Is output order in free recall based on the strength of the memory trace or on the subject strategies and depth of coding employed?

If output in free recall proceeds from items with strong to those with weak memory traces then Craik's (1970) finding is paradoxical. He reported that probability of recall at a later test increases with output position in the first recall attempt. Does recall order depend on depth of encoding and retrieval strategy rather than trace strength? Examination of recall order and the serial position of the first item retrieved in a conventional free recall experiment supported the hypothesis that the Craik effect occurs because on the initial test the items output early are superficially encoded while the last few items are semantically processed. By preventing superficial encoding and encouraging semantic processing the Craik effect was eliminated. There was no evidence for a trace strength explanation of recall order.