Study your flashcards anywhere!
Download the official Cram app for free >
 Shuffle
Toggle OnToggle Off
 Alphabetize
Toggle OnToggle Off
 Front First
Toggle OnToggle Off
 Both Sides
Toggle OnToggle Off
 Read
Toggle OnToggle Off
How to study your flashcards.
Right/Left arrow keys: Navigate between flashcards.right arrow keyleft arrow key
Up/Down arrow keys: Flip the card between the front and back.down keyup key
H key: Show hint (3rd side).h key
A key: Read text to speech.a key
74 Cards in this Set
 Front
 Back
Internal Validity

Permits conclusion that there is a relationship between IV and DV


Threats to Internal Validity

History (external events)
Maturation Boredom Previous testing Regression toward the mean Experimenter Expectancy 

Best way to increase internal validity

Random assignment


External Validity

Permits generalizability of results


Threats to External Validity

Interaction between selection and treatment (rx doesn't generalize to same pop)
Interaction between testing and rx (rx only works if there's a pretest) History Hawthorne Effect (tendency of Ss to beh differently when being observed) Order effects (repeated measures studies) 

Types of Designs

True experimental (Ss randomly assigned to IV)
Quasiexperimental (Ss not randomly assigned) Correlational (Vs not manipulated and no causal relationship assumed) Developmental research (assessing Vs as a fx of dev over time, longitudinal 

Types of Designs cont.

Time Series (DV measured several times at regular before and after rx is administered)
Single subject (ABA, ABAB) Qualitative (descriptive) 

Scales of measurement

Nominal (unordered categories; gender)
Ordinal (ordered, rank) Interval (successive but no absolute 0; IQ) Ratio (has absolute 0; weight, time) 

Parametric Stats Assumptions

Normal distribution, homogeneity of variance, independence of observations


Ttest

Parametric stat, comparison of 2 means


Oneway ANOVA

Parametric stat, one DV and more than two groups, yields an F value, determining whether population means differed


Post hoc tests

Used in oneway ANOVA, ex. Tukey and Scheffe, pinpoints exact pattern of differences of among means b/c F doesn't do this


Scheffe

most conservative of posthoc tests, minimizes Type I error but highest Type II error


Type I error

Finding a difference when there isn't one


Type II error

Finding no difference when there is one


Factorial ANOVA

2 or more IV and one DV
Can't interpret main effects when there's an interaction 

MANOVA

multiple DVs and at least one IV


Nonparametric Stats

For nominal or ordinal data, distribution free, less powerful, includes Chi Square, Mann Whitney U, Wilcoxon Matched Pairs, Kruskal Wallis


Chi Square

Used to analyze nominal data (compares observed frequencies of observations within nominal categories to freqs that would be expected under the null)
Cautions: all observations must be independent (no before/after study) 

Mann Whitney

Compare two indep grps on a DV measured with rank ordered data
Alternative to ttestfor independent samples if nonparametric data 

Wilcoxon Matched Pairs

compare 2 correlated gps on DV measured with rank ordered data
Alternative to ttest for correlated samples if nonparametric 

Kruskal Wallis

Compares 2/more indep grps on a DV with rank ordered data
Alternative to one way ANOVA 

Negative skew

Most scores are high (to the right), but a few extreme low scores.
Mean is lower than median, median lower than mode Means easy test, ceiling effects 

Positive skew

Most scores are low (to the left), but a few extreme high scores
Mean is higher than the median, median higher than the mode Difficult test; floor effects 

Variance

Average of sq differences of each observation from the mean


Standard Deviation

Sq Rt of the variance


Stanine

Divide distribution into 9 = intervals, with 1 lowest and 9 highest


Standard Error of the Mean

Provides index of expected inaccuracy of sample mean


Statistical Decision Making

Four Possibilities:
1) True null retained (correct, no difference between IV) 2) True null rejected (incorrect; Type I error; say difference when isn't) 3) False null rejected (correct; find difference that does occur) 4) False null retained (incorrect; Type II error; there is a difference) 

Onetailed test

Predict that direction of means differ


Two Tailed test

Don't predict direction of difference


Power

Probability of rejecting null H when it's false (probability of not making a Type II error)
Increases by largr N, 1 tailed test 

Pearson r

correlation between two continuous variables


Square Pearson r

Percentage of variability i one measure that is accounted for by variability in other measure (coefficient of determination)


