Forecasting studies of crude oil price depends heavily on historical data. These data should be stationary, consistent, and homogeneous when they are used for frequency analyses or to simulate a model. To determine whether the data meet these criteria, the researcher needs a simple but efficient screening procedure. Such a procedure is described in this paper. A time series of crude oil price data is volatile if its statistical properties are unaffected by choice of time origin. The basic data-screening procedure presented in this research is based upon split-record tests for stability of the variance and mean of such a time series. Although the stability of these two properties indicates only a weak form of volatility, this is enough to identify a non-volatile time series or to select those parts of a time series that are acceptable for use. We employed statistical package for social science to simulate the stated screening procedure to find out how good the data is for further forecasting procedures. It was found out that, after employing all the techniques, the data is found to be rich for forecasting of crude oil price for petroleum industry in Malaysia.