Stats M12: Intro to Methods in Statistics for Social Sciences

Stats M12: Intro to Methods in Statistics for Social Sciences

Spring 2001

Instructor: Robert Gould

Office: MS 6151

Phone: 310-206-3381

email: rgould@stat.ucla.edu

web page: www.stat.ucla.edu/~rgould

Course Web Page: www.stat.ucla.edu/~rgould/m12s01

Office Hours: M 10-10:50, Thursday 11-11:50, or by appointment (send email to make appointment)

TAs: Yan Xiong, xiongyan75@hotmail.com

Monica Almeida, marruda@ucla.edu

Do Not Throw This Away

You will need to refer to it for the rest of the quarter.

``The time may not be very remote when it will be understood that for

complete initiation as an efficient citizen of one of the new great

complex world wide states that are now developing, it is as necessary to be

able to compute, to think in averages and maxima and minima, as it is now

to be able to read and to write."

H.G. Wells, Mankind in the Making

Purpose of Course

Contrary to popular opinion, Statistics is not a stagnant collection of numbers invoked to buttress one viewpoint or the other. Instead, Statistics is a collection of tools and methods that are used to collect, interpret, and analyze things that can be measured. (And some branches of Statistics worry about what things, exactly, can and can't be measured.) In both our professional and personal lives we must often have an opinion on a subject, and a sound opinion is based on a sound analysis of the evidence. Statistics is a method for making sound opinions based on evidence.

This course highlights Statistical methods favored in the Social Sciences, but we will also spend much time on some fundamental issues and concepts in Statistics. These include variability, uncertainty (which is maybe the same thing as variability), causal inference, reliability, presentation of results, inference.

"Social Sciences" here are broadly defined. This class contains students interested in geography, social science, and anthropology. I hope to use examples drawn from these disciplines because doing so does two purposes: it teaches you about how statistical tools are applied to problems in the discipline you're interested in, and it helps build your interest in Statistics if you can get involved in the problems we discuss. To do this, I need your help: please bring to my attention any data sets or data-related problems you are interested in discussing or exploring. As the course progresses, you'll have a better idea of what sort of examples we could discuss.

Structure

As you've no doubt noticed, this is a 5-unit course. The course has three "structured" components:

1) Lecture

Meets three times a week, MWF 1pm. I will present the material I think most important, and give you a chance to ask questions. I will ask some questions of my own.

2) Discussion Section

Meets once a week in smaller groups with your TA. This is your chance to ask questions in a slightly more personal setting.

3) Computer Lab

Meets once in a week in the Statistics computer lab in Boelter 9413 (on the roof on the eastern side of the building.) There will be weekly assignments to help you practice data analysis while you learn Stata. The assignments will require some time outside of the Lab meeting. These sessions will be supervised by your TA.

There are also unstructured components that are vital to your success:

1) Study: you should read the appropriate sections of the book before coming to class. And you should do the homework in small chunks throughout the week. This class is structured so that each lesson builds upon the previous one. If you don't understand one lesson, you will have a harder time understanding the next. Keep up with the reading and homework to help prevent this.

2) Office Hours: Both I and the TAs will hold weekly office hours. Office hours are very important to me because they give me the only real feedback that I can use to adjust the pace and content of the course. They should be a regular part of your study diet, too, because they'll help you focus your study. The secret to having a productive office-hours meeting: do a little studying before hand so that you know what it is you don't know. The better you are able to express your confusion, the better we'll be able to tackle it.

3) Group study: Many people find it very productive to do homework in groups. By all means you should do so if you like. We can organize study groups through our homepage. Email me if you are interested.

Grading Policy

Your grade will be based on

• Two midterms (30% + 30% or 0%)

• Weekly Labs (20%)

• Final Exam (60% or 0%)

• Quizzes (20 %)

• Homework Bonus (explained below)

We'll drop the lowest score in each category. No make-ups are late work allowed at all for any reason. Your final grade will be based on (both) midterms OR the final, but not both.

Quizzes

Quizzes will be given each week in section, beginning the second week of class. The lowest score will be dropped. Collectively they count for 20% of your grade.

Midterms

The midterms are on Monday, April 30 (5^th week) and Friday, June 1 (9^th week). Make-up exams are not allowed.

The midterm emphasizes concepts discussed in class and covered in the homework. You will see problems that do not look exactly like the homework problems, although the homework is the best preparation for the midterms.

Final

The Final Exam is on Thursday, June 15, 11:30am - 2:30.

