WSU COM 309 FINAL fully solved and updated

What is research? - An attempt to answer a question/make a discovery 4 Ways of Knowing: Tenacity/Tradition - It has always been true - You eat turkey on Thanksgiving Tied to existing beliefs 4 Ways of Knowing: Authority - Someone (who should know) says so - Doctor - Parent Hard to change beliefs 4 Ways of Knowing: Intuition/Logic - Truth is self-evident/based on your experience 4 Ways of Knowing: Scientific Method - - Always open to new info - Incremental learning How often are papers retracted from journals? - 0.004% 2 Problems with Everyday Ways of Knowing - 1. we have existing beliefs, knowledge, etc that affect how we process info 2. inconclusive results - Absence makes the heart grow fonder - out of sight, out of mind 6 Characteristics of Scientific Method - 1. Public 2. Objective 3. Empirical 4. Systematic and Cumulative 5. Predictive 6. Science is self-correcting 6 Characteristics of Scientific Method: Public - - must be available for others to scrutinize - ability for others to replicate your work 6 Characteristics of Scientific Method: Objective - - procedures and rules must be followed - try to be as free of bias as possible - results know best 6 Characteristics of Scientific Method: Empirical - - can be tested, measured - types of definitions: - conceptual: - Grit-- firmness of charcter - intelligence: ability to learn - Operational: Procedures to measure a concept - Grit: setbacks don't define - IQ: spatial visualization, reasoning, etc. 6 Characteristics of Scientific Method: Systematic and Cumulative - - research should build on previous work - nothing is ever proven w/ science - goal is to develop theories and laws 6 Characteristics of Scientific Method: Predictive - science should help predict the future - science helps to explain, describe and explore 6 Characteristics of Scientific Method: Science is Self-Correcting - bad results/ideas will eventually be rejected/modified 6 issues researchers find when picking a topic - 1. Can problem be investigated? 2. is there potential harm to subject? 3. can the data be analyzed? 4. is the problem significant? 5. can the results be generalized? 6. what costs and time involved? 2 types of questions proposed by researchers - 1. Hypothesis: statement of testable, expected relationship between two concepts -ex: individuals who watch more tv see the world as a scary place 2. Research Question: when the researcher is unsure of the relationship between two concepts - ex: is there a relationship between social media use and exercise? 7 problems associated with developing research questions - 1. too broad 2. in comm we ask comm questions 3. underlying assumptions driving questions 4. make sure only one question at a time DON'T DO 5. we already have the answer--different topic 6. what if questions 7. should questions... implies something should happen 7 steps of the research process - 1. select topic 2. review extant research 3. develop questions 4. design a study 5. collect data 6. analyze and interpret data 7. inform others Focus Group - guided group discussion - recorded and transcribed - used a lot by marketers - analyze data 7 focus group steps - 1. define the problem 2. sample selection 3. determine # of groups 4. study mechanics - when, where, how, incentive? 5. prepare materials 6. conducting focus group(s) 7. analyze data 2 types of material for focus group - 1. consent form 2. protocol what steps are in a protocol? - - welcome - intro - consent form - questions/discussion - conclusion and thanks 5 types of Main Questions - 1. opening questions - ice breaker 2. intro questions - general impressions of topic 3. transition questions - guide participants toward key topic 4. key topics - 2-5 questions that address heart of the issue 5. ending questions - summarize, ask to clarify purpose of probing questions - follow-up questions - designed to elicit specific info 8 other types of questions - 1. leading 2. testing 3. steering 4. obtuse 5. factual 6. "feel" 7. anonymous 8. silence Moderator - - Leader of the group - keeps group on track - encourages equal participation - keeps people comfortable Common moderator problems - 1. personal bias - don't push convo where you want it to go 2. need to please client - same as personal, but for clients needs/bias 3. need for consistency - Don't ignore things that "don't fit" Where should moderator sit? - in front - position that reinforces leadership where should notetakers sit? - be present but inconspicuous how should the transcript look? - complete record of verbal comments - include ums, ah, pauses, grammar errors - don't include body gestures Internal Validity - - control over research situation to rule out alternative explanations - of interest to quantitative research External Validity - extent to which you can say that your results will be found across populations, settings and time - of interest to quantitative research Key differences between Positivism and Interpretive Perspectives - Paradigms: Positivism Perspective - - uses quantitative methods - uses tools to measure reality and test assumptions using stats - ex: surveys, experiments, etc. Paradigms: Interpretive Perspective - - Uses qualitative methods - understand how individuals interpret events and create meaning - ex: ambiguous movie endings, experience at WSU etc. Paradigms: Critical Perspective - - uses qualitative methods - examines power structures and their influence on society/media - men and women's role in society Assumptions of Qualitative data collection - - reality is socially constructed - everyone's experience is unique, can't be generalized - value held i deep understanding of situation Inductive Thinking - Specific observations to generalization (interpretive) Observation | Pattern | Hypothesis | Theory Deductive Thinking - general to specific (positivism) Theory | Hypothesis | Observation | Confirmation 4 checks for Qualitative data - 1. Use of multiple methods - Interviews, focus groups, etc. 2. Audit Trail - How data was collected (describe) 3. Check with members - Ask them if descriptions are correct 4. Check with research team - Everyone on same page Structured VS in-depth interviews - Structured - self explanatory In-Depth - Open, probing 3 types of in-depth interviews - 1. Life history - Autobiographies 2. Event centered - understanding what researcher can't directly observe (natural disaster) 3. Situational - understanding from multiple settings, situations and people (school) Interviews are great when... - 1. there are clear interests 2. you want to look past events/situations that are difficult to study 3. there is a shorter time frame 4. you want lots of perspectives Issues to consider before interviewing - 1. be clear about your motives 2. anonymity 3. final say (participant) 4. money- some people won't talk unless paid 5. logistics- when, where, how, who, etc. alternative ways to get data (aside from interviews) - solicited narratives - ask them to write about topic log-interview - log what they do in a day personal docs - diaries, journals, letters, etc 5 things to be aware of in an interview - 1. power dynamics - don't want too much of a hierarchy 2. Judging people - don't do it 3. talking too much 4. not paying attention - don't do it 5. coming off as cold - be sensitive field observations can vary in terms of... - 1. Extent to which you're involved in the group 2. Extent to which the group knows you're researching 4 levels/quadrants of field observation - 1. overt participatory - group knows, you work for them 2. overt non-participatory - observe group, don't work for them 3. covert participatory - work for group, and observe them without them knowing 4. covert non-perticipatory - they don't know you're observing them, don't work for them 3 types of Field Notes - 1. descriptive 2. analytic - your analysis using theories - expand later 3. autobiographical - your own behaviors and emotions 6 things to take notes of during field observation - 1. setting 2. participants 3. dialog 4. own feelings 5. own actions 6. what you don't understand 4 important steps to field observation - 1. choose research site 2. gain access 3. sampling - decide # of groups/settings 4. exiting - exit plan (esp. with covert) Case Study - examines a single "system of action" techniques of case study - - Field observation - textual analysis - focus groups/interviewing qualitiative research - exploratory - focus group - interviews - case studies - observation quantitative research - objective measurements and stats - surveys - experiment - content analysis conceptual (constitutive) definition - using words to describe a concept operational definition - procedures to measure a concept conceptual fit - you want fit between your operational and conceptual definitions - i.e. how your describe it to fit with how you measure it EXAMPLE conceptual (violence) ----------------------------------------------------- operational (# of hits, punches, etc) internal validity (and valid measure) - are you measuring what you think you are? metaphor: hitting the target = valid reliability (and reliable measure) - can you measure it consistently? metaphor: hitting the same spot = reliable 5 assessments of validity: face validity - the measure seems to look good on the face of it (it looks right) - should be done by expert not tied to study 5 assessments of validity: content validity - the measure captures the full range of meanings/dimensions of the concept - i.e. ALL aspects of concept 5 assessments of validity: criterion/ predictive validity - the measure is