Psych 325 M. MerritExam 2 Study Guide: Chapters 3 & 4Potential Chapter 3 Multiple Choice Question Topics- Internal validityo The extent to which a set of research findings provides compelling information about causality o When a study is high in internal validity, we can confidently conclude that variations in the independent variable caused any observed changes in the dependent variableo Low internal validity means… Not sure of cause Questionable findingso Experiments Strong argument for cause Control for other causeso Psychologists obsess over this- External validity/generalizabilityo The extent to which a set of research findings provides an accurate description ofwhat typically happens in the real worldo When a study is high in external validity, or generalizability, we can confidently conclude that the findings of the study will apply to other people, other physical or social environments, and perhaps even other cultureso We want our research to… Generate general principles Be able to be replicatedo External validity/generalizability Do findings generalize to all people or only sample study Do findings generalize to all situations or only one studied?o Low external validity means Limited generalizability Still potentially useful, though- Relationship of internal to external validityo More control requires more artificial situationso More representative situations lead to less controlo As one goes up the other goes down- Construct validityo The extent to which the independent and dependent variables in a study truly represent the abstract, hypothetical variables of interest to the researchero Has to do with whether the manipulated and/or measured variables in a specific study accurately reflect the variables the researcher hoped to manipulate or measureo To the degree that a researcher’s operational definitions do a good job of converting the abstract to the observable, construct validity is high- Threats to construct validity (confound, random error, poor measurement)o Confounding variable Shifts along with manipulation Effect is due to this, not IV Good operational definitions helpo Random error Unseen, non-theoretical problem Strict protocols (instructions) helpo Poor measure Measure is in the wrong ballpark Assessing construct validity helps- Conceptual validityo Refers to how well a specific research hypothesis maps onto the broader theory that it was designed to testo Is your hypothesis any good? Valid way to approach theory? Appropriate? Worthy of study?o If not, your hypothesis is not a true test of the theory- John Stuart Mill’s principles of causality (covariation, temporal sequence, eliminating confounds)o Covariation Cause and effect correlate Must change togethero Temporal sequence Timing must be right Cause must precede effecto Eliminating confounds Must rule out other explanations- Third variable problemo Third variable is causing the effect you observe between two variables- Definition of measuremento Systematic way of assigning numbers or names to objects and their attributes- Different types of measures: verbal, behavioral, behavioroid, physiologicalo Verbal Verbal response i.e. self-esteem scaleo Behavioroid real behavior i.e. avoiding someoneo Physiological Biological processes i.e. brain imaging- Theoretical constructs, indicators, operationalizationo Theoretical constructs Theory says they exist “Latent” or hidden variableso Indicators  What we can observe Result of theoretical constructso Operationalization  Finding good indicators of theoretical constructs- Type of measurement scales: nominal, ordinal, interval, ratioo Nominal Group membership- Assign names, numbers- Assess differences- Examples: demographic characteristics, license plate number, male or female in SPSS, social security number, personality traits, marital status, hair color, race, DSM coding systemo Ordinal  Rank ordering on dimension- Assess differences, magnitude of attributes- Intervals aren’t equal- Examples: class rank, year in school, height, birth order, favorite pet order, year in school, finishers in a race, favorite brand of computero Interval Ordered with equal intervals- Assess differences, magnitude of attributes, equal intervals - May have arbitrary zero- Examples: IQ, attitudes, temperature, ACT score, change in a subjective score (depression), zero is not a true zero, anxiety scales, weight losso Ratio Interval scale with true zero- Has all four principles- Examples: weight, time, multiplication and division are okay, height, reaction time to a novel stimulus, yearly income in dollars, number of beers consumed, number of shots taken- Reliabilityo An assessment of consistency and repeatability o If what it measures doesn’t change, a perfect measure gives the same score every timeo Too much variability=error- Inter-observer agreement/inter-rater reliability: what does it mean, when is it used, what makes it pooro Different people agree i.e. beauty pageant judges more judges=bettero Used whenever a score is based on a judge’s observation Study behavior or products of behavior (such as writing samples) Criteria are established for judges- Internal consistency: what does it mean, when is it used, what makes it pooro Agrees with itselfo Used whenever multiple items are combined to create a score Common with self-report scales Can be used with observationso Are the items in a scale measuring the same thing? High correlation with each other means items are similar Low correlation means items aren’t consistent with each other- Bad measure- Measuring different constructs- Can’t understand items More items lead to better reliability- More indicators- Less vulnerable to chance- Temporal consistency/test-retest reliability: what does it mean, when is it used, what makes it pooro Used whenever you assess something that doesn’t change on a regular basis Scores shouldn’t change over time Doesn’t apply to things like moodo Test-retest reliability Over short time spans (i.e. two weeks), do you get the same score?- Same test, same people, different times- High correlation=consistency- Low correlation means the score is changing over timeo Construct may not be tableo May be vulnerable to mood effectso May be difficult to understand items- Focus on major personality traits Happy mood=bad temporal