Chapter 1 the study of collocations

4.2  Results of the Main Analyses 
 
In this section the results of the analyses performed to address each of 
the two hypotheses are described. 
4.2.1  Hypothesis 1: There are patterns of development in collocational 
knowledge across proficiency levels 
 
To address Hypothesis 1, the three sets of data were analysed separately:  
 
i)  For the free production data, tokens of the correct use of the thirty-seven 
types of collocation were recorded.  Lack of, or incorrect use of, a particular 
type were scored as 0.  The data were entered as the sum of tokens of correct 
usage of each collocation type by each subject.  
ii)  For the translation data, the mean accuracy of response to each of the six 
types of collocation repeated across groups was calculated.  
 
iii)  For the blank filling data, the mean accuracy of response to each of the 
eleven types of collocation repeated across groups was calculated. 
 
 
247

---

 
The data for these analyses were examined and were not found to be 
normally distributed.  This is due, in the case of the elicited production 
measures, to the fact that means for accurate responses to some types were 
calculated on the basis of a small number of responses to tokens, thus 
restricting the possible range of scores on these types.  In the case of the essay 
data the mean use of many types of collocation did not follow the normal 
pattern of distribution within and across groups.  This justifies the use of non-
parametric Kruskal-Wallis tests, followed by Dunn's multiple comparisons 
procedures to address the first hypothesis regarding between-group differences 
in accuracy and use of collocations.