Skip to main content

Table 5 Reliability of two methods using Kappa statistics for observations reported by both methods (3897 observations in 186 children)

From: Comparison of fieldworker interview and a pictorial diary method for recording morbidity of infants in semi-urban slums

Variable

P0

Pe

PI

BI

Kappa

BAK

PABAK

Diarrhoea

0.962

0.940

−0.938

0.004

0.360

0.364

0.924

Fever

0.896

0.853

−0.841

0.040

0.290

0.291

0.792

Cough/cold

0.674

0.504

−0.253

0.237

0.342

0.302

0.347

Healthy

0.674

0.490

0.164

−0.219

0.361

0.329

0.347

  1. Po –Proportion observed.
  2. Pe-Proportion expected.
  3. PI = Prevalence index (difference in the probability of agreement cells (a-d)/N).
  4. BI = Bias index (difference in proportion of disagreement cells (b-c)/N).
  5. BAK is Bias-adjusted kappa. It is calculated by substituting the average values of the discordant cells instead of their original values in the regular kappa estimation formula. K = (Po- Pe)/ (1-Pe).
  6. PABAK = Prevalence-adjusted bias-adjusted kappa = 2*(Po)-1.
Back to article page

BMC Public Health

ISSN: 1471-2458

Contact us