A comparative analysis of spelling skills utilizing a computer-based spelling assessment instrument

Vita.The purpose of this research was to develop a computer-based instrument which more thoroughly measures spelling performance. A computer-based format can provide additional quantitative and qualitative data that can direct instructional interventions by providing a complete assessment of spelling ability. Quantitative advantages result from automatically collecting data which measure speed, hesitancies, accuracy, number of attempts, self-reported confidence of the student and a record of self-correction activities. Once we have developed this type of profiling ability, the test results would provide an informed basis for diagnosis and prescription, two necessary components of remediation. This study was conducted using 90 fifth grade students. Stratified random sampling was used to obtain an balanced sample based upon ethnicity (African American, White, and Hispanic), gender and program placement (normally achieving, learning disabled, and gifted/talented). Three assessment instruments were used: (1) the spelling subtest of the Test of assessment instrument (based on the spelling subtest of the TOWL) developed for use in this study and administered by computer. The major findings of the current study indicate that the global results obtained from the spelling test were equivalent for both paper/pencil and computer-based formats. However the computer-based version provided much more information which, upon further examination, revealed significant differences among the three groups of spellers (NA, LD, G/T). There were significant differences among the groups for confidence level for correctly spelled words, confidence levels for incorrectly spelled words, confidence level changes, test scores, interletter latencies, and the keystroke times for correctly spelled words. Results indicate that the learning disabled students took more time at the decision points before, during, and after the words were spelled. Recommendations for further refinement of the instrument and its application to diagnosis are included. Using this improved assessment instrument, further testing of the spelling models and their implications could be done, possibly leading to more accurate diagnosis. Effective remediation is based upon accurate diagnosis