Chapter 1 the study of collocations

Table 30. Summary of the results for the Kruskal-Wallis tests and post-hoc 
analyses for the blank filling data 
     Dunn’s 
Procedure: 
Mean 
Rank 
Differences 
Type 
K-W 
1    vs.    2 
2    vs.   3 
1    vs.   3 
1. Noun Prep 58.03 
 
95.566<146.883*  146.883<171.62*  
95.566<171.628* 
5. Adjective Prep 19.673 
160.791>110.53* 110.532<143.64* 160.791-143.64 
11. SV(O) Prep O  14.711  115.747<137.5* 
137.5    <161.02*  115.747<161.02* 
23. SV(O) Adverb  39.351  102.495<149.42*  149.42  -161.97 
102.495<161.97* 
24. SV(O)wh 22.03 
111.758<139.91* 
139.915<162.53* 
111.758<162.53* 
27. Verb N (creat.)  29.988 108.967<132.78*  132.787<172.8*  108.967<172.8* 
33. Verb Adverb 77.892 
116.731-108.138 108.138<190.69* 116.731<190.69* 
34. Noun Noun 26.577 
167.885>114.82* 
114.824<131.98* 
167.885>131.98* 
36. Prep Det N 17.299 
131.505-118.064 
118.064<165.38* 
131.505<165.38* 
37. Phrasal Verb 26.57 116.176-125.872 125.872<172.73* 116.176<172.73* 
*: Significant at the .05 level 
< or >: Direction of the difference 
 
4.2.1.3.1  Kruskal-Wallis Analyses for the Blank Filling Data 
 
 
270

---

 Type 
1. 
Noun Preposition - Results of the Kruskal-Wallis test of 
responses to Noun Preposition collocations show a significant main effect for the 
factor Group (Kruskal-Wallis c
2
 (2, N = 275) = 58.03, p = .0001).  The mean 
accuracy of responses to Noun Preposition collocations increases across groups 
(see Table 31).  The results of the post-hoc Dunn's multiple comparisons 
procedure show significant differences in the mean accuracy of response to this 
type of collocation between all groups (see Table 30). 
 
 Type 
5. 
Adjective Preposition - Results of the Kruskal-Wallis test of 
responses to Adjective Preposition collocations show a significant difference 
between conditions (Kruskal-Wallis c
2
 (2, N = 275) = 19.673, p = .0001).  The 
results of the post-hoc Dunn's multiple comparisons procedure show 
significant differences in the mean accuracy of response to this type of 
collocation between Groups 1 and 2, and between Groups 2 and 3, but not 
between Groups 1 and 3 (see Table 30).