COMM88 Lecture Notes - Mullin PDF

Preview

Summary

Mullin...

Description

4/3/18 Professor: Dolly Mullin Office: SSMS 4117 Phone: 805-893-2750 Email: [email protected] Office Hours: Thurs 1-2:30pm or by appointment Attend section this week or you will lose your spot in the class! When you email your TA Give your name, class (comm88), and specific section info (enroll code, time, room) Buy textbook, The Process of Social Research 4/5/18 Ways of Knowing AKA “Epistemology” (the study of knowledge) Some “truths” - How do we know? Vegetables are good for you People who are similar to each other tend to like each other Some “Everyday” Ways of Knowing (& their problems) Method of tradition/tenacity Method of authority Problem with BOTH: authorities and handed down truths can be wrong Method of observation Surface level version: Personal experience More rigorous version: “Baconian empiricism” Problems: Inaccurate observation Selective observation Overgeneralization Method of intuition/logic Surface-level version: common sense More rigorous version: “platonic idealism” Problems: Incorrect premises Illogical reasoning Everyday ways of knowing can lead to conflicting ideas about “truth” The Scientific Method Combines platonic idealism with baconian empiricism logic/intuition-> constructing theories observation-> gathering data Communication science: use empirical observations to test theories about common processes Unique Characteristics of Science How is science different from the other everyday ways of knowing) Scientific research is public Published in peer-reviewed journals Opportunity to replicate studies Science is empirically rigorous Conscious, deliberate observations Many observations Science is objective control/remove personal bias

-

Explicit rules, standards, & procedures

4/6/18 What is it you want to know? How different types of advertising affect the selling potential to different age groups Like different sounds, colors, images Spirituality vs connectedness someone feels with people around them I believe people will feel more connected because when you are religious/spiritual you tend to have a community of others supporting you If people that are more prone to buy something after they see an advertisement are more happy or more sad in their everyday life/feeling of connectedness Do they believe buying something will give them happiness? Are they trying to fill a void or do they just really think the product is necessary? Accents vs perceptions of attractiveness/likability/intelligence Media messages affect attitudes/memories Family communication patterns Swearing vs perceptions of attractiveness/likability/intelligence Men vs. women writing can people tell? 4/12/18 Using Theories in Research Theory A scholars attempt to explain some aspect of life (how/why events/attitudes occur) Includes set of concepts and their relationships What are concepts Terms for things/ideas We (researchers) must define them EX: Social Cognitive Theory (Bandura) We learn by watching model behavior Motivation increases learning Concepts are studied as variables They have variations that can be measured EX. Motivation Compare rewarded as punished model Vary the amount of reward (or punishment) EX. Learning Degree of memory for the behavior Whether imitate or not Theories propose relationships between variables EX. Modeled behavior leads to learning EX. Motivation (rewarded vs. punished model) affects learning (imitation or not) From theory and/or prior findings, we derive a testable hypothesis: A specific prediction about the relationship between variables in your study EX. Social Cognitive Theory… to make a prediction about the effects of TV violence H1: Violent TV viewing will produce more aggressive behavior than will non-violent TV viewing WHat are the variables involved here? H2: Rewarded TV violence will lead to more aggression than will non-rewarded TV violence Scientific theories/hyps should be falsifiable Able to be tested empirically There is some result that (if you got it) would show that you are wrong

-

Falsifiable statements? Boys play video games more than girls Fairies inhabit the forest NOT FALSIFIABLE What about SOcial Cognitive Theory? Falsifiable Note- you can never prove theories true- only gain support/evidence What if theory or previous research does not lead to a specific prediction/hypothesis? Eg previous findings conflict/inconclusive Pose research question instead of hyp Ex. Rq: to what extent will children imitate the behavior of a TV character whom they do not like Rq: will there be gender differences in child imitation of violence? Testing a Hypothesis: AN Example Researcher A Social cognitive theory: children learn behavior by watching models behave Hypothesis: watching tv violence will increases kids aggressive behavior How much violent TV viewed How much aggression on playground Researcher B Catharsis Theory: watching others behave allows purging of pent up feelings Hypothesis: watching tv violence will reduce kids aggressive behavior Watch one of four clips (0, 5, 10, 20) Number of hits on toys 4/17/18 Types of Relationships Between Variables Associations between variables X is connected/related to Y EX: the more tv violence children watch, the more aggressive they are OR aggressive kids watch more violence than do non-aggressive kids Causal relationships between variables X influences/affects/changes Y Different Methods for Testing Different Relationships Survey/observational research Tests associations Great for external validity Poor causality Experimental research The Research Process- defining concepts and variables Independent variable Dependent variable 4/19/18 Measurement- operationalizing variables (both IV and DV) (continued) Types of measures Physiological measures EX. BP, brain imaging, cortisol (stress hormone) Behavioral measures Self-report measure

