U.S. patent application number 14/562218 was filed with the patent office on 2015-06-11 for methods for improving test efficiency and accuracy in a computer adaptive test (cat).
The applicant listed for this patent is ACT, INC.. Invention is credited to Wugen Dai, Lingyun Gao, Lisa Gawlick, Nancy Petersen, Changhui Zhang.
Application Number | 20150161902 14/562218 |
Document ID | / |
Family ID | 53271749 |
Filed Date | 2015-06-11 |
United States Patent
Application |
20150161902 |
Kind Code |
A1 |
Gawlick; Lisa ; et
al. |
June 11, 2015 |
METHODS FOR IMPROVING TEST EFFICIENCY AND ACCURACY IN A COMPUTER
ADAPTIVE TEST (CAT)
Abstract
A method for test item selection is provided that includes a
computer implemented test battery having at least two or more
sections with a plurality of test items. An ability estimate is
calculated from an earlier section(s) of the at least two or more
sections and an initial item and subsequent items for a subsequent
section are selected from the plurality of test items based upon
the ability estimate(s) from the earlier section(s). Use of a more
informative initial ability estimate in the item selection process
can improve interim ability estimation accuracy and item selection
efficiency while keeping item exposure and item usage rates at
acceptable levels.
Inventors: |
Gawlick; Lisa; (Iowa City,
IA) ; Zhang; Changhui; (Coralville, IA) ;
Petersen; Nancy; (Solon, IA) ; Gao; Lingyun;
(Coralville, IA) ; Dai; Wugen; (Iowa City,
IA) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
ACT, INC. |
Iowa City |
IA |
US |
|
|
Family ID: |
53271749 |
Appl. No.: |
14/562218 |
Filed: |
December 5, 2014 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
61912774 |
Dec 6, 2013 |
|
|
|
Current U.S.
Class: |
434/362 |
Current CPC
Class: |
G09B 7/00 20130101 |
International
Class: |
G09B 7/00 20060101
G09B007/00 |
Claims
1. A method of test item selection comprising; providing a computer
and a computer implemented test battery comprising: a plurality of
test sections in the computer implemented test battery, each test
section having a set of test items selected from a plurality of
test items; an item selection process; administering one test
section from the plurality of test sections to an examinee using
the computer, wherein the computer receives the examinee's
responses to the set of test items in the one test section; and
informing the item selection process for at least one subsequent
test section based on scores from the one test section or previous
test sections.
2. The method of claim 1 further comprising: calculating an initial
ability estimate from the examinee's responses to the set of test
items in the one test section.
3. The method of claim 2 further comprising: basing the item
selection process for the at least one subsequent test section on
the initial ability estimate.
4. The method of claim 1 wherein the computer implemented test
battery comprises a computer adaptive test battery further
comprising related tests or test sections.
5. The method of claim 1 further comprising: matching item
difficulty in the at least one subsequent section with an estimated
examinee ability from the one test section or previous test
sections.
6. The method of claim 1 further comprising: improving interim
ability estimation accuracy and item selection efficiency by
informing the item selection process with ability estimates from
the one test section or previous test sections.
7. A method of test item selection comprising; providing a computer
and a computer implemented test battery comprising at least two or
more test sections having a plurality of test items; administering
one section of the at least two or more test sections to an
examinee using the computer, wherein the computer receives the
examinee's responses to a set of the plurality of test items for
the one section; calculating an initial ability estimate for the
examinee's responses to the set of the plurality of test items in
the one test section; and selecting one or more test items from the
plurality of test items to include in a subsequent test section to
the one test section of the at least two or more test sections
based upon the initial ability estimate from at least the one
previous test section.
8. The method of claim 7 further comprising: providing an item
selection process for selecting the one or more test items based
upon the initial ability estimate from at least the one previous
section.
9. The method of claim 7 wherein the computer implemented test
battery comprises a computer adaptive test.
10. The method of claim 8 further comprising: factoring into the
item selection process interrelatedness between the at least the
one test section and the subsequent test section.
11. The method of claim 7 further comprising: minimizing
overexposure and over usage of selected test items from the
plurality of test items by basing item selection in the subsequent
test section on the ability estimate from at least the one previous
section.
12. The method of claim 7 further comprising: minimizing the
plurality of test items to a subset of test items based upon the
initial ability estimate from at least the one previous
section.
