research methods 2 ; appendix A, ch6, ch7

nonparticipant observer

outsider who does not become an active part of setting

observational methods

qualitative and quantitative - almost always non experimental - less artificial--usually in real world - no manipulation or random assignment - researcher's observations create data

systematic observation ex

- 3 year olds on video tape in room -- free play situation - each child taped for 100 minutes, observers viewed tape, coded child's behaviors every 15 seconds CODING SYSTEM: - unoccupied = child not doing anything - solitary play = child plays alone with toys, not interested/affected by activities of other children - together = child is with other children not really doing anything - parallel play = child plays with other children with similar toys but does not play with them - group play = child plays with other children sharing toys/in organized activities as part of group - can find sequence/order in which different behaviors were engaged (found parallel to group lay was frequent

Body of paper

- 4 sections: intro, method, results, discussion

citing references page APA

- Harry and Mark (1991) found this - in one study (Harry, 1991), people learned to do this - people learned to do this (Harry and Mark, 1991). - 3 TO 5 authors: first time = bob, harry, and mark (1991) reported xyz - second time = There is no relationship between X and Y (Bob et al., 1991). - titles of books are italicized, only first word of title is in CAPS,

parts of APA paper

- Title Page - Abstract - Body of the paper - Introduction - Method - Results - Discussion - References - Figures & Tables hourglass shape (general to specific to general)

sampling frame

- actual population of individuals/clusters where random sample will be drawn - bias: if you define population as residents of a city, sampling frame will be list of phone numbers used to contact from 5-9pm --> excludes those without phones, who's number not listed, or cannot pick up then

question wording: yea saying or nay saying

- agree or disagree--person may agree with topic but also may just want to agree with anything you say - should word questions so that consistent agreement is unlikely (i spend most of weekends with friends, i spend most of time with family)

population

- all individuals of interest to researcher (eligible voters in US, all students in college)

participant observation

- allows researcher to observe setting from the inside/experience things the same way as natural participants - has active role - can give valuable data

telephone interview

- almost always done for large scale surveys - less expensive than face to face, quick data collecting - can have computerized phone surveys for cheaper cost (less labor and data analysis costs) - use CATI computer assisted telephone interview system--questions are on computer screen and data directly goes for analysis

theoretical articles

- are about the status of existing theories - might propose changes in theory/development of new theory

rating scales

- ask people to provide how much judgement on agreement, liking, or confidence - ex: students in college should be required -- strongly agree to strongly disagree; how confident are you that she killed him

survey issue

- assume questionnaires and interviews have people who are willing and able to provide truthful and accurate answers

question wording: negative wording

- avoid negatives--do you feel city should not approve women's shelter - agreement with the question = disagreement with proposal which might confuse people = inaccuracy

naturalistic observation issues

- be a participant or nonparticipant - can lose objectiveness if involved--bad for conducting scientific observation--even more hard to maintain if researcher is member or ex-member of group - best not to give researcher reason to criticize people in setting or praise them---> prevent bias - conceal purposes or not - known presence of researcher can change the way people behave normally - nonconcealed is best ethically (invasion of privacy) - can be a participant and nonconcealed only to some - can be a nonparticipant but nonconcealed

abstract

- brief summary of research report - on page 2 - 150 to 250 words - provide info for reader to see if want to continue - talk about four main sections in body (intro, procedure, results, discussion)

naturalistic observation

- called field work/field observation - researcher makes observations of individuals in natural environments--normally social/organizational settings (students/teachers in school, sports team, wolves in a pack in mountains) - need accurate description and objective interpretation with no prior hypotheses

confidence intervals

- can be 95% confident that the true population value lies within this interval from sample result - ex: 3 percentage points using 95% level of confidence = sample survey says 61% prefer to study at home so actual population is about 58 - 64%

naturalistic observation limitations

- can be hard to do - most useful when investigating complex social settings to understand them and develop theories based on observations--but not that good for studying well defined hypotheses for specified conditions - cannot be scheduled for time/place--not like lab - procedures are not well defined/same for each participant, data analysis not planned in advance - pattern of events always changing, some important/unimportant but has to all be recorded anyway - analysis after study is not easy--must sort through every incident over multiple times - hard to make hypotheses to explain data and make sure data is consistent

mail surveys

- can be mailed to home or business - much cheaper than contacting people selected for sample - but low response rates--thrown out, forgotten, bored, distracted--no one to help if confused

appendix APA

- can include entire questionnaire used in survey

results APA

- clear description of analyses - state probability level used to make decision about statistical significance, which stat test used (t test) - summarize all findings of stats test used to test hypothesis - NO interpretation of results

