SOCIOLOGY                                                                         FACULTY OF ARTS

SCHOOL OF SOCIAL SCIENCES                                     

INTRODUCTION TO QUANTITATIVE RESEARCH METHODS

SOCY 2038

SECOND SEMESTER 2005

COURSE OUTLINE

Lectures:

 Tuesday 12pm -1pm, Haydon-Allen G052

Wednesday 12pm -1pm, Haydon-Allen G052

Lecturer:

Dr. Joanna Sikora

Office: Haydon-Allen 2151

Phone: 61254574

Second Semester 2005

Australian sociology does not have as strong a tradition of quantitative research as sociology in other countries, e.g. the USA, but a large share of sociological literature published in Australia and about Australia is based on quantitative methodology. To be able to understand this literature, its findings and limitations, it is necessary to know the basic concepts used by quantitative sociologists.

Quantitative methodology emphasizes the ability to use sociological theory to derive testable hypotheses, a skill valuable not only in academic research, but also in policy development and evaluation.

The knowledge of computer software for analyzing survey data is a desirable labor market skill. Many Arts graduates are reluctant to engage with numerical data and hence those who combine the strengths of Arts education with basic statistical and computing numeracy often are preferred by employers. Learning SPSS -- the most widely used Software Package for Social Sciences -- may be useful in your future research or employment. Even if your employer uses a different statistical package, the principles of operation of SAS, STATA etc are similar so mastering one of them is a good introduction to others.

Gaining practical experience with data analysis lays foundations not only for future independent research, but also for developing skills in critical analysis of survey research conducted and publicized by others. As many of us had only limited experience with mathematics we will learn mostly by hands-on exercises. We will conduct our own analyses to find out what Australians think about premarital sex, desirable patterns of employment for women with children, policy aimed at reducing inequality, the role of trade unions and genetic engineering -- the use of fetal tissue in medical treatment. This will provide a comprehensive overview of factors affecting the outcome of survey research.

Quantitative Research Methods is a second/third year 6 credit course.

The course is designed to introduce students to basic research methods and techniques used in social sciences, with particular emphasis on applications of statistics in sociology. It involves both research design and basic statistical analysis.  Its contents can be divided up roughly into three parts:

1) philosophical foundations of quantitative methodology, 

2) applications of descriptive statistics in sociology and

3) inferential quantitative analytical techniques.

These areas will be studied interdependently throughout the semester through lectures, discussions and hands-on exercises. As many students find the third part of the course different to other courses and at times "out of their comfort zone", we will devote a substantial amount of time to data analysis and building up our confidence in interrogating numerical data. The goal is to not only to learn basic statistical concepts, and to generate output in statistical software packages but also to be able to structure, describe and interpret quantitative analyses. The emphasis is on sociological applications of quantitative research, but the methods and techniques discussed in this course are used outside sociology, in psychology, economics, policy analysis, market research and journalism.

Important

While no mathematical background is required for this course, most people will find it more demanding than other sociology units as it incorporates material which is not typical for Arts courses. Even basic applications of quantitative methodology require a sound practical understanding of fundamental concepts which can be achieved only by completing a certain number of exercises both in class and out of class. This is why there are 4 contact hours in most weeks of the course and students should not expect to be able to progress in the course material based only on reading and listening to the lectures. As with learning to drive, listening to and reading descriptions of how to do it is rarely enough. Practice is the key to success!

The details of the topics of the course are in the lecture and tutorial outlines. Further material will be provided in our WebCT (http://webct.anu.edu.au/). It is assumed that students have no previous experience with social research methods.

After completing the course students should know the basics of sociological research methods so that they better understand much of the quantitative literature. They should be able to carry out independently sociological research at a basic level.

The course is also designed as a pre-requisite for more advanced units in sociological research, namely Intermediate and Advanced Methods of Social Research which may be offered in the future subject to demand. Students wanting to specialize in research methods are advised to consider these later year sociology units, as well as other research methods units offered in the Social Research Methods major.

Students will develop a good practical understanding of the logic of sampling, statistical inference and other statistical tools used commonly in sociology.

The course will rely on the examination of Australian studies and survey data. There will be periodic assignments and tutorial discussions on topics such as theory construction and theory testing, establishing causality, measurement, quality and types of data, interpreting and constructing tables, basic data analysis and computer experience. We will learn how to construct a data set and conduct analyses in SPSS.

