258

In search of research methodology for Applied Linguistics

Rahman Sahragard
(Shiraz University)

Abstract
There are a few issues that are the concern of this paper. First, when it comes to research paradigms,
the novice researcher may be confused as to what kind of research has been undertaken to do in terms
of the underling philosophy. Even when known, the same researcher opts for one which is more in
fashion or may lead to more tangible results. Second, up to the early1990s, the consensus among
language researchers was to follow in the footsteps of scientific research. So a large number of studies
on language dealt with issues such as 'research design', 'population and sample selection', 'treatment',
etc. The tendency among the new generation of researchers, however and the insurgence of post-
modern movement has made it inevitable for the applied linguists to look more at the other end of the
continuum which deals with individual cases, small practices, and local events with no tendency for
generalization. This approach has not been wholly recognised by some researchers, however. To them
any deviation from the procedures of scientific research cannot gain the status of real and useful
investigation. The present writer believes that these are just extremes and a solution not the solution
lies in adopting procedures that look at all aspects of the phenomenon under investigation. Finally, in
spite of the clear advantage of having multiple perspectives over single-method studies, in practice, a
majority of studies favour the latter.

The Challenge of Paradigms
The two paradigms mentioned in the abstract are quantitative and qualitative
methodological paradigms (Hammersly, 1992). The present study seeks to draw on these
paradigms in a complementary (if not equally balanced) fashion. An essential preliminary point is
to note that the choice of paradigm depends fundamentally on the purpose of the research
envisaged and on the nature and focus of the research questions. This choice includes, of course,
the increasingly common practice of combining methods within and across paradigms in what
might be termed paradigm mixes and paradigm blends.
The quantitative paradigm is derived from positivism that holds that the world is made up
of observable, measurable, and quantifiable facts (Glense and Peshkin, 1992). The essential aim is
to explain causal relationships and to have generalised knowledge. In contrast, the qualitative
paradigm is rooted in the humanities, which focuses on holistic information and interpretative
approaches (Husen, 1994). Both the quantitative and qualitative paradigms consist of sets of
different research methods that are founded on separate philosophical assumptions about social
life (Hammersley, 1995). There are important epistemological differences between the two
approaches in that they use diverse principles and knowledge about the social world (Bryman,
1988). The essential aim in quantitative research is to produce knowledge that is as value-free as
259
possible and that can be generalised (Borg and Gall 1989). Quantitative researchers try to find out
particular explanations and predictions to generalise from a sample to a population (Glense and
Peshkin, 1992). In contrast to this approach, qualitative research adopts a phenomenological
perspective in viewing reality as a social phenomenon that develops through an individual or co-
operative social definition of the situation. The aim of qualitative research is to understand and
view reality from the actors perspective (Firestone, 1987). Qualitative research gives more
emphasis to the study of subjects holistically (or participants as they are now more commonly
referred to, implying more interaction between perspectives and experiences) rather than to the
generalisation of the results. Therefore, the main purpose is to compile a body of knowledge that
is unique to an individual case or specific context (Borg and Gall, 1989). There is a range of
qualitative methods for data elicitation, but generally, the two most frequently used methods are
observations and interviews. In qualitative research, the research design focuses on participation
and interaction with individuals or groups of people (Glesne and Peshkin, 1992). If examination
of social interaction and first-hand information about social processes is required, participant
observation and interviewing are essential and appropriate research methods. In this kind of
research, researchers try to reduce the distance between themselves and the case (Bryman, 1992).
Therefore, qualitative research provides rich, context-bounded and naturalistic information
(Bullock, Little and Millham, 1992). If it has an ethnographic emphasis, it aims to understand and
interpret participants interpretations of their social worlds. For some, qualitative research may be
judged by its capacity to emancipate, empower or otherwise make free these participants and its
validity can be supported by techniques of member validation (Seale, 1999). However, this
assumes particular kinds of research and participants, and there is no general consensus on the
political desirability of such an orientation because there is no general consensus on either politics