Cross-sectional

- collect data at only one point in time

longitudinal

- collect data from same people at multiple time points (waves)

statistical records

- collected by public and private organizations (US Census Bureau has the most) - ex: see if high temperature raises aggression

sampling error

- confidence interval gives you info about likely amount of error--obtained score may deviate from true score from sampling error

nonprobability sampling: haphazard sampling

- convenience sampling; take them where you find them--even if doesn't represent sample well - select sample of students from school any way convenient - EX: standing on a busy street during rush hour so that you get as much people as possible without moving, even though most of those people are probably middle class and white collar, skewing your data's representation for a population.

wording of questions--issues

- difficult to understand question - unfamiliar terms, vagueness, imprecise, bad grammar, misleading info in question, memory overload (did your mom, dad, full blood sister, full blood brother, daughter, son ever have heart attack or stroke) -- QUESTION is too long and too many relatives to keep track of, two different diagnoses, maybe dont know what stroke is)

probability sampling

- each member of population has specific probability of being chosen - needed to make precise statements about specific population - con: can be more costly/difficult to get list of members - pros: representive of population

nonprobability sampling: quota sampling

- researcher chooses a sample that reflects the numerical composition of various subgroups in the population -- NOT RANDOM - ex: you want 10 freshman, 10 sophomores, 10 juniors, and 10 seniors, but only find 8 sophomores on the street so you go to a sophomore class to find more.

probability sampling: simple random sampling

- every person of population has equal probability of being selected for sample - ex: school--list of all students in school, randomly select from list; computer generated telephone numbers to call = RANDOM SAMPLE - You can use a coin toss or a random number table - if the target population were asian students at GW--you would take a list of all the asian students at GW, order them alphabetically by last name, number them, and use a random number generator to choose from the list that correlated with the random number.

constructing questions to ask: defining research objectives

- first determine explicitly the research objectives - what do they want to know? - tie survey questions to research questions - straightforward and easy to answer

footnotes and tables apa

- footnotes added NOT in body of text but in one page at end of paper - THEN table on separate page at end of paper - THEN figures on separate page

refining questions

- give questions to small group and have them think aloud--use info /advice/responses to improve questions

discussion APA

- go from specific to general - Restate findings, Discuss findings in context of literature review/theory, Limitations, Future directions - discuss implications of the results - can summarize original purpose and expectations, then if results were consistent or not - say how your findings contributed to problem being investigated - explain why if results didn't agree with expectations - good to criticize own study--point out problems (time, money, confounding variables) - apply results to real life--larger issues

survey research

- has questionnaires and interviews to ask people to provide info about themselves (attitudes, beliefs, demographics / age, gender, income, marital status) - PARTICIPANTS provide data

survey archives

- have data from surveys that are available to analyze - have major polling organizations (universities)--part of the Inter-University Consortium for Political and Social Research (ICPSR) which makes survey archive data available or part of General Social Survey (GSS) - surveys range from attitudes, life satisfaction, health, religion, education, age, gender, race - now available online - CHEAP--dont need to do big national samples

case studies

- have descriptions and interpretations of research with specific individual/group/organization

qualitative approach example: describing behavior

- have many focus groups of 8-10 teens - discuss about their perceptions and experiences with work in world - talk to you in their own words/thinking - record with video or audio tape, have a transcriber, observers for note taking - qualitative descriptions = themes from discussions and how teens conceptualized them - use language and images - teen employment example: researcher works in fast food restaurant as a trainee

probability sampling: cluster sampling

- identify clusters of people to sample from those clusters - everyone in cluster is in sample - ex: randomly select from list of classes then get everyone in class in sample - randomly choosing 10% of nursing homes in the country and making everyone in that nursing home answer your survey, instead of randomly getting 10% of people from every nursing home in the country to fill out a survey

focus group intervew

- interview with group of 6-10 brought together for a period of 2-3 hours - members selected with particular knowledge or interest in the topic - participants usually get gift if spend time and cost to attend - questions ask whole group, open ended, can have group interaction/trigger comments - problem: one person tries to dominate discussion or be hostile to others - recorded to find themes of consensus and disagreement--usually do many discussions - time consuming costly but provides a lot of info

face to face interview

- interviewer and respondent meet for interview - can go to person's home or office for interview - expensive, time consuming small sample size

interviewer bias

- interviewer could subtly bias respondent's answers by showing approval or disapproval for certain answers - different interviewers (attractiveness, age, race) might influence answers - answers could be interpreted so matches with interviewer's expectations - should train interviewers to limit bias

sample size

- larger sample size will reduce size of confidence interval = more likely to yield data that accurately reflects true population value

david winter

- leadership study - what makes a leader great/popular - theories of presidential greatness: leader appeal (charisma, qualities), leader situation match (characteristics of successful leaders depond on situation), Leader-follower match (own identities/conflicts match with followers)

experimental design

- manipulations create data

semantic differential scale

- measure of meaning of concepts - respondents rate any concept (person, object, behavior, idea) on 7 point adjective scale - ex: smoking cigarettes good ------- bad (evaluation--wise foolish, kind mean) strong ------- weak (potency--hard soft, large small) active ------- passive (Activity--excitable calm, slow fast) - ex: abortion, weed, governor, university

