Stats Flashcards
IV
manipulated by researcher, presumed to be the agent of change
DV
measured by researcher to determine if IV has an effect
Quasi-independent variable
IV in quasi-experiemtne (using existing groups rather than random assignment in determining condition)
Variance
Sum of squared deviations from the mean, divided by N-1. Less susceptible to extreme values/outliers
Standard deviation
Square root of the variance
r-squared (single predictor), R-squared (multiple predictors)
Proportion of variation accounted for in one variable through linear relationship with another (or others). Not good for sample-to-sample comparisons. Reflects a reduction in error.
Eta-squared
Proportion of variance accounted for in one variable thru relationship (not necessarily linear) with another (or others)
Squared factor loading
Proportion of variance accounted for in one variable by a factor
Beta weight
Standard regression coefficient
Coeffeicient of Nondetermination
One minus r-squared; proportion of variation in the dependent variable not associated with independent variables
Chi-square: Cramer’s phi
Strength of relationship between two variables in a contingency table
t-test: Cohen’s d
Difference between two group means in terms of a standard deviation (control group or pooled)
ANOVA: eta-squared, omega-squared
Proportion of variation in the DV accounted for by the IV
Correlation: r-squared
Proportion of variation in one variable accounted for by the linear relationship with another
p value
The level of significance, or the probability that the null hypothesis is false
Kappa Coefficient
Used to evaluate inter-rater reliability
Coefficient Alpha
Stats used to assess the internal consistency reliability
Pearson’s r
A correlation stat used primarily for two sets of data that are of the ratio or interval scale; it is the most commonly used correlational technique
Pooled variance
The weighted average of two sample variances. Provides better estimate of population variance than either sample alone.
Mean Squared Within (MSW)
A measure of error variation used in ANOVA
Moderator variable
A variable that affects the magnitude of direction of the relationship between the independent variable and the dependent variable
Mediating variable
A variable explaining the process by which the IV affects the DV (therapy affects depression by creating a more positive self-image, which then lessens depression)
Outcome variable
The dependent variable for a prediction in an experiment; it should be clinically relevant
Suppressor variable
Lowers or covers the relationship between variables
Criterion contamination
Occurs when the operational or actual criterion includes variance that is unrelated to the ultimate criterion.
Chi-Square test
Examines frequency distribution of categorical variables such as political party affiliation or eye color. Non-parametirc, does not require normality.
Goodness-of-fit
One-way Chi-Square test for examining frequency distribution of one IV. May use expected frequencies (like expected percentage)
Test for independence
Two-way Chi-Square test for examining contingency table for 2 variables to determine wether they are independent (un-related). Requires counts, not percentages and requires a count of at least 5.
T-test
An inferential statistical procedure used to test whether the means of two groups are equal to each other.
The t-test is more powerful (more likely to reject the null hypothesis) when:
Larger sample size(s); larger mean difference; smaller sample variation
One-sample t-test
Tests they hypothesis that a single sample mean is different than a specific hypothesized value
Independent-samples t-test
Test the hypothesis that two unrelated samples are different from each other
Related or dependent-samples t-test
Tests the hypothesis that the difference between two related samples (pre/post scores, scores of siblings) is not equal to 0 (samples have different means)
Main effect
Arising from ANOVA terminology; represents the effect of an independent variable on Y averaged across the (main or interaction) effects of other independent variables.
Interaction
The circumstance in whciht the impact of one variable on y is conditional on (varies across) the values of another predictor
Kolmogorov
An uncommon stat that utilizes ordinal or ranking data.
Power
The probability of rejecting a null hypothesis that is false
ANOVA post-hoc tests
If a significant difference exists, a post-hoc test will be a more focused examination of which means differ from which. Scheffe (conservative) Tukey’s HD, Fisher’s LSD (liberal)
Factorial or n-way ANOVA
n represents the number of IVs or factors. Used when examining the effects of two or more IVs.
Mixed-design ANOVA
Multiple IVs including both within-subjects (time) and between-subjects (condition) factors. Ex. pre/post test with control condition.
Kruskal-Wallis
An alternative test to the one-way ANOVA that can be used to compare two or more independent groups
Randomized Block ANOVA
A statistical test that controls for the effects of extraneous variables by grouping (“blocking”) the subjects based on the variable and then assigning each subject to one of the interventions; the confounding variable is therefore handled as if it were an independent variable.
Cluster sampling
Sampling technique involving naturally occurring groups (clusters)
Stratified sampling
Sample drawn from each stratum; main objective is improved precision
Multistage sampling
More complex form of cluster sampling. Population divided into start at highest level, then sample s drawn and stratified at lower level; procedure repeated until at lowest hierarchical level
Systematic sampling
A simple, random sampling of each stratum of the population. Additional variables may also be stratified, such as gender.
