intense positive regard and social skills training. h period, nmeasured in terms of three categories: nizophrenic (improved), became more schizophrenic (got dependent variable is the worse), or showed no change. The different proportions of improvement among these four treatments can be tested for significance with the usual patient's six-step procedure. Step 1: State the Hypotheses In the case of the two-variable chi-square test, the null hypothesis is that there is no association or correlation between the two variables-that is, the way that one of the variables is distributed into categories does not change at different levels of the second variable. Stated yet another way, the null hypothesis asserts that the two variables are other. Hence this test is often called the chi-square test for independence. For this example, the null hypothesis is Ho method of treatment and degree of parents reverse seems ivorce. nd that he next independent of each decide independent. As usual, the alternative hypothesis is the vorce is improvement are have to negation of Ho; HA: method of treatment and degree of improvement are not would oblems, independent. in each Step 2: Select the Statistical Test and the Significance Level The data consist of the frequencies in categories arranged along two dimensions, so the two-way chi-square test is appropriate. As usual we mind us and this ariables expects Dr.Null will set alpha = 05 Step 3: Select the Samples and Collect the Data Eighty schizophrenics meeting certain criteria (not responsive to previous treatment, more than a certain number of years on the ward, etc.) are selected and then assigned at random to the four treatments, with the constraint that 20 are assigned to each group. After 6 months of treatment, each patient is rated as having improved, having gotten worse, or having remained the same. The data are displayed in Table 20.6, a 3 x4 contingency ects that he ratios the ratio h row of r finding table. Therapy Economy Row Sums 30 Token Group Psychodynamic Rogerian Therapy Therapy Observed 12 28 8 Frequencies 5 4 22 3 6 3 Improved No change 14 2 N= 80 6 20 8 Got worse 20 Column sums 20 ed which over who