Validation and Development of a CAT Version of Accessibility Instrument

Thursday, April 3, 2014
Exhibit Hall Poster Area 2 (Convention Center)
Elena A. Boiarskaia1, Weimo Zhu1, James H. Rimmer2 and Sangeetha Padalabalanarayanan2, (1)University of Illinois at Urbana–Champaign, Urbana, IL, (2)The University of Alabama at Birmingham, Birmingham, AL
Background/Purpose:

The Accessibility Instrument Measuring Fitness and Recreation Environments (AIMFREE) was created to assess the accessibility of fitness/recreation facilities for people with physical disabilities within the U.S. The length of AIMFREE makes administration of the full version rather time consuming, thus a Computer Adaptive Test (CAT) version is in development. This study validated AIMFREE and assessed the instrument’s readiness for CAT.

Method:

AIMFREE consists of 389 items grouped by 14 subscales (access routes/entrance areas; bathrooms; elevators; equipment; hot tubs/whirlpools/saunas/steam rooms; information/signage; locker rooms/showers; parking; policies; professional support/training; programs; telephones; water fountains; swimming pool). 235 fitness facilities in the U.S. were assessed using AIMFREE.

Analysis/Results:

The psychometric quality of the items was assessed using a three-facet Rasch model. 10 items were deleted due to lack of variation (all “Yes” or all “No” responses) and 6 items were flagged as outliers with Outfit statistics greater than 2. A new model without outliers showed good model-data fit (Infit and Outfit statistics were all close to one) and a good range of item difficulty (-4.89 to 6.53 logits) and precision (SE = 0.31). Facilities’ accessibility scores ranged from -1.85 to 2.90 logits (M = 0.82, SD = 0.20). Simulations were conducted to test whether the remaining 373 AIMFREE items are ready to be used as an item bank for CAT. The results showed that a content balanced CAT version consisting of 50 items would yield a 0.83 correlation to the ability scores estimated using the full version.

Conclusions:

AIMFREE is ready to be used for assessment of accessibility of fitness and recreation facilities both as a full and CAT version. Criterion-based scoring established by a panel of experts is being developed.

RERC RecTech is funded through grant H133E110007 from the National Institute on Disability and Rehabilitation Research