The details of assessment will be discussed in class. There will be 3 assignments and 3 small assessable tasks and a final examination worth at least 30 percent of the course mark. Please note that this course requires systematic work including completing a number of assessable and non-assessable exercises in addition to readings, so planning ahead is strongly recommended. The body of knowledge in the course is strictly cumulative, i.e. most people will not be able to fully understand later topics if they have not completed work on earlier topics.

The due dates for the assignments will be strictly enforced, and three percent per day (of the marks for the particular assignment) may be deducted for assignments presented after the due date. If an assignment due on Friday is handed in on the following Monday the penalty is 9 percent. THERE WILL BE NO EXTENSIONS GRANTED IN RETROSPECT. This means that if you forget to inform me about your medical certificate before an assignment is due, you will be penalized.

 We will discuss the format of the final examination. It may be available as a three-hour sit down examination or as a 24 hour take-home examination.

 Tentative due dates for assignments:

 Note: Assignment 2 will cover a substantial amount of material

While written guidelines for assignments will be provided, the experience has shown that usually supplementary, in-class guidelines are generated in response to students' questions. Lectures will be taped and available in WebCT as audio files and thus missing lectures cannot be considered an excuse for not incorporating these additional guidelines. However, tutorials and labs will be not be taped and hence it is very important that you attend them regularly.

 We will also complete a small number of tutorial or lab tasks during the course. These are "assignments" that do not require much time to complete but are essential for monitoring smooth progress in the course. Usually no written feedback will be provided, unless areas of concern are identified.

Timing of assignment feedback will depend upon the size of the class. I expect to be able to return all assignments within no more than 10 working days from the date of submission.

Normally there will be two one-hour lectures per week, at 12 pm on Tuesdays and Wednesdays in HA G052 (Building 22)

Ten fifty-minute long lab sessions will be held in COPLAND G025. See weekly schedule on p. 13. After 18 July 2005 use http://arts.anu.edu.au/tutorials/ to sign up.

Lab Group 1  Tuesday  2:00:00 PM  3:00:00 PM  COP G025 (Building 24) 

Lab Group 2  Wednesday  1:00:00 PM  2:00:00 PM  COP G025 (Building 24) 

We will have 6 tutorials. Consult the weekly schedule to make sure you know in which weeks they will be held. Note that tutorials in this course involve discussions and completion of exercises rather than discussions only. After 18 July 2005 use http://arts.anu.edu.au/tutorials/ to sign up.

Tutorial Group 1 Tuesday 1:00:00 PM 2:00:00 PM CRISP G015 (Building 26)

Tutorial Group 2 Wednesday 11:00:00 AM 12:00:00 PM COP G039 (Building 24)

1.       A short  reading brick  will be available for purchase from the School office at the beginning of the semester

2.       Norusis, Marija J. 2004. "SPSS 12.0 Guide to Data Analysis", 1/e New Jersey: Prentice Hall ISBN: 0131478869

Note:  Brick readings are from the following sources:

1) Agresti A. and Finlay B. 1997. Statistical Methods for the Social Sciences (3rd Edition). Upper Saddle River, New Jersey:: Prentice Hall. ISBN: 0135265266

2) Babbie. E. 2004.The Practice of Social Research With Infotrac (10th Edition). Thomson-Wadsworth ISBN 0534620280

3) Neumann. W.L.2003. Social Research Methods. Qualitative and Quantitative Approaches. Boston: Pearson Education ISBN 0205353118

4) Salkind Neil J. 2003. “Exploring Research.” Upper Saddle River, New Jersey: Prentice Hall (Fifth Edition). ISBN 0130983527

Therefore if you have access to these books, you do not need to purchase the brick.

Some students may benefit from additional literature. Here are some recommendations.

 1. Statistical Methods for the Social Sciences (3rd Edition)

by Alan Agresti, Barbara Finlay

Hardcover: 643 pages Publisher: Prentice Hall; 3rd edition (March 18, 1997)

ISBN: 0135265266, available in the Co-op bookshop  

I think this is one of the best books around, though some find it a bit dry. Its main advantage is a thorough coverage of most techniques you are likely to come across in quantitative sociology.