-

Levels of measurement Nominal (categorical/discrete): variable is measured merely with different categories Ex. gender (M/F); ethnicity; yes/no Q’s; TV use (Hi/Lo); persuasive appeal (humor/non) Nominal measures are for comparing differences Between manipulated IV conditions in experiments Between existing IV groups in surveys Ordinal: variable is measured with rank ordered categories Ex. rank top five favorite TV shows; most to least important political issues Interval: variable is measured with successive points on a scale with equal intervals Ex. measure of attitude about parenting: parents should talk openly with their children about sex (circle a number 1-7 from strongly disagree to strongly agree) Ratio: interval measurement with a true, absolute zero point Ex. time in hours, weight in lbs, age in years, test scores(if from 0 possible), etc Interval and ratio measures are both “continuous” variables Allow you to capture more variation Can always collapse to categories later, if need be Allow you to compare means (avg’s on DV) Allow you to test continuous relationships: Positive: The more X, the more Y Negative: the more X, the less Y Using Questionnaire Items as Measures Common for IV’s and DV’s in surveys Common for DV’s in experiments (IV is a manipulation into groups) Types of Questionnaire Items Open-ended Respondents give their own answers to Q’s Closed-ended Respondents select from list of choices Choices must be mutually exclusive Choices must be exhaustive Some close-ended formats Likert-type items: Respondents indicate their agreement with a particular statement Ex. i feel uncomfortable when people are arguing (1-5, strongly disagree to strongly agree) Other response options also possible (oppose/favor, not at all/very much, almost never/almost always) Semantic differential Respondents make ratings between two opposite (bipolar) adjectives Ex. this candidate seems: (rate from honest to dishonest; rate from trustworthy to untrustworthy) Composite measures Use multiple items combined to measure one variable (i.e. create an index/scale) Ex. candidate credibility This candidate seems: (honest-dishonest; trustworthy-untrustworthy; sincere-insincere) For each subject, add together (or average) scores of all 3 items into an overall credibility score How Good is your Measurement Reliability and Validity Reliability of Measurement Does your measure (of the variable) have consistency? Assessing reliability For measures using questionnaire items:

-

Inter-item reliability Look at internal consistency of similar items in a scale/index (composite measure) Want a high Cronbach’s alpha score

4/24/18 Measurement- Operationalizing Variables (both IVs and DVs) Assessing Validity Subjective types of validity Face validity The measure looks/sounds good on the face of it Content validity The measure captures the full range Predictive validity The measure is shown to predict scores on an appropriate future measure Ex. SAT scores (your potential to achieve) Convergent validity The measure is shown to converge Construct validity The measure is shown to be related to measures of other concepts that should be measured 4/26/18 Sampling (how we select participants (or other units) for a study Sample: a subset of the target population (who/what you want to report about Ex. target pops: voters, facebook users, married couples, juries, football fans, business owners, etc OR TV Shows, magazine ads, blog posts, etc Representative Sampling (probability sampling) Sample should be a “miniature version” of the target population Allows you to generalize results to that pop KEY is random  selection Everyone in pop has an equal chance of being included How  representative is it? Will always be “Sampling Error” Sample data will be slightly diff from pop because of chance alone (aka “random” error) Statistically, this is known as the “margin of error” Ex. national poll N= ~1,000-> +3% ^sample size v margin of error Representative Sampling Techniques Simple random sampling Select elements randomly from population Listed pops: random #’s table Phones: random-digit dialing Systematic random sampling From a list of the population, take random starting point, then select every nth element, until cycle through entire list Similar results as SRS Stratified sampling For getting pop proportions even more accurate Divide pop into subsets (“strata”) of a partic var Usually stratify for demographic vars Ex. sex, race, political party

