Introduction

This paper discusses the use of the repertory grid technique (RGT) as a credible research tool in the field of teaching English as a foreign language (TEFL), specifically for research focused on issues related to teacher beliefs. As an example of its use, we describe a case study concerning foreign language teachers’ cultural perceptions about good pedagogy. Because research into culture is particularly fraught with methodological difficulties (^{Baldwin et al., 2006}), the topic serves as a useful focus to illuminate some of the advantageous features of the RGT.

The role of culture in second and foreign language teaching and learning is well established, and research on the topic is extensive. In the last decade or so, at least four literature reviews on the subject have been published (^{Álvarez-Valencia, 2014}; ^{Lessard-Clouston, 2016}; ^{Risager, 2011}; ^{Young et al., 2009}). Most empirical studies comparing and contrasting cultural views on pedagogy (language or otherwise) have typically relied on questionnaires, often supported by structured interviews, using previously validated and standardized items (e.g., ^{Clark-Gareca & Gui, 2019}; ^{Liu & Meng, 2009}; Organization for Economic Cooperation and Development [OECD], 2009; ^{Pawlak, 2011}; ^{Schulz, 2001}). These data elicitation techniques are not without problems. Questionnaires have been criticized for their susceptibility to common method variance, which occurs when respondents’ answers do not genuinely reflect their authentic views but are instead influenced by the instrument’s design (^{Gorrell et al., 2011}). Research has identified several ways in which this kind of method bias can pollute questionnaire data (^{Podsakoff et al., 2003}).

To mitigate the inherent defects of questionnaires, many researchers support them with structured interviews. Such interviews, however, are subject to their own limitations, including the possible downsides of “subjectivity, the generalisation of the findings, conscious and unconscious biases, influences of dominant ideologies and mainstream thinking” (^{Diefenbach, 2009}, p. 875). When data from questionnaires and structured interviews are used together, problems arise in aligning data derived from the two methods. Poor alignment can be attributed to

the complexity and instability of the construct being investigated, difficulties in making data comparable, lack of variability in participant responses, greater sensitivity to context and seemingly emotive responses within the interview, possible misinterpretation of some questionnaire prompts, and greater control of content exposure in the questionnaire. (^{Harris & Brown, 2010}, p. 1)

The RGT is a type of structured interview associated with the field of personal construct psychology (PCP). Combining the best features of both qualitative and quantitative techniques, the grid technique militates against some of the weaknesses of both. First, the structure of repertory grid interviews helps mitigate method bias. As with any qualitative interviewing method, there is always a danger that the RGT researcher may impose constructs or lead participants. However, because the main role of the interviewer is to focus and clarify a participant’s responses rather than guide the interviewee through a series of predetermined questions, the potential for “leading the witness” is reduced. Second, unlike elicitation techniques that necessitate post-hoc thematic analyses, the repertory grid process brings the most essential themes of an interview immediately to the fore. Finally, because RGT data is amenable to statistical analysis, it allows researchers to uncover patterns in participant responses which reflect psychological relationships within the “construing systems” of both individuals and groups (^{Fransella et al., 2004}, p. 81).

The following is a consideration of the RGT as a methodological instrument. Specifically, the paper describes an analysis of the commonalities and individuations of the pedagogic beliefs of nine language teachers at a public university in central Mexico in order to illustrate how the RGT can profitably be employed in the field of applied linguistics and TEFL. First, an overview of the RGT is provided. Second, several studies in the field of TEFL that have utilized RGT interviews are reviewed. Third, literature concerning the impact of cultural beliefs on language pedagogy is presented to situate the subsequent discussion of repertory grid analysis. Finally, a case study is presented as one example of how the grid technique can be used to capture information about cultural dissimilarities in teachers’ pedagogic beliefs and, by extension, to capture information concerning pedagogic beliefs in general.

The Repertory Grid Technique

The RGT is a specific type of interview utilized to analyze the content and structure of the implicit theories that people rely on to construe reality. Of the methodologies associated with ^{Kelly’s (1955}, 1963) theory of PCP, the RGT is the most well-known.

Although repertory grids were initially developed by Kelly for use within the field of clinical psychotherapy and primarily focused on the individual level of analysis, scholars in other disciplines have adopted its premises and employed its methods to understand belief systems at the collective level (^{Jankowicz, 1987}). Kelly’s (1955) writings on “commonality” and “sociality” explicitly address the tendency of groups to create tacit theories of the world. People, of course, define themselves in overlapping ways as members of ethnicities, genders, economic classes, age cohorts, and professional or occupational groups. All such sub-cultures build on a shared perspective that orders their respective “fields of experience to provide identification and solidarity for its members” (Kay, 1970, as cited in ^{Diamond, 1982}, p. 401). As ^{Wright (2004)} points out, when individual constructions are brought together, “certain underlying collective frames of reference emerge that reflect a sense of common understanding and shared meaning” (p. 354). ^{Sechrest (2009)} argues that uncovering these frames likely has more definite implications for research than any other area of Kellian theory.

