skip to main content skip to footer

English as a Foreign Language in Ten Countries. Factors Associated With Individual Student Differences in Achievement in English: The Regression Analyses ESL

Massad, Carolyn E.
Publication Year:
Report Number:
ETS Research Memorandum
Document Type:
Page Count:
Subject/Key Words:
Belgium, Chile, Educational Environment, English as a Second Language (ESL), Factor Analysis, Family Environment, Federal Republic of West Germany, Finland, Hungary, Individual Differences, Israel, Italy, Language Proficiency, Motivation, Multiple Regression Analysis, Netherlands, Sweden, Thailand


A regression analysis was used to determine the extent to which achievement in English as a foreign language could be predicted. ESL students in 10 countries were studied. A relatively non- technical and brief description of the procedures used in analyzing the data is given for readers who are not statistical specialists. The regression analysis was carried out in stages by adding blocks of variables in an order reflecting chronology of impact on the student. The blocks were 1) home circumstances and sex of student, 2) type of school and type of ESL program, 3) four time variables—years studied English, grade in which currently enrolled, grade began English, and age, 4) school, teachers, and student variables, 5) kindred variables--attitudes and practices measured by questionnaire responses which were related to achievement but whose causal influence was partly in doubt, and 6) score on the word knowledge test. The criterion variables selected were English reading and English listening. Conclusions include: 1) The data about the home environment and type of school or program allow a fairly good prediction of a 14-year-old student's achievement in both English reading and listening, but the predictive value of these variables falls considerably by the end of secondary education. 2) School factors (school, teacher and student variables) were very important predictors for students at the end of secondary school, but very little consistent pattern was found in the data, 3) Kindred variables are important predictors for the older students but not for the 14-year-olds. Only a motivational factor tends to emerge as a consistent predictor. 4) Work knowledge tends to be a predictor but is the least important of the groups of variables in the analysis.

Read More