< - Back to  Measurement Matrix 

Measurement Evaluation Grids (MEGs):

 

The purpose of the Measurement Evaluation Grids (MEGS) is to provide information for summary reviews of measures. Unlike other reviews, the focus of the MEGS and these reviews is on the use and psychometric qualities of the measures with respect to ethnically diverse groups.  

 

The goals are to a) select the measures that have been used among minority elders; b) summarize and critically evaluate available information on the measurement characteristics of these measures; c) discuss issues related to generalizability (external validity) of results; d) highlight areas in which knowledge is lacking; e) suggest areas in which ethnic differences may impact the measurement process; f) propose areas of future research for the measurement of the domains analyzed in minority elders; and g) provide recommendations for researchers wishing to use existing measurement instruments in their research.

           

Currently there are 40 MEGS on cognitive measures, 10 on affective suffering and 13 on quality-of-life.  As new information becomes available in the literature MEGS will be added, and reviews updated. Summarized below are the criteria for evaluation and elements contained in the MEGS. (See the website for examples (URL:   http://www.research-hhar.org/SubMeasure/ ).  

 

MEGS were developed in collaboration with the Measurement and Methods Cores of the Resource Centers for Minority Aging Research, sponsored by the National Institute on Aging, National Center for Minority Health and Health Disparities and the National Institute of Nursing Research (P30 Ag: 15272, 15281, 15292, 15294, 21677, 21684).

 

The elements described below may or may not be available for each MEG, depending on the study and article reviewed.  For example, some may focus on differential item functioning (DIF), and may not include information about reliability or normative data.

 

Basic possible elements for inclusion in the MEG:

Characteristics of the study population

Age composition;

Gender;

Race/ethnicity (whenever possible race/ethnicity was defined based on 2000 US Census criteria). When available, subgroups within major ethnic/racial categories (e.g. Korean, Dominican, Haitian, Mexican, Cuban, etc.) were identified;

Socioeconomic indicators;

Living status;

Geographic location

Design

Sample size for each of the subgroups for which results are reported;

Sampling methods;

Response rate;

Method of recruitment

Instruments description

Purpose of measurement (as described by authors) and domains addressed; Conceptual framework of domain(s) and subdomains;

Application of method (clinical, research, survey or screening test);

Instruments: names, sources, and versions for standard instruments, or brief descriptions and copies of instruments (if available) for non-standard instruments;

Format and design of measure (literacy or reading level; number of items, length of time to complete, response categories);

Language of administration and details of translation methods, if applicable;

Methods of administration (self-administered, interviewer or expert rater);

Numerical characteristics of scales (nominal, ordinal, interval or ratio);

Scoring (range, meaning of high score, scoring rules, how missing data was handled);

Availability of translations and source;

Questionnaire burden: Number of items in each scale and Time requirements);

Measurement characteristics of the instruments:

Reliability in development and other samples

Test-retest,

Interrater,

Internal consistency

Validity

content,

construct (convergent and divergent),

concurrent,

predictive);

Sensitivity to change (responsiveness);

Measurement equivalence: Conceptual

qualitative analyses,

focus groups;

cognitive interviews

Factorial invariance for racial/ethnic/cultural subgroups

dimensional,

configural,

metric,

intercepts,

residuals

Differential item functioning evidence

assumptions

presence

magnitude

impact

Evidence of use of measure in diverse populations (Changes in format and design (from original) If so: rational for changes; psychometrics for changes; Modifications in definition for subgroups; Language other than English (method of translation); Method of recruitment; Data collection method; Response rate; Sample characteristics (demographics, socioeconomic indicators); Results by groups (differences by groups, if any)).

Information to be abstracted from each publication

The information collected for each article will be summarized in table format (see example).  

Critical evaluation of reported test characteristics in minority elders: A protocol for critically evaluating the measures in terms of psychometric properties and differential item functioning was developed.

Summaries by instruments: In addition to presenting results article by article, instrument specific-summaries, including all studies on minority elders employing identical measurement instruments, can presented in tabular format, and used as the basis for expert reviews.

< - Back to  Measurement Matrix