Contrasting Sensitivity and Specificity
A Program written by: Moira McPherson, William Montelpare, and Brent Faught
Lakehead University
Gold Standard Test
Field Test
1
condition present
+
condition absent
-
condition present
+
a
1
=
b
1
=
condition absent
-
c
1
=
d
1
=
Gold Standard Test
Field Test
2
condition present
+
condition absent
-
condition present
+
a
2
=
b
2
=
condition absent
-
c
2
=
d
2
=
Test-wise Computations
Results
Field test 1
:
Sensitivity (a
1
/ a
1
+c
1
)=
Field test 1
:
Specificity (d
1
/ d
1
+b
1
)=
Field test 1
:
positive predictive value (a
1
/ a
1
+b
1
)=
Field test 1
:
negative predictive value (d
1
/ c
1
+d
1
)=
Field test 2
:
Sensitivity (a
2
/ a
2
+c
2
)=
Field test 2
:
Specificity (d
2
/ d
2
+b
2
)=
Field test 2
:
positive predictive value (a
2
/ a
2
+b
2
)=
Field test 2
:
negative predictive value (d
2
/ c
2
+d
2
)=
Contrasts are based on the chi-square distribution with 1 degree of freedom.
Chi Square Between Test Computations
Results
Comparing Sensitivity Estimates
Field test 1 versus Field test 2
Comparing Specificity Estimates
Field test 1 versus Field test 2
Comparing Positive Predictive Values
Field test 1 versus Field test 2
Comparing Negative Predictive Values
Field test 1 versus Field test 2