 2) The Practice of Social Research With Infotrac (Practice of Social Research) by Earl R. Babbie 10th Edition ISBN 0534620280

Beloved and used by many students around the world, this is by far the most popular introductory social research textbook. In conversational style Babbie introduces qualitative and some quantitative methodology, providing an extensive discussion of conceptual development of surveys and experimental research. The book is accompanied by excellent ancillary resources including an interactive website (http://www.wadsworth.com/cgi-wadsworth/course_products_wp.pl?fid=M2b&product_isbn_issn=0534620280&discipline_number=25)

3) The Cartoon Guide to Statistics

by Larry Gonick, Woollcott Smith , New York: HarperCollins, 1993, ISBN: 0062731025

Unfortunately this title is out of print so it may be available only in second hand bookshops. Consider buying if you come across it. This book has no special social science focus but is witty and unusual so if you enjoy a humorous approach you may find it useful.

I encourage you to come and see me (please send a WebCT email to arrange a time) to discuss, clarify or expand on any issues in the course organization, delivery or material. Often one or two visits will suffice to eradicate confusion and ensure best results in the course.  I cannot give you any real help fifteen minutes before the due assignment time, or before the examination so please schedule your visit well ahead of time. In most cases I am also unable to provide comments on drafts of assignment sent by email a day or two before they are due. Please make sure you obtain comments on early drafts in person and at least three days before the due date.

My preferred mode of contact is WebCT e-mail!

E-mail: "Joanna Sikora" in WebCT

Note: Using the WebCT email will ensure a quick response as I know the messages are genuine. I generally do not open messages,sent to joanna.sikora@anu.edu.au, from unfamiliar email addresses  especially if they contain attachments. To send an email you need to log to http://webct.anu.edu.au/ using your user id and password (we will demonstrate how to do in during the first lecture) and then click Email, Compose Mail Message, Browse and choose my name from the top of the list. 

Regular logging onto your WebCT account to keep up to date with online materials is the course requirement. Minimum one login per week is required, two logins at 3-4 day intervals are recommended. Come and see me if you think you will find it difficult to meet this requirement.

Plagiarism ‘is copying, paraphrasing or summarizing, without acknowledgement, any work of another person with the intention of representing this as the student’s own work’.

Every year cases of plagiarism are identified; consequences may be severe and are always very unpleasant for all involved. Therefore critically examine your drafts with a view to proper referencing and the use of your own wording except for quotations.

For details of the current ANU plagiarism policy see

http://arts.anu.edu.au/student_information/current/rules/plagiarism.asp

Essays which are mostly collections of paraphrased excerpts from the literature will not earn good marks as they lack original input. This is even if the paraphrased quotations are scrupulously referenced, so ensure you execute your own analytical voice in each piece of written work submitted for assessment.

Due August 12

The purpose of this assignment is to demonstrate the role theoretical perspectives play as frameworks for empirical quantitative research. Choose an academic article reporting an empirical quantitative study and summarize it focusing on theoretical underpinnings of data analysis.  Answer the following questions:

1.       What is the main research question addressed in the paper?

2.       What theories does the author /do authors use? How are these theories applied to the investigation of the research problem?

3.       Would Babbie classify your chosen study as primarily (1) deductive or inductive, (2) nomothetic or idiographic? Why?

4.       What are the key concepts investigated?

5.       How have these concepts been operationalized (i.e. measured)? Reconstruct their operational definitions.

6.       What data is the study based on? Is the data of high quality? Why?

7.       What are the key hypotheses?

8.       What variables are used in the study?

9.       What does the study conclude?

Use a concise review format to answer these questions, i.e. focus on providing the information required by the questions above maintaining a narrative which makes your statement a structured entity. Introduce each issue addressed and use subheadings where appropriate. Do not assume that the reader knows what questions you are answering. Start each paragraph with an appropriate transition statement. Be careful to avoid copying fragments of the article you are summarizing. The fewer direct quotations the better. Rely on your own narrative to describe it. Do not attempt to convey the details of analytical techniques you find hard to understand. Instead focus on the research question, the concepts involved in it and their links to the variables.  If the list of analyzed variables is very long, name only the key ones and briefly mention the others without including the whole list.

You may find that some of the papers will have no explicitly formulated hypotheses and if this is the case in your chosen study try your best to unravel implicit hypotheses. Bear in mind that testable hypotheses are statements which have a potential to be confirmed or refuted by the analysis.

