Premium
A STUDY OF THE STATISTICAL FOUNDATIONS OF GROUP CONVERSATION TESTS IN SPOKEN ENGLISH 1
Author(s) -
Liski Erkki,
Puntanen Simo
Publication year - 1983
Publication title -
language learning
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 2.882
H-Index - 103
eISSN - 1467-9922
pISSN - 0023-8333
DOI - 10.1111/j.1467-1770.1983.tb00536.x
Subject(s) - lexis , matriculation , pronunciation , linguistics , psychology , test (biology) , statistical analysis , conversation , mathematics education , statistics , mathematics , paleontology , philosophy , biology
This paper is intended as one statistical step towards achieving objectivity in group oral testing in English. The test and the marking system adopted were developed by Folland and Robertson (1976). The data under study comprise sample results from tests taken by 698 Finnish university students at the University of Tampere. The testees' performance was evaluated by noting errors under four categories—pronunciation, grammatical structure, lexis, and use—and “pluses” under use and lexis. The primary aim of this research has been to bring out statistical factors which may be useful in developing tests of spoken English. Distributions of errors are shown and studied in relation to variables such as number of utterances spoken by the testee, sex, mark in a matriculation examination, and group factors. The results show that most errors were made in grammatical structure and fewest in use. More talkative students showed better proficiency than less talkative when compared on the basis of the number of errors per ten utterances. Generally considered, the girls' proficiency was better than the boys', but this difference disappeared when the mark in the matriculation examination was taken into account: The mark was positively correlated with language proficiency and the girls' marks were significantly higher than those of the boys.