“…This finding implies that if, for a given one-semantic-constraint cloze item, potential responses of a given subject are ordered by associative strength (e.g., R 1 , R 2 , …, R n ), then the addition of semantic constraints eliminates certain responses but does not alter the order of the remaining responses (…, R i , R i + j , …). This evident resistance of response order to experimental manipulation is strikingly consonant with the findings of a somewhat similar study by Weaver, Kingston, and Dinnan (1971).…”