Based on my experience with the introduction of ECD, I suggest
that people who plan to initiate the use of ECD with test developers
who have not previously used it consider the following steps. 1) Do
not promise more benefits than ECD can deliver. Item writing will
not become effortless, nor will pretesting become unnecessary. 2) Be
clear that more work is required at the beginning stages of test
development than people are used to doing and that increases in
efficiency will not immediately be apparent. 3) Point out the many
ways in which ECD is similar to traditional test development and
provide realistic rationales for the differences. 4) Introduce the
vocabulary of ECD as it is needed rather than all at once. Provide
rationales for the use of the new terms. 5) Do not assume that
reading articles about ECD is sufficient. Hands-on training is required.
If possible, the instructor should be an experienced test developer
who has credibility with the participants. 6) Provide examples of the
tools of ECD that are clearly relevant to the work that will be done.
7) Build extra time in the schedule. Expect and encourage iteration.
Not everything that is attempted will work the first time. 8) As work
proceeds, try to ensure cooperation and communication across the
layers of the process. 9) If possible, go through the entire ECD process
with one test rather than attempting to introduce ECD in many tests
simultaneously. 10) After the completion of each major segment of
the work, hold an “after action review” to determine what worked
and what caused problems. Discuss how to avoid the problems in the
next iteration.
Based on my experience with the introduction of ECD, I suggestthat people who plan to initiate the use of ECD with test developerswho have not previously used it consider the following steps. 1) Donot promise more benefits than ECD can deliver. Item writing willnot become effortless, nor will pretesting become unnecessary. 2) Beclear that more work is required at the beginning stages of testdevelopment than people are used to doing and that increases inefficiency will not immediately be apparent. 3) Point out the manyways in which ECD is similar to traditional test development andprovide realistic rationales for the differences. 4) Introduce thevocabulary of ECD as it is needed rather than all at once. Providerationales for the use of the new terms. 5) Do not assume thatreading articles about ECD is sufficient. Hands-on training is required.If possible, the instructor should be an experienced test developerwho has credibility with the participants. 6) Provide examples of thetools of ECD that are clearly relevant to the work that will be done.7) Build extra time in the schedule. Expect and encourage iteration.Not everything that is attempted will work the first time. 8) As workproceeds, try to ensure cooperation and communication across thelayers of the process. 9) If possible, go through the entire ECD processwith one test rather than attempting to introduce ECD in many testssimultaneously. 10) After the completion of each major segment ofthe work, hold an “after action review” to determine what workedand what caused problems. Discuss how to avoid the problems in thenext iteration.
การแปล กรุณารอสักครู่..
