Textbook Notes (367,759)
Canada (161,375)
Sociology (1,513)
SOC200H1 (51)
Chapter

CH 8, 9, 10, 11, 12 textbook notes

7 Pages
115 Views
Unlock Document

Department
Sociology
Course
SOC200H1
Professor
Margaret Gassanov
Semester
Fall

Description
CHAPTER 8 – EXPERIMENTS Classical Experiments  1. HISTORY – during experiment historical events may occur,  3 major components confound experimental results o Independent and Dependent Variable  ex. Police beating Aboriginal chief during experiment  IV = dichotomous variable (present or not)  2. MATURATION – subjects age during experiment  use to contract between results in dependent variable  Participants grow tired, sleepy, bored, hungry, etc.  Variables must be operationally defined  3. TESTING – Subject more sensitive to issues o Pretesting and Posttesting  give more thoughtful/socially desirable answers  PRETEST – measure dependent variable in subjects before  4. INSTRUMENTATION – Different measures in pre/posttest stimulus (IV) exposure  5. STATISTICAL REGRESSION – Interview an extreme sample  POSTTEST – remeasure dependent variable in subjects after  ex. low score, can only get better or stay the same, can’t stimulus (IV) exposure get worse o Experimental and Control groups  6. SELECTION BIASES – Wrongfully assumed preventative groups  EXPERIMENTAL GROUPS – group of subjects administered  7. EXPERIMENTAL MORALITY – a sample dies experimental stimulus  8. CAUSAL TIME ORDER – ambiguity about causality of IV and DV  CONTROL GROUP – resemble experimental group closely except  9. DIFFUSION OR IMITATION OF TREATMENTS – subjects not administered stimulus communicates  used to detect stimulus effects  Pass ideas/information, affect the experiment  HAWTHORNE EFFECT – any impact of research on the subject of  10. COMPREHENSION – give something extra to one group study  ex. more attentive to experiment group vs. control group  i.e. unintended effects  11. COMPENSATORY RIVALRY – groups know there’s different  Double Blind Experiment – neither the subjects nor the experiments know  work harder to compete w/ other group which is control group vs. experiment group  12. DEMORALIZATION – one group feels deprived  ex. educational experiments, demoralized group rebel Selecting Subjects o Sources of External Validity  PROBABILITY SAMPLING – select 2 probability samples represent of total  EXTERNAL VALIDITY – possibility conclusions drawn from pop. experiment can’t be generalizable to the “real” world o do so using a sampling frame  SOLOMON’S FOUR-GROUP DESIGN o 2 groups must resemble each other  Group 1: Pretest  stimulus  Posttest  result 1 o seldom used  Group 2: Pretest   Posttest  result 2  RANDOMIZATION – Randomly assign experimental subjects to  Group 3:  stimulus  Posttest  result 3 experimental/control groups  Group 4:   Posttest  result 3  Matching  Expected Results o Subjects matched based on similarities in 1+ variables  1. a change should be observed in result 1 o 1 of pair in experimental group, other in control  2. no change should be observed in result 2  3. result 1 and result 2 should be different (1 effected) Variable on Experimental Design  4. result 3 should be different than result 4 (1 effected)  Pre-Experimental Research Designs (Avoid the following) o ONE-SHOT CASE STUDY – no pretest, IV not measured Other Experiments  ex. Survey people that watched the Aboriginal History movie  NATURAL EXPERIMENTS – Occur in the course of social events in the real answer questionnaire on prejudice level o ONE GROUP PRETEST-POSTTEST DESIGN – has pretest but lacks world o outside controlled settings control group  WEB-BASED EXPERIMENTS – use web as a place to conduct social science  can’t determine effects of IV o STATIC GROUP COMPARISON –no pretest  ex. Survey 2 groups, 1 watched Aboriginal movie, other didn’t Strength and Weakness  Pros:  separate groups may have diff initial prejudice levels o Scientific Rigor  Validity Issues in Experimental Research o Isolation of IV and its effects on DV o Sources of Internal Validity: 12 sources  INTERNAL VALIDITY – possibility conclusions drawn from o ease of replication  Cons experimental results may not accurately reflect what went on in o ARTIFICIALITY – What happens in experiment context may not reflect the experiment itself what happens in the real world  internal validity threatened whenever anything other than the experimental stimulus can affect the dependent variable CHAPTER 9 – SURVEY RESEARCH Topics Appropriate to Survey Research  use hypothetical situations  Survey  descriptive, explanatory, exploratory purposes o best for collecting original data on pop. too large observe directly Questionnaire Construction  need probability sampling  results reflect larger population  General Questionnaire format  ex. measure large population attitudes and orientations o poorly laid out questionnaire  missed questions, confusion of  Uses questions as means of operationalizing concepts desire data’s nature, unanswered questions o good question quality = essential o Should be spread out, uncluttered  Unit of Analysis = individual people o Shouldn’t squeeze several questions into a line, abbreviate, or use o extendable to groups or