Academia.eduAcademia.edu

Figure 3 – uploaded by Saeed Mehrpour

See full PDFdownloadDownload figure
Table 3: Descriptive statistics for the four sets of raw scores MS = Major/Short, ML = Major/Long, NS = Non-major/Short, NL = Non-major/Long

Table 3 Descriptive statistics for the four sets of raw scores MS = Major/Short, ML = Major/Long, NS = Non-major/Short, NL = Non-major/Long

Related Figures (4)

Table 1: Descriptive statistics and reliability indexes for the two versions of the test As can be seen from the above table, the means obtained for the short and the original version of the test (20.70 and 20.22, respectively) are quite close to each other. In order to see whether the marginal difference existing between the means obtained for the short and original version of the test was statistically significant, we ran an independent t-test. The following table shows the result obtained from this statistical analysis.
Table 1: Descriptive statistics and reliability indexes for the two versions of the test As can be seen from the above table, the means obtained for the short and the original version of the test (20.70 and 20.22, respectively) are quite close to each other. In order to see whether the marginal difference existing between the means obtained for the short and original version of the test was statistically significant, we ran an independent t-test. The following table shows the result obtained from this statistical analysis.
Table 4: Two-way analysis of variance for the effect of test length and the students’ field of study and their interactions As the table indicates, there was no statistically significant difference between the performances of the students on the two versions of the test. No interaction between the English major and non-English majors (indicated as 'group' in the above table) and the type of test was observed. A possible reason for the non-significant effect of the test length on he performance of the students participating in this study is that the students, as stated above, were all advanced language learners who had been learning the language formally for at least 4 years. Such being the case, the lengths of the passages included in the test which were reduced to almost two thirds of their original lengths had not affected the students’ reading performance significantly. Alderson (2000, p.65) lends further support to his finding saying that, “text variables only have a crucial role when materials are conceptually more difficult or unfamiliar and when readers are relatively less able” (emphasis added). In fact, the results might have been different if the two groups of students were of different levels of proficiency.
Table 4: Two-way analysis of variance for the effect of test length and the students’ field of study and their interactions As the table indicates, there was no statistically significant difference between the performances of the students on the two versions of the test. No interaction between the English major and non-English majors (indicated as 'group' in the above table) and the type of test was observed. A possible reason for the non-significant effect of the test length on he performance of the students participating in this study is that the students, as stated above, were all advanced language learners who had been learning the language formally for at least 4 years. Such being the case, the lengths of the passages included in the test which were reduced to almost two thirds of their original lengths had not affected the students’ reading performance significantly. Alderson (2000, p.65) lends further support to his finding saying that, “text variables only have a crucial role when materials are conceptually more difficult or unfamiliar and when readers are relatively less able” (emphasis added). In fact, the results might have been different if the two groups of students were of different levels of proficiency.
Table 5: Readability indexes obtained for the passages Based on this table, we come to the point that the shortened versions of the passages turned out to be more difficult than the original passages in two cases (passages # | and 2), though his difference was not very high, and only in the case of the third passage it led to an increase in the readability of the passage. This being so, the students must have had a better performance on the original version of the test than its shortened version. However, as the comparison of the means obtained for the performance of the students on the two versions of the test and the four sets of raw scores revealed, in all cases the students had a better performance on the shortened version of the test as compared to the original version, hough the difference was very marginal. These contradictory findings render support to he issue of crudity of readability indexes and the fact that readability can not be taken as a good measure for determining the difficulty of texts (Chastain, 1988). As noted already, the Flesch formula is a function of the length of sentences and the size of words in a text, and nothing else. As Fulcher (1997, p. 501) states, “these factors are not, in themselves, what makes a text difficult or easy; but they are predictors of text difficulty.” In fact, important text variables, such as specialist use of lexis, conceptual and propositional complexity, textual organization, and syntax are not taken into account in estimating readability through Flesch formula (Fulcher, 1997). And, as mentioned previously, there are more important factors such as reader factors which contribute to the comprehensibility of texts.
Table 5: Readability indexes obtained for the passages Based on this table, we come to the point that the shortened versions of the passages turned out to be more difficult than the original passages in two cases (passages # | and 2), though his difference was not very high, and only in the case of the third passage it led to an increase in the readability of the passage. This being so, the students must have had a better performance on the original version of the test than its shortened version. However, as the comparison of the means obtained for the performance of the students on the two versions of the test and the four sets of raw scores revealed, in all cases the students had a better performance on the shortened version of the test as compared to the original version, hough the difference was very marginal. These contradictory findings render support to he issue of crudity of readability indexes and the fact that readability can not be taken as a good measure for determining the difficulty of texts (Chastain, 1988). As noted already, the Flesch formula is a function of the length of sentences and the size of words in a text, and nothing else. As Fulcher (1997, p. 501) states, “these factors are not, in themselves, what makes a text difficult or easy; but they are predictors of text difficulty.” In fact, important text variables, such as specialist use of lexis, conceptual and propositional complexity, textual organization, and syntax are not taken into account in estimating readability through Flesch formula (Fulcher, 1997). And, as mentioned previously, there are more important factors such as reader factors which contribute to the comprehensibility of texts.
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved