This is true of the studies cited in Table 4, although many of the studies conducted to obtain reliability and validity data for the construction of standardized tests at least used materials that were equivalent with regard to content and format. The data presented for Spearritt's sixth-grade study are means of some 36 correlation coefficients between nine measures of auding and four measures of reading. In his study, the r of .49 is for girls and the r of .52 is for boys.

The relationship that emerges from these data, when r means at each grade level are computed and plotted as a function of grade level, is shown in Figure 6. The relationship between auding and reading grows as grade level increases up to the fourth grade, and remains fairly constant thereafter at around .58 to .60. Thus, although these data represent a conglomerate of procedures, materials, and subjects, the relationship that emerges between grade level and correlations of auding with reading is consistent with the expectation that correlations among auding and reading test performance should increase as school grade increases and children acquire adequate reading decoding skills. Further credence is lent to the data of Figure 6 when we consider the data from Loban's (1964) longitudinal study displayed in Figure 5. Both sets of data agree in suggesting an increase in the relationship between auding and reading test performance as school grade increases, and children acquire reading decoding skills for using print to develop meaning previously available only by auding.