چكيده
One of the aims of this thesis was to explore the area of metadiscourse and its go-togetherness with readability by means of an analysis of different text types. The analyzed text types were selected from a range of 249-299 propositions. These texts consisted of 20 narrative; 11 argumentative; and 14 expository texts. A comparative chi-square analysis for investigation of the focused text types was conducted which did not show any significant differences among the three text types in terms of their frequencies of explicit verbal metadiscourse modalities. For determination of the degree of go-togetherness of explicit verbal metadiscourse frequencies and readability indices of the texts, idea densities as the representatives of readability indices were calculated using CPIDR software. The results of Spearman Rank order correlation indicated no significant relationship between the use of explicit verbal metadiscourse and idea densities of texts. Moreover, One Way ANOVA analysis was performed to compare readabilities of texts calculated by idea density formula. The results indicated significant differences between argumentative and narrative texts with higher means and as a result higher densities and lower readabilities of narrative texts. For qualitative analysis of readability, three sets of questionnaires each testing content, motivation, quality of language and format as the constructs affecting readability, were developed. The overall Cronbach Alpha reliability of the piloted questionnaires was 0.88. Each questionnaire tested perceptions of readabilities for each text type. Finally, for comparing the results, Kruscall Wallis test was performed which showed significant differences among the perceptions of learners on text difficulty. Then, Mann Whitney U test analyses were conducted to determine exact points of difference. The overall results indicated that the participants had considered argumentative texts as the most readable and expository texts as the least readable texts. Moreover, narrative texts were perceived as less readable than argumentative and more readable than expository texts. As can be understood only the finding concerning higher perceived readability of argumentative texts in comparison with narrative ones, has been in agreement with the results of idea density calculation. One of the implications of the present study in syllabus design and material development is selection of appropriate texts for teaching EFL learners. Also, the present study can help writers improve the readabilities of their texts using factors other than the current formula
