- Bowman, D.T. and Rawlings, J.O. (1995). Agron. J., 87: 147-151
- Yates, F. and Cochran, W.G. (1938). J. Agri. Sci., 28: 556 - 580
Crop performance trials are conducted to provide unbiased relative performance data for breeders, growers and extension personnel. This paper presents a study that provides a criterion level of precision for rejection of data. In the present study historical data on sunflower crop were used to determine relationship between productivity and error variance by regression log(variance) and log(mean) for the variable of primary interest. A nonsignificant relationship dictates a suitable criterion for rejection of the doubtful experimental data from the pooled analysis point of view. In the study of sunflower the results indicated that if its observed EMSS is greater than 40584.4, then the data from such experiments should not be used with the other sets of data and is worth to discard it to avoid misleading conclusion.