Chapter 1 the study of collocations

5.2.2  Blank-Filling Data 
 
 
Cued production of collocations was also measured in a blank-filling test.  
Each sentence contained an English collocation with one part missing.  Subjects 
were required to provide the missing part of each collocation.  The collocations 
included in the blank-filling test could not be translated directly into the learners' 
L1, Greek.  The results for this set of data are discussed below. 
 
 
 
 
5.2.2.1  Between-Group Differences 
 
 
The results in the blank-filling data revealed that for Type 1. Noun Prep, 11. 
SV(O) Prep O, 23. SV(O) Adverbial, 24. SV(O) wh-word, 27. Verb Noun (creation), 33. 
Verb Adverb, 36. Prep Det Noun, and 37. Phrasal Verb collocations, subjects were 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
30 

---

significantly more accurate in supplying the correct collocation as their level of 
proficiency increased. 
 
The textbook analysis (see Chapter 3) also showed that all but two of the 
above collocation types exhibit a similar pattern of increase across the three 
textbooks.  For example, tokens for Types 1. Noun Prep, 24. SV(O) wh-word, 27. Verb 
Noun (creation), 33. Verb Adverb, 36. Prep Det Noun, and 37. Phrasal Verb increase as 
the level of language proficiency increases from TWE1 to TWE3.  The students' 
exposure to these collocation types could have influenced their performance to the 
blank-filling test, i.e. the more frequently students were exposed to a particular 
type of collocation, the more accurate they became in their knowledge of 
collocations of this type. 
 
The subjects' performance on two collocation types showed a U-shaped 
pattern of acquisition.  For Types 5. Adjective Prep and 34. Noun Noun, subjects in 
Group 1 were more accurate than subjects in Group 2, who were also less accurate 
than subjects in Group 3.  A look at the specific collocations shows that the level of 
difficulty increases with proficiency.  For example, Type 34. Noun Noun 
collocations for Group 1 were 'post office' and 'phone number', for Group 2 'traffic 
lights', and for Group 3 'curriculum vitae'.  Group 3 subjects were more accurate in 
responding to the 'curriculum vitae' collocation than subjects in Group 2 were with 
the 'traffic lights' collocation, even though 'curriculum vitae' is less frequent than 
'traffic lights' in everyday speech.  The analysis of the textbooks shows that 'traffic 
lights' appears twice in TWE2, while 'curriculum vitae' appears only in TWE3, four 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
31 

---

times.  Again, the amount of exposure to a specific collocation appears to influence 
performance. 
 
Performance on the Adjective Prep collocations also increased as the level of 
proficiency increased.  For example, some of the Adjective Prep collocations for 
Group 3 were: 'competent in', 'fond of', successful in', 'married to', 'unsure about', 
'similar to', 'slow in', 'capable of', 'regardless of'.  Group 2 subjects were tested on 
the following Type 5 collocations: 'full of', 'sympathetic to', 'engaged to', 'upset 
about'.  Despite the fact that Group 3 students were faced with a larger number of 
Type 5 collocations compared to Group 2 students, they were more accurate in 
supplying the correct collocations.  Both Noun Noun and Adjective Prep collocations 
are lexical collocations.  It appears that students at the initial stages of ESL learn 
specific lexical collocations, possibly as unanalysed chunks, and hence they are 
relatively accurate with respect to selected lexical collocations.  As their proficiency 
increases and their grammatical knowledge develops, their relative accuracy in 
lexical collocations declines and they become better in grammatical collocations (in 
the translation test too, intermediate level students were better in SV(O) Prep O 
collocations).  At the post-intermediate level, the subjects' overall accuracy 
increases, and they once again become more accurate in lexical collocations.  Such a 
U-shaped development in L2 learners has been reported in previous linguistic 
studies too (see McLaughlin 1987, 1990; Lightbown et al. 1980). 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
32 

---

 
Overall, accuracy improves in the same fashion as in the essay and 
translating tasks.  Subjects in Group 3 performed the best and subjects in Group 1 
performed the worst. 
 
The implicational analysis for the blank-filling data had a low coefficient of 
scalability (C
scalability
 = .33), and thus there is little evidence of a stable acquisition 
order.  Again, the U-shaped learning patterns probably contributed to the low 
scalability of the blank-filling data (see Hatch & Lazaraton 1991:216).