or research styles.
In contrast to such qualitative research processes, quantitative research refers to the
implementation of measurements and in gathering and analysing them assigns numerical values to
the research subjects or to their attitude, opinions, and other attributes (Bullock et al, 1992). It
involves several data gathering and analytic methods such as survey techniques, experiments,
structured observation, content analysis and parametric and non-parametric statistical analysis. It
emphasises causality, measurement and generalisability (Bryman, 1992). Both quantitative and
qualitative paradigms rely on different epistemologies. Furthermore, both of them provide
different kinds of data that also each have advantages and disadvantages. Because of this, most
research depends on one type of data and leads to selective and limited information. Each type of
method serves research purposes with different data that are used to explain particular sides of a
subject or reality. This is also the main limitation of relying on a single type of data. Therefore, a
combination of both qualitative and quantitative methods in the same research project is desirable.
The use of paradigm mixes and blends may be held to consolidate methodological strengths and
offset limitations.
260
Denzin (1994) reiterates that the social world is socially constructed and its meaning, to
the observers and those observed, is constantly changing (p. 6462). Therefore, to find valid
information for all perspectives seems difficult. On the other hand, to get some valid perspectives
on reality with a single method can be questionable, certainly from the counter-viewpoints of
contrasting research perspectives. In addition to this, Patton (1990) mentions that fieldwork is not
a single method or technique (p. 244). It comprises a variety of methods such as surveys,
experiments, structured observation, ethnography and interviews. Each of them is used to gather
different kinds of data. Therefore, using multiple sources can give a wide range of information,
unlike the use of a single method. Patton (1990) also points out that each method has particular
strengths and weaknesses. As single methods may not provide enough information for the
purposes of validity, multiple sources or combinations of data types may advance validity and
strengthen conclusions. Denzin (1989) argues that the combination or integration of qualitative
and quantitative research methods in the same study can be a more acceptable approach than using
a single one. He refers to this integrative approach as triangulation in research. Triangulation
uses multiple sources of data collection and provides a deeper understanding of social phenomena.
Therefore, triangulation can be a solid alternative to single-method validation, provided that data
sources are comparable or that data collection procedures focus on some common or comparable
area (Denzin and Lincoln, 1994). In addition to this, it extends the quality of data and
trustworthiness of results (Robson, 1993). Denzin (1989) expands the general triangulation
definition and claims that triangulation is the use of multiple methods in the study of the same
object (p. 236), and further suggests that there are four basic types of triangulation: 1. Data
triangulation, 2. Investigator triangulation, 2. Theory triangulation, and 4. Methodological
triangulation. In addition to this classification, Patton (1990) suggests analyst triangulation, that
is using multiple analysts to review findings. Combining methods in this way invokes other
difficulties, of course, such as greater commitment of researchers (and participants) time and
effort, and the need to master more methods and to ensure methodological compatibility of focus
and object.
Denzin (1989) points out that by selecting dissimilar settings in a systematic fashion,
investigators can discover what their concepts have in common (p. 237). Patton (1990) also
indicates that triangulating data sources includes comparing and cross-checking the consistency
of information derived at different times and by different means (p. 467).

Common Methods of Research: Naturally occurring data
Anthropologists and sociolinguists have repeatedly called for collection of naturally
occurring data (see for example Wolfson 1983, 1986; Roberts, Davies and Jupp 1992). The
advantages of gathering natural data as summarised by Cohen (1996:391-2), are:
1. The data are spontaneous.
2. The data reflect what the speakers say rather than what they think they would say.
261
3. The speakers are reacting to a natural situation rather than to a contrived and possibly unfamiliar
situation.
4. The communicative event has real-world consequences.
5. The event may be a source of rich pragmatic structures.
And the disadvantages (idem.) are:
1. The speech act being studied may not occur naturally very often.
2. Proficiency and gender may be difficult to control.
3. Collecting and analysing the data are time-consuming.
4. The data may not yield enough or any examples of target items.
5. The use of recording equipment may be intrusive.
6. The use of note taking as a complement to or in lieu of taping relies on memory. (Bardovi-Harlig
and Hartford, 1993 cited in Cohen, 1996)

