Class Notes (835,600)
Canada (509,275)
Sociology (2,104)
Lecture

2Z03 LECTURES 5-9.docx

12 Pages
101 Views
Unlock Document

Department
Sociology
Course
SOCIOL 2Z03
Professor
Art Budros
Semester
Winter

Description
TH LECTURE 5 - OCTOBER 15 SAMPLING Def: the process of selecting observations for study Terms: - Population and sample: Population (group you want to know something about) – the larger groups we want to generalize, or to make inferences, about; SAMPLE – drawn from the population, this is the subgroup we study directly o Population (N=2 million) o Sample (n=500) - Representativeness & Generalization: Representativeness – exists when the sample accurately reflects the population, that is, when key traits in the population appear essentially in the same proportion in the sample. In this situation, the findings involving the sample can be generalized to the population Population: Gender (F=49%, M=51%) Sample: “ “ “ o Run into issues with representativeness of sample; using sample to get at the people in your population (the sample does/does not represent the people in the population) You know its representative when the sample appears more or less the same at the population (need same key traits as people in pop.) Nonrepresentative “n” unable to generalize: Population: Gender (F=49%, M=51%) Sample: 40% 60% Non-representativeness: Sampling bias & sampling error: Sampling bias exists when the sample is NOT representative of the population & it exists because a faulty strategy was used to select the sample; sampling error exists when the sample is NOT representative of the population, BUT it exists because due to change: due to random factors, sample traits deviates from what actually exists in the population *sometimes conscious sometimes subconscious bias towards certain people/groups/etc and if you do have this problem = your sample will not be reflective of the population…..EVEN if you do everything to be bias or follow error**you will then have to make adjustments, IF you followed proper procedures. Sometimes these errors are just by chance!* Elements & sampling frames: elements (physical lists of people in the population)- the unit (e.g people, small groups) about which info is collected; sampling frames – the actual list of units (elements) in the population Probability sampling: All members of the population, or all elements in the sampling frame, have an equal chance of being selected in the sample; when probability sampling is used, a sample should be representative of the population from which it is selected Simple random sampling: each population member has an equal chance of being included in the sample. Seldom used, since it can be inefficient and laborious ( babie says this: prof says it’s the MOST common!) - EX1: 10% sample from population of 100 firms: firm #60, 70, 24, 44 (look in back of Babie text to see random numbers table) *has put this on exams before* look through table to find your random sample Stratified Random Sample: if a subpopulation has elements with key yet rare traits (e.g language), we can divide the population into relevant subpopulations (e.g blacks & whites) & then random sample each subpopulation; this ensures that key rare traits will be present in the sample Ontario: French (2%); Spanish (1%); Chinese (3%) Sample: French (2%); Spanish (1%); Chinese (3%) 100x0.02, 100x0.01 etc to find how many people n=100 C. Systematic sampling: select every __ element in the sampling frame; efficient & accurate (these samples closesly reflect simple random samples 1. sampling interval: standard distance between elements selected in the sample: population size (eg. 100)/sample size (eg. 10) = 10, so we select every 10 firms in the sampling frame 2. Random start: to avoid bias, we select the first element randomly. With a sampling interval of 10, we begin sampling one of the first 10 firms in the sampling frame. To decide which of these 10 firms we start with, we pick one of these firms through randomization 5, 15, 25, 35, 45, 55, 65, 75, 85, 95 (n=10) 8, 18, 28, 38, 48, 58, 68, 78, 88, 98 (n=10) 2Z03 Lecture 6 – November 5 th Know 3 kinds of probability sampling, and 2 kinds of nonprobability for exam! Nonprobability sampling: no sampling frame, so deliberately select specific for analysis a) Judgemental/purposive: select subjects on the basis of ones knowledge of the population & of the nature of the research a. Studies of drug users, sect members i. Drug users in Hamilton  researcher knows quite a bit about these people  has expertise and uses that information to decide who to interview b) **Snowball: identify person in the network of interest, interview the person & ask the person for the name of another potential interviewee. Build a sample suing this strategy through “snowballing” a. Studies of sexually abused women; political elites i. Interview people and ask that person “who should I interview next” this is how the sample would be created ii. Only stop when you get redundant answers ie: getting the same names Cases when you know quite a bit about your ‘population’ – knows who they want to sample, researcher DOES make the decisions *Back in 40’s/50’s – researchers who were interested in “who made important decisions” about communities  people wanted to understand how these decisions were made - 2 theories  power elite theorists (Mills -who makes big decisions? The elite), the pluralist theory (Robert Doll – organization of society, anyone can make an impact) EXPERIMENTS (not that important for us, but wants to touch on the other ways to research) Experiments: does an experimental stimulus produce change in subjects. Experimental stimulus = IV, change = DV 1) Classic Experiment a. DV Time 1 Stimulus DV: Time 2 b. Experimental group pretest Yes Post-test c. Control group pretest No Post-test - Assign 1 group to Experimental group, Assign 2 group to Control Group - Group 1 gets “stimulus”, group 2 gets basic health care - Nominal level IV = stimulus (Yes and No, can’t be ordered) - At DV Time 