Knowledge of the critical value for soil erosion data is useful in terms of soil conservation and management. An experiment was carried out in Universiti Pertanian Malaysia on Padang Besar soil series (Petroferric tropudult). The bare plot soil loss and rainfall characteristics were determined after every storm for a year. Soil loss and erosivity index (K.E. > 10) were fitted with different models. Various approaches were then adapted to study their suitability in obtaining critical values for soil erosion data. In this paper, an example showing the application of techniques and their principles is given. Results indicate that all the four procedures provide similar critical values despite their different principles. It is further showns that new Cate-Nelson technique and simplified ANOVM (a- 0.001) result in higher R (0.84**) than the best quadratic model (R2 = 0.54**). However, due to the principle of obtaining critical values for soil erosion data, it may suffice to use simplified ANOVM (Ct = 0.001).