Table 2.

Summary of Measures Used to Assess Health Numeracy

Measuresa	Reliability	Validity	Test–retest	Reading level (grade)	Components	Administration time	Administration method/response format	Original study population	Correlates
Composite measures
TOFHLA (Parker et al., 1995)	α = .981; Spearman–Brown = .92	Correlated with REALM: .84 (p < .001); correlated with WRAT-R (reading): .74 (p < .001)	NS	Reading items: 4.3–19.5; numeracy items: 9.4	50 reading comprehension and 17 numeracy questions based on actual hospital materials	Up to 22 min	Reading: paper and pencil/multiple choice; numeracy: subject views prompts and verbally responds to interviewer questions/open response	Public hospital patients; predominantly African American, low income, less than high school education	Physical health outcomes, mental health outcomes, b disease knowledge, health services utilization
TOFHLA–Spanish (Parker et al., 1995)	α = .98; Spearman–Brown = .84	See Note.	NS	Reading items: 4.3–19.5; numeracy items: 9.4	50 reading comprehension and 17 numeracy questions based on actual hospital materials	Up to 22 min	Same as TOFHLA–English	Public hospital patients; predominantly Hispanic, low income, less than high school education	Same as TOFHLA–English
TOFHLiD (Gong et al., 2007)	α (overall) = .63; α (reading) = .65; α (numeracy) = .59	Correlated with REALD-99: .82 (p < .05); TOFHLA: .52 (p < .05); REALM: .53 (p < .05)	NS	Reading items: 7.0–17.0; numeracy items: 11	68 reading comprehension and 12 numeracy questions based on actual patient materials	20–30 minc	Same as TOFHLA	Caretakers of pediatric dental patients; predominantly female, Caucasian, well educated; wide range of income	Child’s oral health status
Integrative measures
Newest Vital Sign (Weiss et al., 2005)	α = .76	Correlated with TOFHLA: .59 (p < .001); REALM: .41 (p < .001)	NS	NS	Nutrition label accompanied by 6 questions involving arithmetical computation and problem solving	About 3 min	Subject views nutrition label and verbally responds to interviewer questions/open response and multiple choice	University-affiliated primary-care patients; mainly Hispanic and Caucasian; mean years of education = 12.7	NS
Newest Vital Sign–Spanish (Weiss et al., 2005)	α = .69	Correlated with TOFHLA–Spanish: .49 (p < .001)	NS	NS	Nutrition label accompanied by 6 questions involving arithmetical computation and problem solving	About 3.5 min	Same as Newest Vital Sign–English	Hispanic university-affiliated primary-care patients; mean years of education = 10.7	NS
Nutrition Label Survey (Rothman et al., 2006)	KR20 = .87	Correlated with WRAT-3 (math): .67 (p < .001); REALM: .47 (p < .001)c	NS	8–10c	24 questions pertaining to interpreting actual food labels; involves arithmetical computation and problem solving	About 30 min	Subject views nutrition label and verbally responds to interviewer questions/open response and multiple choice	University-affiliated primary-care patients; predominantly Caucasian; socioeconomically and educationally diverse	Income, education, literacy, numeracy
Numeracy Assessment (L. M. Schwartz et al., 1997)	α = .56–.80c	NS	r = .72c	7c	3 numeracy questions pertaining to probability and converting metrics (percentages and proportions)	2 minc	Paper and pencil/open response	Older female veterans; predominantly Caucasian and low income; most with more than high school education	Risk perception, utility assessment
Numeracy Scale (Lipkus et al., 2001)	α = .70–.75	NS	NS	NS	11 numeracy questions covering orders of magnitude, probability, converting metrics, arithmetical computation	Up to 30 min	Paper and pencil/open response and multiple choice	Community volunteers; predominantly Caucasian with more than high school education	See Peters, Dieckmann, et al. (2007) numeracy scale
Expanded Numeracy Scale (Peters, Dieckmann, et al., 2007)	α = .83	High numerate comprehended more and made more high-quality choices in one of two studies; correlated with S-TOFHLA: .51.	NS	NS	15 numeracy questions covering risk magnitude, probability, converting metrics, arithmetical computation	6–15 minc	Paper and pencil/open response and multiple choice	Community volunteers; predominantly Caucasian and low income; wide range of educational attainment	Education, self-rated health, SF-8 physical, SF-8 mental, health literacy
Medical Data Interpretation Test (L. M. Schwartz et al., 2005)	α = .71	High-numerate, high-literate, better educated people scored higher than people lower on those characteristics; physician experts performed better than postgraduates	r = .67	NS	20 multiple-choice questions designed to test the ability to compare risks and put risk estimates in context	NS	Paper and pencil/multiple choice	Volunteers from the community and local hospitals; wide range of income and educational attainment	NS
Numeracy Assessment (anticoagulation control; Estrada et al., 2004)	NS	NS	NS	NS	6 numeracy questions pertaining to probability, converting metrics (percentages and proportions), arithmetical computation, problem solving	NS	Paper and pencil/open response and multiple choice	Patients from anticoagulation management clinics; predominantly Caucasian; 21% with less than 8th-grade education	Health outcomes (anticoagulation control)
Asthma Numeracy Questionnaire (Apter et al., 2006)	α = .57	Correlated with S-TOFHLA: .47; REALM: .41	NS	NS	4 numeracy questions covering arithmetical computation, percentages, problem solving	About 5 minc	Subject views questionnaire while interviewer reads aloud/open response and multiple choice	Asthma patients from urban clinics; largely female and African American; wide range of educational attainment	Hospitalization, emergency room visits
Diabetes Numeracy Test (Huizinga et al., 2008)	KR20 = .95	Correlated with REALM: .54 (p < .001); WRAT (math): .62 (p < .001); Diabetes Knowledge Test: .71 (p < .001)	NS	6.5c	43 items pertaining to arithmetical operations, ratios, multistep mathematics, time, counting, hierarchy	20–46 min	Paper and pencil/open response	Convenience sample of diabetic patients from academic and Veterans Administration health centers; most with less than 9th-grade numeracy	Income, education, insulin adjustment for carbohydrates, insulin adjustment for blood glucose
Diabetes Numeracy Test–15d (Huizinga et al., 2008)	KR20 = .90; KR20 = .89	Correlated with Diabetes Numeracy Test: .96, .97; REALM: .54, .52; WRAT (math): .65, .61; Diabetes Knowledge Test: .73, .67	NS	6c	15 items pertaining to nutrition, exercise, blood glucose monitoring, oral medication, insulin administration	10–15 min	Paper and pencil/open response	Convenience sample of diabetic patients from academic and Veterans Administration health centers; most with less than 9th-grade numeracy	Income, education, insulin adjustment for carbohydrates;c insulin adjustment for blood glucosec
Subjective Numeracy Scale (Fagerlin, Zikmund-Fisher, et al., 2007)	α = .82	Correlated with Lipkus et al. (2001) numeracy scale: .63–.68	NS	NS	8 questions covering perceived mathematical ability and preference for presentation of statistical information	About 5 min	Paper and pencil/multiple choice	Volunteers recruited from local hospitals; predominantly Caucasian; wide range of educational attainment	NS
STAT–Interest Scale (Woloshin et al., 2005)	α = .70	Correlated with Medical Data Interpretation Test: .26 (p = .006)	r = .60	NS	4 statements and 1 question pertaining to interest in medical statistics	NS	Paper and pencil/multiple choice	Volunteers from the community and local hospitals; predominantly Caucasian; wide range of income and educational attainment	Intention to use statistics for health decision making
STAT–Confidence Scale (Woloshin et al., 2005)	α = .78	Correlated with Medical Data Interpretation Test: .15 (p = .04)	r = .62	NS	2 statements and 1 question pertaining to confidence understanding medical statistics	NS	Paper and pencil/multiple choice	Volunteers from the community and local hospitals; predominantly Caucasian; wide range of income and educational attainment; multiple choice	NS