Random selection
Drawing a sample from a population in such a way that each member has an equal probability of being selected
Experimental Design
Researc in whcih random assignment is used to place subjects in groups that will receive different aspects of the variable in question.
Proband
AKA patient zero; the first family member to seek professional attention for a disorder
Normal distribution
mode = median = mean
Positively skewed distribution
mode < median < mean
Negatively skewed distribution
men < median < mode
ETS scores
mean of 500 and standard deviation of 100
Distribution of scores to either side of the mean
50% falls to either side of the mean. 34% between the mean and one standard deviation, 14% between the one and two standard deviations, and 2% between two and three standard deviations from the mean
What percentage of scores falls within one standard deviation from the mean?
84%
What percentage of values are within two standard deviations of the mean?
95%
Uniform distribution
Equal frequencies across distribution (i.e. a block)
J-shaped distribution
Skewed, but without a tail on distribution with a mode
Actuarial Data
Grove and Meehl: demographic data related to risk calculation of births and deaths. (10% better than clinical judgment)
Larger sample sizes are preferred in order to:
Ensure sample adequately represents population; reduce sampling error; increase statistical power
Standard Error
Standard deviation of a sampling distribution
Sampling Error
Difference between an obtained sample statistic and the corresponding population parameter
Systematic Error
Measurement bias leading to measured values being systematically too high or too low.
Standard Error of the Mean
Estimates how much the sample mean will deviate from the population mean due to sampling error. When population size goes up, sample size goes down, and the Standard Error of the Mean will go up
Standard Error of Measurement
Constructs confidence intervals for test scores. How much the score is expected to vary from the person’s actual ability. Standard deviation of observed values around predicted values on a regression line (measure of prediction error). Higher standard deviation is higher SEE.
Central Limit Theorem
Increaseing the size of the random sample N drawn from a population will cause the distribution of the sample means to form a more normal distribution, with a mean equal to the population mean, and a standard deviation (i.e. standard error of the mean) equal to the sample standard deviation divided by the square root of N
When a constant is added to or subtracted from a variable
Measures of central tendency (median, mean) change similarly; measures of variability (range, standard deviation) remain the same.
When you multiply or divide by a constant
measures of both central tendency and variability change
When you add, subtract, multiply, or divide by a constant…
the shape of the distribution remains the same; correlations with other variables remain the same
Attenuation
Decrease in correlation coeffecient reflecting relationship between two variables due to measurement error in one (or both)
Attenuation correction formula
Estimates true correlation between two variables; requires correlation coefficient and a reliability coefficient for each variable. If absolute value of result is greater than one, round to one (Pearson’s r ranges from -1 to 1)
Measurement Error
Error in the employed values of a variable due to the presence of distorting influences on the assessment, such as momentary distractions, error in recording or understanding, influences of other variables on responses to particular items
2 types of construct validity
Convergent (tests correlated with tests of similar trait) and divergent (tests not correlated w/tests of unrelated traits)
2 types of criterion-related validity
Concurrent validity (test correlated with criterion measured at same time; SAT and high school GPA) and predictive validity (test correlated with criterion variable measured at a future time; SAT and college GPA)
Content validity
Adequacy of test in measuring all facets of a construct or trait
External validity
Extent to which experimental findings may be generalized from the lab to the world at large
Internal validity
Ability to assert that observed effects are attributable to an independent variable rather than confounding variable
5 main threats to external validity
Interactions of different treatments; interaction of testing and treatment; interaction of selection and treatment; interaction of setting and treatment; interaction of history and treatment
Threats to internal validity
History, maturation, testing, instrumentation (changes in measure may cloud results), regression to the mean, selection, mortality (dropouts), interactions with selection (any of above threats may interact with selection and be mistaken for treatment effects) and ambiguity about the direction of causation
Confidence interval
Range of values centered at sample statistic used to estimate the population parameter with a confidence of (1-a) percent
Standard Error of Estimate
An index of the degree of variability of the data points about a regression line, determined to be the square root of the sum of squared deviations of the points about the line divided by (N-2)
Familywise alpha
running multiple tests, each with its own alpha. Results in a familywise alpha for the entire set approximately equal to the sum of all the alphas. May correct (Bonferroni-adjustment) or run alternative test (MANOVA to replace multiple ANOVAs)
Changing-Criterion Design
A single-case experimental design that demonstrates the effect of an intervention by showing that performance changes in increments to match a performance criterion
Reliability
The extent to which a measure or test is consistent and repeatable. Necessary, but not sufficient, for validity.
