Applying the Theory of Planned Behaviour to Account for Students’ Choice of a Target Accent (Part 2)

Bearing in mind the importance of attitude in sociolinguistic research and its huge theoretical potential for accounting for various language behaviours, it is surprising to see numerous misconceptions concerning this construct and its conceptualization as well as criticism as to its role in predicting and explaining speech behaviour (cf., for instance, Cargile, Giles 1997: 195; Edwards 1999: 109; Ladegaard 2000: 229–230; Garrett 2001: 630; Soukup 2012; Taylor, Marsden 2014). The author claims that attitude research can still prove very insightful and helpful in sociolinguistic theory building, but to do so, one needs to reconceptualize attitude along the reasoned action approach on the foundations of which the theory of planned behaviour rests. The theory posits that attitude is one of the three general predictors having a sufficient explanatory and predictive power in the case of most human behaviours. The major goal of the present article is to report on a study attempting to apply the theory of planned behaviour to explain why students of English being given an alternative to choose either an English or American accent as a target model to learn opt for one and not the other. The second goal of the article is to discuss the role of language attitudes in determining students’ decisions. Part 2 of the article elaborates on the main study as well as includes a brief discussion followed by suggestions for further research. 200 KRZYSZTOF PRZYGOŃSKI 5. The main study – introduction The main study was conducted with a view to answering two research questions concerning respectively: (1) the assessment of the potential of the three TPB variables (predictors) for explaining a language-related behaviour and (2) the actual importance of attitude in determining students’ behaviour. To specify, the study consisted in establishing the strength and nature of the correlations between the three variables and students’ choices of the target accents: British (Received Pronunciation – RP) and American (General American – GA) ones. Since the theory of planned behaviour posits that individuals behave largely the way they do because of their attitudes, social pressure and perceived behavioural control, the three predictors were hypothesized to offer a valuable insight into the causes of students’ course enrolment decisions and, in this way, account for their choices. It was taken that the greater the differences in the values assumed by a particular variable with respect to the two accents (for instance, attitude to GA vs. attitude to RP), the more likely it was that this variable impacted on students’ choices. In addition, more extreme values of variables as opposed to those “more neutral” were assumed to point to their greater role in influencing students’ decisions. Importantly, it was also presupposed that in some cases a decision to learn a given target accent could have resulted from very low values assumed by the variables related to the declined pronunciation model rather than a very favourable evaluation of the chosen accent. When it comes to respondents, they were first-year students from the English Department at Adam Mickiewicz University. At the very beginning of the first semester, they could choose between the two accents that were taught there: American English and British English. The accents may be thought to correspond, respectively, to what has been labeled as Received Pronunciation and General American, i.e. models characterized by being made up of standard pronunciation features of the two varieties of English. Because the percentage of students who enrolled on the British pronunciation course was higher (around 60%), it can be assumed that this accent enjoyed a greater popularity among the population of students. This state of affairs translated into a situation in which there were altogether more “British” groups than “American” ones. The majority of respondents could be regarded as young adults, highly motivated aspirers with over 10-year experience of English language learning who had achieved at least a B2 level of competence when entering the university. All of the respondents were asked about their perceptions of their speaking a certain accent and not simply about their decontextual perceptions of the accents in general. In this way, all questions were made relevant to their enrolment decision and subsequent language-related behaviour (learning and speaking this accent). As regards the measurement instrument, it took the form of a written questionnaire distributed to students during their regular classes. It was prepared after conducting two pilot studies the aim of which was a selection of the most relevant semantic-differential scales. The final number of questionnaires accepted Applying the theory of planned behaviour to account for students’ choice ... (Part 2) 201 for a statistical analysis amounted to 53 forms from the “British” and 44 from the “American” groups. The theory of planned behaviour offers two methods of measuring attitude, subjective norm (SN) and perceived behavioural control (PBC). The one applied in this research consisted in constructing relevant scales probing into specific subcomponents of the three behavioural predictors (for details, see Ajzen 2002).1 Attitude here included adjectival pairs that were classified as experiential (attractive – unattractive, pleasant – unpleasant, interesting – uninteresting, satisfying – dissatisfying, natural – unnatural), instrumental (useful – useless, advantageous – disadvantageous, helpful – unhelpful) and general (e.g. the good – bad scale). Subjective norm, in turn, encompassed injunctive and descriptive components. The injunctive one related to students’ perception of what they should do because of other people’s opinions (e.g. the one of parents, teachers and peers) or because of a conviction that something was expected of them. The descriptive item concerned their personal belief about the general popularity of a given accent, i.e. which one was more commonly used and learnt by people in Poland, especially by those who were important to the respondents or those who could be their role models. The components of PBC pertained to students’ “sense of self-efficacy with respect to performing” a given behaviour and his/her perceived control over doing it (controllability) (Ajzen 2002: 7). In this study, the former concerned students’ perception of their own aptitude for learning the accents and the latter related to their perception of the ease in doing so thanks to the accessibility of teaching materials, well-qualified staff, contacts with native speakers and opportunities for travel to the country where the accent was spoken. All of the three variables were measured by means of seven-point, semanticdifferential scales2 in which the value 7 was given to the most favourably perceived items constituting a given variable, 1 to those indicating the least favourable perception and 4 to neither favourable nor unfavourable perception. In order to avoid automaticity in students’ responses, a decision was made to vary the most favourable end-points of the scale so that they sometimes came on the left and sometimes on the right side of it. Attitude was measured by means of 9 bipolar scales whereas subjective norm and perceived behavioural control by two (one for each subcomponent). This was largely due to the greater complexity of the construct of attitude in comparison to the other two predictors. The statistical significance of the data presented in Tables 4 and 5 was verified by a T-test for dependent and independent variables. The level of statistical significance was established on the level of α = 0.05. Consequently, a given difference in students’ perceptions of the two varieties of English can be regarded as statistically significant for α > p. 1 Alternatively, the variables can be measured by means of relevant normative, control and attitudinal beliefs (Ajzen 1991). 2 The whole questionnaire was in Polish and the scales concerning attitude to speaking a given accent were the following: przyjemne – nieprzyjemne, satysfakcjonujące – niesatysfakcjonujące, ciekawe – nieciekawe, dobre – złe, przydatne – nieprzydatne, naturalne – nienaturalne, użyteczne – nieużyteczne, atrakcyjne – nieatrakcyjne, korzystne – niekorzystne. 202 KRZYSZTOF PRZYGOŃSKI 6. Presentation of results To begin with, it must be reiterated that the following analysis of research results is based on an assumption that the more extreme evaluation of a given scale and, by extension, of a given predictor there was, the more likely it was to have impacted on students’ choice of a pronunciation model to learn. Accordingly, when comparing the differences in the values of particular scales constituting attitude (see Table 4), one can come to a conclusion that experiential items (scales 1, 2, 3, 6, 8) of attitude contributed the most to students’ from the “American” groups favourable evaluation of American English and their unfavourable perception of British English. This is especially true about the scales referring to pleasantness (the difference reached – 3.32) and naturalness (4.36). As regards the instrumental component (items 5, 7, 9), one can easily discern that students from “American” groups believed that their speaking British English would be for them only slightly unhelpful, useless and disadvantageous (values close to the “neutral” value of 4) and that American accent would be for them in this respect quite helpful, useful and advantageous. As for the influence of social pressure (SN) on the students’ choice of a target accent, it seems that the injunctive component did not contribute to their perception of social pressure to speak either of the two accents. We can draw such a conclusion because there is no a statistically significant difference in perceiving which behaviour (speaking which accent) was expected of them. Despite this fact, the students did believe that