. Author manuscript; available in PMC: 2008 Apr 10.

Published in final edited form as: Brain Lang. 2006 Oct 27;100(1):53–68. doi: 10.1016/j.bandl.2006.09.001

Table 5.

Means for QPA and CIU indices from unpracticed (Snowman ) narrative for each subject before and after period of SentenceShaper use. Values in bold are within one standard deviation of the controls' mean.

QPA Index		ER	MR	HD	KR	NA	SL	Controls^*
Words in sentences/total	baseline	0.57	0.66	0.82	0.9	0.93	0.84	.95 (.09)^&
	post	.72^a	.85^a	.91^a	.99^a	0.89	1.00^a
Mean sentence length	baseline	5.12	4.5	5.59	7.32	7.63	10.77	9.55(2.3)^&
	post	4.19	7.47^a	8.88^a	7.19	6.09	10.33
Prop sent well-formed	baseline	0.41	0.14	0.05	0.32	0.47	0.54	.85 (.12)^&
	post	0.65	.76^a	.63^a	0.33	0.55	0.67
Prop closed class/total	baseline	0.41	0.37	0.41	0.48	0.51	0.49	.54(.04)
	post	0.49	.49^b	.56^a	.57^a	0.51	0.50
Obligatory	baseline	0.7	0.11	0.33	0.76	0.85	0.86	.99 (.02)
determiners	post	0.67	.97^a	.97^a	1	0.86	0.96
Median length	baseline	4	3.5	5	6	7	7	8.17(1.39)
utterances	post	4	7	6	5.5	5	7
*CIUs/total words*	baseline	0.49	0.48	0.58	0.45	0.58	0.46	.84(.05)^&
	post	0.54	.59^a	.76^a	0.43	0.66	.64^a

Control data from Rochon et al., 2000 (N=12); figures marked '&' represent new data collected for this paper (N=10)

post SentenceShaper performance significantly improved over baseline, p<.01

post SentenceShaper performance significantly improved over baseline, p<.05