interactions few pages as possible  Respondent – Person who provides data for analysis by responding to a  demoralized after finished first page when there’s more o Respondent shouldn’t be forced to reread b/c confusion survey questionnaire  Some surveys false pretense –ex. telemarketers o Respondent shouldn’t be forced write in cramped spaces  Formats for Respondents Guidelines for Asking Questions o Make genuine boxes, professional look o Print coded number, circle appropriate number, avoid boxes  Variables operationalized as questions  to gather data o can ask directly or written down and given to respondent  easier to identify code for surveyor and participant  Questionnaires – instrument containing questions & other items  CONTINGENCY QUESTIONS – a question intended for some respondents o used in surveys, to solicit needed information o Contingency questions isolated from other questions off to side or enclosed in box  Choose appropriate Question Forms o Questions and Statements  instructions on top of each page for contingency questions  many questions in questionnaires also statements = intended o ex. if Question A == True: Do Question B; else Skip Question B  can show range of statements ask respondent Likert items  Question B == contingency question o Open-ended questions – respondent give own answer  MATRIX QUESTIONS – series of questions using same answer categories  most use in qualitative in-depth interview o ex. rows of Likert Items with the same scale  answers must be coded o Pros: o Close-ended questions – respondent select an answer  efficient space use  most popular survey research method b/c:  respondents answer faster  increase comparability of responses  response uniformity, easier processing  may over look important responses o Cons:  close-ended questions guided by 2 requirements:  req. structure item to fit into matrix format  1. responses exhaustive (cover all possible response)  foster a “response-set” i.e. answering all “Yes”  avoid by alternating statements representing different  2. responses mutually exclusive (no overlap) o 1. Make Items Clear & Unambiguous orientations and make statements short & clear  Precise so respondent knows exactly what’s asked  Ordering Items in a Questionnaire o 2. Avoid Double-Barreled Questions o Appearance of one question may influence answer to later questions  DOUBLE-BARRELED QUESTIONS – a question w/ multiple parts  ex. Start w/ questions relating to crime, then an open-ended question on severe social problem likely mention crime  ex. Agree/Disagree but another answer is viable  General rule: “and” in question = double-barreled question o Less educated respondents more influenced by ordering o 3. Respondents Must be Competent to Answer o Avoid by randomizing order of items  continuously question whether respondent able to provide  problem: respondents see questions as chaotic & worthless  need respondents switch attention btwn topics information’s reliably o 4. Respondents Must be Willing to Answer o Can anticipate item’s effects, interpret results meaningfully  ex. undecided response b/c opinion in minority o Self-administered survey:  guarantee anonymity to avoid overwhelming undecided  no negative/threatening starting questions  Demographic data request placed at end o 5. Questions should be Relevant to most respondents  asked irrelevant question higher chance skewed results o Interviews: demographic questions first  include “Not Sure”, “No Opinion” option  Questionnaire Instructions o 6. Short Items are Best  unambiguous & precise o Short introduction, orient respondents  respondents unwilling to study item to answer it  especially needed if questionnaire incl. a variety of questions o Have basic instructions for completing survey  should req. quick read, understand intent, select/provide answer w/o difficulty  have instructions on expected answers –ex. in box X want T/F o 7. Avoid Negative Items to avoid easy misinterpretation o Open-ended questions give guidelines on answer length  avoid words such as “not”, “prohibited” o limit rank-ordering questions b/c need reread  Pretesting the Questionnaire o 8. Avoid Biased Items and Terms  no ultimate true meaning for any concepts studied o pretest w/ diverse set of 10 people  wording effects the responses to the question  participants don’t need to be a representative sample  Bias – the quality a measurement device misrepresents what is  ask for completion, not error checking actually being measured in a particular direction o Precoding instructions for data processing  Biased question = leading, loaded  ex. identification of attitude/position w/ a prestigious Self-Administered Questionnaires person/entity can cause bias response  1 of 3 methods of administering a survey  Aware of social desirability of questions/answers o respondents asked to complete questions by themselves o most common method is mail survey  people answer question through filter of what will make them look good, especially for interviews  may mailed to house, researcher pick up or vice versa  to avoid:  Usually Questionnaire accompanied w/ Letter of Explanation, a Self-  imagine the feeling of giving each question/answers  