In the same line, Wolfson (1983) stresses using ethnographic fieldwork as the only
reliable method of data collection about the way speech acts function in interaction. Ethnographic
fieldwork (or observation) can be done using notes, recording on audio or videotapes, etc.
According to her, ethnography is the most satisfactory source of reliable and varied data for
research in pragmatics. There are practical difficulties in obtaining naturally-occurring recordings
as data, and Wolfson acknowledges that this kind of research involves spending a considerable
amount of time for collecting data and transcribing them. Stubbs (1983) states that a minimum of
20 hours is normally required to transcribe a 50 to 60 minute conversation. This is for a normal
transcription without any special signs, markings or tables. The depth of the detail of the
transcription, clearly, is based on the kind of investigation and the information one hopes to
recover from the transcript. Nonetheless, where there are time restrictions, the time spent on
transcription may limit the amount of data that can be transcribed.
Olshtain and Cohen (1983) also point out that sometimes the object of the research does
not occur when recording: some theoretically important speech acts may be infrequently used in
natural contexts. It may take several hours or sessions for recording sufficient instances to be
collected.
In observation of naturally occurring data, there is also the problem of the observers
paradox, as referred to by Labov (1972). He believes that people may not speak naturally when
they notice that they are being observed or recorded. Therefore, the presence of a tape recorder or
an observer has an undesirable effect on the quality and the quantity of data obtained. In such
situations, the subjects perceive the context of observation as formal and adjust their speech style
accordingly, thus presenting the observer with some invalid data. One way of avoiding this
problem is to shift informants attention from form to content by introducing topics of high
involvement (Labov, 1966). Such topics override the formal constraints in the production of
naturally occurring interaction. Another way of side-stepping this problem has been secretly
recording the interaction and obtaining the subjects consent to use the data afterwards. This
262
approach poses major ethical issues, though, and should be avoided. Milroy (1987) argues that
people have the right to refuse to have their voice recorded, as they might express reservations
about being photographed. Besides ethical issues involved in secret recording, there are some
practical constraints as well. A researcher, who apparently spies in gleaning data, finds the job
quite arduous and may eventually be deprived of access. Secondly, a well-hidden microphone is
not likely to receive sound as well as it otherwise would, which may result in data of doubtful
quality and value (Labov, 1981). The same is true for a video camera that may not be able to
capture the full scope of the situation. The focus and scope of a video recording is very limited
and depends on the camera-holders perception (Saville-Troike, 1996).
So if a study uses naturally occurring data through ethnographic field notes, this
overcomes some of the practical and ethical difficulties, but leads to a reliance on memory, note-
taking skills, and selectivity, besides the researchers ability to note or recall not only words but
also contexts. The researcher also needs to see such fieldwork as learnable skill, to treat it
reflexively as a non-obvious method, and to recognise that there are still lingering difficulties
regarding obtaining participants consent after noting their words (Emerson et al., 1995). However,
such difficulties are recognised and writing field notes is, after all, part of a longstanding tradition
in anthropology and sociology, if much less so in applied linguistics. As with all research methods,
it has its inherent constraints and problematic aspects.

Participant-Observation
The main difference between participant-observation and other types of observation is
that in participant-observation the researcher is a native member of the society or has been
immersed there for a substantial period (Saville-Troike, 1996). This method is a common method
of collecting data. According to Saville-Troike (1996:119), the key to successful participant-
observation is freeing oneself as much as humanly possible from the filter of ones own cultural
experience. This requires cultural relativism, knowledge about possible cultural differences, and
sensitivity and objectivity in perceiving others.
Saville-Troike (1996) further elaborates that the necessary step toward participant-
observation is for the researcher to be involved in group activities frequently over a period of time
and take up a role in the group. In this way, the researcher acquires the necessary background
knowledge and develops relationships of trust with members of that community. He or she then
can benefit from testing hypotheses about rules of communication by breaking them and
observing or eliciting reactions (Saville-Troike, 1996).

