The literature is reviewed for four combinations of evaluation measures and methods: traditional methods with traditional measures (T-Meth/T-Mea), nontraditional methods with traditional measures (N-Meth/T-Mea), traditional measures with nontraditional measures (T-Meth/N-Mea), and nontraditional methods with nontraditional measures (N-Meth/N-Mea). What is traditionally measured is cognitive achievement before and after a treatment (i.e. the program). The method used is traditionally the administration of an achievement test whose score is afterwards statistically manipulated. The approach is research-oriented in that an hypothesis is made before the evaluation begins. It is frequently difficult to separate what is measured ("measures") from how it is being measured (method). The N-Meth/N-Mea category contained the largest number of examples in the literature and may be described as one in which materials other than tests and direct observation are examined by other than statistical methods. "Triangulation" has been defined as the systematic use of different methods to cut down on validity threats. The use of a variety of techniques increases reliability by cross-checking. Triangulation seems to combine the advantages of both the traditional and nontraditional approach. (RL)