if Addressed stamped envelope for return  Should track rates of return embarrassed, reword the answer/question o graph 1: number returned per day Telephone Surveys o graph 2: cumulative percentage of returned  1 of 3 methods of administering a survey  returned questionnaires should be scanned, assigned ID number o 3. surveys conducted by telephone  Follow up mails encourage participation, increase return rate  Years ago social-class bias b/c excluded poor from surveys o 1 original questionnaire mail w/ 2 follow ups 2 – 3 weeks apart  Sampling problem w/ unlisted numbers  RESPONSE RATE (%) – number of people participating in a survey divided o Survey sample = telephone directory, omits rich by/ own request by the number selected in the sample  Pros: o aka. completion rate or self-administered survey return rate o Time and money advantages o Higher response rate = less chance for response bias o Sometimes respondents more honest giving socially disapproving  b/c non-respondents may differ w/ respondents other than just answers if not looked directly in the eye willingness to participate  Con: o 50% = adequate, 60% = good, 70% = very good o Misleading sale campaign now encouraged most to hang up o Demonstrated lack of response bias more important that high  CATI – Computer Assisted Telephone Interviewing response rate o Automated and random dialing o Computer displays to questions and accepts answers Interview Surveys  1 of 3 methods of administering a survey Technologies and Survey Research o survey administered by interviews in face-to-face encounters  CAPI – Computer Assisted Personal Interviewing  INTERVIEW – data collection encounter w/ one person (interviewer) asks o Similar to CATI but face to face questions of another (respondent)  CASI – Computer Assisted Self-Interviewing o by face to face or by telephone o researcher bring computer to respondent’s home o Most interview surveys require more than one interviewer  CASQ – Computerized Self-Administrated Questionnaire  Higher response rate than mail surveys (80-85%) o completely electronic, answer questionnaire = data o interviewer presence generally decreases “I don’t knows”, “no  TDE – Touchtone Data Entry answers” b/c interviewers can probe for answers o Respondent calls research organization, answer computerized  PROBE – solicit a more complete answer to a question questions w/ telephone key presses  Probe must be neutral, can’t affect response’s nature  VR – Voice Recognition, similar to TDE but accepts spoken words  ex. “anything more?” o Interviewer can also clarify confusing answers Online Surveys  Interviewer can observe respondents while asking questions  Web surveys may not be representative of general population o ethical issue: respondent not being aware that such information o depends on goal of study, the sample used gather  respondent not be fully informed about the study  Online survey higher completion rates than paper survey  interviewer should be a neutral medium  Pros: o shouldn’t affect respondents perception of a question or the answer o responses written, already organized for analysis o dress in a fashion similar to interviewees o Survey sites tracks who participated/refused o be pleasant, relaxed and friendly  Cons:  show genuine interest to get to know the respondent w/o o Survey sites doesn’t allow complete question format freedom appearing to be spy o not all population are good fit for online survey o record answers exactly as given with no summaries, paraphrase, or correct bad grammar Strength and Weaknesses of Surveys  will not know coding until read ~100 responses  Pros: o add marginal comments of interpretations, reasoning behind them o Survey useful describing larger population o prepare specifications, explanations for questions o Develop operation definitions from actual observations  Interviewer must study questions, read w/o error like natural convo.  vs. experiments, need commitment in advance o prepare specifications, explanations for questions o Strong on reliability  more than 1 interviewer, req. training & supervision o Repurpose data, allow secondary analysis  cheaper and faster than doing original survey  Cons: o Req. standardization, fit round pegs into square holes o Weak on validity b/c artificiality o Secondary analysis data collection doesn’t reflect purpose of study CHAPTER 10 – UNOBSTRUCTIVE RESEARCH  UNOBSTRUCTIVE RESEARCH – study social behavior w/o affecting it  ex. frequency the word “love” appears in a book o can be qualitative/quantitative o LATENT CONTENT – underlying meaning of communication  ex. holistic judgment of the book regarding the message “love” Analyzing Existing Statistics o Conceptualization and creation of code categories  Uses existing official or quasi-official data, not secondary analysis  Conceptualization and Operationalization typically involve o no ethical approval req. but info may be constrained interaction of theoretical concerns and empirical observations  Used as supplemental source of data  inductive, deductive methods should be used  ex. Durkheim’s Study of Suicide
More Less

Related notes for SOC200H1

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