# PSY395 test 3

## Terms

undefined, object

copy deck

- Set of all people, objects, or events of interest to the researcher
- population
- A variable that divides the population into mutually exclusive segments
- stratum
- e.g., gender, SES, politics
- stratum examples
- A single member of the population
- population element
- A subset of the population used in an experiment
- sample
- A count of all the elements in a population
- census
- 2 goals of sampling
- Maximize external validity, minimize threats to internal validity
- If you can specify for each element of the population the probability that it will be included in the sample, you are using a...
- ...probability sample
- Makes representative sampling plans possible
- probability sample
- Allows investigators to figure out which findings are likely to differ from actual population
- probability sample
- Can specify size of sample needed if they want a specific degree of certainty
- probability sample
- A distribution of sample means
- sampling distribution
- The discrepancy between the sample and the population
- sampling error
- Specification of the population from which elements are drawn to form a sample
- sampling frame
- Divide population into strata and take a simple random sample in each subgroup
- stratified random sampling
- Can oversample for a particular group if you want more statistical precision for that group
- stratified random sampling
- Representative of both population and key subgroups
- stratified random sampling
- Divide population into geographic clusters, randomly sample clusters
- cluster random sampling
- Use when population is spread out
- cluster random sampling
- Combination of stratified and cluster
- multi-stage sampling
- Does not involve random selection, there is no way to estimate the probability each element has of being included in the sample
- nonprobability sampling
- Hard to know whether population is well-represented
- nonprobability sampling
- e.g., college students, clinical practice samples
- examples of convenience sampling
- One or more specific groups being sought
- purposive sampling
- e.g., people in a mall with a clipboard looking for young Caucasian females
- purposive sampling
- Sampling most frequent or â€œtypicalâ€ person
- modal instance sampling
- Sample of people with known expertise
- expert sampling
- Select people nonrandomly according to some fixed quota
- quota sampling
- Represent major characteristics of a population by sampling proportional amount of each characteristic
- proportional quota sampling
- Specify minimum number of sampled characteristics you want in each category
- nonproportional quota sampling
- nonproportional quota sampling is similar to...
- ...stratified sampling
- Use when you want to include all views, but it doesnâ€™t matter if theyâ€™re presented proportionally
- heterogeneity sampling
- Opposite of modal instance sampling
- heterogeneity sampling
- Useful for brainstorming
- heterogeneity sampling
- Only research that supports causal inferences
- randomized experiments
- strength of randomized experiments
- internal validity
- weakness of randomized experiments
- lower external validity
- People bring them to the study, itâ€™s not possible to manipulate them
- individual difference variables
- Variables that the experimenter can manipulate or expose people to
- experimental variables
- e.g., suburban all-boys private school vs. inner city coed public school
- examples of confounds
- e.g., theft-ice cream sale relationship
- example of a third variable
- An unintended effect on the DV caused by some feature of the experimental setting, not the IV
- artifact
- Reduces impact of alternative explanations/confounds for effect of IV on DV
- random assignment
- Used after we have a sample, and before theyâ€™re exposed to treatment
- random assignment
- Compare differences among groups
- between-subjects experimental design
- Each subject experiences one level of IV
- between-subjects experimental design
- Both groups get pretest and posttest
- Pretest-posttest two group design
- Rules out selection and maturation as threats to validity (2 designs)
- Randomized two-group design, pretest-posttest two group design
- Provides check on history and instrumentation threats (2 designs)
- Randomized two-group design, pretest-posttest two group design
- Independent measures t-test
- Randomized two-group design
- Repeated measures t-test
- Pretest-posttest two group design
- 2 controls, 2 experimental groups
- Solomon four-group design
- One of each gets pretests, one of each does not, all get posttest
- Solomon four-group design
- ANOVA
- Solomon four-group design
- 2 IVs, presented in combination (X1/Y1, X1/Y2, X2/Y1, X2/Y2)
- Between-subjects factorial design
- measure differences in subjects over time
- within-subjects
- Each subject experiences all levels of IV
- within-subjects
- 2 IVs, one within and one between
- mixed design
- Researcher manipulates something by accident
- procedural confounds
- Measure does not map onto construct
- operational confounds
- Preexisting differences between individuals
- Selection threat to internal validity
- Effects of time on individual
- Maturation threat to internal validity
- Events that affect the study
- History threat to internal validity
- Changes in measurement
- Instrumentation threat to internal validity
- May result from experienced raters, fatigued raters, changes in a survey
- Instrumentation threat to internal validity
- Participants leave study, maybe at differential rates
- Mortality threat to internal validity
- Changes in time with the intervention
- Selection by maturation threat to internal validity
- The degree of resemblance between laboratory operational definitions and some targets/objects outside the lab
- mundane realism
- The extent to which manipulations or measures are truly perceived in the intended ways by the research participants
- experimental realism
- What might happen
- basic research
- controlled setting
- basic research
- what does happen
- applied research
- real-life setting
- applied research
- Concerned with between-treatments variance
- experimental research
- Derives hypothesis from theoretical premises and tests it
- experimental research
- Treat everyone the same
- experimental research
- Try to control for individual difference
- experimental research
- Goal is to predict variation within a treatment
- correlational research
- Many factors that may affect DV are free to vary
- correlational research
- Treat people differently
- correlational research
- Manipulation â€œhappensâ€ to the subjects
- impact studies
- e.g., Milgram, Zimbardo
- examples of impact studies
- Set of conditions is provided and subject makes a judgment
- judgment studies
- e.g., spousal/family interactions
- observational studies
- demand characteristics
- Personality and situational strength, power of the lab environment
- At least one IV is manipulated, but participants are not randomly assigned to all conditions
- quasi-experimental design
- Nearly impossible to make causal inferences
- nonrandomized designs
- Groups are nonequivalent before experiment begins
- nonrandomized designs
- Divide groups by IV, measure each group on DV, control doesnâ€™t have IV
- Static-group comparison design
- Selection is serious threat to internal validity, temporal precedence hard to establish
- Static-group comparison design
- Examine several groups at one period
- Cross-sectional design
- Follow same groups across many measurement periods (longitudinal)
- panel design
- Examine change over time for same group of people
- panel design
- Divide on DV, give treatment (IV), measure on DV, control doesnâ€™t get intervention
- Pretest-posttest nonequivalent control group design
- Selection is a threat, but pretest helps give insight to extent of threat, temporal precedence is clear
- Pretest-posttest nonequivalent control group design
- Extension of pretest-posttest
- Replicated interrupted time-series design
- May attempt to match groups to deal with lack of random assignment, makes groups dependent
- Pretest matching in quasi-experiments
- Doesnâ€™t control for regression toward the mean
- Pretest matching in quasi-experiments
- Evaluation of process: What is it and how does it work?
- Formative evaluation
- Evaluation of outcomes: Does it work?
- summative evaluation
- Do participants find program to be valuable (similar to face validity)
- reactions criteria
- Do participants learn/understand the information that the intervention is designed to impart?
- learning criteria
- Do participants change behavior as result of program?
- behavioral criteria
- Is organization more successful as a result of intervention?
- results criteria