3 social motives

- need for achievement/excellence --> moderate risk taking, using feedback, and entrepreneurial success - need for power/impact/prestige --> need formal social power, might do impulsive actions such as aggression, drinking, and taking extreme risks - need for affiliation/close relations with others --> need interpersonal warmth, self-disclosure, and good adaptation to life - use these social motives to code the leadership study! david winter - found that personality and popularity depend on situation and is very complex

graphic rating scale

- needs a mark along continuous 100 millimeter line ex: how would you rate that movie? not very enjoyable to very enjoyable - then use ruler to measure out 0 to 100 range and put a mark on the line

why APA

- norm for psychology papers - following guidelines makes it easier for reader to understand/evaluate report

case study

- observational method that provides description of individual (person, setting, school, neighborhood) - harder to generalize - similar to naturalistic observation - can be description of a patient by a psychologist - can present the individual's history, symptoms, behaviors, reaction to situations, response to treatment - usually done of someone with rare, unusual, noteworthy condition to provide unique info - EX: S--amazing ability to recall info ; Genie--girl tied to a chair in isolation, never spoken to so had no language abilities, RM--severe limbic system damage--tested to see if he could detect violations of social exchange rules and precautionary action rules--he did well on precautionary action rules for hazardous behavior but badly on social exchange rules due to his damage

Graham: aggression

- observed specifically aggression in bars in large city late on weekend nights - also naturalistic research (did not try to influence what was happening in the settings)

references apa

- on a new page - include only sources mentioned in article

optimists vs pessimists (negative events)

- optimists: expect positive outcomes, can cope with difficulty ( in negative events: external, unstable, specific -- does not let negative events affect them) - pessimists: expect negative outcomes, think they cannot cope with difficulty ( in negative events: internal, stable, global) - pessimists go see doctor more (sick more often) - optimists will be healthier

coding systems

- part of systematic observation - if researcher needs to decide which behaviors are of interest-->>choose setting to observe behaviors and develop coding system - used to measure behaviors - ex: hostility = blames other, provokes partner - affection: expresses concern, agrees with partner

personal administration of questionnaires to groups/individuals advantage

- people can ask questions - get captive audience at once who is likely to complete (college class, orientation)

response rate

- percentage of people in sample who actually finished survey - if asked 1000 and 500 finished, then 50% - indicates how much bias will be in final sample

probability sampling: stratified random sampling

- population is divided into subgroups (strata) and use random sampling techniques to select members from each strata/group - divide members into dimensions (sexual attitudes divided into age gender and education) - if you wanted to find out if students who stayed after midnight in the library were doing better in school. If you had existing data that students who sleep over 8 hours a night on average tend not to stay in the library past midnight, then you would divide the students into sleep over 8 hours a night on average and those who don't. Then randomly select the same amount from each group.

reactivity

- possibility that presence of observer will affect people's behaviors - can be reduced by concealed observation (small microphones/camera help even if participant knows about recording) - can be reduced by allowing time for people to get used to observer and equipment

formatting questionnaire

- professional, attractive, neat, no spelling errors - easy to identify questions and response choices - use consistent scales - prepare sequencing of questions (most interesting/important first to motivate responders)

naturalistic observation goals and requirements

- provide complete and accurate picture of what happens in the setting, not test hypotheses made prior to study - first: describe setting, events, persons observed - second: analyze what is observed - need to keep detailed field notes (write/dictate/record regularly)--carefully watch/listen - best to report by two or more people on several different occasions to support interpretations - normally qualitative but can be quantitive if about income, family size, education level, age, gender

method APA

- provides reader with detailed info about how study was conducted - have enough info so reader can replicate study - can have subsections (overview, participants, design and materials, and procedure) - participants --#, age, sex, ethnicity, other characters, how they were recruited, incentives used - materials -- special equipment needed, have operational definition of variables, info on reliability and validity, response options - procedure -- how study was conducted/collected data, describe step by step (randomization, controlling extraneous variables), measurement techniques, describe if any manipulation or deceiving - ex: each student in class invited at least 10 people to complete an online survey about x, y, and z. Participants were directed to a url. No compensation was provided.

