Researchers typically report the reliability of a scale in a table that reports simple descriptive statistics for the study’s variables such as means, standard deviations, and intercorrelations. In these tables, coefficient alpha estimates are usually reported on the diagonal of the correlation matrix, within parentheses. Such an approach appears in Table 7.1.
Variables | Mean | SD | 1 | 2 | 3 |
---|---|---|---|---|---|
1. Authoritarianism | 13.56 | 2.54 | (.90) | ||
2. Helping others | 15.60 | 3.22 | .37 | (.78) | |
3. Financial giving | 12.55 | 1.32 | .25 | .53 | (.77) |
Note. N = 200. Reliability estimates appear on the diagonal. |
In the preceding table, information for the authoritarianism variable is presented in both the row and the column that is headed “1.” Where the row headed “1” intersects with the column headed “1,” you find the coefficient alpha for the authoritarianism scale; you can see that this index is .90. In the same way, you can find the coefficient alpha for helping others where row 2 intersects with column 2 (α= .78) and you can find the coefficient alpha for financial giving where row 3 intersects with column 3 (α= .77).
When reliability estimates are computed for a relatively large number of scales, it is common to report them in a table (such as Table 7.1) and make only passing reference to them within the text of the paper when within acceptable parameters. For example, within the section on instrumentation, you might indicate:
Estimates of internal consistency as measured by Cronbach's alpha all exceeded .70 and are reported on the diagonal of Table 7.1.
When reliability estimates are computed for only a small number of scales, it is possible to instead report these estimates within the body of the text itself. Here is an example of how this might be done:
Internal consistency of scale responses was assessed by Cronbach's alpha. Reliability estimates were .90, .78, and .77 for responses to the authoritarianism, helping others, and financial giving subscales, respectively.
3.21.162.242