Interviews
Semi-structured interviewing lies between structured and unstructured interviews. It
refers to a situation in which the interviewer sets up a general structure for the interview, but
details this structure by asking extra, spontaneous questions depending on the interaction process
263
during the interview. The interviewer extends questions using prompts, probes, and follow up
questions to get the interviewees to clarify and expand their answers (Drever, 1995). This is more
flexible than structured interviewing and provides the opportunity to probe and expand the
interviewees responses. The interviewer remains almost free to build up on interaction within a
particular subject area, to ask questions spontaneously and to establish conversation relating to
particular predetermined issues (Patton, 1987).
Cohen and Manion (1994) define such interviews as a two-person conversation initiated
by the interviewer for the specific purpose of obtaining research-relevant information, and focused
by him on content specified by research objectives of systematic description, prediction, or
explanation (p. 271).
The aim of interviews is neither to test a hypothesis nor to evaluate a program. The broad
purpose of interviewing is an interest in understanding the experience of other people and the
meaning they make of that experience (Seidman, 1991:3). In a semi-structured interview, the
main focus is the other individuals stories and thoughts (Seidman, 1991). By interviewing, the
interviewer gains access into another individuals world in order to perceive, learn and understand
the meaning of the topic from the other persons perspective (Patton 1987). The interview is
therefore a very helpful method, especially where it is used to learn something that cannot be
directly observed. Kvale (1996) emphasises the importance of obtaining the interviewees
meaning of the underlying general themes not only through words; the interpretation of the themes
of the interviewees voice intonation, eye contact, vocalisation, and facial expressions that reflect
the emotional circumstance is also very important.
Broadly speaking, semi-structured interviews involve activities such as asking open-
ended questions, listening, and recording the answers given by the interviewee. In addition to
being active in the talk, the interviewer has to be able to be sensitive, attentive, and understanding
(Patton, 1987). Because a semi-structured interview has a largely conversational style, there is the
possible disadvantage that what respondents say is, in part, influenced by preceding contributions
to the jointly constructed talk by the interviewer. Further, there is a tension between naturalness
and consistency which needs to be recognised in conversation-type interviews: the more natural
the conversation is, the less consistent the contribution of the interviewer is likely to be across
interviews; the more consistent the attempts of the interviewer to control the contributions, the less
natural the interview will seem to the interviewee. Consistency is, of course, generally valued in
research but as a principle of interviewing it needs to be reconciled with naturalness. Naturalness
is quite crucial to interviews designed to promote respondents expression of experience or elicit
how they understand key concepts. Without such naturalness, any data elicited may be less
authentically what respondents believe or know and more of a construct of the interview situation.
Yet it may be recognised that most research interviews are, in fact, constructs in this way: given a
different context of conversation many respondents would give different replies. Like other
264
disadvantages of other research methods, the limitations of this general construct can only be
recognised and, perhaps, compensated by the use of complementary methods.
According to Keats (1994:3004) there are several factors that affect the consistency of the
research interview: the structure of the interview schedule; the development of listening skills;
probing and empathy; the control of bias; training of interviewers; coding, scoring; and analysis of
the responses.
An obvious further disadvantage of interviews is a possible lack of validity in responses.
It is likely that some respondents may evade telling the truth or might provide biased information.
They may do this in order to present themselves or their responses to conform to what they
perceive as the purpose of the interview or what the interviewer expects. Because of this,
interviews should be used as complementary methods alongside other methods. Another important
concern is the ethical issues in analysing interview data. It is the researcher who reduces data,
refines them, codes them and analyses them according to those codes. So what is considered
important and valuable to collect, analyse, and report relies mainly on the researchers point of
view rather than that of participants (Merriam, 1988). Thus, a report may fail to represent the
actual information which might be seen if investigated by other researchers or by the participants
themselves.