nonprobability sampling: purposive sampling

- purpose is to obtain a sample of people who meet some predetermined criterion - you want to know about moves so go to theater--can filter for people under 30 or those with kids - ex: you are doing a study about dietary habits so you sit in whole foods for your survey.

question wording: loaded questions

- question is written to lead people to respond in one way (do you want to eliminate waste excess, do you want to reduce public school budget--words like rape, waste, immoral, dangerous can influence way people respond and bias conclusions)

questionnaire

- questions in written format and answers are written down - pros: less costly than interview, allow respondent to be anonymous if no name, SSN, drivers license - cons: require reader to be able to read and understand questions, can be boring (no motivation) - can be done in person, mail, internet, computer

question wording: double barreled questions

- questions that ask two things at once - should we give money for X and Y? -- could have different attitudes for X and Y

interviews

- require interaction between people - more likely to agree to answer questions (easier to convince them to participate than mailed questionnaire) = higher response rates - motivate to complete all questions - interviewer can clarify any problems and ask follow up questions

why surveys?

- research tool used to ask people about self - provides data about issues in society/for studying behavior (customer satisfaction, student services) - ex: college survey from telephone survey on politics - without surveys, would only have stories heard or letters written by customers - allows researchers to study ways attitudes/behaviors change over time

labeling response alternatives

- respondents decide meaning of the choices that are not labeled - if strongly agree _ _ _ strongly disagree --> can assume that middle is neutral, left is agree, right is disagree - less than twice a week, twice a week, 4 times a week, 6 times a week, once a day = HIGH FREQUENCY - less than once a month, once a month, once every 2 weeks, once a week, more than once a week = LOW FREQUENCY

open ended questions

- responders are free to answer whatever they want - need time to categorize and code--more costly - but can provide valuable insight into what people think--see how they naturally view their world/think

sampling participants

- selecting people to study from population of interest - selected randomly to estimate characteristics of the population as a whole

title page

- separate page, numbered page 1 - short, informative title of report - author and affiliation (university) - can have author note in lower half page (how to contact info, acknowledge assistance of others/funding) - have running head in caps (summary of title) in header

nonverbal scales for children

- smiley or sad face for level of pain

most common response set

- social desirability / faking good --> leads to person answering in the most acceptable way -- the way most people "should" respond or way it reflects most favorably - worst when about violence/aggression, drugs, sex - can reasonably expect participant to be honest if promise confidentiality, tells purpose of study, and provides feedback

Sylvia Scribner practical thinking research

- studied how people in different jobs made decisions and solved problems--NATURALISTIC research - drove on 3am milk route, helped cashiers with receipts, watched machine operators--recorded how they went about performing jobs - photographed/collected copies of all written materials they wrote or read, all devices they used to work (scales, thermometors) - found that milk drivers need to do complex math--so they have learned to use efficient strategies to solve problems at work

apa writing style general info

- tell a story--have beginning middle end - assume reader has general psych knowledge but nothing in depth about your topic - use active voice not passive (I TOOK the test, not I was given a test to take) - usually use third person - reduce bias - say participants, subjects, or respondents - be specific when referring to groups of people (korean american vs asians) - double spaced, 1 inch margins, have headers with page number, times new roman or Arial

clarity apa

- think about intended audience--assume they are unfamiliar with your topic/methods of study - don't use jargon incomprehensible - can use an outline to organize ideas - paragraphs have topic sentence and related to it

introduction APA

- title at top of page - very general intro to topic - problem under study (give reader appreciation of context and significance of topic being studied) - literature review (say how previous research/theory is connected to your research problem) - rationale and hypotheses of the study (state what variables you are studying and what results you expect) - how is your info new/important - hypothesis: specific research prediction, What relationship do you expect to find between variables

equipment

- to directly observe in a classroom or interactions, use paper and pencil measurements--observe and record at same time - but video recording is more common now--advantage to provide permanent record of behavior to code later - use stopwatch to time duration

administer survey

- use written questionnaire (respondent reads q and writes down answer) or interview (ask q and record response personal verbal interaction)

internet surveys

- very easy--have open and closed end questions - immediately sent to researcher when done - hard to sample people--only get people interested in the topic - is getting easier to organize people willing to take surveys--and obtain samples of people with particular characteristics (job, status, age) - use social networking sites, email discussions, bulletin boards, chat rooms to exchange ideas/info - problem: results might not be similar to what would have been with traditional methods - ambiguity of characteristics of responders (misrepresent age, gender, ethnicity)

