The purpose of this study is to describe and evaluate a multilingual automated essay scoring (AES) system for grading essays in three languages. Two different sentence embedding models were evaluated within the AES system, multilingual BERT (mBERT) and language-agnostic BERT sentence embedding (LaBSE). German, Italian, and Czech essays were…

Computer simulations were run to measure the relationship between testlet validity and factors of item pool size and testlet length for both adaptive and linearly constructed testlets. Making a testlet adaptive yields only modest increases in aggregate validity because of the peakedness of the typical proficiency distribution. (Author/SLD)

Hierarchical (adaptive) and linear methods of testlet construction were compared. The performance of 2,080 ninth and tenth graders on a 4-item testlet was used to predict performance on the entire test. The adaptive test was slightly superior as a predictor, but the cost of obtaining that superiority was considerable. (SLD)

Over 2,000 writing samples were collected from four undergraduate institutions and compared, where possible, with scores on a multiple-choice test. High correlations between ratings of the writing samples and multiple-choice test scores were obtained. Samples contributed substantially to the prediction of both college grades and writing…

To study whether different forms of the Scholastic Aptitude Test (SAT) used since the mid-1970s varied in their correlations with academic performance criteria, 1975 and 1985 forms were administered to 1,554 and 1,753 high school juniors, respectively. The 1975 form did not have greater validity than the 1985 form. (SLD)