Students often ask to take the exam early. These requests will not be granted. If you must leave before the final exam date, then your score will be based on the midterms.

Labs

You will meet with your TA once a week in Boelter 9413. These labs are designed to give you practice with the concepts discussed in class. You will have the opportunity to work with a professional-level statistical software package to analyze data. As the quarter progresses, the labs will be increasingly "open-ended". By the end, you should be writing well-reasoned, organized reports and expositions of your analyses.

You MUST have a Bruin On-line account. These are available for free through the Office of Academic Computing on the 4^th floor of the Math Science Building.

Software: We will use and teach Stata. It is available on several computer labs throughout campus, and you can also purchase it for your home use if you like. (This is by no means required.)

On-campus work: There are various places on-campus where you can work on a computer.

• Statistics Computing Lab, Boelter 9413. (iMacs) Hours and information at http://www.stat.ucla.edu/undergraduate/icl/

• CLICC: Powell Library (Macs and PCs) http://www.clicc.ucla.edu/

• OAC Health Sciences: CHS A1-242 (Macs and PCs) http://www.oac.ucla.edu/labs

• OAC Commons Lab: MS 4328 (Macs and PCs) http://www.oac.ucla.edu/labs

Off-Campus: You can buy Stata for use on your home PC or Mac, if you like. You need to order directly from Stata, and it will be delivered within two days to UCLA. Follow the directions available at http://www.oac.ucla.edu/training/stata/how_to_get_stata.htm. You can also get pricing information here.

Due: Labs are due one week after they are assigned. Turn them in at the next lab session.

Homework

Homework will be assigned weekly, and collected on Friday. Homework will be graded 0 (not turned in or many missing problems), 1 (mostly complete), and 2 (complete). If you have a total of 17 or more, then your letter grade will be raised a half-grade (e.g. B to B+). If you have a 7 or less, it will be lowered a half-grade. Any other score results in no change. The one exception to this is for students whose score on quizzes and exams gives them an A; an A+ will be given only for exemplary achievement. NOTE: This means you must turn in ALL of the HWs to get the ``bonus''. The bonus is very generous, and so I'm very strict about turning in homework on time. There will be NO late homework accepted. If you forget to turn homework in, then you can rest assured that it will have no affect on your grade, but also you won't get the bonus.

Late homework will not be accepted for ANY reason. Late HW is any HW turned in after 2pm Friday.

I can't guarantee that all of your homeworks will be returned in time for you to study for the midterm or final. If you are concerned about this, I recommend you make copies before turning them in.

Course grade

Grades are curved so that the median score in the class is a B-. Midterms are

not given a letter grade, and so I can't tell you what your letter grade is based on your midterm alone. However, if you need to know for a letter of recommendation or application of some sort, I'll make an estimate. Just to emphasize this point: your grade is based on your total points, and so looking at any single exam (or quiz) will not be that helpful in helping you predict your grade. The best thing is to look at your total points with respect to the median total points of the class.

Your letter grade will be made available on the last day of class. You may use this to decide whether to take the final.

Regrades

You might have a complaint or concern about the grading of your midterm. If you think there is something that might affect your grade on the midterm, then you will want to submit it for a regrade. The rules are as follows: you must submit your requests in writing by the NEXT lecture after the midterm is returned. This requires that you attend lecture so that you can pick your midterm up. Your written request should clearly point out which items are

being challenged, and why. (Do not just say "Number 3" but instead something like "Number 3 was marked incorrect, but I don't know why because...."

Please don't hesitate to ask questions or challenge the grading AFTER the deadline; but don't expect the grade to be changed. It's important that you understand the material, so please feel free to discuss it at any time.

Materials

The required text is Introduction to the Statistics: The craft of data collection, description, and inference, 2^nd ed. Monrad, et. al.

Optional: Stata (only if you have a home PC or Mac), Bruin On-line account (to dial-in from home)

Activ-Stats (a software tutorial that makes a very good reference source.) This is available at the Bookstore.

HOMEPAGE

You are responsible for information that appears on the homepage for this course. Check the homepage as your first stop for answering questions, particularly questions about this course. The address is http://www.stat.ucla.edu/~rgould/m12s01

In particular, be sure to check the "announcements" section, which will be updated as needed. It should be your first stop.

One Final Word

Never hesitate to ask a question, either in lecture, section, during office hours, or over the homepage. Contrary to popular opinion, there are stupid questions, but I promise to never let you know if you ask one, and promise that there's an (intangible) reward even for stupid questions. So feel free to ask, because the worst that can happen is you can have it answered.