Flashcards in PSY395 Exam 3 Deck (55):
Aggregate of all of the cases that conform to some designated set of specifications.
One or more specifications that divide a population into mutually exclusive segments.
A single member of a population.
A count of all the elements in a population.
A subset of the population.
One can specify for each element of the population the probability that it will be included in he sample.
No way to estimate the probability each element has of being included in the sample and no assurance that every element has some chance of being included.
Simple Random Sampling
A process that gives each element in the population an equal chance of being included in the sample, but also every combination of the desired number of cases equally likely.
Representative sampling plan
A plan that ensures the odds are great enough so that the selected sample is sufficiently representative of the population.
Divide the population into two or more strata and then take a simple random sample form from each stratum.
Groups or sets of elements using simple or stratified sampling, then randomly choose a cluster.
Using cases at hand as a sample.
A selection of a sample that is a replica of the population to which one wants to generalize. But, uses a convenience sample to meet the quota.
Handpicking cases to be included in the sample.
-following an election district that in the past has been predictive of the election.
A small initial sample :snowballs: into a sample large enough to meet the requirements of research design and data analysis.
Hard to get groups/samples (drug users).
Any instrument that asks people to report their own attitudes, feelings, perceptions and beliefs.
Goal of designing a study
-Attitudes and beliefs?
-Facts and demographics?
--"Does the sun revolve around the earth?"
--How often do you exercise/go to church?
Wording Problems on Surveys
Question wording, question order
Identity of survey taker/interviewer
Simple wording changes
Loftus & Palmer, 1974
'homosexuals' vs 'gay man and lesbians'
Final judgments and behaviors are assimilated or become more similar to an initial anchor value.
Do you have a family history of cerebrovascular accident? Stroke. It's a fucking stroke.
Avoid asking multiple things at once
pay AND job conditions
interesting AND useful
Loaded (Leading) Questions
Designed to elicit a certain response.
-Should the mayor spend more tax money to keep the city's excellent streets in super shape?
Does it seem possible or does it seem impossible to you that the Nazi extermination of the Jews never happened?
--Affirmative response because of confusion about the word 'never' here mixed with impossible
Does it seem possible to you that the Nazi extermination of the Jews never happened, or do you feel certain that it happened?
--Only 1% reported that it might not have happened.
Questions where agreement means disagreement
-Do you think that Congress should not pass a new gun control law?
Questions that assume an answer to the question they ask
There are many people who believe that campus parking is a problem. Are you one of them?
Question Order Problems
Contrast Effect: order results in greater differences in responses
Assimilation effect: responses are more similar as a result of their order
Context effect: could get very different answers to question about "How satisfied are you with your life?"
Identity of Survey Taker
A black vs. white interviewer asking about racial attitudes
-More likely to get socially desirable responses, not honest ones for certain combinations
Acquiescence Response Bias
Tendency to check the same response category in a series of similar questions (yea-sayers, nay-sayers)
-Resolve by including some reverse-coded items
Responding to look good, not to respond accurately.
-Add inducements for accurate reporting (Bogus Pipeline - trick into believing researcher can read true feelings)
-Include filler items to Mask/Obscure true purpose
-Include catch items ('I have never lied')
-Cross check with Behavior measures (self-reports)
-Avoid self-reports altogether (instead use implicit/covert measures - the measurements of attitudes using unobtrusive techniques like physiological measures or behavioral observations)
Requires an unstructured, free response
-Thematic Apperception Test (TAT)
---unreliable, not valid, lots of training required
-Written Text Question (requires coding scheme that is qualitative or quantitative
Closed-ended (Fixed-coice) Question
Asks for structured, fixed response
-set of fixed options
--Categorical: no meaningful numerical values assigned to responses (probably nominal)
--Dichotomous: clear-cut yes or no, good for symptom checklists, public opinion surveys, but limits ability to do statistics
--Rating Scales: Likert rating scales (5 or 7)
--Semantic Differential Scale: respondents rate any concept using 7-point, evaluated on one of three attitudinal dimensions (evaluation, activity, potency)
--Graphic Rating Scale:like Likert, but responses are more continuous, precise (placing X on continuum)
Anonymity/reduced bias, low cost, wide geographic area, convenient
low response rates/biased sample, no control over conditions when survey is completed, can't answer questions of participants, limits on question types
Fast/inexpensive, no limits on question types and visual aids, contingency questions possible
Overused/poorly conducted, sampling, conditions when survey is completed
Quickly/w/large staff, sequence of questions, computers speed data entry, random digit dialing makes random sample more likely
Selection bias, limits on question types
High response rate, long/complex surveys, clarify/answer questions, control over order of questions, control/report conditions
Expensive, sampling problems (homeless), interviewer bias
--Decide what info/how to administer
--Write a draft
--Pretest, edit, finalize procedures
--Put demographic questions at the end
Key components of experiments
1. Manipulated independent variable (at least 2 conditions, control of extraneous variables)
2. Statistical testing of hypotheses (ANOVA, differences due to chance is
Benefits and limitations of experiments
Benefits: inferences for causality, theory testing, interactions between variables
Limitations: low generalizability, a few variables at a time in an experiment, difficult/impossible, must know a lot about phenomenon prior, relatively expensive.
Quasi-experiments (Grouping Variables)
IV = GV (not manipulated)
---Gender/ethnicity, pre-existing attitudes, personality or trait variables
Why manipulate more than one IV?
More efficient/more control.
Results valid across wide range of conditions are more valuable.
Identify where the results are not valid (C moderates relation between A & B).
Use theory, want it to be reliable/valid.
Be aware of ceiling/floor effects
Facilitated Communication: when knowledge of the facilitator is controlled for, the autistic child's performance is found to depend on tactile cuing by the facilitator.
---These could be seen for what they are if a control group is included.
Each person in your sample has an equal likelihood of being in any condition ----causality
Each person in the population has an equal likelihood of being in your study. -----generalizability
Is the study logic solid?
Did the experimental stimulus make some significant difference in this specific instance?
--Establish with random assignment to condition and/or order control of extraneous variables
-ambiguous temporal precedence
-selection (cohort effect)
-additive/interactive effects of threats to internal validity
External Validity (generalizability)
To what populations, settings, independent variables, and dependent variables can this effect be generalized?
To what people can you generalize?
Allows you to assume that the results of your study are true for some larger group of people.
Allows you to generalize to the POPULATION from which you RANDOMLY SELECTED (only).
What are the limitations of the experimental setting compared to real life? - research setting different from other settings, variables operationalized.
Threats to external validity
Interaction of the causal relationship with units (does effect found in one population hold for another population).
ICR over treatment variations (effect for one treatment may not hold with other variations).
ICR with outcomes (does treatment hold across different DVs)
ICR with Settings (does treatment hold across different settings).
The experimenter has created differences not just on the independent variable, but other variables too.
Reactivity (Hawthorne Effect)
Participant's behavior is influenced by he knowledge that they are being observed.
Cues available to participants in an experiment that enable them to figure out what is expected the experimenter. May alter their natural behavior.
Experimenter Effects (Observer Bias)
Treating Ps differently depending on the condition they are in.
-Use "double-blind" experiments