Which Machine Scoring Engine to Use: IntelliMetric vs. Approximator

January 15, 2019

In 1997 Vantage Learning’s IntelliMetric® was the first Automated Essay Scoring System to reach and exceed human-level performance.  Having scored over a billion essays, IntelliMetric® is the “Gold-Standard” in automated essay scoring.  With immediate scoring, accuracy, consistency, and reliability greater than human raters, IntelliMetric® is the leading essay scoring platform available today.  This Web-based tool is accessible  anywhere users have access to a computer and the internet and is capable of scoring both long and short answer responses in more than 20 languages.


IntelliMetric scores writing assignments regardless of level of ability, ensuring that scores and feedback are relevant to the individual’s ability and/or grade level.  Whether used for academic purposes, certification testing, or job screening, users/examinees are able to submit persuasive, informative, and/or narrative essays against a library of trained prompts.  IntelliMetric is able to score against six, eight, and twelve point rubrics.  In academic settings, IntelliMetric scoring can be used for instructional support as well assessment.


The technology used in machine scoring provides administration of open-ended prompts, and provides immediate feedback and scores holistically and against writing domains including Focus and Meaning; Content and Development; Organization; Language Use, Voice, and Style; and Mechanics and Conventions thereby providing multiple determinants of the writer’s ability.  IntelliMetric’s automated essay scoring relies on Artificial Intelligence (AI) to emulate the process used by human raters, to achieve scoring accuracy equal to or greater than expert raters.  IntelliMetric provides the advantages of traditional expert scoring methods, scoring both short answer and extended responses immediately.


The process used to train IntelliMetric incorporates the initial human scoring of a given prompt.  Raters tally 300-500 submissions with scores equally distributed across the score scale.  These  results are then fed into the scoring engine.  The AI provides analysis of more than 400 semantic, syntactic and discourse-level features to formulate the scoring process.  Our technicians work with clients in support of the process; training prompts is standard practice for our writing team.


For analysis and scoring of “un-trained” responses, we offer a second scoring engine called Approximator® which provides a reliable estimate of the candidate/student’s writing ability.  Approximator assesses the quality of writing adjusted for grade level and genre specific essays.  This tool does not require the training and set-up associated with IntelliMetric; Approximator scores with analogous accuracy to IntelliMetric without the cost of development and training required for client/teacher-specific prompts.  Further, it provides immediate scoring of prompts created by instructors at your institution.


The Approximator is grounded in a similar background and concept to prompt-specific IntelliMetric scoring models.  Rather than being prompt-specific however, Approximator has been trained with a “pooled approach” covering an assortment of prompt genres encompassing multiple grade ranges (Elementary, Middle, High School, Higher Ed, and Higher Ed Proficient).


For use in Learning Management Systems or grammar improvement software as with our CorrectEnglish™ Writing Support Platform, Approximator’s immediate scoring feedback can be used for client/teacher-created prompts.  Use by instructors/professors eliminates the drudgery associated with checking commonly made mistakes in grammar and spelling, allowing more time to focus on students’ critical writing abilities.


IntelliMetric and Approximator scoring of essays and subsequent associated feedback dramatically reduces the cost and time required to evaluate student and professional writing.  Writing improves with practice.  Machine scoring tools improve the instructional process by providing more frequent and immediate feedback to writers; encouraging improvements in students’ writing ability by communicating their strengths and weaknesses.  Allowing multiple essay submissions with instantaneous, detailed adaptive feedback; our machine scoring tools encourage improvements in a user’s writing ability through repetition.


An optional Originality Detector ensures that students properly format and cite their writing.  For assessment, the Originality Detector, via a detailed web and/or database search, readily identifies plagiarized work.  The Legitimacy detection feature of IntelliMetric® flags submissions if they are too short to grade, off-topic, repetitious, have inadequate development, contain too many unknown words, have major syntax errors, copy the question, or are inappropriate or contain messages of harm.  The essay scoring tools we offer can be aligned with a publisher’s or institution’s LMS via a Single Sign On (SSO) or API.  All integrations provide immediate results and/or the option to have reports transferred by FTP.


We welcome a discussion on how our Automated Essay Scoring can be of immediate and ongoing benefit your organization or institution.  Please contact us


Further Reading:

IntelliMetric marketing website

IntelliMetric Blog

White Paper on Machine Scoring