Generating a Growth-Oriented Partial Credit Grading Model for the Force Concept Inventory

Traditionally, multiple choice assessments are graded in a dichotomous manner, where selecting the correct option for a question awards 1 point and the selection of an incorrect option awards 0 points. As a result of this grading scheme, all incorrect response options are treated as being equally in...

Full description

Saved in:
Bibliographic Details
Published in:Physical review. Physics education research 2019-12, Vol.15 (2), p.020151, Article 020151
Main Authors: Eaton, Philip, Johnson, Keith, Willoughby, Shannon
Format: Article
Language:English
Subjects:
Academic Ability
Achievement Gains
College Students
Colleges & universities
Error analysis
Feedback (Response)
Grading
Item Response Theory
Logistics
Mathematical models
Parameters
Physics
Questions
Science Achievement
Science Tests
Scores
Scoring
Student Development
Teachers
World Views
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Traditionally, multiple choice assessments are graded in a dichotomous manner, where selecting the correct option for a question awards 1 point and the selection of an incorrect option awards 0 points. As a result of this grading scheme, all incorrect response options are treated as being equally incorrect regardless of potential differences in their relative correctness, intentional or otherwise. We propose a partial credit grading model for the Force Concept Inventory (FCI) that allots points to the incorrect responses. This was done using slope parameters from the two-parameter logistics nominal response model (2PLNRM), a model from polytomous item response theory (PIRT). The resulting scores from the partial credit model represent student growth towards a proper Newtonian mindset, as measured by the FCI. Observations indicate that this model accounts for student progression through prominent misconceptions (i.e., impetus) as their world views become more Newtonian. As expected, we find that student total scores increase as a result of the model, but the average overall gains on the assessment are essentially unchanged. The data used in this analysis were maintained and organized by PhysPort and included about 20 000 responses from first semester introductory physics courses at multiple universities. Ultimately, this partial credit model allows instructors to more accurately gauge the growth of their students over the course of instruction. Additionally, as a result of these partial credit scores, we are able to identify potentially malfunctioning questions on the FCI that may be sources of error in measuring student abilities.
ISSN:2469-9896
2469-9896
DOI:10.1103/PhysRevPhysEducRes.15.020151