PR -terms #3 Flashcards
The degree to which scores from the first iteration of a test or evaluation are correlated to subsequent
iterations of the same test or evaluation.
Test–retest reliability
The degree to which two tests in a given subject area, given to one group of test takers, are correlated
Equivalent forms reliability
The degree to which scores or evaluations from two or more people are correlated. A large reliability coefficient suggests a higher
Interrater reliability
The degree to which scores from one-half of a test or evaluation are correlated to scores or evaluations from the second half.
Split-half reliability
are values such as IQ and personality that we know exist but are not tangible.
constructs
we’re investigating the degree to which an instrument measures what it claims to measure.
Construct Validity
investigates whether the construct you’re measuring is highly correlated with a similar construct.
Convergent validity
demonstrates the construct validity of the instrument we are evaluating.
discriminant validity,
asks the question “Is the entire content area I want to measure covered?”
Content Validity
It would be necessary to evaluate each question to make sure it is relevant to the subject matter taught.
Item validity
is concerned with how well the results of a survey or test you develop correlate with a previously validated instrument
Criterion validity
as it is sometimes called, categorical or discrete data, is simply a value that we count. (For example, if we are interested in determining the number of males and females we work with, the data value “gender” is nominal in nature.)
Nominal data
it is sometimes called, rank
data. (An example we are all familiar with is the use of grade point averages to assign positions within a class)
Ordinal Data
It is the first of two types of data that are called quantitative or continuous data. By this, we mean that a data value can hypothetically fall anywhere on a number line within the range of a given data set.
Interval Data (interval-level data include temperature, aptitude scores, and intelligence quotients.)
differs from interval data because it does have an absolute zero point and the various points on the scale can be used to make comparisons between one another.
Ratio data