Here are some articles you can choose from, but feel free to search for suitable articles yourself (eg. American Sociological Review publishes a lot of fine quantitative articles):

All the articles are available in the ANU library either in hard copies or as electronic documents

Baxter, J. 2000. "Barriers to equality: men's and women's attitudes to workplace entitlements in Australia." Journal of Sociology 36:12-29.

Baxter, J. 2002. "Patterns of change and stability in the gender division of household labour in Australia, 1986-1997." Journal of Sociology 38:399-424.

Baxter, J. and E. O. Wright. 2000. "The glass ceiling hypothesis - A comparative study of the United States, Sweden, and Australia." Gender & Society 14:275-294.

Bean, C. 2002. "Political personalities and voting in the 1999 Australian Constitutional Referendum." International Journal of Public Opinion Research 14:459-468.

Bulbeck, C. 1999. "The 'nature dispositions' of visitors to animal encounter sites in Australia and New Zealand." Journal of Sociology 35:129-148.

Evans, M. D. R. and J. Kelley. 2002. "National pride in the developed world: Survey data from 24 nations [Review]." International Journal of Public Opinion Research 14:303-338.

Evans, M. D. R. and J. Kelley. 2004. "Subjective social location: Data from 21 nations." International Journal of Public Opinion Research 16:3-38.

Evans, M. D. R., J. Kelley, and R. A. Wanner. 2001. "Educational attainment of the children of divorce: Australia, 1940-90." Journal of Sociology 37:275-297.

Graetz, Brian and Ian McAllister. 1994. "Crime and Deviance." Pp. 298-324 in Dimensions of Australian society. South Melbourne: Macmillan.

Marks, G. N. and J. McMillan. 2003. "Declining inequality? The changing impact of socio-economic background and ability on education in Australia." British Journal of Sociology 54:453-471.

Pakulski, J. and B. Tranter. 2000. "Civic, national and denizen identity in Australia." Journal of Sociology 36:205-222.

All essays must be typed, double-spaced, with 1" top, left, right, and bottom margins in 11 or 12-point font.  Avoid using odd fonts that may be difficult to read (Times New Roman is preferred).  Pages must be numbered.  Your full name must be on the cover of the assignment.  Title your assignment and state clearly in the introduction the title of the chosen article.  The recommended length is about 1000 words.

Descriptions of Assignment 2 and 3 and Tasks 2 and 3 will be available later in WebCT, distributed and discussed in class.

Please note that some alterations to the lecture, tutorial and lab schedule are possible. Hence students are advised to check their WebCT accounts, as notifications of changes will be emailed to students as well as announced in class.

These are the guidelines provided by The Faculty with regard to assignment submission:

1. Written work submitted after its designated submission date (without an extension) will have marks deducted at the rate of 3% of the total possible mark per day.

2. Written work submitted fourteen days (i.e. two weeks) after the designated submission date (or extension date) will be given a zero mark.

3. An extension of time for submission of written work may be sought from the lecturer/tutor-in-charge in the event of unforeseen circumstances preventing timely submission of written work. The appropriate request forms are available from the cabinet outside Room HA 2078.

4. It is possible that deadlines may be set in individual negotiations with students where appropriate.

5. Students having difficulty meeting deadlines should see their lecturer/tutor BEFORE the assignment is due.

IT IS ESSENTIAL THAT ANY STUDENT ANTICIPATING A LATE SUBMISSION, CONSULT THEIR TUTOR OR LECTURER AS SOON AS POSSIBLE IN ADVANCE OF THE DUE DATE

You will have a round-the-clock access to your grades as they appear in your WebCT grade book. You have 2 weeks after an assignment or homework paper is returned to notify me about any discrepancies between your marks as they appear on your paper and in your WebCT records. Hence it is crucial that you collect your work in a timely fashion. Usually no homework is accepted in an electronic format. The best strategy is to start a folder in which you will keep all the graded papers until the end of the semester. Assignment and homework papers will be brought to class ONCE and if you are absent on that day, you need to collect your paper from the School Office in COP2147.

Any suggestions provided in the assignment feedback have to be taken into account in subsequent assignments. For instance if I point out a logical error in percentaging a table included in an assignment and a similar error is repeated in the subsequent assignment your mark will be reduced.

