|
SY 4520 Methods of Sociological Research M/W, 10:20-noon, Room B-135 Prof. Jacob Heller, Ph.D. - email: hellerj@oldwestbury.edu Office Hours: M/W 1:30-2:30, and by appointment HOME | READINGS | ASSIGNMENTS | GRADES | COURSEBLOG on Monday, April 28, 2008... We talked about some of the practical aspects of writing paper #7, using the example of someone interested in understanding the effects of harsh punishment (e.g. chopping-off hands, as in Saudi Arabia). We discussed the ways in which a comparison between Saudi Arabia and the United States would be both good and not so good, for studying this topic. The idea in historical/comparative methods is to find suitably comparable cases (whether in the same time - perhaps Saudi Arabia compared to Syria?) or over time (perhaps Saudi Arabia now compared to Saudi Arabia shortly after independence) - cases that minimize the differences and begin to "isolate" the crucial independent variable(s). Only if we compare Saudi Arabia's punishment scheme to the punishment scheme in another country with similar religious, legal and cultural variables will we be able to parse the effects of punishments as severe as dismemberment. This again highlights the importance of a clear hypothesis for good research design: what factors would be more important for understanding the causes or effects of hypothesized crucial independent variable? I also distributed the assignments for papers #7 and #8. Topics for paper assignment #8 are due on Wednesday, 4/30. I reminded the class that APRIL 29 is the LAST DAY TO WITHDRAW FROM THIS CLASS. on Wednesday, April 23, 2008... We talked still more about content analysis, and began to talk about historical methods. I promised to post the assignment for paper #7 by Thursday, April 24. on Monday, April 21, 2008... We talked at much greater length about content analysis, I distributed a short reading in class (an example of historical methods), and I pushed back to due date for paper assignment #6 to Monday, April 28. on Wednesday, April 16, 2008... Students sat for Midterm #2, and I distributed paper assignment #6. on Monday, April 14, 2008... Today I collected Paper #5, and returned those paper assignments #4 that I received on time. We talked for some time about Content Analysis (which was the subject of the Baker reading distributed last week - well, Baker used content analysis to collect her data). Students volunteered different kinds of human communication, from print to voice to art to music (e.g. the Andrews Sisters) to the Internet. I talked about the ways in which radio was, in its time, as powerful as the Internet has become now (I mentioned the dissemination of the photos from Abu Ghraib prison [WARNING - this links to some very disturbing images] as a recent event that could not have happened without the Internet, and mentioned the power that radio gave to political leaders, including Winston Churchill, who rallied the British before, during and after "the blitz"). Content analysis collects data from various forms of human communication, and can be either qualitative or quantitative in nature. We talked about: how images of women's bodies might influence both obesity and anorexia; how different stories of first sex in Germany and the US can help us understand not only the different cultural attitudes towards sex in those countries, but our own American attitudes towards sexuality; how analysis of the letters that soldiers wrote home during wartime can help us understand the role of war in people's lives, as well as how families related in different historical periods - etc. After our discussion of content analysis, I distributed an outline of suggested study topics for the second midterm exam, which will be held on Wednesday, 4/16. That exam will have three parts, each weighted roughly equally: multiple choice, identifications and an essay. on Wednesday, April 9, 2008... I collected paper assignment #4, and we talked about intensive interviews - and then the class "interviewed" me about my drinking behavior. I talked a bit about group interviews (focus groups). I distributed a new reading based on data collected using content analysis, which needs to be read for Monday (maybe another quiz?!?!). on Monday, April 7, 2008... I returned the unobtrusive observations, and we talked about Whyte's research methodology. I distributed paper assignment #5, which is due on Monday, April 14. on Wednesday, April 2, 2008... We talked about the Whyte reading (I gave a quiz on it) and about participant observation in general. I collected homework assignment #3. on Monday, March 31, 2008... I was home sick, and had to cancel class. on Wednesday, March 26, 2008... We talked at some length about the structure of research articles (a senior thesis has the same basic structure), with reference to the Marquart reading. The abstract serves a number of purposes, as does the review of the literature. The Marquart reading doesn't have any substantive findings, and instead discusses methodology. In our discussion I tried to draw some of the connections between his evaluation of his own methods and the ability to obtain good data about a subject in which we are interested. (I made a painful analogy between difference data collection methods and different ways to obtain food.) With the full menu of data collection methods available (surveys, experiments, interviews, full and unobtrusive observations, content analysis, historical/comparative analysis and secondary analysis), students in the senior seminar need to understand each method, its applicability and its practicality for thesis research. I assigned homework assignment #3, but because of technical problems (my laptop won't recognize the campus server), was unable to distribute the written assignment in class; it is available HERE. on Monday, March 24, 2008... We talked somewhat more about paper assignment #3, the ethical rules governing research on human subjects (see chapter 9 in the reader packet for more details) and talked in greater detail about the research process and the initial steps in conducting research: ideas, hypothesis, literature review. I spent a good deal of time talking about the importance of the lit review, and its interaction with initial ideas, hypotheses (if any) and method. on Wednesday, March 12, 2008... We talked a little bit about the Milgram reading, and I showed the admittedly old and spotty video recording of the Milgram experiment. Before the film, though, I talked at some length about the nature of experiments (which are only one kind of research study), and how - even though sociologists and criminologists rarely conduct experiments - we try to approximate the comparative nature of experimental studies. We