Abstract:
The Combined English Language skills assessment in a Reading Context, or CELSA, is a standardized, multiple-choice cloze test intended for use with adults and high school students in open enrollment as well as academic ESL/EFL programs. It is appropriate for making placement decisions and evaluating the effectiveness of instruction through pre- and post-testing, but should be used in conjunction with some measure of ability with spoken language since CELSA measures grammar and reading alone. The test is available in beginning, intermediate, and advanced levels, each of which have two forms. Most test items have a conversational format with four choices possible as responses. Distractors reflect common errors made by second language learners and correct answers were determined by a survey of native-speaker responses. A test form may be used for retesting after 2.5 months, but the equivalent form may be used after only 6 weeks. The Kuder-Richardson Formula 21 shows a reliability across test forms of .90. Tests may be scored by hand or by machine and answer sheets may be duplicated. Scores are interpreted by the test administrator using a formula that takes into account the number of levels in a program of instruction. The result is the designation of an instructional level for each examinee that corresponds with the specific program in which they are enrolled, with seven being the maximum number of discriminable levels. A study of the reliability and validity of CELSA involved ESL programs at eight different institutions. Reliability coefficients ranged from .88 to .97. CELSA was cross-validated on a variety of other ESL tests with correlation coefficients ranging from .65 to .94. The User's Guide includes references for studies used during test development as well as the validation study used afterwards.