longitudinal study/panel study

- way to study changes over time - same people are surveyed at two or more points in time (two wave panel, three wave panel) - variables of time 1 and time 2 etc - ex: exposure to sexual content on TV and teen pregnancy over time -- surveyed three times over 3 yrs

non probability sampling

- we don't know probability of any particular member of population being chosen (arbitrary/nonsystemic) - con: don't ensure that sample accurately represents population, likely to introduce bias in sample --> if Fox news asked viewers to do survey (conservatives) ; if passionate topics are asked about (abortion, taxes, war--brings certain types of responders) - cheap, convenient (students will participate if need credit from school)--can be more representative (outweigh the possible bias)

apa citing

- whenever you directly quote, paraphrase, or use someone else's idea - Smith and Jones (2009) found that trees are green. - Trees are green (Smith & Jones, 2009). - "et al." = and everyone else --> 3-5 authors - references: first line is at the front, rest of lines are indented

written and mass communication records

- written records: diaries, letters from historical societies, ethnographics of other cultures, public documents (politician speeches, discussions by internet users) - mass communication: books, magazines, movies, TV programs, newspapers

number of response alternatives

- yes/no, agree/disagree - can give more alternatives for people to express opinions (5 or 7 point scale from strongly agree to strongly disagree -- very positive to very negative)

question wording: simplicity

-question should be easy to understand and respond to - avoid unfamiliar terms and jargon - if need to, can provide description of content or be more descriptive (less vague) of what you want to know

3 types of survey questions

1. attitudes and beliefs: focus on ways people evaluate and think about issues (are you happy with how police responded to your call, evaluate professor) 2. facts and demographics: ask people what they know about themselves/situation--age, gender, ethnicity, income, marital status, employment -- depends on survey (health survey, car products) 3. behaviors: focuses on past behaviors or intended future behaviors--how many days did you exercise last week, how many children do you plan to have

closed ended questions

limited number of response alternatives are given--list to choose from - more structured--easier to code and alternatives/choices are same for everyone

archival data pros and cons

allow to study interesting questions that couldn't be studied any other way - valuable - but records may be hard to obtain (placed in lost storage or destroyed) - never sure of accuracy of info collected by someone esle

Systematic observation

careful observation of one or more specific behaviors in particular setting (is it race and economic status?) - only interested in specific behaviors, not global - observations are quantifiable, usually has prior hypotheses about behaviors

literature reviews

describe RELEAVANT past research in specific area of psych--integrate research findings, evaluate current status of research on topic, and point to new directions for research

quantitative approach example: describe ways in which lives of teenagers are affected by working

develop questionnaire--ask teens to complete - ask numbers of hours they work - type of work - levels of stress - school grades - drug usage - then can do statistical analysis (% of teens who work and age) - teen employment example: examine data from state department of economic development

Method issues

equipment, reactivity, reliability, sampling

sampling

extended periods give more accurate and useful data than single short observations - short data can be disorted by short term trends (particular meal served, illness, recent event (death)) - sample over breakfast, lunch for 6 weeks--randomly choose person to be observed for a 3 minute period for 10 days both meals

qualitative research

focuses on people behaving in natural settings and describing their world in own worlds - emphasizes collecting in depth info on few individuals or within limited setting - conclusions based on interpretations made by investigator

explanatory style

how people explain to themselves why they experience a particular event - internal vs external -- how one explains the cause of event (your fault or something else did it) - stable vs unstable/transitory -- how one explains the extent of the cause (consistent vs changeable) - global vs specific/local -- how one explains extent of effects (generalizable or restrained -- do well on all tests or just this one)

reliability

refers to degree to which measurement reflects true score, not measurement error - reliable measures are stable, consistent, and precise - use 2 or more raters to code behavior - indicated by high agreement among raters--very high agreement is best - can train observers to use equipment and be checked for agreement with previous observers

empirical studies

reports of research that were conducted by the writers/authors of the report

content analysis

systematic analysis of existing document - needs researchers to make coding systems to quantify info from documents

response set

tendency to respond to all questions from particular perspective rather than to provide answers that are directly related to questions - can affect usefulness of data from self reports

quantitative research

tends to focus on specific behaviors that can be easily quantified (counted) - usually have larger samples - conclusions based on statistical analysis of data

psychobiography

type of case study where researcher applies psychological theory to explain the life of someone--usually someone historically important

archival research

uses previously compiled information to answer research questions - do not actually collect original data - analyze existing data (stats from public records--divorce petitions filed, reports, letter contents, database