The development of typewriting tests for measuring student achievement at various performance levels
The problem of this study was to develop typewriting tests to be used as pretests and posttests in typewriting and Vocational Office Education programs. The tests included both objective and performance tests and were designed to evaluate students at various performance levels.
The research was conducted during the summer of 1973 and the winter and spring of 1974. Approximately 2,000 students in over 30 secondary schools and colleges in the state of Tennessee participated in the research.
Equivalent forms of the various tests were constructed from an outline, and the outline and test items were evaluated by professional people. Other procedures included trial testing and revision of instruments prior to final testing. In addition, scoring criteria were developed based on the judgment of professional people and obtained through interviews and review of professional literature.
Major findings of the study were:
1. The reliability coefficients of .86 and .89 met the criterion set for the objective tests that the tests should have a coefficient of .85 or greater.
2. The objective tests met the criterion that the two forms were to contain at least 50 percent of the items with a discrimination index of .30 or greater, with less than 5 percent of the items below .10. Neither of the forms had test items with an index below .10; and, Form A and Form B yielded 51 and 68 percent, respectively, of items with an index of .30 or greater.
3. Form A of the objective tests did not meet the criterion established for the percentage of difficulty indexes (less than 5 percent of the items should have an index of above .90 or below .10). Form A had 7 percent of the items with an index below .10; Form B had a total of 4 percent of the items with an index below .10; therefore, Form B met the criterion.
4. No significant differences were found at the 10 percent level between the two forms of objective tests; therefore, the criterion was met for similarity of equivalent forms.
5. The criterion that significant differences should be found (5 percent level) between performance levels was not met for Timed Writings, Errors, and Timed Writing Score test items of the performance tests. However, significant differences were found for Letters, Tabulation, Manuscripts, and Typewriting Numbers in Context test items; therefore; the criterion was met fer these test items.
6. The criterion that no significant difference should be found between mean scores of similar test .items on equivalent forms (10 percent level) was met for all test items of the performance tests.
Thesis75b.C426.pdf
6.15 MB
Adobe PDF
8e2cb4d8913eef964d431afcbb1ca90f