Inferencing and syntactic complexity as two determinants of comprehension difficulty

Abstract

This investigat:i.ondemonstrates that school-leavers" experience the greatest amount pf difficulty in answering comprehension questions that demand inferencing related to stylistic effects in written passages. It also demonstrates that the texts use,d in Transvaal Education Department written comprehension examinations and the questions<asked on the texts ar,enot syntactica.lly difficUlt to.process. A brief introduction to comprehension testing in the Republ)~c of •South Africa is followed bY an explanation as to how 2077 scripts from the Transvaal province were seleqted and treated" to establish which questions ma't:riculatioU candid~tes found most, difficult and which the easiest. \\t'he items most ca.ndidates<,failed, and those for which most scored 70% er'more, were then categorised accol:'d.in,•tg•o two , different taxonomies"" to,reveal that the c(}f.ficultquestioIj~ reguire inferencing procedures that easy questions do not. ,/, ,;;:-~ . , ... . .... , _ _ _ ,f,~", ' __ ' ' , , ." ", ,,' ,',", " __ ~~;;;o<'; T,O,conf-rm that comprehensJ.on'·dJ.ffl.cultyfor school"'leavers stems ~rom what they are asked to infer, and not from the inhere\~t syntactic complexity of the t~,ts or questions the:mse:l)ves,the syntax of both,the difficult and easy items is scrutinized. The inferencing tasks demanded of candida~es in the six ~ost difficult questions are desc:.;:ih~d in detai~l."and,sonte,of the findings of this investigation are related·to ongoing research in Britain. The study questions. whe~;:::(er c(;)mp::.~ehertsa1b0inlity is best tested by the kinds of g:jlestir\lclRit'nldSic1atesfound :most difficult and suggests ways teachers may use to prepare candidates for a comprehem~ion examination of this kind.