Pointbiserial coefficient

correlates one continuous V with one dichotomous V


Phi coefficient

corerlates 2 dichotomied Vs


Spearman's rho

correlates 2 rank ordered Vs


Regression

When 2 Vs correlated, constructs equation to est the value of a criterion (outcome) V on the basdis of scores on a predictor (input) V (2 or more Vs used to predict scores on one criterion)
Results in Multiple R 

Multiple R

can be squared, called coefficient of multiple determination (proportion of variance in criterion V accounted for by combination of predictor Vs)


Stepwise regression

Goal is to come up with smallest set of predictors that maximizes predictive power


Canonical Correlation

Used to calculate relationship btwn 2/more predictors and 2/more criterion Vs


Discriminant Function Analysis

Used when goal is to classify individuals into groups based on their scores on multiple predictors


Partial Correlation

Used to assess relationship btwn 2 Vs with the effects of another V partialed out


Zeroorder correlation

Correlation btwn 2 Vs determined without regard for any other Vs, converse of Partial Correlation


Structural Equation Modeling

Calculating pairwise correlations btwn multiple Vs, purpose is causal modeling, uses path analysis, LISREL


Normal curve

Could also be referred to as probability distribution


Z score

Obtained by subtracting sample mean score from an obtained score and dividing the result by the sample SD


Relationship of percentiles and zscores

Constant difference between raw or z scores will be associated with variable differences in percentile scores, as a function of the distance of the two scores from the mean.
Closer to mean, more difference in z or raw scores 

T scores

Mean = 50, SD = 10


Extreme scores

Take care in interpreting: percentile is an extrapolation
Use estimated prevalence value to determine whether interpretation of extreme scores may be appropriate 

Normalizing test scores

Best to add test content, rather than statistically transforming nonnormal scores into a normal distribution


Reliability

Consistency of a measurement of a given test, includes
Internal consistency TestRetest reliability Alternate form reliability Interrater reliability A reliability coefficient is interpreted directly as the percent of variance accounted for (don't square it) 

KuderRichardson reliability coefficient

Measure of internal reliability of a test;for items with yes/no answers or heterogeneous tests where split half methods must be used


Cronbach's Alpha coeffiecient

Measure of internal consistency (average intercorrelation between test items), used for tests with items that yield more than two response types


Adequacy of reliability coefficients

.90+ very high
.80.89 High .70.79 Adequate .60.69 Marginal <.59 Low 

Spearman Brown

Can be used to estimate the effects of lengthening or shortening a test on its relability coefficient


Standard Error of Measurement

Index of error in measurement
Estimation of confidence interval around obtained score Lower standard deviation and higher reliability, the lower the SEM 

Standard Error of Estimate

Estimation of confidence interval around estimated true scores


Validity

Does test measure what it was intended to measure


Face validity

Extent to which test appears to measure what it is supposed to measure; could affect test taker motivation


Content related validity

Systematic evaluation of test by experts  relevance, representativeness


Convergent and Discriminant Validity (Construct)

Correlate with tests of similar and dissimilar constructs


Concurrent and Predictive Valdity (Criterion)

Concurrent used to identify existing diagnoses or conditions
Predictive used to determine whether a test predicts future outcomes 

Size of validity coefficient

Rarely exceeds .3 or .4


Sensitivity

Proportion of testtakers with positive attribute who are positively identified by the test


Specificity

Proportion of testtakers with negative attribute who are correctly identified by the test


Positive Likelihood Ratio

Combines sensitivity and specificity into a single index of overall test accuracy indicating the odds that a positive test results has come from a positive examinee


Positive Predictive Power (PPP)

probability that an individual with a positive test result has the condition of interest


Negative Predictive Power (NPP)

Probability that an individual with a negative test result does not have the condition of interest


Bayesian statistics

Application of methods for deriving predictive power and othe related indices of confidence in decision making


Reliable Change Index (RCI)

Indicator of the probability that an observed difference between two scores from the same examinee on the same test can be attributed to measurement error (i.e. to imperfect reliability)


Standard Error of Difference

Standard deviation of expected testretest difference scores about a mean of 0 given an assumption that no actual change has occured


Relationship of Reliability and Vaidity

Reliability places a ceiling on validity
High reliability does not guarantee validity Reliability is necessary but not sufficient for validity 

Limitation of age and grade equivalents

Highly sensitive to minor changes in raw scores
Does not represent equal intervals Do not uniformly correspond to norm referenced scores 