Project Background: Participants, papers, and statistical analysis is drawn from an earlier study of 64 undergraduates (Haswell, "Documenting Improvement in College Writing," Written Communication 17.3 (2000). Study of statistical outliers in a distribution of essay-performance is novel approach, I think, at least for college writers.

Time Frame: Finished by March 15, 2008

Research Question or Hypothesis: Is justified to exclude from assessment findings test-takers whose performance lies a certain degree beyond the norm (as is often done in standardized testing firms)? My hypothesis is that the performance of these outliers may lie within normal bounds of writing growth.

General Research Approach: Mixed / Multi-Modal

Participants and Setting: College students writing essays under assessment contexts as first-year and third-year undergraduates.

Research Methods: Creation of scatterplots for each of eight measures; selection of students whose essays fall beyond a certain standard deviation from the mean on the measure; critical analysis of their essays; comparison of their performance with their earlier (or later) performance and with the performance of students writing near the mean; judgment of their performance in light of college-span development in the measure.

Data / Information Sources: Student-written essays, analysis of eight measures of essay quality, statistical distributions of the measures (N=64)

Design Comments: Profiling, as is sometimes used in ESL studies

Funding: none

Support: none

Intended Audience: Lower angels, principalities, and archangels--no higher.