13. The method of claim 8 further comprising: improving interim
ability estimation accuracy and item selection efficiency by
informing the item selection process with ability estimates from
the one test section or previous test sections.
14. A method of test item selection comprising; providing a
computer and a computer implemented test battery comprising at
least two or more test sections having a plurality of test items;
administering one section of the at least two or more test sections
to an examinee using the computer, wherein the computer receives
the examinee's responses to a set of the plurality of test items
for the one test section; calculating an initial ability estimate
from the examinee's responses to the set of the plurality of test
items in the one test section; minimizing the plurality of test
items to a subset of test items based upon the initial ability
estimate from at least the one previous test section; and selecting
a next test item from the subset of test items for the subsequent
test section.
15. The method of claim 14 further comprising: adapting the next
test item selection to scores from at least the one previous
section.
16. The method of claim 14 further comprising: calculating an
initial theta for the subsequent test section based on an
examinee's ability from at least the one previous test section,
wherein the at least one previous test section is related to the
subsequent test section.
17. The method of claim 14 further comprising: selecting the next
test item from the subset of test items to include in the
subsequent test section to the one test section of the at least two
or more test sections based upon ability estimates from at least
the one previous test section.
18. The method of claim 14 further comprising: factoring
interrelatedness between the at least two or more test sections
into the next test item selection.
19. The method of claim 14 further comprising: minimizing
overexposure and over usage of selected test items from the
plurality of test items by basing next test item selection in the
subsequent test section on an interim ability estimate arrived at
from the initial ability estimate.
20. The method of claim 7 further comprising: improving interim
ability estimation accuracy and item selection efficiency by
informing the next test item selection with an interim ability
estimates resulting from the initial ability estimates informed by
ability estimates obtained from one test section or previous test
sections.
Description
CROSS-REFERENCE TO RELATED APPLICATIONS
[0001] This application claims priority under 35 U.S.C. .sctn.119
to provisional application Ser. No. 61/912,774, filed Dec. 6, 2013,
which is hereby incorporated in its entirety.
BACKGROUND
[0002] I. Field of the Disclosure
[0003] The present disclosure relates to computer adaptive testing.
More specifically, but not exclusively, the present disclosure
relate to methods for improving ability estimation accuracy and
item selection efficiency in a Computer Adaptive Test (CAT).
[0004] II. Description of the Prior Art
[0005] In educational assessments, a test battery is usually
composed of several related sections based on content categories.
The inter-relationships among different sections can be used to
improve the test efficiency in a Computer Adaptive Test (CAT)
scenario.
[0006] Therefore, a primary object, feature, or advantage of the
present disclosure uses an examinee's ability estimate from an
earlier section to inform the selection of the initial item and the
subsequent items in a later section.
[0007] To date, a number of studies have been conducted on
selecting the initial item based on an examinee's scores from
earlier related tests using a variety of methods. However, none of
the methods used in these studies have provided satisfactory
results. Furthermore, using prior ability estimates to drive the
starting point for a subsequent test section has not been widely
used and practiced.
[0008] Therefore, it is another object, feature, or advantage of
the present disclosure to provide a method for using the ability
estimates from a previous test or test section(s) as prior
information to inform the selection of the initial item of a
subsequent test or section of a test.
[0009] Another object, feature, or advantage of the present
disclosure is to improve methods for item selection and ability
estimation in a CAT for accurately measuring an examinee's
ability.
[0010] A still further object, feature, or advantage of the present
disclosure is to provide methods for initial item selection based
on ability estimates from a previous section.
[0011] One or more of these and/or other objects, features or
advantages of the present disclosure will become apparent from the
specification and claims that follow.
SUMMARY
[0012] The present disclosure improves test efficiency and accuracy
in a Computer Adaptive Test (CAT).
[0013] One exemplary method is for test item selection. One example
of such includes a computer and a computer implemented test battery
having a plurality of test sections. Each test section has a set of
test items selected from a plurality of test items. An item
selection process is also provided. At least one section from the
plurality of test sections is administered to an examinee using the
computer. The computer is configured to receive the examinee's
responses to the set of test items in the one section. The item
selection process for at least one subsequent test section is
informed based on scores from the one test section or previous test
sections.