shown to predict scores on some other future measure - ex: 10th grade IQ ---- college GPA 5 assessments of validity: concurrent validity - measure discriminates between groups that should differ on key variable - ex: test for athleticism separates those who exercise from those who don't 5 assessments of validity: construct validity - the measure is shown to be related to other concepts that it should be related to (and not to those it shouldn't be) - ex: grit should be related to effort, grades, determination grit should not be related to shyness, anger, etc. ways to test reliability: stability - consistency of measure at different points in time How to test: test retest... same measure on same people at different time ways to test reliability: internal consistency - measure assigns similar values ways to test: A. Split-half: first 5 Q's get similar scores as second 5 B. cronbach's alpha: stats that assess whether items get at the same concept ways to test reliability: eqivalency - cross-test - develop two versions of a scale ways to test reliability: intercoder - see if results can be reproduced by others 5 assessments of validity - 1. face validity 2. content validity 3. criterion/predictive validity 4. concurrent validity 5. construct validity independent variable - cause - predictor dependent variable - effect - outcome nominal measure - variable is measured with categories - *numbers here are meaningless* categories must be: - mutually exclusive/no overlap - exhaustive/everything must fit in category weakest form of measurement, discreet ordinal measure - measured in rank-order categories - tells you placement, but not how far apart interval measure - measurement points with equal distance between them Ex: strongly neutral strongly 1 2 3 4 5 6 7 agree disagree - no true zero, continuous ratio measure - measurement points with equal distance between them - true, meaningful zeros - continuous Ex: how much tv..... 0hrs 1hr 2 hrs 3hrs 4hrs 5hrs 6hrs likert scale - - agreement - frequency - satisfaction ex: strongly---agree---neutral---disagree---strong agree disagree semantic differential scale - motivated 1 2 3 4 5 unmotivated uninterested 1 2 3 4 5 interested involved 1 2 3 4 5 uninvolved uninspired 1 2 3 4 5 inspired isomorphism - whether the measure corresponds with an actual number (reality) - ex: temperature means something *social science is bad at having theses correspondents - attitude/beliefs don't line up well with reality sampling - describes how you select your subjects representative/probability sampling - used to generalize to a larger population non-representative/ non-probability sampling - used when you are looking for a specific characteristic - can't generalize population - the entire group you want to study census - when the entire population is measured - ex: us census of poverty simple random sampling - - obtain a list of all population members - assign numbers to all members - randomly select numbers until desired sample size is reached ex: Phone sampling: random digit dialing systematic random sampling - - obtain a list of entire population - assign numbers to all members - randomly select start position in the list - select every "nth" element from list Stratified Random Sampling - - used to get representations from a certain group - divide population into "strata" or groups - Strata: variables/characteristics that are important ex: hair color, gender Multi-stage cluster sampling - - randomly select "clusters" - ex: zip code - randomly select participants within clusters *use this when you don't have a list of people* - can be combined with stratified margin of error - statistical estimate of what population looks like - how close your results are to actual population confidence level - the certainty that your results fall in a given margin of error sampling error - your sample will never be a perfect representation of the population because of chance - ex: flip a coin 10 times, unlikely that you land 5H and 5T - too many/too few non-sampling error - - measurement error A. random error: survey questions give you different results across samples - hard to detect B. systematic error: results wrong, but always in the same direction - could be correctable non-response error - you can select people randomly, but you cant make them participate - only really generalize to people who are willing to respond non-probability sample - any method where a member of a population doesn't have an equal chance of being selected *typically used in qualitative data* - most cases can't be generalized non-probability examples - 1. convenience sample: when subjects are selected base don availability to the researcher, not purpose the project 2. volunteer sample: individuals volunteer 3. purposive sample: research