Browse Prior Art Database

Essay Exam Grading Assistant

IP.com Disclosure Number: IPCOM000247062D
Publication Date: 2016-Jul-31
Document File: 2 page(s) / 64K

Publishing Venue

The IP.com Prior Art Database

Abstract

Disclosed is a computer-implemented method for assisting the grading of essay exams. The method comprises: receiving a first exam answer and a second exam answer; analyzing the first exam answer and the second exam answer to identify a plurality of common cognitive elements; and annotating the identified common elements of the second exam answer.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 51% of the total text.

Page 01 of 2

Essay Exam Grading Assistant

Major education assessment systems include a computer-scored multiple-choice exam and an essay exam. While there are many pedagogical advantages to the essay exam, grading remains a time and labor intensive process, and thus, is seldom used in larger classes. Moreover, current manual essay grading likely produces inconsistent results, particularly when grading occurs over several nights, because the teacher's mental state naturally varies.

A number of automated essay scoring systems exist. The general approach is: (1) hand-grade a few hundred exams; (2) create a regression model from the hand-graded sample by selecting a set of features that are believed to be highly correlated with essay quality, such as essay length (in number of words) or text coherence; (3) apply the regression model to the rest of the exams.

The novel contribution is a computer-implemented method for assisting the grading of essay exams. The method comprises: receiving a first exam answer and a second exam answer; analyzing the first exam answer and the second exam answer to identify a plurality of common cognitive elements; and annotating the identified common elements of the second exam answer.

A law school, for example, has many commercial topic outlines, treatises, hornbooks, and specialized dictionaries to use as a training corpus. It is also common for law students themselves to produce detailed outlines for a particular professor; outlines from high scoring students in previous years could provide additional training material for the way a particular professor teaches a particular course.

The method embodies two primary workflows and works with an existing cognitive analysis and machine learning system. The following explanations are in the context of a law school.

The basic idea is that a professor initially grades one particular issue-rule-analysis-conclusion (IRAC) analysis. The machine learning system then goes through the re...