TABLE 2-1 Results of a Diagnostic Test of Deception of Accuracy (A) = 0.90, with Hypothetical Populations of 10,000 Examinees with Base Rates of Deception of 1 in 2 and 1 in 1,000

2-1A Sensitivity of 80 percent (i.e., threshold set to correctly identify 80 percent of deceivers)

 

Criminal Suspects

(Base Rate: 1 guilty of 2)

National Laboratory Employees

(Base Rate: 1 spy of 1,000)

Result

Guilty

Not Guilty

Total

Spy

Not Spy

Total

“Failed” test

4,000

800

4,800

8

1,598

1,606

“Passed” test

1,000

4,200

5,200

2

8,392

8,394

Total

5,000

5,000

10,000

10

9,990

10,000

2-1B Sensitivity of 50 percent (i.e., threshold set to correctly identify 50 percent of deceivers)

 

Criminal Suspects

(Base Rate: 1 guilty of 2)

National Laboratory Employees

(Base Rate: 1 spy of 1,000)

Result

Guilty

Not Guilty

Total

Spy

Not Spy

Total

“Failed” test

2,500

170

2,670

5

340

345

“Passed” test

2,500

4,830

7,330

5

9,650

9,655

Total

5,000

5,000

10,000

10

9,990

10,000

2-1C Sensitivity of 20 percent (i.e., threshold set to correctly identify 20 percent of deceivers)

 

Criminal Suspects

(Base Rate: 1 guilty of 2)

National Laboratory Employees

(Base Rate: 1 spy of 1,000)

Result

Guilty

Not Guilty

Total

Spy

Not Spy

Total

“Failed” test

1,000

19

1,019

2

39

41

“Passed” test

4,000

4,981

8,981

8

9,951

9,959

Total

5,000

5,000

10,000

10

9,990

10,000



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement