Competing theories of analogical reasoning have disagreed on the relative contributions of surface and structural features to the access of previously read base stories when one is reading a current cue story. A key limitation of the prior work was that surface and structural feature overlap between bases and cues was not manipulated precisely. The present study systematically manipulated the number of surface and structural matches to determine their relative effect on access. Results involving reminding and reading-time measures suggest that surface and lower-order structural features affected access about equally, at least when a higher-order relation (HOR) was shared between a base and cue story. When a HOR was not shared, surface feature overlap continued to affect access while lower-order structural features had a less reliable effect. Models of access might need to be adjusted to account for these phenomena.