It is also your responsibility to notify me about any errors regarding your record in WebCT tutorial group listings and any other lists which may be created during the course.

IMPORTANT: It is the student’s responsibility to retain a copy of their submitted homework, which should be presented in case of any dispute with the instructor.

Assignments turned in by close of business on the due date are on time -- others are late. This includes assignments left in the School’s Office or elsewhere. Note that an assignment is submitted when the office received it, so please do not place them in my mailbox; do not slide them under my office door etc.

Unfortunately there is no such thing as a perfectly consistent notation across social statistics textbooks. We will use mostly Marija Norusis and Alan Agresti’s notation and this summary should help you navigate in the world of symbols. The list below contains the key symbols and formulae but a few additional ones will appear in our discussions of correlations and regressions. Do not expect to understand everything on this list. We will gradually introduce and explain these concepts.

Note that Norusis assumes that Latin letters denote sample statistics, and Greek letters denote population parameters. Agresti distinguishes one more group of symbols: Greek letters with subscripts which refer to probability distributions of particular variables i.e. information we would get if we kept drawing samples from the same population again and again.

                                                                    μ Greek letter mu

Y is our variable of interest

n is sample size

see M. Norusis p. 198

                                                                                    σ small Greek letter sigma

                                                                        Note that Agresti denotes standard error as

For our purposes both these formulas mean that standard error can be found by dividing the sample standard deviation by the square root of the number of cases

These two formulas are equivalent, the only difference is that the formula to the left is in more general form and the formula to the right assumes we are dealing with the sample mean   which is an estimate of the population value, which we contrast with the population value specified in by our null hypothesis.

The z-score for a given value Y expressed as the number of standard deviations that y falls from μ

Hypotheses about the proportion

H₀: π= π₀  The null hypothesis

H_a: π > π ₀     or    H_a: π < π ₀       or      H_a: π ≠ π ₀  The alternative hypothesis also called the research hypothesis

Hypotheses about the mean

H₀: μ= μ₀  The null hypothesis

H_a: μ > μ ₀      or      H_a: μ < μ ₀      or     H_a: μ ≠ μ₀   The alternative hypothesis also called the research hypothesis

A range of values – an interval – that includes the population parameter with certain probability (usually 95% or 99% because that is by tradition what people like to use). For example in the normal distribution of the working hours amongst college graduates, with standard error of 0.5 hours, 95 % of all possible sample means will be within 1.96 standard errors, or approximately 1 hour from the unknown population mean. (see Norusis p.236)  and 99% of all possible sample means will be within 2.58 standard errors or roughly 1.5 hours from the unknown population mean.

-zσ μ +zσ

-1.96σ μ +1.96σ                 95% confidence interval based on normal distribution

-2.58σ μ +2.58σ                 99% confidence interval based on normal distribution

The value to the left of ≤ is the lower confidence interval bound.

The value to the right of ≤ is the upper confidence interval bound.

PART ONE: FOUNDATIONS AND BASIC TERMINOLOGY IN QUANTITATIVE METHODOLOGY

Lectures 19, 20 Jul

Introduction to the course. How sociology uses statistics and surveys? The link between the theory and empiricism.

Quantitative approach originated from a specific area of social philosophy dominated by positivism and later neo-positivism. These schools of thought equipped quantitative sociology with several ontological assumptions regarding the relationship between our knowledge and reality that is the theory and observables. Building on the fundaments of these axioms positivists developed a model of scientific explanations based on generalizations and deterministic relations. So do quantitative sociologists really believe that there are "objectively existing social facts" organized into universal relationships which can be "objectively" determined as in physics or biology? Our first week will be devoted to answering this question as well as explaining concepts such as paradigm, operationalization, induction, deduction, dependent and independent variables.

IMPORTANT NOTE: After the first and the second lecture I will distribute short self-completion survey questionnaires which you will have to distribute to 5 people (as different with regard to gender, age, education, occupation and income as possible). Detailed instructions will be provided with questionnaires. The completed questionnaires are needed for our first lab session and first assessment item, so it is important you do them on time.

READINGS:

1. Reading Brick Text 1: Earl Babbie: Human Inquiry and Science and Paradigms,
   Theory and Research, pp. 2-29

2. Norusis: Chapter 1 Introduction and  Chapter 2 An Introductory Tour of SPSS.

No tutorial or lab this week.