want always to compare different groups. Clearly we usually cannot ethically (or practically) create the conditions we want to study, but if we can find already existing situations that approximate the conditions we want to compare, we can still do research and test hypotheses about things like what kinds of incarceration practices work better to rehabilitate convicts, what kinds of factors lead to social revolutions or even how children develop under adverse circumstances. I distributed paper assignment #3, which is due on March 26 (and is one of the "optional" paper assignments). I also distributed information about fall, 2008 registration and advising. Advising appointment sign-up sheets will be available on the Monday we come back from spring break (March 24), in the C-2 Tower in Academic Village, opposite the Sociology Department office (C-220). Have a safe, relaxing and productive spring break... on Monday, March 10, 2008... Today we talked about the Humphreys reading, and I tried to tie it together with both the Sex in America reading and the article I distributed in class today ("A Behavioral Study of Obedience." by Stanley Milgram). Together, the three readings encompass the idea of rigorous random sampling, participant observation and the experimental model while raising some important ethical issues. I spent most of the remainder of the meeting discussing the five main ethical areas: autonomy, privacy, non-maleficence, beneficence and justice, as well as the role of Institutional Review Boards (IRB's). (See also chapter 9, in the reader packet, for details about research ethics). on Wednesday, March 5, 2008... I returned some more midterm exams, I distributed paper assignment #2 (optional) and we talked a good deal more about sampling. We talked about the importance of random sampling and the size of a sample for generalizability. Assuming the sample is random, it's not the relationship between the size of the population and the size of the sample that is most important, it's the absolute size of the sample. The larger the sample size, the better the sample will represent the population from which it's drawn, but it is possible to have statistically significant findings with samples as small as 30 or 35, and the improvement from a sample of 1000 to a sample of 10,000 is not very great. That is, a population of 100,000,000 can be studied just as accurately as a population of 10,000 with a random sample of the same size. Thus, though Kinsey's samples were very large, the way he selected them (non-random sampling, really a sample of convenience) made his data invalid for describing the general population - he was unable to generalize beyond his sample. Because random samples are often difficult or impossible to obtain, we also talked about convenience sampling and snowball sampling - which relies on social networks to find people (e.g., drug dealers, prostitutes, etc.). Snowball sampling is not representative sampling, but it can provide important information about structures and institutions, rather than individual-level data. So a snowball sample can be a kind of informant sampling. Even with random sampling, response rates can skew - or bias - the data. If only 50% of the randomly selected people who received our questionnaire responded to it, our sample isn't random any more because we don't know why those who didn't respond didn't respond - and their non-response might be related to the kinds of information we're collecting. Good researchers strive to be objective, both in the way they look at issues, questions and data, and in how they select their samples. on Monday, March 3, 2008... I returned midterm exam #1 to all the students who were present and we went over the multiple choice questions, the kinds of answers that were better and worse for the identifications and the most common errors in the problem set. We began talking about the reading by Michael, Gagnon, Laumann and Kolata in terms both of studying sexual behavior and sampling and generalizability. We talked about the previous studies of sexuality that the authors discussed (Kinsey, Hite, Redbook, Masters & Johnson, etc.) and how that part of the reading constituted a review of the literature (or, more simply, a literature review). We talked about random samples, samples in general, populations, statistics, parameters, why people might not answer honestly to questions (about sex, but also about other things), norms and normative responding, generalizability, self-selection and the difference between the commonsense way we use many of these words and the specific meanings they have when used in the context of research methods. We talked about the value of studying something as private as sexual behavior, and the need to be open to the wide variety of human behaviors without making value judgments - the need to be objective. Sex is important for understanding disease transmission, emotional lives, reproductive trends, teenaged pregnancy, contraception efforts, information about abortion-related behaviors, etc., as well as interesting on its own merits. Even so, until the Sex in America study, there were no reliable or valid studies about human sexuality in the United States; even Kinsey's massive study failed to be generalizable. The result was that researchers were providing unsubstantiated and invalid information to the general public as if it were authoritative. The damage that poorly done research can do may be greater than other kinds of misinformation. on Wednesday, February 27, 2008... Students sat for the first midterm examination. on Monday, February 25, 2008... We had an in-class review for the first midterm examination. on Wednesday, February 20, 2008... We went over a couple of problems from the second homework assignment, talked about the logic behind percentaging tables (we want to be able to compare the categories of the independent variable in terms of the dependent variable, for example: age --> # of facebook.com friends, we expect people who are older to have fewer facebook.com friends than people who are younger). (The professional networking site the name of which escaped me in class is LinkedIn.) We also talked briefly about the importance of operationalizing concepts like "a friend" or "friendship." I sketched a scattergram (also called a scatterplot) on the board to illustrate how relationships between quantitative variables look (our tables have been only for categorical variables). I tried to stress the importance of understanding why we percentage tables as well as why we percentage them the way we do. The idea of learning concepts by rote or by memorization is antithetical to higher levels of learning: the point is not simply to