Abstract
Purpose. The purposes of the study were: (1) to isolate a set of variables which could be used to predict text difficulty in Malay and (2) to construct a practical, empirically based readability formula for Malaysian prose.Procedure. This was a text analysis study in which grade level designations of textbook passages served as the dependent variable. Two textbooks each were chosen from language, social science and physical science textbooks used in Malaysian schools from grade two to grade eleven. Six 100-word sample passages were randomly selected from each book. Ten structural variables were selected as independent variables. Tabulations for the ten independent variables were made based on each of 36 passages at each grade level. Thus, there were 360 observations at each grade level and 3,600 in the entire study. Stepwise multiple regression analyses were performed in order to identify the regression equations which best predicted the grade levels of the textbooks. In order to identify patterns of relationship among variables, the ten independent variables were factor analyzed. All analyses were completed using SPSS.Findings. In the factor analysis, two major factors were extracted, a "vocabulary" factor and a "sentence" factor.The results of the multiple regression analyses indicated that two simple variables, syllables per 300 words and average sentence length could be used to predict grade level designations in all the three subject areas. Coefficients of determination ranged from .84 to .93. Three separate readability formulae were derived, one for each subject area.Conclusions. (1) Structural variables are excellent predictors of text difficulty in Malay. (2) Using 300-word samples yields better estimates of readability than 100-word samples.Recommendation. In addition to implementing the use of readability formulae in Malaysia, recources should be dedicated to "explaining" the specific variables associated with readability predictions. High correlations alone do not provide such explanations.