Three techniques for univariate outlier identification are: Extreme Studentized Deviate(ESD), the Hampel identifier and the Rousseeuw identifier. The latter two are robust statistical techniques. The purpose of the paper is to determine how these outlier identification techniques perform under varying conditions. An experimental design along with two different Monte Carlo simulations provides insights into the problem. Under certain assumptions it is shown that the ESD identifier performs well with very small data contamination and the robust Hampel and Rousseuw identifiers perform better with large samples and with multiple outliers.