Table 4. Logistic Regression Analyses Regressing Participants' Responses Applying Surface and Deep-structure Nonlocal Chunks as Predictors.

	Regression coefficient	p	t (79)
Nonlocal dependency (intercept)	0.280*	0.000	5.85
Nonlocal bigram with two intervening (word)	0.003	0.660	0.44
Nonlocal bigram with three intervening (word)	0.009	0.290	1.06
Nonlocal bigram with two intervening (class)	0.006*	0.000	4.49
Nonlocal bigram with three intervening (class)	0.000	0.873	0.16
Local versus Nonlocal	0.019	0.757	0.31

^*: p<.05.