Introspection
Introspection is a way of collecting data only about ones own speech community
(Saville-Troike, 1996). As such it is an important skill to develop not only for data collection itself,
but also for finding answers about language and culture from the perspective of both the
researcher and the subjects. The researcher has to differentiate between beliefs, values, and
behaviours. This exercise in itself will provide information and insights on the group and on the
individuals (Ibid. p.118).
Saville-Troike proposes two steps to follow for developing introspective skills: 1. Asking
individuals to formulate very specific answers from their own experience to various questions
about communication [and 2. Recognising] the significance of differences between answers which
reflect cultural ideals or norms and the real, or what actually occurs (Ibid. p. 118). The
distinction between ideal and real should not be interpreted negatively or in true and false senses;
these are just two levels of behaviour. The important thing about them is that the researcher should
refer to them and treat them accordingly. Thus, even when researchers are sure they know about
patterns of language use in their own speech community, it is important to check hypotheses
developed on the basis of their own perceptions with the perceptions of others, and against
objective data collected in systematic observation (Ibid. p. 119).
Wolfson, however, (1989) argues against the use of native speakers perception in the
data. According to her, native speakers have the competence to use their own language
appropriately and also to pass judgement on other speakers usage such as children or learners that
265
may break the rules. However, even though speakers are able to explain whether an utterance or
usage is acceptable, they cannot predict reliably the whole range of possible expressions that
speakers will use (Wolfson, 1989). Wolfson (1983) argues that this may be because the knowledge
of the sociolinguistic rules of their language is below the level of conscious awareness (p. 83).
However, this is less likely to be the case when native-speakers who are themselves linguists
introspect.
Brouwer, Gerritsen and DeHaans (1979) study is among the first to point towards the
inadequacy of native speakers intuitions. They looked into the ways in which mens and womens
speeches differ in Dutch. The researchers hypothesis was that mens and womens speech differ
with regard to the following: the number of words used in dealing with a task, the use of
diminutives, polite forms and expressions indicating insecurity such as repetition, hesitations, self-
corrections. The list of these differences was based on the authors own intuitions and the existing
sociolinguistic literature on sex differences in speech. The authors secretly recorded 309 women
and 278 men buying a train ticket in the central station in Amsterdam. Contrary to their
expectations, the data revealed that differences in speech had to do with the gender of the
addressee rather than with that of the speaker. It is important to point out that even such
supposedly transparent aspects of language use as differences in mens and womens speech have
been found, once empirical analysis is done, to be very different from what native speakers would
have expected (Wolfson 1989:41).
According to Blom and Gumperzs (1972) informants in a study of code-switching in a
small town in Norway, the strongly favoured use of the native dialect Ranamal signified a positive
sense of local pride and represented an acceptance of the communitys common culture and
heritage. However, the recording of a friendly gathering showed that speakers code-switched
considerably between the local dialect and Bokmal, the standard variety, mainly for stylistic
reasons, for example when a speaker wanted to validate her status as an intellectual. This study
demonstrates how native speakers actual linguistic usage can differ widely from their self-reports
and how native speakers may mistakenly believe their ideal to be their actual usage. It is important
to point out, however, that even faulty perceptions, are unquestionably valuable and important in
many ways (Wolfson, 1989:44) because arguably what respondents believe about language is an
important element in the total linguistic picture of what people do with language. However, the
pitfalls inherent in such perceptions and beliefs are recognised and remedied. Intuitions enable
speakers of a language to recognise inaccuracies and inappropriate speech behaviour as well as
give them insight into the meanings behind various means of expression (Wolfson, 1989).

Written Data: Questionnaires
There are different forms of questionnaires. Some are used in quantitative studies to
gauge learner variables, both cognitive and affective such as Strategy Inventory for Language
266
Learning (Oxford, 1986), Attitude Motivation Test Battery (Gardner, 1960, 1968), and Self-
efficacy Questionnaire (Sadighi et al, 2003).
Other forms used in qualitative studies are meant to elicit language data. These are also known as
a discourse completion test or task. Levenston and Blum (1978) first developed this procedure to
study lexical simplification. Blum-Kulka (1982) then adapted this procedure to investigate speech
act realisation. Since then DCQs have been widely in use mainly in testing foreign language
learners performance (see, for example, Blum-Kulka and Olshtain 1986, Beebe 1985, Takahashi
and Beebe 1987, Eslamirasekh 1993, Maeshiba et al. 1996). In DCQs subjects can provide written
or spoken responses to prompts. A more fashionable method is videotaping subjects while they act
out a response in a role-play. Based on observation of naturally occurring data, researchers might
set up situations in their DCQs. This would reduce one aspect of this artificiality. The written
procedure usually involves presenting a prompt and space for a response. For example:

