ACTFL and LTI Introduce Groundbreaking Automated Scoring System for the AAPPL Spanish Presentational Writing Component

AAPPL Machine Scoring

ACTFL and Language Testing International (LTI) proudly announce the launch of an automated scoring system for the ACTFL Assessment of Performance toward Proficiency in Languages (AAPPL) Spanish Presentational Writing (PW) component. This pioneering innovation represents the culmination of nearly ten years of dedicated research and development, and a significant advancement in language assessment rating systems which will enhance the accuracy and efficiency of the AAPPL evaluation process.

The AAPPL PW component evaluates the spontaneous writing abilities of learners in grades 5-12, without access to revisions and editing tools. It serves as a crucial measurement of written language performance, guiding goal setting, curriculum development, and instruction.

The automated scoring system, also known as machine scoring, will be implemented for Spanish PW tests, in conjunction with human raters, beginning September 15, 2023. This integration of machine scoring alongside human raters will mean that all Spanish PW tests will receive a machine score combined with those provided by certified human raters to produce a final score. In addition, the rating process will be more streamlined by identifying tests that require additional human rating. This approach will benefit AAPPL test takers and users by providing an additional rating while we simultaneously continue to improve our quality assurance processes. The machine scoring system assesses the AAPPL PW responses based on several parameters, as described in the ACTFL Performance Descriptors, including functions (global tasks the learner can perform in the language), context/content (situations and topics the learner can comprehend and discuss), text type (comprehension and production capabilities for performing functions), and comprehensibility (accuracy and vocabulary use).

LTI's machine scoring system has undergone rigorous training using rating data from ACTFL-certified AAPPL raters. Extensive research, conducted by Dr. Erik Voss of Teachers College Columbia University, has validated the accuracy of machine scoring compared to human raters over multiple years of administration.

ACTFL and LTI remain committed to maintaining currency and relevance in language testing and contributing to the advancement of research in non-English, second language machine scoring procedures.

More information on the research used during development and answers to frequently asked questions will be available on the ACTFL website soon.