The RGT is a “two-way classification of data in which [entities or] events are interlaced with abstractions” (Shaw, 1984, as cited in ^{Zuber-Skerritt, 1988}, “The repertory grid” section, para. 2). Repertory grids “reflect part of a person’s system of cross-references between their personal observations and experience of the world . . . and their personal classifications or abstractions of that experience” (^{Zuber-Skerritt, 1987}, p. 604). Using Kellian nomenclature, personal observations and experiences are denominated “elements”; abstractions of experiences are denominated “constructs.”

Elements are a set of events and entities external to the interviewee. In clinical psychology, for instance, elements might be family members. In marketing, elements might be a set of different cars, or vacation destinations, or cellphones. In the current research, participants were asked to think about their past teachers. The choice of “past teachers” as grid elements is premised on ^{Lortie’s (1975)} theory of the “apprenticeship of observation,” which denotes the internalization of teacher roles, identities, and practices that occurs over the course of an instructor’s time as a student. These beliefs about teaching constitute what have been referred to as “folk pedagogies,” a term which emphasizes the cultural dimension of how students come to understand teaching (^{Joram & Gabriele, 1998}).

Constructs, ^{Kelly (1955)} asserted, are the personal theories that arise when humans compare or contrast any two entities. Humans develop hypotheses based on these theories which, in turn, are tested through on-going “experiments” (i.e., interactions) with their environments (^{Beail, 1985}; ^{Fromm, 2004}; ^{Hardison & Neimeyer, 2012}). In other psychological approaches, these theories may be variously referred to as personality, attitudes, habits, reinforcement history, information coding system, psychodynamics, concepts, or philosophy (^{Fransella et al., 2004}). Borrowing ^{Jerrard’s (1998)} denotation, in this paper constructs are defined as a “basic dimension of appraisal” (p. 41).

The constructs in ^{Kelly’s (1955)} model are always bipolar. When elicited during an RGT interview, the two poles are designated the emergent pole and the implicit pole. The emergent pole refers to the original comparison or contrast given by the participant; the implicit pole is elicited by asking the participant to provide what they believe to be the semantic opposite of the emergent pole. Because each construct is bipolar, it is possible for participants to use a numerical scale to evaluate each element. That is, each completed grid can be thought of as a “personal differential questionnaire” (^{Tomico et al., 2009}, p. 57) that participants can use to numerically rate elements in terms of constructs, allowing for a variety of statistical analyses to be conducted.

The Impact of Cultural Beliefs on Language Pedagogy

The widely acknowledged idea that socio-cultural forces influence teachers’ pedagogical beliefs and professional practices is encapsulated in the term situated cognition (^{Brown et al., 1989}). This concept is based on the notion that knowledge is always developed within a given context. Teaching and learning are never neutral acts: They are inseparable from their socio-cultural settings (^{Brown et al., 1989}; ^{Lave, 1988}; ^{Lave & Wenger, 1991}). Through classroom activities, teacher models, and peer influence, students are apprenticed into a particular culture of learning that reflects wider cultural assumptions (^{Lave, 1988}). Teachers are likewise enculturated. The anthropologist Conrad Kottak (2004, as cited in ^{Read et al., 2009}) defines enculturation as

the process where the culture that is currently established teaches an individual the accepted norms and values of the culture in which the individual lives. The individual can become an accepted member and fulfill the needed functions and roles of the group. Most importantly, the individual knows and establishes a context of boundaries and accepted behavior that dictates what is acceptable and not acceptable within the framework of that society. It teaches the individual their role within society as well as what is accepted behavior within that society and lifestyle. (p. 52)

There is considerable empirical evidence to support these ideas. For instance, the OECD’s Teaching and Learning International Survey (2009) compared perspectives on pedagogy in 16 OECD and seven partner countries. Findings indicated that the influence of culture and pedagogical traditions on teachers’ beliefs and practices is “exceptionally high” (p. 96). ^{Schleicher (2018)}-summarizing the OECD’s most recent 2018 survey of teaching and learning-reaffirmed the cultural dimensions of teaching, noting that “the meaning of teacher professionalism varies significantly across countries, and often reflects cultural and historical differences” (p. 29).

Language education, like all education, is a cognitively situated activity. Whether overtly or covertly, a process of “cultural scripting” (^{Stigler & Hiebert, 1999}) encourages both teacher and students to conform to the socio-cultural practices of their educational environment.

