Skip to main content

Table 1 Demographic characteristics and the percentage of missing data on cognitive test scores in the Atherosclerosis Risk in Communities Neurocognitive Study (ARIC-NCS), 2016–2017

From: The effect of missing data and imputation on the detection of bias in cognitive testing using differential item functioning methods

 

ARIC-NCS

Age—Mean (SD)

79.8 (4.7)

Sex—Female—N(%)

2142 (59.8)

Race—Black—N (%)

837 (23.4)

Hearing Impairment—Impaired—N (%)

2614 (73.0)

Vision Impairment—Impaired—N (%)

180 (17.9)

Education: Less than HS—N (%)

428 (12.0)

Education: HS or equivalent—N (%)

1475 (41.2)

Education: Beyond HS—N (%)

1673 (46.8)

Cognitive Items—% Missing in Observed Data (N)

 Boston Naming Test (30 item)

8.2% (270)

 Category Fluency (Animals)

0.5% (17)

 Delayed Word Recall

1.6% (58)

 Digit Symbol Backwards

8.8% (289)

 Digit Symbol Substitution Task

3.9% (136)

 Incidental Learning

4.5% (154)

 Logical Memory 1

9.2% (303)

 Logical Memory 2

9.3% (306)

 Phonemic Fluency (Sum of 3 Trials)

1.5% (54)

 Trail-Making Test A

3.8% (130)

 Trail-Making Test B

19.1% (573)

  1. HS High school, the sample size of each DIF analysis can be calculated by subtracting the number of records with missing data on a given cognitive test from the total sample size, as each DIF analysis started from the reference scenario of no missing data.