ANÁLISIS DE LA POTENCIA DEL ESTADÍSTICO C MEDIANTE SIMULACIÓN

Abstract

C statistic power analysis through simulation. The C statistic (Young, 1941) constitutes a suitable analysis alternative for interrupted brief or short-time series used in behavioral designs. For this reason, a Monte Carlo simulation study is carried out to determine the statistical power of the C statistic for each of the usage strategies proposed by Tryon (1982). The series generated allow to evaluate the limitations of the C statistic and to compare the three possible analysis alternatives when the baseline phase shows a trend, for different change patterns (i.e. comparison of series using the regression technique, comparison of series with direct data, and analysis of the set