Use http://arts.anu.edu.au/tutorials/ to sign up for one lab and one tutorial time. The sign up tool will not be available before July 18^th.

Lectures 26, 27 Jul

 "Surveyology" in sociology - how to construct social survey questionnaires

Social sciences rely on a number of methods one of which is survey. Historically sociology in particular countries, research centers and sub-disciplines has leaned either towards qualitative or quantitative approaches. While many sociologists advocate the triangulation of methods i.e. employing all of them to best account for complexity and dynamics of social life, some feel that one particular methodology is superior regardless of the research question. We will review these standpoints and identify types of research questions suitable for survey research. "When to use surveys" and "how to design a survey study" will be our two areas of focus. The lectures will present you with a range of practical criteria which should be met by a well-designed survey study. The tutorial will provide an outlet for your own reflections on good and bad survey questionnaires.

READINGS:

1. Reading Brick Text 2: W. Laurence Neumann, Survey Research, pp. 30-52
2. Norusis: Chapter 3 Sources of Data.

 Lab 1 -- July 26 or 27

Data entry. You will have to bring your completed questionnaires to be able to participate in this lab. Questionnaires will be distributed in the first week after lectures. Task 1 will be described in this lab. If you have never used Microsoft Excel before, try opening a worksheet before this lab. Make sure you know how to save files on the ORACLE (see http://students.anu.edu.au/StudentITGuide/6oracle.asp)

Tutorial 1 -- July 26 or 27: Good and bad questionnaires

Task 2 – correcting a poorly written questionnaire will be distributed in this tutorial.

Lectures 2, 3 Aug

Units of Analysis, Types of Data, Types of Samples, Conceptualization and Measurement, Operationalization, Validity and Reliability

Principles of survey design and questionnaire construction are invaluable but by themselves insufficient even to read survey-based reports. To fully understand how and why surveys take on particular formats, you need a good understanding of several aspects of research design. A review of the concepts introduced in the first week will precede a review of units of analysis, types of data, types of samples, conceptualization and measurement, operationalization, validity and reliability. A good orientation in the choices of sampling techniques is a key prerequisite to knowing under what conditions inferential analyses can be meaningfully utilized. The understanding of the difference between nominal and operational definitions clarifies why at times different high quality studies provide apparently different results. Why do some surveys seem to contain large numbers of rather repetitive questions? How is it possible to ascertain what exactly people had in mind when ticking off boxes? In standardized mail surveys one cannot ask respondents for further clarifications so how to know what in the data is reliable information and what is simply a random noise?

We will also lay foundations for your skills in choosing optimal analytical techniques given the data in hand. The secret is to understand the amount of information locked in different measurement levels (nominal, ordinal, interval and ratio).

 READINGS:

Reading Brick Text 3: Agresti and Finlay and Salkind, pp. 54-83

Lab 2 -- August 2 or 3

We will focus on generating frequency distributions and descriptive statistics. We will learn how to open the Output, Syntax and Data windows in SPSS. You can use either pull-down menus or syntax (i.e. text commands) to utilize SPSS. Syntax is preferable and we will discuss reasons for this. We will define variable names and labels and value labels. Make sure you know the difference between them.

Tutorial 2 -- August 2 or 3

All you need to know about types of data and descriptive statistics

Bring a scientific calculator to this class. We will complete a number of exercises in computing descriptive statistics.

PART TWO: LIES, BIG LIES, DESCRIPTIVE AND INFERENTIAL STATISTICS

Lectures 9, 10 Aug

Descriptive Statistics: Tables, Graphs, and Measures of Central Tendency

There are plenty of unreliable statistical analyses and reports around us and there are a few good quality ones. How to sort the wheat from the chaff? The first step is to develop a good knowledge of most frequently used statistical summaries: the mean, median and standard deviation. What they are and what they can and cannot tell will be the focus of lectures. To fully benefit from their contents you will need a sound understanding of the material from the previous week i.e. levels of measurement (nominal, ordinal, interval and ratio) and how these overlap with the  quantitative/qualitative and continuous/discrete classifications of measurement (Page 74 in the reading brick).