-

Select randomly from each strata to get right proportions of the pop Need prior knowledge of pop proportions Increases representativeness Bc reduces sampling error (for the stratified var) But more costly and time consuming Multistage cluster sampling Useful for populations not listed as individuals First randomly sample groupings (“clusters”), then randomly sample individual elements within each cluster Reduces costs But sampling error at each stage So, for all 4 kinds if Representative Sampling Will always have sampling error But can generalize findings to the larger target population (assuming done properly) What should avoid: Systematic error (sampling bias) Systematically over or under represent certain segments of pop Caused by Wrong sampling frame Low response rate/response bias Improper weighting Non-representative sampling Cannot generalize results Can only make conclusions about participants in sample Typical of experiments and qualitative research Convenience sample Select individuals that are available/handy Purposive sample Select certain individuals for special reason (their characteristics, etc.) Volunteer sample People select themselves to be included Quota sample Select individuals to match demographic proportion in population network/snowball sample Select individuals, who contact other similar individuals, and so on… 5/3/18 Survey Research (aka observational/correlational studies) Primary Goals identify/describe attitudes or behaviors (in a given population) Can examine one point in time or track over time Cross-sectional Surveys One sample, variables measured once  Longitudinal Surveys Vars measured more than one point in time Panel- same people each time (eg Nielsen ratings families) Trend- diff random samples from same pop (eg poll ‘Americans’ every 10 yrs re church-going; track “likely voters” over the course of election campaign) Cohort- diff samples, but of same “cohort”

-

(eg survey class of 2015 every 5 years re their employment since graduation) Examine relationships between the att/beh variables measured Does X predict/relate to Y Do many factors together predict Y Ex. pew poll on marijuana legalization IV:generation (age group) DV: support for legalization Ex. Does exposure to alcohol ads (X) predict teen drinking (Y) Administering Surveys Self-Administered Questionnaires Mail surveys, online or emailed questionnaires, handouts, diaries Relatively easy and inexpensive No interviewer influence Increased privacy/anonymity BUT must be self-explanatory, very low response rate! Ways to increase response rate Have inducements Make it easy to complete Include persuasive cover letter and/or do advance mailing Send follow-up mailings Interview Surveys Personal/face-to-face More flexible (can probe for depth) Higher response rate BUT more potential for interviewer influence, higher costs 5/8/18 Interview Surveys Telephone Quickest results Compared to face-to-face: reduced costs, more privacy, more efficiency Compared to self-administered: more detail possible, better response rate But what about call screening and cell phones Experience Sampling Send text messages to Ps, they link via phone to survey online (can also use survey apps) Ps answer Qs about their experiences/feelings in the moment Can improve accuracy of self-reports Allows for longitudinal/panel data Understanding Data in Survey Research Depends on hyps and how IVs/DVs measured Examining Differences IV is categorical (nominal/discrete): Ex. comparing age groups; heavy vs light talkers; tinder If DV is continuous (interval or ratio): (DV uses Likert, semantic diff items, etc) Compare mean (average) DV scores for the different IV categories Examining continuous relationships Both IV and DV are continuous (interval/ratio data): Compute a correlation Correlation Magnitude of Relationship (strength)

R ranges from 0 to 1 So, SUrvey Research has 2 Causality Problems Causal direction problem (time order) Does X cause Y or does Y cause X Third variable problem (other explanations) 5/10/18 Survey Research To establish causality… Vars must be related (X correlated w Y) Okay so far- surveys can show that EX: time studyinbg^, GPA^ Must establish time order (IV happen before DV) Must rule out other explanations or causes Survey/Correlational Research has Two Causality Problems Causal direction problem (time order) Does X cause Y or does Y cause X Ex. studying might affect GPA, or GPA might affect studying Third Variable Problem (other explanations) Does some third variable explain the X/Y relationship? Getting Closer to Causality To help solve 3rd variable problem Partial correlation Measure potential 3rd variable Statistically partial out (control for) effects if those 3rd variables To help solve causal direction problem: need a longitudinal study Cross lagged panel design Experimental research OLIVIA START HERE Purpose To test hypotheses of cause and effect Goal is to establish internal validity Willing to sacrifice external validity Key Elements to a True experiment Manipulation of IV(s)... Divide into different “conditions” Ex IV: New painkiller drug, half of Ps get drug/other half do not ...while controlling all other variables Subjects in each condition treated the same, etc Examine effects on Dependent Variable (DV) Compare measures (mean scores) for subjects in each condition and see if diffs exist EX DV: amount of perceived pain (eg on 1-10 scale) Random Assignment of participants (Ps) to conditions Everyone must have an equal chance of ending up in either condition Why important? Makes groups equal before  manipulation Types of Experiments Design notation X: IV manipulation (treatment/induction) Y: DV measure

