Abstract
The statistical analysis of short time series designs is influenced by the presence of serial dependence. Therefore, it is important to correctly estimate the first-order autocorrelation in behavioral data. The empirical bias is an indicator generally used to evaluate the adequacy degree of an estimator. This paper presents the Bias program, a Monte Carlo simulation pro gram tha t generates fir st-order autoregressive processes and calculates the bias in three autocorrelation estimators (r 1, r 1+ and r1 ') for different values of the lag-one autocorrelation parameter and sample siz es. The prog ram has been designed with MATLAB programming language and it runs on IBM-PC compa tible computers with a 486 or later processor.