You promised to return a textbook to your classmate within a day or two, after photocopying a chapter.
You kept it for almost 2 weeks.
Classmate: Im really upset about the book because I needed it to prepare for last weeks class.
You:
(Cohen, 1996:390)

This kind of DCQ is open-ended but other controlled formats may also be followed, such as
presenting the subjects with multiple responses and asking them to choose the best possible
answer. If the respondents cannot decide, they then may provide their own answers.

The most obvious advantage of using DCQs over methods such as ethnographic observation is
that a sizeable corpus of data can be collected in a relatively short time since a DCQ can be
administered to large groups simultaneously (Cohen, 1996). Beebe and Cummings (1985)
conclude that DCQs are highly effective means of :

1. gathering a large amount of data quickly;
2. creating an initial classification of semantic formulas and strategies that will occur in natural
speech;
3. studying the stereotypical, perceived requirements for socially appropriate (though not always
polite) response;
4. gaining insights into social and psychological factors that are likely to affect speech and
performance; and
5. ascertaining the canonical shape of refusals, apologies, partings, etc., in the minds of the speakers
of that language. (p. 13)

267
However, data elicited using DCQs do not satisfactorily represent the following:

1. the actual wording used in real interaction;
2. the range of formulas and strategies used (some, like avoidance, tend to be left out);
3. the length of response or the number of turns it takes to fulfil the function;
4. the depth of emotion that in turn qualitatively affects the tone, content, and form of linguistic
performance;
5. the number of repetitions and elaborations that occur; or
6. the actual rate of occurrence of a speech act-e.g., whether or not someone would naturalistically
refuse at all in a given situation. (Beebe and Cummings 1985:14)

Among other advantages of DCQs is that it is possible to compare a large number of responses. As
many responses are collected through the same prompt, significant factors like age, sex, status,
and role-relationships between the participants and the DCQ situation they find themselves in are
held constant. However, it is very difficult to control such factors in ethnographic methods.
A much reported disadvantage of DCQs is the relative unreliability of results (Beebe and
Cummings, 1985). Beebe and Cummings discovered differences between written and oral role-
plays. Among other differences, they found that written role-plays bias the response towards less
negotiation, less hedging, less repetition, less elaboration, less variety, and ultimately less talk (p.
3). This finding very clearly indicates that researchers should be wary of assuming that a DCQ
will yield similar or comparable results whether delivered in the written or the oral mode or
whether using open or multiple-choice formats (Rose, 1994; Hinkel, 1997).
A number of studies have shown that the DCQ method is an effective way of detecting
differences in the performance of learners in a foreign language and native speakers responses.
For instance, Faerch and Kasper (1989) and Rintell and Mitchell (1989) mention that second
language learners are prone to more extensive responses than native speakers. Faerch and Kasper
argue that such responses should be considered in a positive light rather than as a proof of
inadequacy because they indicate that learners are trying to avoid misinterpretation, giving
prominence to the maxim of manner (be clear, be brief, be orderly) over that of quantity. Holmes
(1992) states that lack of experience in analysing the socially relevant variables in another culture
means learners inevitably draw on formulaic responses, as well as on their mother tongue (p.123).
She adds that it is likely that the differences between role-plays and natural conversation are
greater in native speakers than foreign learners because of this lack of experience.
Aside from DCQs being a very effective means of collecting great amounts of data in a
short time, they also provide an excellent complement to ethnographic studies by corroborating
results (Wolfson, Marmor and Jones, 1989). Detailing the advantages of oral DCQs, Rintell and
Mitchell (1989) state that the subjects have the opportunity to say what and as much as they
would like to say, and their spoken language is thought to be a good indication of their natural
way of speaking (p. 251). They also maintain that although their data appear in two modalities,
268
written dialogue and oral role-play, they do not really reflect the differences between spoken and
written language. This is because the task in both of them was to respond with what the person in
each situation would say, and therefore both modes presumably elicit oral responses (although it is
not common to write oral responses to written prompts outside DCQ contexts). A reservation
expressed elsewhere is that the extent to which the subjects responses correspond to what they
would actually say in real life is a factor that has to be considered in the analysis. It is doubtful if
this would be a close correspondence since DCQ contexts are highly reduced versions of any
authentic situation; the latter must therefore contain more variables which potentially affect speech
act realisations. Besides, subjects, especially foreign language learners, may avoid words they
would choose in real life, simply because they cannot spell them, and include words they know
how to spell, but would not consider appropriate. Thus the written format may skew self-reports of
purportive oral responses. The disadvantages of DCQs are therefore that they seem somewhat
artificial or experimental tasks and the precise relation of DCQ responses to those apparently
similar comments that might be made in real situations remains unknown and difficult to ascertain.
Successful DCQ research probably includes giving respondents familiarisation tasks; if it is in a
foreign language it needs to make allowances for the time needed and for learners limitations of
producing speech acts in such tasks.

