Targets are identified faster if they appear at the same level (global
vs. local) as the target on the previous trial. This effect has been
attributed to a controlled attentional process wherein participants at
tend to the level containing the target on trial n - 1, thus facilitat
ing performance if target level repents on trial n. An alternative exp
lanation (Lamb & Yund, 1996) is that automatic activation of level-spe
cific neural mechanisms is responsible. The controlled-process hypothe
sis predicts better performance whenever target level is predictable,
because participants can anticipate, and thus attend to, the appropria
te level. The automatic-process hypothesis predicts better performance
when target level repeats, regardless of target-level predictability.
In the present studies, level repetitions facilitated performance reg
ardless of target-level predictability. The data support the automatic
-process explanation of level repetition and other attentional effects
. The data illustrate how an automatic process can give rise to perfor
mance that seems to reflect more cognitive or strategic mechanisms.