IMPORTANT:  At this stage of the course you need to start exercises to consolidate systematically the material covered. In addition to our assignments and tasks each week non-assessable exercises will be provided during the lecture and exercise answers will be posted to WebCT. It is up to you how many you complete, but the more the better.

NON-ASSESSABLE EXERCISES:

This week do exercises from Norusis p. 93  Problems: 1,2,3 and p.94 Problems: 4,5,6,7,8,9

READINGS:

Norusis: Chapter 4. Counting Responses,
Chapter 5. Computing Descriptive Statistics and
Chapter 6. Comparing Groups (optional)

Lab 3 -- 9 or 10 Aug

Do you know what Syntax, Data and Output windows are and how to open them in SPSS? Do you know how to generate frequencies and descriptive statistics? How to save a syntax file? Do you know what a codebook is? If yes, we are ready to exploring our data set. We should (1) run frequency distributions (2) explore chart and editing options 3) identify some patterns in our data set.

No tutorial this week.

Lectures 16, 17 Aug

Probability Theory/ Sampling Theory

Interviewing a thousand of two thousand Australians can give an accurate picture of public opinion in the 20-million nation. How is it possible? The probability theory delivers a couple of basic tools (we will engage with only the rudiments) capable of producing educated and precise estimates. Generalizing from the information contained in a sample to the population is the staple of inferential statistics. Human passion for gambling gave birth to the probability theory, which turned out to have more uses that just making some people realize they do not stand a great chance of getting rich quickly by playing Lotto. We will review populations and sampling frames, probability distributions including the special case of the normal distribution to gain an insight in the logic of sampling theory. We will also learn about a tool enabling measurement-independent comparisons of different distributions: Z-scores known also as standard scores.

This information is indispensable for everything that will follow, so make sure you are present and tuned in. The concept of standard error (not to be confused with the standard deviation) will enable us understand why it is possible that there is no difference between two public opinion polls taken on the same day, one of which identified 55% Liberals while the other identified only 43% Liberals. From now on our thinking about social statistics will change, rather than operating within the realm of single numbers, we will always deal with ranges of numbers.

NON-ASSESSABLE EXERCISES:

The lectures will provide you with some non-assessable exercises meant to teach you how to use normal tables. Make sure you complete them. Access Java Applets in WebCT to consolidate your understanding of the Central Limit Theorem

READINGS:

1. Reading Brick Text 4: Agresti and Finlay Probability Distributions, pp. 84-104

2. Norusis: Chapter 10 Evaluating Results from Sample and Chapter 11 The Normal Distribution

Tutorial 3 -- 16 or 17 Aug

Probability in a bag of beans: Sampling Distribution, and the Central Limit Theorem. You will need to bring some graph paper and a set of cooking measuring spoons (this item is optional) to this class.

Lab 4 -- 16 or 17 Aug

Data exploration continued. Data transformations -- Making your variables match your research questions.

Lectures 23, 24 Aug

Contingency Tables, Chi-square Statistics

It is time to make an acquaintance with basic formats of analysis appropriate for different measurement levels. If your research question calls for the use of categorical variables, chances are you will use contingency tables as your preferred analysis presentation format. Knowing which variables are suitable and which are not for contingency tables (cross-tabulations) is the main goal of this week.  We will introduce Karl Pearson who developed the chi-square statistic used to determine whether two or more variables are statistically dependent or independent. The good news for this week is that the chi-square statistic is surprisingly simple to calculate. It will provide us with an opportunity to introduce the concept of degrees of freedom which has to be taken into account when deciding if our variables under investigation are or are not related.

NON-ASSESSABLE EXERCISES:

There will be a couple of non-assessable exercises to consolidate 1) correct percentaging of tables given the research question 2) calculation of chi-square statistic for contingency tables. Make sure you gain a good grasp of correct table percentaging.

READINGS:

Norusis: Chapter 8. Counting Responses for Combinations of Variables
Norusis: Chapter 17. Comparing Observed and Expected Counts

 

Lab 5 --  23 or 24 Aug

How to generate a crosstab (i.e. contingency table) in SPSS? How to produce and interpret a chi-square statistic? What do I do if my variable has too many answer categories?

No tutorial this week.

