Karteikarten
Terms for Final 5
Karten in diesem Stapel (49)
- 1A well-designed experiment:Varies at least one independent variable, assigned participants to conditions, and controls extraneous variables.
- 2Environmental manipulationsModifications to the research setting
- 3Instructional manipulationsVarying the instructions or information.
- 4Invasive manipulationsCreate physical changes in someone through simulation, drugs or surgery.
- 5Control variablesHeld constant to separate causes or eliminate alternatives.
- 6Internal validityEnsuring that the proposed causal variable is responsible for the change in the effect variable.
- 7ConfoundsAn alternative explanation for the research findings.
- 8Design confoundsA second variable that varies systematically with the IV and is an alternative explanation for the results.
- 9Between groups DesignDifferent groups are exposed to different levels of the IV. Randomly assigned or matched. Selection effects.
- 10Pretest only designParticipants are only tested on the DV once.
- 11Pretest/posttest designParticipants are tested on the DV twice, before and after the different IV.
- 12Within subjects DesignWhere all participants are exposed to the IV. Order effects.
- 13Concurrent-measures designParticipants are exposed to all levels at the same time and asked to choose one. (two IVs, one DV)
- 14Repeated-measures designParticipants are exposed to the IV, then DV, and repeated.
- 15Manipulation checksConstruct validity. A question/questions to ensure the experiment is successful.
- 16Pilot testA trial run to test effectiveness.
- 17Experimenter's DilemmaThe more controlled the experiment (and better the internal validity) the less generalizable (and worse the external validity).
- 18Selection effectsWhen one kind of participant is systematically different in one group. Between-groups. Random assignment. Matched.
- 19Order effectsWhen being exposed to one variable changes the outcome of the others. Within-subjects. Carryover, practice, fatigue, sensitization.
- 20CounterbalancingPresenting the levels of the IV in different orders.
- 21Partial counterbalancingWhen some, but not all, possible combinations of order are presented. Latin square.
- 22MaturationWhen an observed change could have emerged spontaneously. Use two groups?
- 23History threatWhen change occurs because of an event outside the experiment. This must vary systematically.
- 24Regression threatRegression towards the mean. Extreme findings will become normal.
- 25AttritionIn a repeated-measures experiment where participants systematically drop out. Their pretest scores might be removed.
- 26Testing threatA repeated measures experiment where scores change because the participants took the test more than once.
- 27Instrumental threatWhen the measuring instrument changes because its been used before.
- 28Observer biasWhen an observer's expectations change the result of the study. It can happen whenever behaviour is the DV. Double-blind.
- 29Placebo effectWhen participants feel like they are getting better because they believe they are getting treatment. Double-blind placebo control group.
- 30Ceiling effectNull findings. When something is too easy and all scores fall on the high end.
- 31Floor effectNull findings. When testing is too easy and the results fall on one end.
- 32Confounds in reverseWhen confounds eliminate the effect of the IV. Nervousness with GPA.
- 33Within-group variabilityNull findings. Noise or error variance creates overlap and makes it hard to find differences. Measurement error, individual differences, and situation noise.
- 34Measurement errorIs part of the DV (plus true score). Create reliable, precise measurements, and more instances of it.
- 35Individual differencesBaseline mood could be difference or experimenter. use a within-groups design or a matched-groups. Add more participants.
- 36Situation noiseWhere the participant gets distracted. Control the setting.
- 37Inferential statisticsDetermine if the difference in group means are significant.
- 38Systematic variabilityTreatment and confound variance. Has to do with group membership. If everyone varies systematically because of a confound you are in for a bad time.
- 39Unsystematic variabilityIndependent of group membership, so within groups. Situation noise, measurement error and individual differences can make it hard to detect.
- 40PowerThe probability that the study will show a statistically significant effect from the actual population.
- 41Effect sizeSize of the relationship using eta squared or omega squared figure out the proportion of variability. Cohen's d determines the size of difference between means
- 42t testA statistical test used to evaluate the size and significance of the difference between means. Error variance?
- 43Rejecting the null hypothesisThe conclusion that the IV did have an effect.
- 44Failing to reject the null hypothesisWhere researchers conclude the IV had no effect. Could be explained by error variance, and so this does not give us information.
- 45Directional hypothesisOne mean is expected to be greater than the other (one tailed)
- 46Nondirectional hypothesisWhere the mean is just expected to differ (two-tailed)
- 47Type 1 ErrorWhere the researcher wrongly concluded the null hypothesis is false (false alarm). More important. Alpha.
- 48Type 2 ErrorWhen the researcher fails to reject the null hypothesis (miss). Beta. Power is used to prevent this and detect IV.
- 49P ValueThe probability that the relationship is not by chance.
Informationen
- Beschreibung
- Simple experiments, confounds/obscuring variables, inferential statistics
- Erstellt von
- rsprinse
- Aufrufe
- 1
- Bewertung