ABSTRACT
If the flexible and nonflexible treatments do not differentially affect the student's ability to construct a good research design, then we should find the numbers of plus signs for the two groups differing only as a result of sampling error. The hypothesis may be tested by X2 • (This is another application of the l for correlated observations discussed in Chapter 10). For the data above we have
x2 2: (lfo - /..1 .5)2 (4.5)2 (4.5)2 = -- + -- = 4.05 f, 10 10 x2 = 4.05, df = I, p < .05.
