EDE Quiz #1 (Ch. 4&5)

NCE

-Normal Curve Equivalents -Standard scores with a mean equal to 50 and a standard deviation equal to 21.06. Although the standard deviation may first appear strange, this scale divides the normal curve into 100 equal intervals.

Correlation Coefficient

-Numerical indexes of these relationships. They tell us the extent to which any two variables go together-that ism the extent to which changes in one variable are reflected by changes in the second variable. -Used in measurement to estimate the reliability and validity of a test.

Explain three factors that can affect a test's validity.

-Reliability: all valid tests are reliable, unreliable tests are not valid, and reliable tests may or may not be valid. -Enabling Behaviors: Skills and facts that a person must rely on to demonstrate a target behavior or knowledge. EX: A student must be able to write to complete an essay question. -Systematic Administration Errors: Tests need to be administered according to standardized procedures.

Variance

-S2 or σ2 is a numerical index describing the dispersion of a set of scores around the mean of distribution. -Distance between each score and every other score in the set

What does a reliability coefficient of .75 tell you about true-score variability?

This means that 25% of the variability in the distribution is attributable to error. (1.00 is perfect reliability, .00 is totally unreliable).

Percentile

-Derived scores that indicate the percentage of people whose scores are at or below a given raw score. (Example: if bill earns a percentile of 48 on a test, Bill's test score is equal to or better than 48 perfect of the test takers)

Compare and contrast item reliability, stability, and interobserver agreement.

-Item reliability: it is not possible to administer all possible test items of interest. Instead, testers use a sample of items from all the possible items. Sometimes the test items accurately represent their performances on all test items (reliable), other times they do not (unreliable) -Stability: how accurate/reliable the test is over time. -Interobserver Agreement: different observers may interpret things differently. For example, students who are saying the alphabet, one examiner may score the student a 90 and the other may score the student a 70.

Age Equivalent

-Means that a child's raw score is average (the median or mean) performance for that age group. -Expressed in years and months (Example 7-1= 7 years, 1 month old)

Grade Equivalent

-Means that a child's raw score is the average (the median or mean) performance for a particular grade. -Expressed in grades and tenths of grades

What is the difference between simple agreement and point-to-point agreement, and when might you use each appropriately?

-Simple Agreement: calculated by dividing the smaller number of occurrences by the larger number of occurrences and multiplying the quotient by 100. (Example: 2 teachers observes a student on 20 occasions to determine how frequently he is on task during reading instruction. The results show that the first teacher observes 12 times when he is on task, the second teacher observes 10 observations. The simple agreement is 83% because 100x(10/12)). -Point-to-point agreement: a more precise way of computing percentage agreement because each data point is considered. It is calculated by dividing the number of observations for which both observers agree (occurrence and nonoccurence) by the total number of observations and multiplying the quotient by 100. Since the agree 14 times, the point-to-point agreement is 70% because 100x(14/20).

IQ

-Standard scores with a mean of 100 and a standard deviation of 15. -A person earning an IQ of 85 scored one standard deviation below the mean, whereas a person earning an IQ of 115 scored one standard deviation above the mean.

What is a standard error of measurement?

-Text error -The average standard deviation of error distributed around a person's true score. - They are seldom calculated -SEM = the standard deviation of the obtained scores (S) multiplied by the square root of 1 minus the reliability coefficient. -SEM provides information about the certainty or confidence with which a test score can be interpreted. -When the SEM is large the uncertainty is large - we cannot be sure of the individual's score. -When the SEM is small, the uncertainty is small - we can be more than certain of the score.

Z Score

-The most basic standard score is the z distribution. In the distribution of z scores, the mean is always equal to 0. -Z scores are interpreted as being x number of standard deviations above or below the mean.

Why is the acculturation of the parents of students in the normative samples important?

-The socioeconomic status of a student's parents is strongly related to that student's scores on all sorts of tests, including intelligence, achievement, adaptive behavior, and social functioning. -The children of middle- and upper- class parents have tended to score higher on such tests.

Explain evidence of validity based on test content.

-The test questions and procedures for test administration and scoring -A test's content validity is determined by the appropriateness of the items included, the importance of items not included, and the way in which the items assess the content.

Explain evidence of validity based on relations to other measures

-Validity: the degree to which evidence and theory support the interpretations of test scores entailed by proposed uses of tests. -Questions of validity are specific to the individual student. -Example: ELLs are not tested in native language so their test is inappropriate.

Compare/contrast 2 scales of measurement

1. Ordinal Scales: order things from better to worse or worse to better. (For example: good, better, best, or novice, intermediate, expert). The magnitude of difference between values in unknown and unlikely to be equal. (You cannot determine how much better an intermediate performance is than a novice performances). Thus, ordinal scores cannot be added together or averaged. 2. Equal-interval Scales: also order things from better to worse, but the magnitude of the difference between adjacent values is known and is equal. (Examples include: the measurement of time, length, weight, and so forth). Because the differences are equal, they can be added, subtracted, multiplied, and divided.

Mean

Average of scores

Mode

Score most frequently obtained

Median

The one in the middle of a number line. Where 50 percent of the scores above and 50 percent are below

Explain the concept of measurement error.

There are two different types of error: 1. Systematic or Predictable Error (also called bias): is error that affects a person's (or group's) score in one direction. Bias inflates people's measured abilities above their true abilities. (example: only using MC when boys do better on MC). 2. Random Error: its direction and magnitude cannot be known for an individual test taker. Reliability refers to the relative absence of random error present during measurement.