09.12.2012 Views

I__. - International Military Testing Association

I__. - International Military Testing Association

I__. - International Military Testing Association

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

_____._._ .- _ .-... -.._- -..._ . . _ _.<br />

The Effect of Reading Difficulty<br />

on Correspondence Course Performance<br />

Dr Grover E. Diehl (ECI)<br />

During the 1989-90 academic year the Air Force Extension Course<br />

Institute(EC1) broadly examined the impact on reading level on<br />

the correspondence courses in Career Development Course (CDC)<br />

First reading grade levels (RGL) of CDCs were<br />

~~~~~%$ using Che FORCAST method. The FORCAST method involved<br />

the manual counting of words in samples of text. These RGLs here<br />

then examined to determine whether the RGLs had increased significantly<br />

on a year by year basis and also with 1982 data to<br />

determine whether there were significant differences between the<br />

samples. Next, RGLs were correlated with end-of-course performance<br />

by percent of first time exam failures and then by proportion<br />

of overall course failures. Following this, the FORCAST<br />

RGLs were correlated with target RGLs prepared by the Air Force<br />

Human Resources Laboratory and with computer generated RGLs<br />

using a Flesch-Kincade formula.<br />

A basic intervening variable in the assessment of reading difficulty,<br />

however, was the fact that personnel and Air Force jobs<br />

were matched during enlistment processing so that the most intellectually<br />

demanding skills were peopled with the most intellectually<br />

able personnel. One way around this problem was to<br />

calculate difference scores between the RGL targets and the obtained<br />

FORCAST RGLs -- a measure of perceived difficulty of the<br />

material to the student -- and correlate this with failure rate.<br />

This procedure treated student ability as a covariate with a<br />

corresponding reduction in the error portion of the predication<br />

equation, without the necessity of using analysis of variance.An<br />

analysis of difference scores constituted the last question to<br />

be addressed.<br />

Findings<br />

FORCAST RGL and Edition Date. No statistically significant association<br />

was found between the FORCAST reading level and the edition<br />

date of the materials (a period of about 12 years). The<br />

Pearson Product Moment Correlation coefficient (r) of FORCAST<br />

RGL with edition date was .0742 (N = 215, p =.279). To check<br />

for possible curvilinearity, a scatterplot was prepared which<br />

suggested a completely random occurrence pattern. FORCAST reading<br />

level did not vary in a linear way from year to year.<br />

Difference Between RGLs Sampled in 1982 and 1990. There was apparently<br />

sufficient variation within the samples to be statisti-<br />

128

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!