05. You do not need x1 and x2 to calculate the Z statistic because they give you the p-hat values directly, but anyway since p-hat = x/n, you have x1=8 and x2=4. > -9.67-->Finally, in this question once again I'm confused about what are my n > and x values. I think 21 and 7 are my x's. Right. > But which ones are my n's because > I have tried a couple of different ways and i still dont get the correct > answer. n1=38, n2=78, p-hat1 = 0.5526316, p-hat2 = 0.08974359, p-hat = 0.2413793. The book's answer of z = 5.47 is correct. ----------------------------------------------------------------------------- > > I believe that Professor Brunner mentioned in > > lecture on Wednesday that the > > exam is heavily weighted on Chapter 5 and i just > > wanted to confirm with you > > as to whether I heard correctly. Thank you for > > your help. Not true! It's heavily weighted on the material from chapter 5 ONWARD -- meaning Chapters 5,6,7,8,9,and part of 10. Be guided by the homework assignments. ----------------------------------------------------------------------------- > > I'm slightly worried about this type of questoin for the exam, because I > really have no idea what to do with all the information. > > The question gives: F= 7.70, numerator df = 3, denominator df = 56, > p-value < .001. Alpha = .05. > > It asks to conduct a significance test, but I'm not sure how to do it? > Any help would be appreciated. Essentially, the test has been done for you already. If p < alpha, reject H0. That's it. ----------------------------------------------------------------------------- > I was wondering, since the exam is open book, are we allowed to bring > in the solutions manual for the textbook aswell? Please say YES! > Nope! ----------------------------------------------------------------------------- > I a little worry about the time we have to finish the exam. I am little > slow writing these exams, and I often use the extra time normally given to > finish it (e.g. 2hour exam, 3hours given). It may not be the case for this > exam...but, if so (and if there are only few people unfinished), can you > give us that extra times finishing our exam in your office or in a reserved > room. > No, I'm sorry, that's not going to happen. To finish the exam, you will need to work hard and fast. I suggest being aware of the clock, and keeping the pace of 3 minutes average per question -- or faster, if you want to re-check any of your work. The theory here is that if you really know this stuff, then you know what to do right away with very little pause for reflection and almost never looking at the book, except for the formulas, which you can find very fast because you have read the text more than once and you are very familiar with it. Or you have made a good formula sheet, or both. Tomorrow, you will probably know the material from this course better than you ever will in the future. If you are shaky and slow tomorrow, your chances of remembering much after a few months are pretty remote, and we have wasted our time. To qualify this a bit, if you have a specific, medically diagnosed disability that prevents you from working rapidly, then you can make special arrangements throught the Accessability Office. If you are in this situation then it is not too late, even though the exam is tomorrow. Good luck! Jerry