Determination of reliability with respect to internal consistency
Reliability coefficient, split-half, Cronbach’s Alpha (Coefficient Alpha), Kuder-Richardson’s Formula (KR20)
Determination of reliability in respect to consistency between alternative forms
Coefficient of equivalence
Reliability in respect to test-retest consistency
Coefficient of stability
What is the Kappa Statistic used for?
To determine inter-rater reliability. For use with nominal or ordinal data.
Eta (n)
A universal measure of relationship that can be used regardless of the from of the relationship
Incremental validity
In psychometrics, the degree to which a test improves on decision than can be made from existing information, such as the base rate of the attribute being measured and other measures that are available.
p-value
In statistical hypothesis testing, the p-value is the probability of obtaining a result at least as extreme as a given data point, under the null hypothesis.
Validity coefficient
The correlation between the predictor test and the criterion variable that specifies the degree of validity of that generalization
ABAB Design
A single-subject design in which a baseline measure of the DV (depression) id obtained (A) before treatment introduced (B), removed (A), and reintroduced (B)
Multiple-Baseline Design
A single-case experimental design strategy where the intervention is introduced across different behaviors, individuals, or situations; the key distinction between this and ABAB design is that this design no treatment is withdraws; rather, the treatment is applied to multiple settings, behaviors, or subjects
Counterbalancing
A method of arranging conditions or tasks for the subjects so that a given condition or task is not confounded by the order in which it appears.
Randomized Block Design
This requires that the researcher divide the sample into relatively homogenous subgroups or blocks (analogous to strata in stratified sampling), and then the chosen experimental design is implemented within each block
Solomon Four-Group Design
An experimental design used to evaluate the effect of pre-testing; a combination of the pre-test/post-test control-group design and a post-test-only design in which a pre-test and the experimental intervention are combined.
Split-plot design
An experimental design that includes both randomized group designs and randomized block (repeated measures designs)
Matched subjects design
Each participant in one sample is matched with a participant in another sample with respect to a specific variable (SES)
Repeated measures design
A research design in which participants appear in each condition
Systemic variance
Variance due to the IV
Experimental variance
Variance due to the DV
Systemic bias
The tendency of a process to favor a particular outcome
Factorial design
A group design in which two or more variables are studied concurrently. For each variable, two or more levels are studied. The design includes the combinations of the variables so that maine effects as well as interaction effects can be evaluated.
Cross-sectional design
The most commonly used version of a case-control design in clinical psychology, in which subjects (cases and controls) are selected and assessed in relation to current characteristics (different from events that happened in the past or the future)
Interrupted time-series design
Repeated measurements are made on participants both before and after a manipulated intervention of a naturally occurring event
Nonequivalent Control Group
A group used in quay-experiments to rule out or make less plausible specific threats to internal validity; the group is referred to as nonequivalent because it is not formed through random assignment in the investigation
Protocol analysis
Qualitative data analysis method involving verbalization of thoughts occurring while completing a given task.
Event coding
A technique used to record the events that led up to the subject’s thinking process
Retrospective debriefing
The act of having a subject describe how he or she determined the solution after working on a problem
Structural Equation Modeling (SEM)
Technique for building and testing statistical models. Uses factor analysis, path analysis, and regression. Two step process: validates measurement model with confirmatory factor analysis and tests structural model with path analysis
Q-Technique Factor Analysis
A statistical analysis that determines how many types of people a sample represents
Discriminant function analysis
Classifies people into criterion groups based on their scores or statuses on two or more predictors.
Partial correlation
Correlation between x and y after removing variation in each shared with a third variable z
Semi-partial correlation
Correlation between x and y after removing variation in x (and only x) shared with a third variable z
phi
2 dichotomous variables
tetrachoric
2 artificial dichotomous variables
contingency
two nominal variables
Spearman’s rho
two ordinal variables
Linear discriminant function analysis
Used to analyze the relationship between variables when there are multiple x variables and one y variable that is categorical
Path analysis
The application of correlational analysis to test models of causality
Logic Analysis
A multivariate technique that uses two or more categorical variables to predict the status of a single categorical variable
Multiple Discriminant Analysis
Several independent variables are used to predict group membership
Cluster analysis
Data is gathered on a number of DVs and statistically analyzed for naturally occurring subgroups without using an “a priori” hypothesis
Communality
Each test in a factor analysis has a communality, which indicates the total amount of variability in test scores that has not been explained by the factor analysis - i.e. by all of the identified factors.
Factor loading
The correlation between a single test and an identified factor.
