Computerized mastery testing system, a computer administered variable length sequential testing system for making pass/fail decisions
First Claim
1. A method of computerized mastery testing for deciding whether an examinee has attained a specified mastery of a given subject by using a computer system means, said method comprising:
- a. determining test specifications;
b. creating a pool of test items which meet the test specifications;
c. determining the dimensionally of the pool of test items;
d. obtaining item response theory (IRT) statistics and assessing IRT model fit for each test item;
e. determining mastery ability values;
f. assembling the test items into testlets;
g. verifying all of the testlets meet testing criteria;
h. selecting loss functions and prior probability of mastery;
i. estimating computerized mastery testing (CMT) cutscores which minimize expected posterior loss;
j. estimating expected error rates and average test length;
k. authoring and packaging the test for administration; and
l. administering presentation and scoring of the testlets in a sequential manner to the examinee.
4 Assignments
0 Petitions
Accused Products
Abstract
A computerized mastery testing system providing for the computerized implementation of sequential testing in order to reduce test length without sacrificing mastery classification accuracy. The mastery testing system is based on Item Response Theory and Bayesian Decision Theory which are used to qualify collections of test items, administered as a unit, and determine the decision rules regarding examinee'"'"'s responses thereto. The test item units are randomly and sequentially presented to the examinee by a computer test administrator. The administrator periodically determines, based on previous responses, whether the examinee may be classified as a nonmaster or master or whether more responses are necessary. If more responses are necessary it will present as many additional test item units as required for classification. The method provides for determining the test specifications, creating an item pool, obtaining IRT statistics for each item, determining ability values, assembling items into testlets, verifying the testlets, selecting loss functions and prior probability of mastery, estimating cutscores, packaging the test for administration, randomly and sequentially administering testlets to the examinee until a pass/fail decision can be made.
269 Citations
25 Claims
-
1. A method of computerized mastery testing for deciding whether an examinee has attained a specified mastery of a given subject by using a computer system means, said method comprising:
-
a. determining test specifications; b. creating a pool of test items which meet the test specifications; c. determining the dimensionally of the pool of test items; d. obtaining item response theory (IRT) statistics and assessing IRT model fit for each test item; e. determining mastery ability values; f. assembling the test items into testlets; g. verifying all of the testlets meet testing criteria; h. selecting loss functions and prior probability of mastery; i. estimating computerized mastery testing (CMT) cutscores which minimize expected posterior loss; j. estimating expected error rates and average test length; k. authoring and packaging the test for administration; and l. administering presentation and scoring of the testlets in a sequential manner to the examinee. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23)
-
-
24. A computerized mastery testing system having a computer system means for testing whether an examinee has attained specified mastery of a given subject comprising:
-
a. means for inputing test specifications and test items into a computer database; b. means for applying item response theory (IRT) so as to obtain IRT statistics for the test items and determine mastery ability values; c. means for assembling the test items into testlets; d. means for verifying the testlets meet testing criteria; e. means for selecting loss functions and the prior probability of mastery; f. means for applying Bayesian decision theory so as to estimate computerized mastery testing (CMT) cutscores which minimize the expected posterior loss and determine the associated operating characteristics; g. means for authoring and packaging the test for administration; and h. means for administering presentation and scoring of an examinee response dependent variable number of the testlets in a random, sequential manner to the examinee.
-
-
25. A method of computerized mastery testing for deciding whether an individual examinee has attained a specified mastery of a given subject by using a computer system means, said method comprising:
- determining test specifications;
creating a pool of test items which meet the test specifications;
determining the dimensionality of the pool of test items;
obtaining item response theory (IRT) statistics and assessing IRT model fit for each test item, wherein the pool of test items is unidimensional and said IRT statistics are estimated for each test item by using a joint maximum likelihood logistic estimation procedure implemented by executing LOGIST calibration program;
determining mastery ability values;
assembling the test items into testlets by using the computerized mastery testing (CMT) testlet assembly procedure;
verifying all of the testlets meet testing criteria;
selecting loss functions and prior probability of mastery;
estimating CMT cutscores which minimize expected posterior loss, wherein the CMT cutscores are determined for a variable-length test by executing the CMT cutscore estimation program;
estimating expected error rates and average test length;
authoring and packaging the test for administration; and
administering presentation and scoring of the testlets in a sequential manner and wherein the testlets are chosen in a random manner for all testlets not as yet administered to the examinee by executing the CMT testlets presentation and scoring procedure.
- determining test specifications;
Specification