In recent years, research with both first and second language students has suggested that the standard nth word cloze -procedure could be improved as a measure of reading comprehension through the introduction of multiple-choice or matching cloze formats. In addition, research conducted mainly with first language students, has suggested that the every nth word deletion procedure should be restricted so as to delete only the lexical words in the passage. In this study, three different deletion systems, presented in a matching cloze format, were evaluated as measures of reading comprehension for seventh grade ESL students from Papua New Guinea. The possible interaction of the three deletion types with these students’ vernacular background was also investigated.