Journal of outcome measurement最新文献

This paper explores the use of Rasch analysis and linear confirmatory factor analysis as methods for investigating the dimensionality of an early childhood test (Gesell School Readiness Screening Test), taking into account the theoretical basis of test construction. The paper presents the results of empirical analyses using both approaches and discusses the theoretical and psychometric considerations that guide the selection and application of each technique.

The purpose of this paper is to present a set of procedures based on Rasch measurement theory for construction an assessment network. An assessment network is defined as a connected system of rater and task banks. Three general classes of data collection designs are presented that can be used to calibrate an assessment network; these are complete, incomplete, and non-linked assessment networks. Carefully constructed assessment networks based on Rasch measurement theory and sound data collection designs provide the opportunity to achieve objective and fair measurements.

This study equates the physical functioning subscales of the Medical Outcomes Study Short Form 36 (SF36) and the Louisiana State University Health Status Instruments (LSU HSI). Data from the SF36's 10-item physical functioning scale, the PF10, and the LSU HSI's 29-item Physical Functioning Scale (PFS), were fit to separate and mixed Rasch rating scale models. Data were provided by a convenience sample of 285 patients waiting for appointments in a public hospital general medicine clinic. Difficulty estimates for a subset of similar items from the two instruments were highly correlated (.95), indicating that the items from the two scales are working together to measure the same variable. The measures from the two equated instruments correlate .80 (.86 when disattenuated for error). Of the two instruments, the PFS's error is lower, model fit is better, and reliability coefficients are higher. Both instruments measure physical functioning, and can do so in a common unit of measurement. Conversion tables are provided for transforming raw scores from either instrument into the common metric.

Rasch analysis was used in this paper to evaluate the Motor component of the FONE FIM, the telephone version of the Functional Independence Measure (FIM). For this purpose, 132 patients discharged from an inpatient geriatric assessment and rehabilitation program were assessed by trained research assistants using the FONE FIM. The results at 5 weeks post-discharge were compared to the observation FIMs (OBS FIMs) done at home 6 weeks post-discharge. These patients had an average age of 79 years and presented with multiple, complex medical problems and significant functional decline. The FONE FIM and the OBS FIM were shown to share a strikingly similar item hierarchy, based on Rasch item difficulty measures. Only bladder management and climbing stairs were misfitting items as indicated by item fit statistics. The same 13-item set and 4-point scales were shown to be psychometrically optimal for both the FONE FIM and the OBS FIM based on the person separation index. Further research is required to address the issue of the optimal item set and scale levels from psychometric and clinical perspectives.

This study used Rasch methodology to pursue three goals. First, we sought to demonstrate the psychometric limitations of the Falls Efficacy Scale (Tinetti, Richman, & Powell, 1990). Second, we addressed these limitations using a simultaneous calibration of the Falls Efficacy Scale and Mobility Efficacy Scale items. Third, we review previous explorations of the self-efficacy construct in relationship to health behaviors and discuss a possible treatment program based on the simultaneous calibrated items and Social Cognitive Theory. Results indicate that responses from the Falls Efficacy Scale fail to assess the higher ends of the self-efficacy continuum. Simultaneous calibration of items improved this lack of scale definition. This initial work in assessing self-efficacy perceptions provides a theoretical framework for planning treatment programs that may be more cost effective than collecting performance measures.

The Sexual Experiences Survey (Koss, Gidycz, & Wisniewski, 1987) is a commonly used instrument for assessing various degrees of sexual aggression and victimization among male offenders and female victims. Rasch analysis was used to transform qualitative raw score observations into objective linear measures using the responses of a national sample of 6,159 higher education men and women across the United States, aged 18-24. This paper supports the construct validity of the survey through evaluation of the item hierarchy, fit statistics, and separation indices. Findings confirm a "dimensional" perspective on rape, suggesting that sexually aggressive behaviors can be scaled along a single continuum from normal to extreme sexual behavior. The item hierarchy reveals an arrangement of sexually aggressive acts in an order of mild to severe, which compares with the one theorized by the authors of the SES. Identity plots demonstrate the validity of using a common set of SES item calibrations to measure both male and female respondents. For interpretation of person responses to the SES, three conclusions are suggested. First, Rasch analysis must be employed to examine item responses effectively. Second, when the survey is administered to a college sample aged 18-24, the item calibrations obtained in this paper can be used to measure offenders and victims. Third, a total raw score-to-measure conversion is not always sufficient to interpret person measures. Instead, a scalogram method needs be added to the Rasch analysis to separate the measures of offenders and victims who complete the survey. Implications for future research are discussed.

This study investigated parameter recovery for the partial credit model using the MULTILOG computer program. Factors studied were the sample size and the number of item parameters, which were manipulated by systematically varying the number of steps per item and the number of items. The findings suggest that the ratio of sample size to number of item parameters being estimated as a "rule of thumb" can be a more complete guideline when the number of steps per item is taken into account. Accurate estimation of ability can be obtained across all conditions, even with sample sizes as small as 250. With regard to estimation of step values, however, more caution is warranted. Accurate estimation of the step values of items which have more categories requires larger sample sizes for a given number of total parameters to be estimated.

Premenstrual Dysphoric Disorder (PMDD) has remained in appendices of the last two editions of The Diagnostic and Statistical Manual of Mental Disorders due to lack of empirical study. Items included in its set of research criteria are considered tentative pending evidence of diagnostic validity. The present study attempts to establish the construct validity of the PMDD criteria using the Rasch method to analyze the validity of individual items as contributors to the diagnosis, in contrast to the usual but less precise approach of using an external validator to establish the diagnostic utility of psychiatric conditions. Analysis of which items best differentiate participants with and without PMDD provides an idea of the relative ability of these items to distinguish PMDD. It is recommended that the areas of anger/irritability, depressed mood, and problems in interpersonal functioning be expanded in further studies and corresponding items added to symptom checklists.

The "motor" (activities of daily living) component of the FONE FIM, the telephone version of the Functional Independence Measure (FIM) was evaluated in a cohort of 132 patients who had been discharged to home from a geriatric inpatient assessment and rehabilitation program. In the current study, Rasch person ability measures were derived from telephone assessments 5 weeks after discharge and in-home assessments 1 week later. Concordance between the modes was shown to be satisfactory for the Rasch measures based on intraclass correlation coefficients. However, the telephone mode consistently generated lower estimates than did the observational mode. This was due to the fact that the telephone mode underestimated motor function for the majority of patients who were at higher levels of cognition and motor function, but overestimated for patients who were at lower levels of cognition and motor function. At the item level, concordance, as determined by Kappa statistics, was better when the FONE FIM responses came from the patient rather than proxy respondents, and when the assessments were done by more experienced rather than less experienced raters. Based on these findings, a mixed strategy, the telephone mode for patients capable of responding to the FONE FIM and in-home assessments for those who are incapable, is recommended.

Figure skating uses the median rank aggregation system for determining medal winners. Unfortunately, the system can be influenced by idiosyncratic ratings made by some judges unlike the skating ability measures from a many-facet Rasch analysis. These measures are constructed to be independent, as statistically possible, of item difficulties, judge severities, and the rating scale structure. A many-facet Rasch analysis was conducted on data from the controversial ladies event at the 1994 Olympics. The results illustrate how the idiosyncratic ratings of the judges were not accounted for by the median rank system, thus biasing the selection of the gold medal winner. All sports that rely on judges' ratings should investigate the use of a many-facet Rasch model in order to bring more objectivity and fairness to the winner selection process.