Lectures 30, 31 Aug

Statistical inference -- Estimating Population Means -- Hypothesis testing

This week we will learn how to correctly formulate and test hypotheses which is the ultimate goal of quantitative analysis. Our first step will be learning how to translate a plain English research question into the language of formal hypotheses. Next we'll find out what pieces of information we need to calculate the test statistic. Then we will learn to find the rejection areas to decide if we can reject the null hypothesis. One secret is that we never "accept" the null hypothesis. Why? We will explain this in the lectures. You will need to review your understanding of the mean and the proportion for a binomial variable. Make sure you remember why the proportion is a special case of the mean.

READINGS:

Norusis: Chapter 12. Testing a Hypothesis about a Single Mean.

Norusis: Chapter 14. Testing a Hypothesis about Two Independent Means.

NON-ASSESSABLE EXERCISES:

At least one example of a hypothesis test will be given to complete at home.

Tutorial 4 -- 30, or 31 Aug

 Hypothesis testing, types of errors, confidence intervals, how to select the right sample size

Lab 6 -- 30, or 31 Aug

SPSS and T-tests for the mean: one sample and independent samples. How to generate and interpret them?

20– 21 Sept   Guest Lectures -- topics to be announced

This week we will have a couple of guest lectures presenting practical applications and extensions of techniques we have covered so far in the course.

No tutorials or labs this week

Lectures 27, 28 September

Relationships between Two Variables: What is Correlation?

Correlation is the most frequently used and abused concept of quantitative analyses. Commonly substituted for a word to denote "a relationship", Pearson Product-Moment correlation is suitable for describing only certain relationship of certain types of variables. We will review the history of this coefficient, its possible applications and the link with ordinary least square regression. Both were developed in the course of Sir Francis Galton's work on eugenics.

NON-ASSESSABLE EXERCISES:

Access Java Applet links in WebCT to explore interactively the properties of correlation coefficients

READINGS:

1. Norusis: Chapter 19. Measuring Association.

Lab 7 – 27 or 28 September

A break from SPSS – Calculations of correlation coefficients

Lectures 4, 5 Oct

Relationships between Two Variables: What is Regression?

Ordinary Least Square regression is a basic modeling technique. We will begin from a bivariate regression analysis and learn about its basic properties. Multivariate OLS regression, i.e. the form that utilizes more than one independent variable, delivers a very powerful analysis which not only identifies determinants but also weighs their importance. For instance we can find out if attitudes to homosexuals in Australia depend more on age or education or residence in cosmopolitan areas of the country etc. OLS analyses are a tool frequently applied in policy development and forecasting thanks to their power to quantify the effects of independent variables i.e. we can estimate by how much our earnings should on average increase for each additional year of education or how much more money per child in the ACT schools should be spent to increase numeracy levels?

READINGS:

Norusis: Chapter 20. Linear Regression and Correlation.

Norusis: Chapter 21. Testing Regression Hypotheses.

NON-ASSESSABLE EXERCISES:

Access Java Applet links in WebCT to explore interactively the properties of regression lines

Tutorial 5 -- Oct 4 or 5

Standards of presenting regression and correlation tables in the sociological literature

Lab 8 -- Oct 4 or 5

How to generate and interpret SPSS OLS regression output?

Lectures 11, 12 Oct

Research Ethics and Politics of Social Research -- The uses of social research

Nowadays it is a truism to say that ethical considerations apply to any quantitative or qualitative study. In fact some people argue that more recently the preoccupation with ethics reached unprecedented form and scope as never have so many ethics committees been working on screening and approving research involving human subjects. So how can survey research be unethical? We will describe a couple of the most infamous cases.

From a different angle some people are distrustful of research based on statistics as statistics are believed to be extremely easy to manipulate and to be used to support any point of view. We will consider the validity of this concern.

READINGS:

Reading Brick Text 6: Kendall and Salkind

Lab 9 – 11 or 12 October

Finding areas under the normal curve and z-scores in SPSS. Review

No tutorial this week

Lectures 18, 19 Oct

Review

This week will be devoted to a review consolidating primarily the body of knowledge related to the logic of inferential statistics in the context of all the techniques reviewed during the course. Exercises will be completed both in the lectures and during the lab. There will be scope for catering to individual needs, so prepare any questions you may have regarding the material covered.

Tutorial 6 -- Oct 18 or 19

Review

Lab 10 -- Oct 18 or 19

Review