How socio-cultural forces influence teacher and student perspectives on foreign language pedagogy has been a fertile area of study in TEFL research (see, for instance, ^{Amiryousefi, 2015}; ^{Widiati & Cahyono, 2006}; ^{Yoo, 2014}). Indeed, the large number of such studies supports ^{Atkinson’s (1999)} assertion that “except for language, learning, and teaching, there is no more important concept in the field of [TESOL] than culture” (p. 625).

Given the sheer volume of research focused on questions of culture, it is remarkable that there are relatively few studies in the field of TEFL devoted to cross-cultural comparisons. The exception here is the research contrasting Chinese and Western beliefs and attitudes about EFL pedagogy (^{Anderson, 1993}; ^{Burnaby & Sun, 1989}; ^{Clark-Gareca & Gui, 2019}; ^{Degen & Absalom, 1998}; ^{Hong & Pawan, 2015}; ^{Rao, 2013}; ^{Shi, 2009}; ^{Simpson, 2008}; ^{Stanley, 2013}; ^{Zhang, 2016}; ^{Zhou et al., 2011}). Other comparative studies, however, are relatively rare (^{Aubrey, 2009}; ^{Can et al., 2011}; ^{Liu, 2004}; ^{Pawlak, 2011}; ^{Richter & Lara-Herrera, 2017}; ^{Rubenstein, 2006}; ^{Schulz, 2001}).

Method

This exploratory study is concerned with the tacit beliefs of foreign language teachers concerning cultural perceptions of good language teaching. It is offered as an example of the usefulness of the RGT, both in terms of the productiveness of elicitation and the utility of subsequent analysis.

Findings and Discussion

In all, the nine participants generated 177 constructs. The average number of constructs among the English teachers was 21; among the Spanish teachers, the average was 26; and among the French teachers, the average was 11. In all, 1,770 pieces of data (i.e., all emergent and implicit constructs plus the participants’ ratings on the constructs) were elicited.

Dominance and Importance Measures

Table 1 displays the relative percentages (i.e., dominance) for each construct category as well as the elicitation order (i.e., importance) for the English, Spanish, and French teachers. The elicitation order index was derived by calculating the mean of the order of all constructs within each construct category. Based on the total number of constructs generated, the order of the constructs was normalized for each participant to a range of 0 to 1: a 0 value reflects the first construct that was elicited, and a 1 value reflects the last construct. A standard normalization formula was used:

X normalized = (b - a) * [ (x - y) / (z - y) ] + a

This can be reduced to

normalized order = order rank - 1 / total constructs - 1

The standard deviations (which are critical for estimating the homogeneity of a category of constructs in the relative order) are included in parentheses after each rank.

Table 1 Dominance and Importance Measures 

Note. The dominance of each category (measured by the relative percentage of total constructs) and the importance of each category (measured by the elicitation order) for the English, Spanish, and French teachers. Standard deviations are displayed in parentheses.

The dominance analysis highlighted major alignments and disjunctures between the three groups. For instance, the relative percentage measures demonstrate that all of the teachers regarded the personal and professional aspects of their work as significant. Constructs in this category, which included “willingness to grow,” “ability to adapt,” and “dedication,” demonstrated a commitment towards instructional excellence and professional development. The French teachers emphasized this aspect of their work: almost 50% of their constructs had to do with their personal and professional orientation. Overall, all three groups shared beliefs about the importance of establishing a classroom atmosphere conducive to learning. The English and French teachers were alike in that both groups generated a significant number of constructs having to do with structuring learning, such as careful planning, organization, and assessment (20% and 15% of total constructs, respectively). The Spanish teachers placed a great deal of emphasis on the importance of pedagogic knowledge (20% of the total constructs generated by this group).

These findings are enhanced by the nuance afforded by an analysis of the importance indices. While the English teachers offered the largest number of constructs related to establishing an atmosphere conducive to learning, according to the elicitation index, content and pedagogic knowledge may be more important to them (these constructs were ranked as 1 and 2, respectively). The agreement between the English teachers regarding their rankings, as reflected by the low standard deviation scores, adds credence to this claim. Pedagogic knowledge was similarly important to the Spanish teachers. Interestingly, the relative percentage here is more in line with the salience of the construct. That is to say, the Spanish teachers both created a high number of constructs associated with pedagogic knowledge and rated these constructs as the most important to their practice. Finally, for the French group, although personal and professional orientation was the most “replete” category (comprised of 49 constructs), “motivating learning” was ranked as the most important construct.

