Open AccessACCOUNTING FOR RANDOM RESPONDING AT THE END OF THE TEST IN ASSESSING SPEEDEDNESS ON THE TEST OF ENGLISH AS A FOREIGN LANGUAGE
Author(s) -
Secolsky Charles
Publication year - 1989
Publication title -
ets research report series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.235
H-Index - 5
ISSN - 2330-8516
DOI - 10.1002/j.2330-8516.1989.tb00337.x
Subject(s) - test of english as a foreign language , test (biology) , psychology , point (geometry) , section (typography) , set (abstract data type) , degree (music) , statistics , mathematics education , social psychology , english language , mathematics , computer science , paleontology , physics , geometry , acoustics , biology , programming language , operating system
The usual assessment of speededness for rights‐only scored tests such as the Test of English as a Foreign Language (TOEFL) does not account for the possibility that examinees respond in a random or patterned fashion to the items at the end of the test as the time limit approaches. Therefore, for TOEFL, speededness measures that are based only on the number of items not reached may underestimate the degree to which the test is actually speeded. The present study represented an attempt to determine if Sections 2 and 3 of TOEFL are truly speeded according to established criteria. Two exploratory techniques employing regression analyses were used in an attempt to account for the possibility that examinees responded randomly or in a patterned fashion to unconsidered items at the end of each section. One technique provided an estimate of the degree to which all examinees truly reached the 75 percent point on the sections. The second technique provided an estimate of the degree to which examinees truly completed the last set of items. Support for the results was obtained from an examination of the number of items not reached and the number of items at the end of the sections to which examinees responded in a patterned fashion. The findings are limited to the extent that one can attribute items not reached and patterned responding to the effect of speededness. Four administrations of TOEFL were studied: two non‐pretest administrations and two longer pretest administrations. The results suggest that Section 3 for pretest administrations may be slightly speeded. It is recommended that more observational or survey methods be used to confirm this finding. If the finding is confirmed, it would be recommended that the TOEFL program investigate ways to increase the amount of time allotted per item.