-

R: Random assignment

5/15/18 Experimental Research Key Elements to a True Experiment manipulation/control + Random assignment = internal validity Types of Experiments Design notation: X: IV manipulation True Experiments Posttest only control group design: R X1 Y (group 1) R X2 Y (group 2) Example R X1 (antismoking ad) Y (beliefs about smoking) R X2 (no antismoking ad) Y (beliefs about smoking) The difference between X1 and X2 is the IV manipulation Y is the dependent measure If get difference between groups means on Y, the IV manipulation caused Variations: more groups, several different treatments Example: IV- diff types of ad appeals R X1 (personal cancer story) Y (group 1) R X2 (cancer statistics) Y (group 2) R X3 (tobacco industry) Y (group 3) Pretest-posttest control group design R Y1 X1 Y2 (group 1) R Y1 X2 Y2 (group 2) Example R Y1(beliefs about smoking) X1 (antismoking ad) Y (beliefs about smoking) R Y1(beliefs about smoking) X1 (no antismoking ad) Y (beliefs about smoking) Difference between X1 and X2 is IV manipulation DV measure before and DV measure after Again, if difference between group means then caused by IV manipulation Possible problem: diffs on Y2 might be result of interaction of manipulation with pretest Solomon four-group design R Y1 X1 Y2 (group 1) R Y1 X2 Y2 (group 2) R X1 Y2 (group 3) R X2 Y2 (group 4) Pretesting: Should You or Shouldn’t you? Useful: To “check” on RA To get info on change But: Not necessary to establish causality Bad idea if treatment/pretest interaction is likely Factorial Designs Purpose: to examine the effects of two or more IVs simultaneously

Factors are IVs Each factor/IV has at least two levels (conditions) What if more than two factors? Music factor: lyrics/no lyrics/none Task factor: math/reading Class level: lower/upper division Factorial designs test for two kinds of effects Main Effect The effect of one IV individually on the DV Ex. for the 2 x 2 To test for main effects, compare “marginal” means of DV for each factor/IV 5/17/18 Factorial Designs Main Effect The effect of one IV individually on the DV Interaction Effect The unique effect of the combination of IVs The effect of one IV depends on the levels of the other IV(s) Example for a Music X Task interaction Music reduces learning when studying reading, but enhances it when studying math A word about factors(IVs)... In one design, can have IVs/factors: Manipulated variables Ex. music exposure, study task Subject variables Ex. gender, personality traits, TV use (hi/lo) Can only make causal conclusions about manipulated IVs (not Ss variables) If no manipulated variables at all, then it’s not an experiment Key Elements to a True Experiment manipulation/control Random assignment Internal validity Posttest only control group design vs pretest posttest control group design vs factorial design Threats to Internal Validity If not a true experiment or if do exp properly, then Alternative explanations for results become possible (ie threaten internal validity) Selection bias History effect Related to pretesting (or measures over time): Maturation testing/sensitization Statistical regression (to the mean) “Pre-Experiments” may sound like experiments (IV manipulation), but no RA and many threats to internal validity One shot case study One group pretest-posttest design Static group comparison 5/22/18 Experimental Research

Key Elements to a True Experiment manipulation/control Threats to Internal Validity In NOT a TRUE experiment or if do exp improperly Selection bias History effect Threats related to pre-testing (or measures over time): Maturation testing/sensitization Statistical regression (to the mean) Mortality Reactivity Effects: Ps’ reaction to being studied, rather than to IV/treatment, influences DV Demand characteristics Hawthorne effect Placebo effect So, how to remove/control these threats Conduct a TRUE experiment RA to proper conditions Be sure to treat groups equally All groups get equal time, attention, etc Experimenter Effect/Bias: Experimenter’s behavior or attitudes, rather than treatment (IV), influence DV How to control exp’er effects? Same thing again (true exp, etc.) but also: Use d...