Test 2 Flashcards
Nominal
Classifies variables into categories Way to measure age: young, old Least amount of information Sex, Male/Female Qualitative
Ordinal
Classifies variables into ranked categories
Ways to measure age: infant, toddler, child, teen, adult, senior.
Ways to measure comm. competence: high competence, medium competence, low competence
Qualitative
Interval
Classifies and ranks, and establishes standard, equal distances between variables
Ways to measure age: 1-10 years, 11-20 years, 21-30 years, etc.
Quantitative
Fahrenheit scale, distance between 17 & 18 is the same as that between 89 & 90.
Ratio
Classfies and ranks, and establishes standard, equal distances between variables, with a true zero where variable ceases to exist
Quantitative
Specific age, aka 20
4 Primary Ethical Guidelines for the treatment of research participants
- Provide free choice
a. voluntary informed consent (cover letter, not forced)
b. child assent (people under 18 parent permission)
c. implied consent (mail survey, send back & get coupon) - Protect right to privacy
a. anonymity: cannot link given response to given participant, mail survey
b. confidentiality: you know participants but modify to keep secret, researcher can identify a persons response but promises to keep secret
c. security of data: locked for 5 years - Benefit people, don’t harm them
a. study natural conditions, do not create them (don’t inject cancer cells)
b. don’t withhold lifesaving treatment
c. debrief: answer all questions & reveal results - Treat people with respect
a. avoid being androcentric (male centered view)
b. avoid being racist
Does every research study intrude on the life of people?
Yes. Could be through reflection (marital satisfaction, childhood trauma, memories)
IRB:
Oversee research at the university level
-Review all research proposals involving human participants, main goal is to ensure that the risks faced by human participants in research are minimal
Sample
Special subset (small portion snapshot) of population observed to make inferences on general behavior
Goal of sampling?
- Generalize results of study
2. Minimize sampling errors (though every study contains error)
Representative Sample
Same distribution of characteristics as the population from which it was selected, characteristics closely matched so you can generalize, so random sample
Bias Sample
Those selected are not typical or representative of the larger populations, chosen for specific reasons, so cannot generalize, so non-random sample
Random Sampling
Relatively few observations and generalize to a much wider population, everyone has an equal change of being selected (names being drawn out of a hat)
-Probability sampling
Non-random Sampling
Lets us control the likelihood of specific individuals being selected for study, participants intentionally picked, and not everyone has an equal change (TA selection)
-Non-probability selection
Non-probability (non-random) Sampling Types
- Convenience: possible participants are there (stand outside of grocery story)
- Purposive/ judgmental: to study a subset that you want (like stratified) (Hispanic, women who smoke, and are divorced)
- Snowball: network
- Quota: representations of different groups (matrix)
Probability (Random) Sampling Types
- Simple random: totally random, draw from the available population, (close eyes and point to a number)
- Systematic: Ordered, structured, choose the nth subject (2,4,6,8,10)
- Stratified: layered (50 freshmen, 50 sophomores, 50 juniors, 50 seniors)
- Multistage cluster: Start with cluster, then do stratified selection (specific organization, western specifically 50 freshman, 50 sophomores, 50 juniors)
Considerations for probability or non-probability
- unit of analysis
- quantitative (want to generalize) vs. qualitative (doesn’t care about generalization)
- sample size
- lit. review
- accessibility of participants
- researchers resources/ permission
*More homogeneous the population is, the more representative the sample will be
True
Measurement
Careful deliberate observations for the purpose of describing objects and events in terms of the attributes composing a variable
Ethics
Professional standards that guide the research procedures acceptable for studying human beings, conforming to the standards of conduct of a given profession or group
Survey Research vs. Survey questionnaire
Survey research is operationalizing variables by asking people structured questions 1.) development of a questionnaire or survey 2.) selection of a sample of respondents 3.) administration of the questionnaire through any of several modes
A survey questionnaire is just a method
Why are surveys popular?
- Can target specific or large groups
- Easy to generalize
- Most widely used
- 2+ variables
- Cheap
- Easy to administer
- Creates confidentiality
Question Types
- Factual vs. Opinion
- Objective Truth vs. Evaluation (survey usually has combo)
- Closed vs. open
- Contingency/Filter (Filter: are you a comm major? contingency: if yes, answer q 1-37)
Contingency/Filter Questions
Filter: (are you a comm major?)
Contingency: If yes, answer q. 1-37
Survey Response Types
- Checklist: select all that apply
- Rank-order: 1-6
- Multiple choice: A-E, exam
- Yes/no: are you a comm major
- Linkert: 5 ft. scale, SA, A, N, D, SD
- Likert Type: Can be any other point, 7, 10, 4 etc. less points, less sensitive
- semantic differential: measures feelings/attitudes, asks respondents to choose between two opposite positions with adjectives on each end
Criteria Question Construction
- Clear question construction
- K.I.S.S., keep it simple, use exact expression, no excessive words
- No leading questions: suggests an answer, says “should”
- no double-barreled; asking for two things, says “and” “or”
- avoid being threatening
Order of Survey Questions
Organizational Patterns
- funnel (broad to narrow)
- inverse funnel (specific to broad)
- tunnel (steady throughout)
- topical: topics
- easy
- sensitive
- demographic items: race, age
Protocol, instructions
Name, organization Nature/purpose Confidentiality Directions Time Thank you
Milgram Study
Unethical toward the participants, benefit people dont harm them (psychological)
Challenges of asking good questions
- Memories can be biased and affected by emotional states
- What people say they do, or will do, are not always related
- small changes in wording result in different responses
- the order of questions affect answers
- respondents answer even when they don’t know about the subject
Unidimensional composite measure
When a variable has a single facet, all of the questions will focus on the single facet
Multidimensional composite measure
Several facets, set of questions focused around each underlying facet or dimension
Response rate
Divide and then multiply by 100, 50% is adequate, 60% is good, 70% is very good
Modes of survey questionnaire administration
Mail distribution and return:+ confidential, Follow-up mailings, -certain populations are more likely to respond,
Standardized interview