samples with certain number of subjects in a certain category primary goal of surveys - identify/describe attitudes or behaviors - examine relationships between variables measured 2 broad categories of survey questions - 1. open-ended: participants generate answers 2. close-ended: participants cheese from answers given 9 question guidelines - 1. be clear 2. keep questions short 3. avoid negatives 4. avoid double-barreled questions 5. avoid leading/biased questions 6. avoid asking difficult questions 7. avoid false premises 8. avoid embarrassing questions 9. response options should be... - match q, mutually exclusive, exhaustive social desirability bias - - over-reporting desirable behaviors - underreporting undesirable behaviors to overcome.... A. emphasize anonymity B. phrase questions to save face what to include in survey intro - 1. introduce yourself 2. introduce purpose 3. risks, benefits, time requirements 4. emphasize that you are NOT selling anything skip pattern - true or false: you should use a wide variety of question styles in a survey - false - use a limited # true or false: group questions using the same style together - true order effects - always consider: will questions that you are asking influence questions that follow? ex: asked first in a survey: worked with obama vs asked second in a survey: worked with a Rep. leader mail survey advantages and disadvantages - ADVANTAGES - cost - length - no interviewer influence - can have visual content DISADVANTAGES - must be self explanatory - low response rate - collection takes a lot of time 5 ways to increase response rates in Mail Surveys - - advance mailings - follow up meetings - university logo - personal touches - offer some sort of compensation telephone survey advantages and disadvantages - ADVANTAGES - quick data collection - moderate cost - ability to clarify - moderate response rate DISADVANTAGES - some interviewer influence - short questionnaire - no visual depictions - cell phones push poll - pretends to be a phone survey but is really to persuade - statements false or bend the truth Advantages and disadvantages of face-to-face - ADVANTAGES - clarification - audio/video possession - reliable/high response rate - length DISADVANTAGES - interviewer influence - expensive - lots of time Advantages and Disadvantages of online surveys - ADVANTAGES - can be the cheapest - quickest - in interviewer influence - wide range response rate - audio/visual content DISADVANTAGES - hard to get a reputable sample - "big brother" concerns - must be self-explanatory - questionable responses content analysis - systematically examining content of communication 10 steps of conducting content analysis - 1. develop research questions 2. define the population 3. select the sample 4. define unit of analysis 5. develop content categories 6. establish quantification system 7. train coders 8. code data 9. analyze data 10. draw conclusions 3 problems that could reduce reliability scores for coding sheet - 1. poor definition of categories 2. unclear instruction given to the coders 3. unanticipated content that is difficult to code Intracoder reliability vs Intercoder reliability - INTRA: agreement for same coder INTER: agreement between coders criteria for causation (3 key) - 1. IV and DV must be correlated 2. cause must precede effect 3. study must account for alternative explanations 8 threats to validity - 1. history 2. maturation 3. experimenter bias 4. testing/sensitization 5. regression to the mean 6. experimental mortality 7. contamination 8. sample bias/non-equivalent groups key elements of an experiment - - manipulation of IV - create "conditions" that isolate the "cause" - measure the effect (DV) after exposure to (IV) - control group-- no treatment - experiment group-- treatment - random assignments of participants *creates equal samples in the different conditions* ROX - R = random assignment O = observation (assessing the DV) X = manipulation/treatment (IV) post-test only control group - -----X------O1 x=play violent VG R ----- -----O2 o=aggression *test to see if Ag1 Ag2* threat to validity: sample bias - no way to assess if groups were different to begin with pre-post control group design - O1-----X-----O3 R x= violent VG O2-----X-----O4 o= fouls - test if difference between O3 and O1 is significantly different that that of O4 and O2 - provides info on change - threat to validity: sensitization solomon 4 group design - - addresses weaknesses in other designs - does two things: 1. see if treatment works (O3 vs O4 and O5 vs O6) 2. see if pre-test had any effect (O4 vs O6) Problems with solomon 4 group - - very time consuming - needs lots of participants repeated measures design - - every subject completes every treatment - every person serves as their own control Problems: