In experiments where interfering activity is given after the list to wipe out the buffer, the advantage of the last few items disappears and they often show poorer recall - the so-called negative recency effect (Craik, 1970; Gardiner, Thompson & Maskarinec, 1974). In experiments where an effort is made to eliminate rehearsal the primacy effect largely goes away and so does the negative recency effect when tested after a delay (Baddeley, 1986). In such studies, where subjects are prevented from forming a rehearsal buffer and forced just to process the item under study, one also sees a diminished positive recency effect. Performance tends to drop off continuously from the end of the list.

An experiment by Glenberg, Bradley, Stevenson, Kraus, Tkachuk, Gretz, Fish, and Turpin (1980) is of this kind. Subjects studied pairs of words for 2 seconds. In one condition they studied 36 such pairs (for 72 items) each preceded by 4 seconds of distracting activity while in the other condition they studied 9 such pairs (for 18 items) each preceded by 22 seconds of distracting activity. The experiment was designed so that subjects spent 216 seconds studying the list. The effect of the distraction was to prevent any cumulative rehearsal and force subjects just to attend to the presented items. In both conditions there was 20 seconds of intervening activity before recall.

ACT-R Model