get through the class (is it?), but to be in the class to learn... Data analysis is necessary so that we can judge whether our ideas about society are valid (supported by the data) or invalid (disproved by the data). The first midterm exam is scheduled for Wednesday, 2/27. It will have three parts, each counting about a third of the grade: (1) multiple choice questions covering important ideas, topics, etc. (2) identifications that will ask for the meaning and a relevant example for important terms or concepts (see especially the italicized terms in chapters 1, 2 and 6) and (3) problems that will involve percentaging and interpreting tables - including explaining why you are doing what you've done. on Wednesday, February 13, 2008... We went over the problems from Homework Assignment #1, and I tried to emphasize the importance of understanding why it is right to percentage tables one way, rather than another. The only reason for going through this process is so that we can use the data we've collected. We percentage "within" each category of the independent variable in order to be able to compare how each category "fares" in terms of the dependent variable. So if we think education influences income, we need to look at what proportion (we've been using percentages) of the people who are more highly educated are earning at higher levels, compared to the percentage of people who are less highly educated. Please see the assignments page for the reading assignment and where to find the problems that are due on Wednesday. on Monday, February 11, 2008... I returned all the papers I'd received by the end of last week (with grades, anonymous identification numbers as well as substantive and technical comments). I talked about how I figured the grades, what the grades mean and so forth; we also looked at the bar graph (not really a histogram, because the "bars" aren't touching one another) and from that segued into a discussion of how to represent data in one variable - histograms (normal or "bell" curves) and pie charts. On the back of the sheet that had the grade distribution, I included an article describing the relationship between drinking diet soda and something called "metabolic syndrome"; this led to a discussion of causality and percentages as a measure of association. The three criteria for causality are (1) association; (2) time order (causes must precede effects) and (3) lack of spuriousness (lack of "fakeness" - there cannot be a third [test] variable that better explains the relationship). I collected data from the class about gender and height, and used it to make a dataset, from which we constructed a 2 x 2 table showing the relationship between gender (men/women) and height (tall/not tall). We found a 75% difference, that is: for this group women were 75% less likely than men to be tall. I asked everyone to complete Homework #1 (on the same sheet as the "histogram" and the NY Times article) to be handed-in on Wednesday; also to read "Testing Ideas With Data" (it comes right after the "Troublemakers" article) up to the section beginning with "Linear Regression Analysis." on Wednesday, February 6, 2008... I collected paper assignment #1. on Monday, February 4, 2008... We went over some concepts from the reading - notably logical positivism and its notions about falsifiability. I used the example of Freud's theories of the mind to illustrate how some things are simply not empirically testable - not falsifiable (how can we measure someone's "id," "ego" or "superego"?). The positivistic notion that theories (and hypotheses) can never be proved true, but can only be proved false means that we often seek disproof - what would it look like if my theory (hypothesis) is wrong? I explained what the null hypothesis is, and how researchers use it to "disprove the opposite" of their hypothesis, and thereby indirectly support their own theories. I hope I made it clear that for the purposes of this course (and sociological research in general), a hypothesis is not simply "an educated guess" (though it may be that, as all); a hypothesis is a relationship between two variables, with x standing for the independent variable and y for the dependent variable. A hypothesis might be represented this way: x ---> y with the arrow (here crudely drawn) representing the direction of the influence. on Wednesday, January 30, 2008... We talked in terms general and specific about issues that came out of the readings - pre-modern and modern; deductive and inductive research; empirical questions and empirical evidence; logical positivism and falsifiability (and logically meaningless statements). We talked about the importance of challenging other people's assertions - the difference between a subjective opinion and a testable statement about reality - as well as challenging our own assertions about reality. I distributed the first paper assignment, which is due on Wednesday, Feb 6 (in class) - any late paper will be downgraded a full letter grade for each calendar day it is late. on Monday, January 28, 2008... Before we began, we had a visit from Saima Anjam, who spoke briefly about NYPIRG, and some of their current campaigns. We had a general discussion about things we think are true - like "NYC police are underpaid" and "the sun is hot" - and how and why (and whether) we should think such things are, indeed, true. We talked about issues of precision in language and the ways in which we (necessarily) rely on experts for much of what we believe to be true in life. I implored everyone to ask "how do we know?" when it comes to assertions of truth, particularly conventional wisdom. Some things we think we know turn out, on closer examination, to be far less certain than we thought (or think?). Some questions cannot be answered definitively - like what flavor of ice cream tastes best. Such questions are fundamentally subjective - they depend on personal perceptions and opinions. Other questions - the kinds with which we will concern ourselves in this course - are answerable; this is because they are empirical questions - questions that we can answer by valid examination of the real world around us. I asked everyone to complete a short reading assignment for Wednesday. I also promised to distribute the first (mandatory) paper assignment on Wednesday, to be due the following Wednesday: Feb 6. On Wednesday, January 23, 2008... I distributed the course syllabus, the readings for the course (a few more will be forthcoming), and talked in general terms about research methods, this course and its place within the Crim and Soc majors, as well as what students will have to do to earn a good grade. I asked students to complete the first set of small assignments by Monday. |