For the “equal groups” condition, both array size, slope = 139 ± 4 ms/item [±standard error], t (286) = 35.25, p <10 −16 , and operation type, slope = 289 ± 14 ms per additional added item, t (286) = 19.41, p <10 −16 , were highly significant (conditional r 2 = 89.5% of variance explained), thus confirming that the maximal savings were observed for nonprime numbers that could be subdivided into equal groups, and that the need to add 1 or 2 imposed an additional toll of almost 300 ms/item for numerosities 5, 7, 10, and 11.
1
—
—