Verification of the Results of a Directed Association Experiment (Limiting Responses to the Comparative Conjunction as ... as) Based on the Data from the English-Language web Corpus iWeb
Description
The aim of the article is to develop a methodology for verifying the results of a directed association experiment (limiting the associations of English-speaking respondents with the comparative conjunction as ... as) based on the iWeb corpus. By testing the methodology, the scientific validity of the following assumptions is tested: (1) corpus data serve as a reliable basis for verifying directed association experiments and (2) the frequency of query word occurrences correlates with the frequency of participants’ responses in association experiments, so that the study of corpus data can in some cases replace the need to conduct association experiments in general. The two-stage research methodology involved a comparative analysis of (1) the results of the association experiment (the first sample) and the material of the dictionary of set similes in English (the second sample) and (2) the data from the first two samples and the data from the third – corpus – sample, which contains the most frequent occurrences of query words in comparative constructions [As+Adj+As+Def. Art.+N], [As+Adj+As+Indef.Art.+N], and [As+Adj+As+0Art.+N]. Correlation between the first and third samples was additionally tested by statistical analysis (χ2 determination). Results. The first stage of the study revealed a noticeable gap between the relevant consciousness (respondents’ reactions) and the linguistic material (dictionary data). This gap is explained by various linguistic and cultural factors, primarily globalisation, the reduction of "cultural memory", and the tendency to save on speech effort. The second stage found that the most frequent occurrences of stimulus words representing figurative similes were more appropriate than dictionary material. In contrast to the latter, the occurences of the iWeb corpus correlate better with the reactions of the association experiment, showing the highest coincidence rate of figurative comparisons among the three samples – 14.4%. However, this indicator is too low to confirm the scientific validity of the above assumptions. Conclusions. The verification carried out was not sufficiently objective. However, since the correlation between the data of the first and third samples was statistically proven (χ2=10.45, which allows us to accept the null hypothesis), the idea of such a verification should not be rejected. Therefore, associative experiments can be replaced by corpus data analysis, but only if the relevance (prevalence) of certain linguistic units or phenomena is established. However, even in this case, it should be borne in mind that the results obtained by studying corpus data will be fundamentally different from the results of experiments.