CHAPTER 4: Understanding Test Scores

Grade Equivalents

Indicate relative test performance in terms of the grade level at which the student's raw score matches the average score earned by the norm group. 5.4= fifth grade student with four months of instruction

They are derived through the use of comparable procedures by the publishers of the various tests used in federal research projects

Main advantage of NCEs

• Compute group statistics • Compare the performance of students who take different levels of the same test • Compute student performance across subject matter • Evaluate gains over time

NCE scores can be used to:

norm-reference scores

Score that is compared to the scores of other people who took the same test (i.e., norm group)

criterion-referenced scores

Score that is measured against a specified level of performance (i.e., a criterion)

Quartile

Similar to percentile ranks, except that they divide the data set into four equal parts instead of 100 equal parts

norm group

The group of individuals whose scores were used to norm a test. They should be representative and current, and have adequate sample size. -Typically include individuals who represent the age and demographic characteristics of those for whom the test is intended.

(1) The larger the normative sample, the more likely it is to represent the population (2) A large norm group increases the stability and accuracy of statistical analysis

The size of a norm group holds importance for the following reasons:

purpose of the test and the use of the results

The type of norm-references score selected is dependent upon the:

percentile ranks, standard scores, grade and age equivalents

Types of norm-referenced scores

CEEB Scores

Used in educational testing (SAT, GRE) and have a mean of 500 with a standard deviation of 100.

Scale scores

Usually three-digit scores that have been converted from raw scores and their meaning varies depending on the test

Percentiles

______________________ are not equal-interval measurements, meaning that they are not equally spaced across all parts of a distribution: they are compressed near the middle of the distribution (where the majority of scores fall) and spread out near the tails (where there are fewer scores).

stratified sampling

involves choosing norm groups based on certain important variables, such as age, gender, education, ethnicity, socioeconomic background, region of residence, and community size.

norms

provide the basis for interpreting test scores

Representativeness

the degree to which the norm group represents the population for which the test was written.

interquartile range

the distance between the first and third quartiles of the data set; contains 50% of all values in the distribution

transformed scores

transformations of raw scores, which are more easily interpretable

standard scores

z scores, T scores, deviation IQs, CEEB scores, stanines, sten scores, and norm curve equivalents are all types of _______________________.

Percentages

• A form of raw score that reflects the number of correct responses obtained out of the total possible number of correct responses on a test. • Symbols for percentage score: % • Example: Shari scored 70 percent on her test. This means she answered 70% of the test items correctly.

Norm-referenced Interpretation

• Compare individual scores to the test scores of group of individuals. • The norm group must be clearly defined. • How well the test taker does in comparison with other individuals (the norm group)

z Scores

• Convert test scores into a standard deviation value, ranging from -3.0 to +3.0. • Tell us instantly how large or small an individual score is relative to other scores in the distribution

Percentile ranks

• Express the examinee's relative position on a norm-referenced test. • Range from 0 to 100 and indicate the percentage of scores that were lower than the examinee's. • Not equal-interval measurements

Normal Curve Equivalent

• Normalized standard score that ranges from 1 to 99, with a mean of 50 and a standard deviation of 21.06. • Used primarily in education, a score of 50 represents the national average of any grade level at the time of year the test was normed.

Standard Scores

• Refers to raw scores that have been converted to an interpretable scale that has a set mean and standard deviation. • Allow individual test scores to be interpreted in terms of the normal curve. • Linear transformations of raw scores, so they retain a direct relationship with raw scores

Meaningful Assessment Scores

• Scores are often difficult to interpret without some method for creating a baseline. • Criterion and norm-referenced interpretations can be used to make sense of individual scores.

Percentiles

• Scores that reflect the rank or position of an individual's test performance on a continuum from 0 to 99 in comparison to others who took the test. • Symbols for percentile rank: PR or %'ile • Example: Shari scored in the 80th percentile on the test, meaning that Shari scored better than 80 percent of the other individuals who took the test.

Stanine

• Short for STAndard NINE • Convert raw scores into values ranging from 1 to 9 • Mean of 5 with a standard deviation of 2 • Have a constant relationship to percentiles in that they represent a specific range of percentile scores in the normal curve

Sten Scores

• Short for Standard TEN • Scores range from 1 to 10, have a mean of 5.5 with a standard deviation of 2

T scores

• Use a fixed mean of 50 and a fixed standard deviation of 10, which eliminates the need for decimals and negative values. • Calculated using the z score as a base. • Often used to report personality test results

Deviation IQs

• Used in intelligence testing; they have a mean of 100 and a standard deviation of 15.

Criterion-Referenced Interpretation

• Uses some standard of performance to interpret an examinee's test results. • Interpreted in absolute terms and often have a cutoff score representing minimum competency. • The content domain must be clearly defined. • What the test taker knows or does not know with respect to a specified content domain. • Example- typical test from a high school teacher (Did the student learn the material?) 94+ A; 86 B; 70 C; etc

Raw score

A test score that represents the number of correctly answered questions coded in some specific manner, such as correct/incorrect, true/false, and so on.

Age Equivalents

A very general score that is used to compare the performance of children at the same age with one another. It is the estimated age level that corresponds to a given score. • Example: A score of 9-3 indicates that a person's performance is similar to the average performance of children 9 years and 3 months old.

Cutoff score

(i.e., the passing score); represents the minimum number of items the examinee must answer correctly in order to demonstrate mastery

Performance categories

-aka achievement levels or proficiency levels, they describe an individual's performance by sorting or classifying scores into categories based on the quality of performance, such as failing, basic, proficient, and advanced, or Level 1, Level 2, and so on -actually specified ranges of a particular score, such as scale scores