The focus of this experiment was on the effects of associative strength on retention of connected discourse, in terms of both single words and strings of words. Also of interest was the short- and long-term retention of two types of information, verbatim and substance. Verbatim information covered words and word sequences identical to those in the test passage. Substance information covered the main ideas or essential points. The subjects were 120 fifth graders from predominantly middle-class communities. Three 172-word passages were prepared which were identical except for the embedded associative pairs. The three passages had the same stimulus words, but the response words were different. Groups of subjects were assigned to one of the following four conditions: one reading, immediate testing; two readings, immediate testing; one reading, delayed testing; or two readings, delayed testing, according to a predetermined random sequence. The subjects were instructed to silently read the passage and remember as much of it as possible. A cued recall test was given immediately or 24 hours later. From the results it was concluded that no significant high-associative facilitation occurred. Furthermore, greater recall was noted across the four verbatim measures and one substance measure in the low-associative condition. (TS)