Semantic Similarity Measures

Semantic similarity is a metric defined by how closely or remotely an individual perceives the distance between the meanings of two (or more) units of language, concepts, or instances (^{Harispe et al., 2015}). When a hierarchical cluster analysis of correlations is applied to the numerical data in a repertory grid, the more that constructs or elements are alike, the closer they approximate a score of 100, which would signify a perfect correlation. Thus, in Figure 1, the construct categories “establish an atmosphere conducive to learning” and “personal and professional orientation” are closely linked (a 95.7% match). This suggests that for the English teachers, an instructor who makes students feel secure, is approachable, and “nurturant” (^{Dunkin, 1995}, p. 24) and is probably also a teacher who tends to integrate their personal and professional identities (^{Sabirova et al., 2016}). For the English teachers, content knowledge and experience were also highly correlated (94.3% match). In terms of elements, the English teachers viewed being a language teacher in roughly the same terms as being any other type of teacher (94.4% match).

Figure 1 Cluster Analysis of English Teachers’ Construct Categories and Supplied Elements 

Figure 2 shows that the Spanish teachers also believed there to be a strong connection between establishing an atmosphere conducive to learning and an instructor’s personal and professional orientation (94.4% match). In comparison to the English participants, the Spanish group, however, viewed the pedagogic characteristics of language teachers as being relatively distinct from the characteristics of the non-language teachers they had identified as elements (85% match).

Figure 2 Cluster Analysis of Spanish Teachers’ Construct Categories and Supplied Elements 

The French teachers viewed both the construct categories and the supplied elements as semantically independent units with relatively little overlap between them. As seen in Figure 3, for this group, the concepts “establishing an atmosphere conducive to learning” and “personal and professional orientation” were also the most semantically similar. However, unlike the English and Spanish groups, these concepts only matched at a relatively low 89.6%. For the French instructors, a “great teacher” and a “great second language teacher” only matched at approximately 75%, suggesting that for this group, language teachers possess several characteristics and beliefs that distinctly separate them from teachers in other fields. To understand these differences better, follow-up interviews would have to be conducted.

Figure 3 Cluster Analysis of French Teachers’ Construct Categories and Supplied Elements 

Conclusion

The repertory grid is an interview technique that explores the structure and content of the implicit theories people rely upon to construe their experiences. As mentioned in the introduction, the technique merges the strengths of both qualitative and quantitative approaches in a way that mitigates a number of methodological difficulties associated with other qualitative, quantitative, and mixed-methods procedures.

In this study, nine language teachers at a public university in central Mexico were interviewed about their conceptions of “good” language pedagogy. The dominance and importance measures demonstrate that all the language teachers viewed the personal and professional aspects of their work as important. All three groups shared beliefs about the importance of establishing a classroom atmosphere conducive to learning. Both the English and French teachers generated a significant number of constructs having to do with structuring learning. In addition, the Spanish teachers placed a high emphasis on the importance of pedagogic knowledge. Findings like this demonstrate the usefulness of the technique in uncovering tacit, pedagogical beliefs, a knowledge of which would, of course, be useful in second language teacher education, particularly in terms of opportunities for self-reflection and monitoring changes in pedagogical perspectives over time.

As this research was premised on creating an example of how repertory grids function, the study must be considered exploratory and illustrative. And as with any such study, the results must be heavily caveated. In the present case, first, the concept of culture is infamously “messy” (^{Fives & Buehl, 2012}; ^{Pajares, 1992}) and thus would have to be carefully disambiguated were this research to advance beyond its current state. Second, there are unresolved questions regarding sample size. The literature is notoriously unresolved as to requisite sample sizes in repertory grid investigations, with different researchers advocating sample sizes of between six and 25 to approximate the universe of meaning within a given population (^{Dillon & McKnight, 1990}; ^{Dunn, 1986}; ^{Ginsberg, 1989}; ^{Hassenzahl & Trautmann, 2001}; ^{Heckmann & Burk, 2017}; ^{Moynihan, 1996}; ^{Tan & Hunter, 2002}). In any follow-on study, the question of proper sample size would have to be carefully addressed. Lastly, were the research to be carried further, whether to rely on pre-formulated categories or to undertake conceptual content analysis is a question that would need to be resolved.

These methodological concerns, however, are largely tangential to the purpose of the current article, which offered a small-scale study to elucidate how RGT interviews are conducted and to present a few of the ways in which the resultant data can be analyzed. As should be apparent, the RGT’s usefulness is in no way limited to the topics explored in this article. It is a data elicitation and analysis approach suitable for any study focused on teacher, student, or shareholder beliefs. It allows for comparisons of beliefs between a variety of people on a wide range of topics. In sum, repertory grids are distinguished for their ease of use, their utility in precisely defining concepts without the need for post hoc analyses, their usefulness in uncovering connections between seemingly dissimilar concepts, their ability to restrain researcher bias, their high degree of validity, and their amenability to several types of statistical analysis grounded in participants’ qualitative and idiosyncratic views of the world.