No announcement yet.
  • Filter
  • Time
  • Show
Clear All
new posts

  • Intra-obsever variability

    Dear colleagues can I still use kap in stata for intra-observer variability? I have two readings by the same reader trying to asses presence of radiologically confirmed pneumonia. With 1 denoting presence of rcp, 0 denoting absence of rcp.Have a look at the data:

    input float(reader1_a reader1_b)
    1 1
    0 0
    1 1
    1 1
    0 1
    1 0
    1 1

    can I do something like;
    kap reader1_a reader1_b?

  • #2
    Cohen's kappa (and other chance-corrected agreement coefficients; for alternative coefficients, see kappaetc from SSC) are one way of looking at the agreement between measures/ratings of the same subject. Whether those different measures represent ratings/scores of different raters/judges/observers or different readings is not relevant to the mathematics. It might be conceptually harder to establish "independence" (however defined) between two readings by the same reader as a plausible Nullhypothesis to test against, though.


    • #3
      Thanks Daniel Klein, so this means I can do;
      kap reader1_a reader1_b as I asked before to asses intra-observer variability for reader1