All stimulus sets contain biases in similarity measures such that grammatical test stimuli resemble training stimuli acoustically more than do non-grammatical test stimuli.
Their approach required a considerable amount of effort and knowledge to manually design patterns that were later used to generate grammatical test items (Chen, Liou, & Chang, 2006).
Their technique was based on manually designed patterns and it was used to find authentic sentences from the Web and transform into grammatical test items.
Three old items used in training were also included to make up 25 grammatical test items.
Among the reasoning tests they used were Baddeley's grammatical reasoning test and verbal analogies test.
Since the adapted grammatical reasoning test was timed, test takers reached only some of its items, obtaining a mean score of 22.89 out of 64 and a standard deviation of 7.46 (range = 7 41).
The correlations reported between the grammatical reasoning test and other intelligence tests are almost as high as the correlations between pure nonverbal measures of intelligence.
For instance, Hartley and Holt (1971) reported a correlation of.70 between the grammatical reasoning test and the AH4, a group test of general intelligence in children.
The grammatical reasoning test correlated at.47 with the verbal analogies test and.20 with paragraph comprehension.
In another study, Furnham and Chamorro-Premuzic (2006) demonstrated that the grammatical reasoning test correlates with the Wonderlic Personnel Test, Raven's Advanced Progressive Matrices (Raven, Court, & Raven, 1977), and General Knowledge Test (Irwing, Cammock, & Lynn, 2001) at.65,.39, and.35, respectively.
For data collection, two pre−/post tests were designed by the researchers: a 35-item grammatical structures test (GST), targeting the simple present, gerund, modal auxiliaries, questions, present perfect, and conditionals/type II and a 30-scenario Discourse Completion Task (DCT) divided equally among the speech acts of apology, request, and suggestion.
