Position on AES: 1. Is completely against AES/automated scoring in “admissions, placement, formation, or achievement testing” (par. 1).
Condon critiques AES/automated scoring on the grounds that these programs cannot do justice to the “writing construct,” and focus instead on surface-level features. Additionally, the sample sizes required by most programs are not realistic examples of student writing, and even the best software is only able, statistically, to poorly mimic the abilities of a human rater. Condon places machine scoring on the lowest point of the progression “ranking, assessing, evaluating.”
- He assumes that the use or rejection of machine scoring is a zero-sum game: there should be no compromises, and the software is not really useful in any context.
- He assumes that the use of machine scoring cannot support good teaching of writing.
He provides some interesting information comparing machine scoring software to human raters (par. 9). He also supplies us with a helpful chart to understand the levels of assessment effectiveness and granularity (Fig. 1).