Note. For the Test of Functional Health Literacy in Adults (TOFHLA) and TOFHLA–Spanish, reliability and validity were calculated for the entire TOFHLA, not for the individual reading comprehension and numeracy sections. Validity of the TOFHLA–Spanish could not be determined because the Rapid Estimate of Adult Literacy in Medicine (REALM) and Wide Range Achievement Test–Revised (WRAT-R) are not available in Spanish. NS = not specified; TOFHLiD = Test of Functional Health Literacy in Dentistry; REALD-99 = Rapid Estimate of Adult Literacy in Dentistry; WRAT-3 = WRAT (3rd ed.; Wilkinson, 1993); S-TOFHLA = short version of the TOFHLA; SF-8 = Short Form (8) Health Survey.

^aThe Health Activities Literacy Tests (Baker, 2006; Educational Testing Service [ETS], 2007) are not included among the numeracy measures in this table because they are proprietary and available only commercially from ETS. However, they bear mentioning because they assess the ability of 16- to 65-year-old adults to understand and act upon health-related written materials. Like the National Adult Literacy Survey, these tests include prose, document, and quantitative literacy items that reflect practical health-related tasks. They address health activities related to health promotion and protection, disease prevention, health care and maintenance, and systems navigation. The Health Activities Literacy Tests are only available electronically. What is unique about the tests is that they adapt to the test taker’s ability and are scored by computer. The full-length Health Activities Literacy Test requires one hour to complete, whereas the shorter “Locator” test takes about 30–40 min to complete. Currently, these tests are used primarily for research purposes (http://www.ets.org/Media/Tests/ETS_Literacy/LITERACY_TECHSPECS.pdf).

^bStudy used original S-TOFHLA.

^cPersonal communication from instrument developer.

^dThe Diabetes Numeracy Test–15 was verified through random split sample analysis; reliability and validity data for the test are given for both subsamples.