[0014] According to another aspect, a method for test item
selection is provided. A computer and computer implemented test
battery are included along with at least two or more test sections
having a plurality of test items. One test section of the at least
two or more test sections is administered to an examinee using the
computer. The computer is adapted to receive the examinee's
responses to a set of the plurality of test items for the one test
section. An initial ability estimate for the examinee's responses
to the set of the plurality of test items in the one test section
is calculated. One or more test items from the plurality of test
items are selected to include in a subsequent test section to the
one test section of the at least two or more test sections based
upon the initial ability estimate from at least the one previous
test section.
[0015] According to still another aspect, a method for test item
selection is provided. The item selection method includes a
computer and a computer implemented test battery having at least
two or more test sections with a plurality of test items. One
section of the at least two or more test sections are administered
to an examinee using the computer. The computer receives the
examinee's responses to a set of the plurality of test items for
the one test section. An initial ability estimate is calculated for
the examinee's responses to the set of the plurality of test items
in the one test section. The plurality of test items are minimized
to a subset of test items based upon the initial ability estimate
from at least the one previous test section. A next test item is
selected from the subset of test items for the subsequent test
section.
BRIEF DESCRIPTION OF THE DRAWINGS
[0016] Illustrated embodiments of the present disclosure are
described in detail below with reference to the attached drawing
figures, which are incorporated by reference herein, and where:
[0017] FIG. 1 is a flowchart of a process for item selection in
computer adaptive testing in accordance with an illustrative
embodiment; and
[0018] FIG. 2 is a block diagram providing an overview of an item
selection process for computer adaptive testing in accordance with
an illustrative embodiment.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0019] The present disclosure provides for various computer
adaptive testing methods. One exemplary method includes a method
for item selection and ability estimation in a computer adaptive
test. The accuracy and efficiency of a computer adaptive test is
improved by making the test more succinct and/or accurate, and thus
more effective. What results is a testing platform using a computer
adaptive test that can estimate an examinee's ability using the
examinee's response information relating to a specific set of
operational test items.
I. Item Selection in a Computer Adaptive Test (CAT)
[0020] In educational assessments, a test battery is usually
composed of several related sections based on content categories.
The inter-relationships among different sections can be used to
improve the test efficiency in a computer adaptive testing (CAT)
scenario. For example, an examinee's ability estimate from an
earlier section can inform the selection of an initial item and the
subsequent items in a later section. The methods used for item
selection and ability estimation in CAT are essential for
accurately measuring an examinee's ability.
[0021] a. Illustrative Embodiments for Item Selection in a CAT for
a Test Battery
[0022] FIG. 1 provides an illustrative flowchart diagram for an
exemplary process for item selection in CAT for a given test
battery. Aspects of the disclosure contemplate interfacing with,
displaying and receiving examinee input via a computer adaptive
testing platform. A workstation, such as the one illustrated in
FIG. 2, may be used to administer a computer implemented test
battery. The workstation may be part of a computer network in which
any embodiment of the present disclosure may be implemented. As
shown, the workstation or computer network (not shown) includes,
for example, a server, workstation, scanner, a printer, a
datastore, and other linked networks. The computer networks may be
configured to provide a communication path for each device of a
computer network to communicate with other devices. Additionally,
the computer network may be the internet, a public switchable
telephone network, a local area network, private wide area network,
wireless network and any of the like. In various embodiments, an
automated computer adaptive testing module, operated, at least in
part, by an operating protocol, may be executed on a server and/or
workstation, similar to the workstation shown in FIG. 2. For
example, in one embodiment or aspect of the application, the server
may be configured to execute a computer adaptive testing module
using an operating protocol individually or co-operatively with one
or more workstations. The scanner may be configured to scan textual
content and output the content in a computer readable format.
Additionally, the printer may be configured to output the content
to a print media, such as paper. Furthermore, data associated with
any one of the computer adaptive testing modules, test script
administration process or interim ability estimation process may be
stored on the datastore. The datastore may additionally be
configured to receive and/or forward some or all of the stored
data. Moreover, in yet another aspect, some or all of the computer
network or workstation may be subsumed within a single device.
Although a workstation is depicted in FIG. 2, it is understood that
the disclosure is not limited to operation within a computer
workstation, but rather, the disclosure may be practiced in or on
any suitable computer network or electronic device. Accordingly,
the workstation depicted in FIG. 2 is for illustrative purposes
only and thus is not meant to limit the present disclosure in any
respect.
