Options
Testing oral proficiency in English at the secondary level
Loading...
Type
Thesis
Author
Tan, Winnie Swee Ling
Supervisor
Read, John A. S.
Abstract
This study set out to measure the reliability of the oral interview as a testing technique, and to investigate the validity of other less direct speaking tests as measures of oral proficiency. The subjects were forty students from an average government school in Singapore. Stratified non-random sampling was employed to ensure that the subjects were representative of the school populations. The raters were eight teachers with varying degrees of teaching experience from four different schools. For the interview, two forms of assessment were investigated -an Analytic method and an Impressionistic method. Rating scales for both methods were drawn up in conjunction with a sample tape with calibrated levels of proficiency. All responses were recorded on tape and rated later.
The results indicated that the Oral Interview was a reliable measure of oral proficiency. Of the two forms of assessment, the, the Impressionistic scoring was shown to be slightly more reliable than the analytic scoring . Generally , the raters were shown to agree substantially in their ratings. It was found that the reliability coefficients increased when the average ratings of a pair of raters were taken.
Claims that the oral interview tends to be subjective, and unreliable in the hands of untrained raters led to the need to consider other forms of testing that can be scored more objectively. Four semi-direct speaking tasks were investigated. They were : Reading Aloud, Picture-Cued Description , Elicited Imitation and situational response. Concurrent validity was established by using the Oral Interview as a criterion measure against which all the other tests were compared. Of these, Reading Aloud and Picture-Cued Description appeared to be valid substitutes for the oral interview. Situational Response was seen as a complementary test incorporating communicative strategies, an area of language proficiency not tested in the the traditional test of oral proficiency. The least effective measure was Elicited Imitation. More research needs to be done in this area before it can be considered in a test of oral proficiency.
The results indicated that the Oral Interview was a reliable measure of oral proficiency. Of the two forms of assessment, the, the Impressionistic scoring was shown to be slightly more reliable than the analytic scoring . Generally , the raters were shown to agree substantially in their ratings. It was found that the reliability coefficients increased when the average ratings of a pair of raters were taken.
Claims that the oral interview tends to be subjective, and unreliable in the hands of untrained raters led to the need to consider other forms of testing that can be scored more objectively. Four semi-direct speaking tasks were investigated. They were : Reading Aloud, Picture-Cued Description , Elicited Imitation and situational response. Concurrent validity was established by using the Oral Interview as a criterion measure against which all the other tests were compared. Of these, Reading Aloud and Picture-Cued Description appeared to be valid substitutes for the oral interview. Situational Response was seen as a complementary test incorporating communicative strategies, an area of language proficiency not tested in the the traditional test of oral proficiency. The least effective measure was Elicited Imitation. More research needs to be done in this area before it can be considered in a test of oral proficiency.
Date Issued
1981
Call Number
PE1135 Tan
Date Submitted
1981