Scientific Approach Flashcards
Causal Inference
Conclusion that a variable or an event influences/changes another variable or event.
Ex.) Fleming’s penicillin experiments
Causal Criteria
- Correlation (Cause -> Effect)
- Time sequence (Cause happens before effect)
- All other explanations ruled out based on other factors.
ex.)
Correlation
Observed relationship of two variables measured for the same individuals; one variable explains (x) or predicts an outcome of the second variable (y).
Correlation Coefficient
Between or equal to 1 and -1, denoted by “r.”
Indicates the magnitude of how accurately two variables are to each other, with 1/-1 being perfect +/- associations, and 0 being no association at all.
Ex.) Job satisfaction and supervisor’s rating of performance, r= +/- .3
External validity
Generality of research, i.e. can it be applied across contexts, in different populations, etc.
Field Study
Empirical data-collection in a natural setting using only measurement - no manipulated variables, yielding only correlation and no basis for causal inference.
Mediating Variable
Factor that conveys the influence of one variable to another in a sequential process.
Ex.) Role conflict and job performance - Stress from role conflict degrades job performance.
Moderator Variable
Factor that defines conditions under which the relationship of 2/more other variables differs.
Ex.) Relationship of job satisfaction and quitting varies with the availability of jobs.
Meta-Analysis
Statistical summary of accumulated scientific knowledge and on specific hypotheses from multiple, published studies that tested them.
ex.) Rosenthal - 346 studies of teachers’ self-fulfilling prophecies of their students.
Quasi-Experiment
Empirical study in a natural setting intended to support causal inference concerning the effects of >/= 1 interventions on >/= 1 variables in 1/more intact groups., no randomization of groups, incorporating repeated measurements and/or multiple comparison groups and/or interventions.
ex.)
Operational Definition
Specific procedure used to measure or manipulate a variable under study.
Randomized Experiment
Empirical investigation intended to find causal inference, incorporating:
- An independent (x) variable defining 2/more groups or conditions.
- Control procedures to make the conditions otherwise comparable, like holding some (extraneous) factors constant.
- Random assignment of each participant or unit to one condition.
- Measuring the dependent (y) variable afterwards.
ex.) Matsui et al. (1981) study of goal-setting in in number comaparison task.
Sample
Deliberately selected subset of a population intended as a representative of the whole population.
ex.)
Statistical Significance
In empirical research, when an observed difference between groups or relationship between variables is unlikely to be random.
p
Self-Report Measure
Procedure for assessing a variable through answers to questions posed to people via interview or questionnaire, relying on their recollections or judgements.