[0023] Using, for example, a test script administration process or
application, a section or item selector is used to select a first
test section from a test battery. The selected test section may be
displayed using one of the aforementioned interface pieces or other
like electronic device, whereby for example a plurality of test
items from a first selected section are displayed. Through a
workstation, computer network, or other like electronic device,
examinees' input, answers or responses are received in response to
the plurality of test items from a first selected section of a test
battery being presented using a test scripted administration
process or application. Operably configured on a workstation, such
as the one illustrated in FIG. 2, a computer network or other like
electronic device is an interim ability estimator using one or more
estimation methods for calculating an examinee's ability estimate
for the first selected section of test items for a test battery. In
the following description, several estimation methods are
discussed, which the interim ability estimator application or
module may use to calculate an examinee's ability estimate for a
first selected section of a plurality of test items in a test
battery. One or more operating protocols on the workstation shown
in FIG. 2, or a computer network, or another like electronic device
may be used to inform selection of an initial test item for next
test section in the test battery based upon an examinee's ability
estimate from the first or previous selected section.
[0024] One comparative method selects a first item of a later
section randomly from a pool of items most informative near the
initial fixed theta value and then uses one or more estimation
methods to estimate an examinee's interim ability. For example, the
maximum likelihood (ML) estimator can be used to estimate an
examinee's interim ability upon responding to each item, which was
then used for selecting next items in subsequent sections of the
test battery. Other exemplary methods herein use a Bayesian
approach. The ability estimates from the first section may be used
as prior information to select the first item of a subsequent or
second section in a test battery. The expected a posteriori (EAP)
estimation method can also be used at every step to obtain an
examinee's interim ability which may then be used for selecting
subsequent items in subsequent sections of a test battery.
[0025] b. Illustrative Application(s)
[0026] In practice, the existing CAT programs involving multiple
test sections usually use the same initial ability for all
examinees without considering the interrelatedness between the test
sections. This practice is limited because it involves arbitrary
choices of ability values and items for each examinee. If the
initial ability estimate is away from the true ability of the
examinee, the CAT procedure will have an inaccurate start. For
example, starting the test with a mid-level difficulty item for a
high- or low-proficiency examinee would take longer to arrive at an
accurate estimate of his/her ability. This, to a large degree,
affects the efficiency of item selection. Moreover, initialization
at the same ability estimate for all examinees leads to first items
in the test that are always chosen from the same subset in the
pool. Hence, these items are overexposed (Van der Linden &
Pashley, 2010).
[0027] In the present disclosure, using an initial ability estimate
predicted by each examinee's performance on earlier related test
section(s) provides an individualized initialization for adaptive
tests. That is, different examinees will start at different initial
ability estimates. Further, using an initial theta predicted by
each examinee's performance on earlier section(s) provides an
individualized prior that is located near each examinee's true
ability. The use of a predicted initial theta and an individualized
prior that is continuously improved during the test using
additional information obtained from the individual examinee will
improve item selection and speed up convergence of the ability
estimates. Finally, the present invention helps improve item
exposure. The empirical initialization of the test entails leads to
a variable entry point to the pool, and hence offers a more even
exposure of its items (Van der Linden & Pashley, 2010). The
item selection and ability estimation procedure proposed in the
present invention can be integrated with the current algorithm for
delivering the CAT tests.
[0028] II. Other Embodiments and Variations
[0029] The present disclosure is not to be limited to the
particular embodiments described herein. In particular, the present
disclosure contemplates numerous variations in the type of ways in
which embodiments of the disclosure may be applied to computer
adaptive testing. The foregoing description has been presented for
purposes of illustration and description. It is not intended to be
an exhaustive list or limit any of the disclosure to the precise
forms disclosed. It is contemplated that other alternatives or
exemplary aspects that are considered are included in the
disclosure. The description is merely examples of embodiments,
processes or methods of the invention. For example, the methods
used to select an initial item in a subsequent section based upon
ability estimates from a previous section are not limited to those
disclosed herein. It is understood that any other modifications,
substitutions, and/or additions may be made, which are within the
intended spirit and scope of the disclosure. For the foregoing, it
can be seen that the disclosure accomplishes at least all of the
intended objectives.
* * * * *