PSYC 3510 1st EditionExam 3 Study Guide Chapters 5-7I. Ways to measure variablesA. Conceptual Definition: Construct; The researcher’s definition of the variable in question at a theoretical level.B. Operational Definition: Represents a researcher’s specific decision about how to measure or manipulate the conceptual variable.1. Ed Diener study operationalizing “happiness”- Ladder of Life2. Other conceptual variables- Religiosity, Gender, Wealth, Well-beingII. Types of MeasuresA. Self-report: Questionnaire or interview (For children, parents or teachers are given questions)B. Observational: Intelligence tests; Physical traces: Stress behaviorsC. Physiological: Recording biological data; Brain activity, hormone levels, or heart rate.1. Equipment- EMG: Facial Electromyography; fMRI: Functional Magnetic Resonance ImagingIII. Scales of MeasurementA. Categorical: Assign numbers to the levels of the variables.B. Quantitative: Meaningful numbers.1. Ordinal: Ranked order; Intervals can be uneven.2. Interval: Represent equal intervals between levels; there is no true 0.3. Ratio: Equal intervals and “true zero” actually means 0.IV. Validity: Measure what is supposed to be measuredV. Reliability: Consistency across studies.A. Test-Retest: Apply to any operationalization; Consistent through timeB. Interrater: Two or more observers come up with very similar findings.C. Internal: Questions are worded differently, but intended to be a measure of same construct.D. Represented by scatterplot and correlation coefficient r- Reliability is Association Claima. Kappa: A statistic used when the observers are rating a sample on a categorical variable. Close to 1.0 means the raters agree.b. Cronbach’s Alpha: Correlation-based statistic; Also coefficient alpha; .70 or higherVI. Validity of MeasurementA. Face: A plausible measure of the variable in question; Subjective judgment; Look at a scale, you know what it is studying.B. Content: Subjective; A measure must capture all parts of the defined construct.C. Criterion: Whether the measure is related to a concrete outcome, such as a behavior, that it should be related to; Example: IQ and the construct of intelligenceD. Known-Groups Paradigm: Researchers see whether the scores on the measure can discriminate among a set of groups whose behavior is already well understood.E. Concurrent: The scale should correlate with other scales measuring the same thing.F. Predictive: Should predict future behavior.G. Divergent: The scale needs to be negatively correlated.H. Discriminant: Two measures of opposite things will be strong but negatively correlated.I. Convergent: Two different scales should have strong correlation.- Studies*Collett and Marsh (1974) - Gender differences in how people avoid bumping into others.*Zimmerman and West (1975) - Whether dominance is reflected in patterns of speech.* Festinger, Riekin, and Schacter (1956) - Participant; Disguised/Undisguised*Simons and Levin (1998) - Change blindness (Miss unexpected things)VII. Reactivity ( Observer effects)A. Demand characteristics, Unobtrusive measures, DesensitizationVIII. BAD ItemsA. Ambiguous, Complicated, Biased, Unclear Alternatives ( Not mutually exclusive or exhaustive), Social Desirability, Fence sitting, and Acquiescence ( No verification key)IX. Question FormatsA. Open-endedB. Likert scaleC. Likert-type scaleD. Semantic Differential FormatX. Improperly worded questionsA. LeadingB. Double-BarreledC. NegativeD. Question OrderXI. Socially Desirable Responding: (Faking Good); Giving answers to make one look better than they really are; Similar but less common than faking bad.XII. Observer effects: When participants confirm observer expectationsXIII. Observer bias: Observers see what they expect to seeXIV. Population of InterestA. Biased (Unrepresentative): Some members have a much higher probability of being included in the sample.1. Purposive: Only certain types of people are includeda. Combination of convenience and self-selection2. Snowball: Participants are asked to recommend acquaintances.3. Convenience: Participants are the easiest to access and readily available4. Self-Selected: Only people who volunteer to participate.B. Unbiased (Representative): All members of the population have an equal chance. 1. Simple random2. Stratified: Particular demographic categories are identified and individuals randomly selected3. Cluster: Selected at random and data is collected from all individuals in the cluster4. Oversampling: Intentional overrepresentation of one or more groups.5. Multistage: Data is collected from a random sample of individuals in the cluster.6. Systemic: Counting off members of population using a randomly chosen
View Full Document