Conclusion
It was earlier pointed out that the adoption of an appropriate method or methods for a particular
research project is of paramount importance. Use of multiple methods (methodology triangulation)
has long been a fashion in social sciences to reduce problems attached to single methods. It is,
however, only recently that scholars in language related studies have opted for combining
different data gathering methods. Kasper and Rose (1999) review 28 cross-sectional and
longitudinal pragmatic studies done up to the year 1998. Among these studies only 2 employed
more than one method. It is important to note, however, that the significant contributions of
different elicitation techniques to different research issues need vigorous empirical scrutiny
(Kasper and Dahl, 1991). As in all data-based research, a good method is one that is able to shed
light on the question(s) under study (Ibid. p. 245).

Much has been said about the adoption of appropriate methodology or methodologies.
Commenting on methods used for collecting data on language use, Kasper and Dahl (1991)
express the view that it is rather hard to collect sufficient instances of data solely through
observation of authentic conversation. Conversely, tightly controlled data elicitation techniques
[namely questionnaires] might well preclude access to precisely the kinds of conversational and
interpersonal phenomena (Ibid. p. 245). As reflected earlier, a remedy to this proves to be
using multiple methods in collecting data.

269
References
Bardovi-Harlig, K. & Hartford, B. S. (1993). Redefining the DCT: comparing open
questionnaires and dialogue completion tasks. Pragmatics and language
learning, vol. 4, pp.143-165.
Beebe, L. M. (1985). Input: Choosing the right stuff. In Gass and Madden (eds.) (1985). Input
in Second Language acquisition. Rowley: Newbury House.
Beebe, L. M. & Cummings, M. (1985). Speech act performance: A function of the data collection
procedure. Paper presented at the Sixth Anuual TESOL and sociolinguistic
Collouim at the International TESOL Convention, New York.
Blom, J. P. & Gumperz, J. J. (1972). Social meaning in linguistic structures: code-switching in
Norway. In J. J. Gumperz & D. Hymes (eds.) (1972). Directions in
Sociolinguistics. New York: Rinehart & Winston.
Blum-Kulka, S. (1982). Learning to say what you mean in a second language: a study of the
speech act performance of learners of Hebrew as a second language. Applied
Linguistics, vol. 3, pp.29-59.
Blum-Kulka, S. (1983). Interpreting and performing speech acts in a second language: a cross-
cultural study of Hebrew and English. In Wolfson, N. & Judd, E. (1983).
Sociolinguistics and language acquisition. Rowley: Newbury House.
Blum-Kulka, S., (1992). The metapragmatics of politeness in Israeli society. In Watts, J., Ide, S.
& Ehlich, K. (eds.) (1992). Politeness in language, Studies in its history,
theory, and practice, Studies and monographs 59, Mouton De Gruyter.
Blum-Kulka, S., House, J. & Kasper, G. (eds.), (1989). Cross- cultural pragmatics: Requests
and Apologies. Norwood, New Jersey: Ablex Publishing Corporation.
Blum-Kulka, S. & Olshtain, E. (1986). Too many words: Length of utterance and pragmatic
failure. Studies in second language acquisition, vol. 8, pp.47-61.
Borg, R. W. & Gall, D. M. (1989). Educational research: an introduction, New York: Longman.
