1 / 15

Advanced Analysis of Physical Functioning Scales and Item Response Theory Models

This study evaluates the psychometric properties of two physical functioning scales: Form C and Form G, comprising 112 and 56 items respectively. It employs classical test theory and item response theory (IRT) to assess reliability, scalability, and dimensionality. Findings include high reliability coefficients (Alpha = 0.99 for Form C, 0.98 for Form G) and strong scalability results. Monotonicity and unidimensionality were confirmed through advanced statistical methods. The analysis utilizes descriptive statistics and confirmatory factor analysis to explore model fit and construct validity.

brian
Télécharger la présentation

Advanced Analysis of Physical Functioning Scales and Item Response Theory Models

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. Physical Functioning Ron D. Hays September 12, 2006 Network Testing ~ 5 or 6pm

  2. Physical Functioning Items • Form C • 112 items (16 HAQ like) • Form G • 56 items (4 HAQ like) • Sub-domains • IADL • Upper • Central • Lower

  3. Analyses Conducted • Descriptive statistics • Classical test theory/reliability • Monotonicity/scalability • Dimensionality • IRT

  4. Classical Test Theory • Form C (112 items) • Alpha =.99 • Range of item/total correlations (corrected for item overlap): 0.40 (able to brush teeth) to 0.85 (able to do yard work) • Form G (49 items) • Alpha =.98 • Range of item/total correlations (corrected for item overlap): 0.30 (open previously opened jar) to 0.87 (help limit climbing flight of stairs)

  5. Monotonicity/Scalability • Monotonicity assessed using Mokken Scale Procedure • No items in Form C or G had monotonicity problems (minimum violation index < .03) • Scalability assessed using Loevinger’s H • Form C • H ranged from 0.50 (health limit you eating a meal in a normal time) to 0.80 (able to run or jog for two miles) • Form G • H ranged from 0.36 (open previously opened jar) to 0.84 (help getting dressed) • Scalability “strong” when H > 0.5

  6. Unidimensionality • Confirmatory factor analysis conducted using Mplus with polychoric correlations and WLSMV estimator • Form C • CFI = 0.973, RMSEA = 0.091 • Factor loadings ranged from 0.74-0.95 • Form G • CFI = 0.926, RMSEA = 0.136 • Factor loadings ranged from 0.54-1.04*

  7. IRT Analysis • Graded Response Model fit using MULTILOG • Form C • Too few observations in some categories for 28 items to estimate SEs • Form G • Too few observations in some categories for 3 items to estimate SEs • So collapsed categories and reran model

  8. Form C GRM Results • χ 2 = 17,087.3 (1-PL); χ 2 = 16,829.9 (2-PL) • Marginal reliability = 0.9924 • Slopes ranged from 1.68 (PFB6) to 4.07 (PFB15) • Thresholds ranged from -2.78 (PFA54) to 3.06 (PFA39) PFB6 (health limit you in eating a meal within a normal time) PFB15 (able to change the bulb in a table lamp) PFA54 (able to button your shirt) PFA39 (able to run at a fast pace for two miles)

  9. Form G GRM Results • χ 2 = 21,019.4 (1-PL); χ 2 = 20,407.7 (2-PL) • Marginal reliability = 0.9803 • Slopes ranged from 1.06 (PFC27) to 5.08 (PFC16) • Thresholds ranged from -4.69 (PFC43) to 3.19 (PFC33) PFC27 (use jar opener or help to open previously opened jars) PFC16 (need buttonhook, zipper pull, other gadget or help to get dressed) PFC43 (able to use hands to turn faucets, etc.) PFC33 (able to run ten miles)

  10. Questions?

More Related