Perhaps poor comprehenders have a specific (or larger)
difficulty in inhibiting information that becomes irrelevant after they
have activated and processed it in their WM, and not necessarily in
preventing irrelevant information from entering the WM (as in the NP
paradigm used here) – even when the text does not impose a memory
load. If so, a measure of resistance to proactive interference would be
a good predictor of text comprehension for poor comprehenders, as in
the De Beni et al. study