2  test what changes have occurred - ^ probably will test on this EX’s to help understand: 1) Medical studies; 2) medical/sociological studies: 3) sociological studies Here’s what you can find out using experiments 1) 1994 – study about medical researchers who used rabbits/mice/chicken embryos…..all had cancers, channelled into experimental group and control group, then applied stimulus. a. Point was to shut off “food supply” to tumor b. EG got protein shot, SG got nothing c. Experiment worked; researchers stunned by results of EG - In 2001, Scientific America(n), prof read that this treatment is still “the best you can have” - In 2010, article from Bono, Ten for the next Ten: “a person (Dr.li) & a word (Angiogenesis) o Most promising areas of cancer research 2) 1989, Stanford, breast cancer study. Women who’d had full-mastectomies a. Some are put into EG, some in CG b. EG got stimulus of a ‘therapy session’ once a week, CG got standard care c. Women in EG lived twice as long as women in CG d. Researchers tried to replicate but never published any work on it – hard to track progress e. In 2008, finally news about the study and that it was replicated and showed same results 3) Gender: sexual harassment  response is to have seminars about sexual harassment  seminars are not helpful; instead do an experiment! a. In a lecture; man came down & talked to prof that the attitude of the person LEADING made a difference LECTURE 7 - Experiments Don’t want systematic differences between people in your experimental and control group - People in each group need to “look the same” - Assign randomly is the best thing to do o This process decides who goes into one group or another o Keeps results from being affected by HOW people get into a group - ANOTHER WAY: matching  make sure the distribution is the same based on certain criteria, but this may create bias via coding Internal Validity In Experiments - Make sure changes are due to the stimulus not something else….want something to be valid and accurate o Make sure that there aren’t external factors that are contaminating your finings o HIGH INTERNAL AND EXTERNAL VALIDITY Some problems that may arise: - Selection bias: “better risk” get channeled disproportionately in the experimental group; this can happen due to chance. It may show that your experiment has worked  but it is not because of the stimulus (OUTSIDE FACTORS) - Experimental Mortality: subjects drop out of study (especially in experiments over time) - Testing: pretest affects posttest; etc (experience one has in pre-test affects the rest of the experiment  superficial changes, not due to stimulus) ***BETTER RISK = certain participants having less BAD external “risk” over others that affect and outcome…..best way to avoid this is randomization and then you have that to “blame” it on” *** External Validity in Experiments Experiments take place in artificial settings, so you always have to wonder if what you saw would happen in real life EX: experiment asked people to make decisions concerning what they would do if they were CEO’s of a business in certain scenario’s  usually using students with no knowledge on the subject, even if you asked CEO’s what they may say might be different from what they actually do  how the setting effects the experiment Sum: Advantages and Disadvantages Advantages: level of control over conditions, replication  witnessing your experiment in front of your eyes Disadvantages: artificiality, would these findings generalize to the real world SURVEYS Interviews: surveys handled through face-to-face interviews….something enticing about having a sample (few thousand people in different countries) to study a large array of things and draw accurate conclusions/data about people in those 3 countries. PEOPLE LIKE THE IDEA OF DRAWING CONCLUSIONS ABOUT LARGE UNITS OF STUDY A. Rules for Surveying (face-to-face): 1) Appearance/style; 2) Prepare; 3) Conducting interviews (eg. Wording, probing); 4) Interview location B. Advantages - High response rates (80%); few “Don’t Knows” and “No Answers”; clarifications; ability to observe; probing. Disadvantages: time consuming; cost; interviewer effect (possibility that the researcher’s presence affects the outcome) PROF TALK about his experience with interviewing: - Need to fit into the community in which you are surveying o “see if you can get through the door” important to look and conduct yourself appropriately o Be friendly, cordial, take interest in your subjects to gain some sort of relationship - BE PREPARED - How you will literally read/express questions in a survey – standardize your phraseology o “matter of fact way” o Watch how you say things as they may change the tone of the situation  Don’t say anything that shows certain personal opinion  Probe when you see fit as you may receive extra information - Location/setting o Think about advantages and disadvantages of a certain setting  EX: Pros of home: comfort, time, reflective of research matter  EX: Pros of an office: visual proof of surveyors reputability   in this case of “economic culture” the home would be a better place Phone Surveys: surveys handled through phone interviews Advantages: cheap; quick; less interviewer effect; less of an interviewer effect (sense of security of not being face-to-face) Disadvantages: persons without phones; unlisted numbers; refusals, screening of calls using answering machines PREPARATION: (prof story) KNOW INTERVIEW SCHEDULE - Don’t want to make a fool of yoursel
More Less

Related notes for SOCIOL 2Z03

Log In


OR

Join OneClass

Access over 10 million pages of study
documents for 1.3 million courses.

Sign up

Join to view


OR

By registering, I agree to the Terms and Privacy Policies
Already have an account?
Just a few more details

So we can recommend you notes for your school.

Reset Password

Please enter below the email address you registered with and we will send you a link to reset your password.

Add your courses

Get notes from the top students in your class.


Submit