Performance testing holds great promise and appeal. However, before jumping into a performance testing solution, the test sponsor should carefully consider what the test is attempting to measure.

While Alpine Testing Solutions believes that performance tests can offer great advantages over typical selective-response tests, the cost of the performance test may far exceed the value it provides.

It is important to first understand what the test must measure and then determine the most appropriate testing vehicle. In some situations a pure performance test is the best option. In others, a hybrid of performance tasks, multiple choice, and advanced item types is most appropriate. And there are many times when multiple choice and advanced item types are all that is needed.

Alpine Testing Solutions provides a reasoned test design methodology using the most appropriate test and item types.

Performance tests are designed to measure an examinee’s ability to perform real-world tasks.  A performance test typically utilizes a testing environment that is representative of the actual job environment, or at least allows the examinee to perform a task in a fashion that is representative of the actual job environment.

There is more to building a performance test than simply creating and delivering a performance environment. While performance tests can provide great face validity, due diligence must be taken with all aspects of test design, validation, and psychometrics or the test sponsor will likely not reap the benefits promised by performance testing.

Alpine Testing Solutions has the expertise to help your organization make good decisions around the design, development, and maintenance of a performance test.

Some examples of the value that Alpine Testing Solutions can provide when developing a performance test include:

Task Clusters. An important output of the job/domain analysis are task clusters (also known as work models). A task cluster is a synthesis of knowledge, skills, and judgments that are implemented holistically within a relevant, real-world context. The task clusters are the foundation and driving force behind the objectives, content weighting, performance scenario specifications, and scoring rubrics. If a performance test is designed from the decomposed knowledge, skills, and judgments that are normal outputs of a job/domain analysis rather than from task clusters, there is a good chance that the performance test is providing no more information about the test-taker than could be obtained from a traditional – and much less expensive – multiple-choice test.

Contact Us

Phone: 801-226-4283

E-mail: Testing_Solutions@alpinetesting.com

(Comments to webmaster@alpinetesting.com)

Increased Reliability. A test-taker typically requires far more time to complete a performance scenario than a multiple choice item. Therefore, a performance test will have far fewer “items” than a multiple choice test. This can negatively impact reliability because reliability is dependent upon the number of items – or scoring opportunities – in the test. A well designed performance test will provide adequate scoring opportunities within each performance scenario to maintain reliability as compared to a multiple choice form. A well designed performance test will also make it possible to “fast forward” a candidate through trivial tasks that enable performance of the important tasks. A fast forward design allows for more scoring opportunities in the time available – therefore allowing us to learn more about a test-taker’s level and breadth of competence.

Innovative Technology Solutions. Lab, simulation, virtual environment, delivery, scoring, data management – the technology issues involved with performance testing can be staggering because there is not a “standard” approach for developing and delivering a performance test as there is with multiple choice tests. The ability to select and integrate the best mix of technologies for a specific performance test will greatly impact the usefulness of the test.

Advanced Data Analysis.  Alpine Testing Solutions has vast experience with classical item statistics, Rasch, and multiple-parameter IRT. We are able to determine the most appropriate approach and model for psychometric analysis of performance tasks, enabling the test sponsor to make informed decisions about the tasks and test.