Research Methods Flashcards
Compares groups of a single IV based on a single DV (e.g., comparing test score by level of education)
One-way ANOVA
Compares groups of 2+ IVs for mean differences on a single DV (e.g., comparing test scores by both level of education and zodiac sign)
Factorial ANOVA (includes two-way)
Compares a DV by both an IV and a covariate (e.g., comparing test scores by both level of education and number of hours spent studying)
ANCOVA
Comparing 2+ DVs by one IV (e.g., comparing test scores and annual income by level of education)
MANOVA
Comparing groups by IV and covariate on 2+ DVs (e.g., comparing test scores and annual income by level of education and number of hours spent studying)
MANCOVA
A type of statistical analysis that extends multiple R for multiple outcome variables.
Correlation between 2+ IVs and 2+ DVs
Canonical correlation
Canonical R
A statistical analysis that explores the naturally occurring groups within a data set.
Cluster analysis
A statistical analysis to use when the Y variable is nominal and two or more X variables are used to place the examinee into a category on Y.
Discriminant function analysis (2 groups)
Multiple discriminant analysis (3+ groups)
A research design that combines the pre-test/post-test control group design with the post-test only control group design in order to evaluate the effects of pre-testing on a study’s validity.
Solomon four-group design
A correlation between a naturally dichotomous variable and a continuous variable.
Point biserial
A correlation between an artificially dichotomous variable and a continuous variable.
Biserial
Random factors that produce a non-representative sample
Sampling error
Non-random factors that produce a non-representative sample
Systematic error
In diagnostic testing, the proportion of the individuals with the condition who test positive.
Formula: True Positives / (True Positives + False Negatives)
A high __________ means that the test is good at identifying people who actually have the condition.
Sensitivity
In diagnostic testing, the proportion of the individuals without the condition who test negative.
Formula: True Negatives / (True Negatives + False Positives)
A high __________ means that the test is good at identifying people who don’t have the condition.
Specificity
The probability that a person with a positive test result actually has the condition.
Formula: True Positives / (True Positives + False Positives)
A high __________ means that if someone tests positive, they are very likely to have the condition.
Positive Predictive Value (PPV)
The probability that a person with a negative test result does not have the condition.
Formula: True Negatives / (True Negatives + False Negatives)
A high __________ means that if someone tests negative, they are very likely to NOT have the condition.
Negative Predictive Value (NPV)
In factor analysis, the amount of variability in test scores that is explained by all of the identified factors.
Communality
How do you calculate variance?
Square the coefficient
Assesses whether a test or measurement adequately covers all relevant aspects of the construct being measured.
A math test on fractions should include questions on addition, subtraction, multiplication, and division of fractions, not just addition.
Content validity
Evaluates how well a test or measurement correlates with a specific outcome.
A new test for predicting job performance should correlate with actual job performance scores.
Includes concurrent and predictive validity.
Criterion validity
Determines whether a test or measurement actually measures the theoretical construct or concept it’s intended to measure.
A test designed to measure intelligence should correlate with other tests of intelligence (convergent validity) and not with tests of, say, physical strength (discriminant validity).
Construct validity
How do you determine a z-score from a raw score?
Calculate how many standard deviations the score is from the mean. If the score is one standard deviation above the mean, the z-score is 1.
A situation in which a response measure is influenced by factors that are not related to the concept being measured. Evidence of this may be observed through correlations of the response measure with variables that are conceptually distinct from that measure.
Criterion contamination
A statistical measure of inter-rater reliability.
Cohen’s kappa
Kappa coefficient
What is the difference between a z-score and a t-score?
They both represent how many standard deviations a data point is away from the mean.
Z is used when you know the population standard deviation and have a large sample size.
T is used when the population standard deviation is unknown and/or the sample size is small.
A problem that occurs in multiple regression when the predictors are highly correlated with one another.
Multicollinearity
