A descriptive analysis of the data was carried out and the difference (d) between the Pz values obtained on the first and second reading were calculated, followed by the mean of differences for each examiner (–d) and for each protocol.
The intra- and inter-examiner reproducibility of the measurements for each protocol was estimated using the
Intraclass Correlation Coefficient (ICC; ρ) and its confi-dence interval (95% CI), considering the absolute agreement
and two-way analysis of variance (ANOVA) model,for single and average measurements (when Pz was calculated after counting one colony and triplicate, respectively).
According to the ρ value obtained, the degree of correlation between measurements was classified according to Fermanian.3