Measurements Exam 2

Correlation

A measure of the extent to which two variables are related to one another in a linear fashion

Positive Correlation

Goes in the same direction

Negative Correlation

Goes in opposite directions

Asymptotic

A function that has an asymptote

Asymptote

Data goes up and eventually levels out

Exponential

Starts slow and eventually shoots up

Quadratic

Increases up to a point and then decreases or decreases up to a point and then increases

Linear

Straight line on scatter plot that best fits

Bivariate Outlier

When a data point doesn’t fit the trends of linear data

Correlation and Causation

Just because 2 variables are correlated doesn’t necessarily imply that one caused the other. X may cause Y. Y may cause X. Z may cause both X and Y.

Coefficient of Determination (r2)

Proportion of variance in one variable explained by the variables in another. Correlation squared

Venn Diagram

Shows overlapping correlations between coefficients. Coefficient of determination is the overlapping part

coefficient of Alienation

Measure of 1 - r2 non-association between 2 variables

Linear Regression

Ŷ = bx + a. Used in criterion related validity. B = slope. A = y-intercept. Ŷ = expected value of y. X = measured variable

Residuals

Difference between observed values and predicted values

Standard Error of Estimate

Standard deviation of the residuals for a regression line. Tells how far, on average, a given predicted score is likely to deviate from an observed score.

Shrinkage

Amount of decrease observed when a regression equation is created for one group and then applied to another

Reliability

Consistency in test scores and freedom of random measurement error

Basic Classical Test Theory Equation

X = T + E or X - T = E. X = observed score. T = true score. E = error → random error → difference between observed score and true score.

Measurement Error

Random and normally distributed

Standard error of measurement (SEM)

Represents the standard deviation of measurement errors for a given test

Reliability as a coefficient of determination

1 - rxx = % random error in test scores

SEM Equation

SEM = square root of 1 - rxx

SEM used for

Determining whether test scores differ significantly from one another. Determining whether test scores differ significantly from a hypothetical true score

Consistency across time

Error is attributed to random fluctuations in performance across occasions→ test-retest

Consistency across different samples of items

Error is attributed to inconsistency in content sampling across different test forms → alternate/parallel forms reliability

Internal Consistency

Error is attributed to inconsistency in content sampling within a given test → split-half and Spearman Brown

Scorer Consistency

Error is attributed to the use of different examiners or scores of the test → scored/inter-rater reliability

Validity

Concerns whether a test measures that we intend it to measure. The accuracy or appropriateness of inferences that we draw from scores

Content-related validity

The extent to which a test adequately samples the content domain → analogous to expert testimony in a criminal trial

Face Validity

Test-takers perceptions concerning whether a test looks or seems valid

Construct Related Validity

The extent to which the label for what the test measures is appropriate

Convergent Construct Validity

Produces scores that correspond with other established measures of the same construct

Discriminate construct validity

Produces scores that differ from established measures of the same construct it purports not to measure

Criterion Related Validity

The effectiveness of a test in a predicting individuals behavior in specific situations → gets at whether a test is a good prediction of those future behaviors

Predictive criterion related model

Predictive data is collected first; criterion data s collected later

Concurrent criterion related model

Predictor and criterion data are collected at the same time

Postdictive criterion related model

Predictor data is collected and compared with past criterion data

Test-Retest Reliability

Estimates are used to evaluate the error associated with administering a test 2 different times. Take test at one specified time and the at a different specified time take it again

Parallel Forms Reliability

Compares 2 equivalent forms of a test that measure the same attribute → use different items with the same difficulty level

Split-half Reliability

A test is give. And dived into halves that are scored separately the compared. Could be split by 1st half and 2nd half or odd-even system

Spearman-Brown

Allows you to estimate what the correlation between the 2 halves would have been if each half had been the length of the whole test

KR20

Calculating the reliability of a test in which the items are dichotomous, scored 0 or 1 (usually right or wrong)

Kappa

Indicates the actual agreement as a proportion of the potential agreement following correction for chance agreement. Best method for assessing the level of agreement between several observers

PHQ-9

Depression screener

GAD-7

General anxiety screener

MDQ

Mood disorder screener

AUDIT

Alcohol use disorder screener

PCL-5

PTSD screener

Y-BOCS

OCD screener

McLean

Borderline personality disorder screener

SCID-5-PD

Personality assessment

ASRS self-report and other report

ADHD adult symptoms

Vanderbilt

ADHD adult symptoms → parent/teacher fill out

Mopet

She created → used to do direct observation of faculty using validated assessment tools