
correlation
 the type (positive or negative) and degree of relationship
 between 2 variables. Does not prove causation.

standard deviation
 average distance that cases within a
 distribution of scores vary from the mean

standard scores
 these describe individual statistics/test
 scores in relation to the bell curve

normal curve/ bell curve
 is a graphical representation of a normal distribution. For anything you measure, results across a large population distribute themselves symmetrically. In a normal distribution, the mean median and mode are all equal and the distribution is symmetrical across
 the mean (one half mirrors hte other). The tails are asymptotic in that they come closer and closer to the horizontal axis, but never touch. 68% fall within
 one SD. 95%fall within 2 SD's and 99% fall within 3 SD's. (Internal percentages= 34%, 13.5%, 2%). skewed right=positively skewed and is caused by extreme scores
 at the high end that pull the mean higher than the median (midpoint). skewed left is opposite.


z score
 an expression of an individual score in a way that
 conforms to standard deviation units. For example, 1.5 is one and a half standard deviation units above the mean and .25 is onequarter of a standard deviation units below the mean. These are useful in comparing scores across different settings , testing situations and tests.

T score
 o sometimes called a McCall T is a standard score resulting
 from a z score transformation T= z(10)+50. The advantage= eliminates negative
 numbers or fractions.

mean
arithmatic average. measure of central tendency

median
 midpoint. the number in the middle of a distribution (15
 numbers, this would be the 8th in order)
 a measure of central tendency

mode
 the most frequently occurring number in a
 distribution.
 measure of central tendency

range
 distance between highest and lowest
 a measure of variability

variance
 how far apart numbers are (how spread out they are
 from the mean)

Standard Error of measurement (SEM)
These are found in standardized tests and are a measure of how much observed scores vary from a true score. The smaller the SEM= the more reliable the test.

3 ways to describe a distribution of a set of scores:


3 measures of central tendency


established by comparison with another
test or measure of the same criteria.

content validity
 is subjective. for example, does the test
 fairly cover what was taught or what experts agree content is appropriate for measurement

correlation coefficient
 identifies numerically the relationship between 2 variables (from 1 to 1 with midpoint of 0 that means no
 correlation)

formative evaluation
evaluation aimed at guiding the nextinstructional step

normreferenced test
results interpreted relative to thesuccess rate of other test takers

criterionreferenced test
results interpreted relative to #of test items answered correctly,without reference to the success of other test takers.

pilot study
used to identify strengths and weaknesses of a tool or process so that they might be revised for future use.

summative evaluation
aimed at summarizing a segment of achievement (at the end of instruction) (ie, generating a grade or other symbol of achievement)

reliability
 how dependable the test is. The consistency
 with which something is accomplished.

validity
 accuracy. Does the test measure what it says it
 measures.

variability
one of three ways to describe a set of scores. Answers the question "How wide are the differences between the scores?" These can include range, standard deviation and variance.