Brouwer, D., Gerriten, M. & DeHaan, D. (1979). Speech differences between men and women:
on the wrong track. Language in society, vol. 8(1), pp.33-50.
Bryman, A. (1988). Quantity and quality in social research. London, Routledge.
Bryman, A. (1992). Quantitative and qualitative research: further reflections on their integration.
In Brennen (ed.). Mixing methods: qualitative and quantitative research,
Avebury, Aldershot, pp.57-78.
Bryman, M. (1989). Cultural studies in foreign language education. Clevedon: multilingual
Matters.
Bullock, R., Little, M., & Millham, S. (1992). The relationshops between quantitative and
qualitative approaches in social policy research. In Brannen J. (ed.), Mixing
methods: qualitative and quantitative research, Avebury,
270
Cohen, A. D. (1996) Speech acts, in McKay, S. L. & Hornberger, N. H. (eds.). Sociolinguistics
and language teaching, Cambridge: Cambridge University Press, pp.383-420.
Cohen, A. D. & Olshtain, E. (1981). Developing a measure of sociolinguistic competence: the
case of apology. Language Learning, vol. 31(1), pp.113-134.
Cohen, L.& Manion, L., (1994), Research methods in education (4
th
ed.). London: Routledge.
Denzin, K. L. (1989) The research act: a theoretical introduction to sociological methods.
New Jersey: Prentice Hall Inc.
Denzin, K. L. (1994). Triangulation in educational research. In Husen, T. & Postlethwaite, N.
(ed.). The international encyclopaedia of education. London: Pergamon,
pp.6461-6466.
Denzin, K. L. 7 Lincoln, Y. S. (1994). Introduction: entering the field of qualitative research. In
Denzin, K. L. & Lincoln, Y. S. (ed.). Handbook of qualitative research.
Thousand Oaks: Sage Publications, pp.1-17.
Drever, E., (1995). Using semi-structured interviews in small-scale research. Glasgow: SCRE
Publications.
Eslamirasekh, Z. (1993). A cross-cultural comparison of requestive speech act realisation patterns
in Persian and American English. In Pragmatics and language learning:
Monograph series, vol. 4, Urbana, University of Illinois, pp.85-103.
Emerson, R. M., Fretz, R. I., & Shaw, L. L. (1995). Writing ethnographic fieldnotes. Chicago:
The university of Chicago Press.
Eslamirasekh, Z. (1993). A cross-cultural comparison of requestive speech act realisation patterns
in Persian and American English, in Pragmatics and language learning:
Monograph series, vol. 4, Urbana, University of Illinois, pp.85-103.
Faerch, C. & Kasper, G. (1989). Internal and external modification in Interlanguage request
realisation, in Blum-Kulka, S., House, J. & Kasper, G. (eds.) (1989) Cross-
cultural pragmatics: Requests and Apologies. Norwood, New Jersey: Ablex
Publishing Corporation, pp.221-247.
Firestone, W. A. (1987) Meaning in method: the rhetoric of quantitative and qualitative research.
Educational Researcher, vol. 16, no. 7, pp.16-21.
Glesne, C. & Peshkin, A. (1992) Becoming qualitative researchers: an introduction. London:
Longman.
Hammersly, M. (1992) Deconstructing the qualitative-quantitative divide. In Brannen, J. (ed.),
Mixing methods: qualitative and qualitative research. Avebury: Aldershot,
pp.39-55.
Hammersly, M. (1995). Opening up the qualitative-quantitative divide. Education Section
Review, vol. 19, no. 1, pp.2-5.
Holmes, J. (1992). An introduction to sociolinguistics. London: Longman.
271
Husen, T. (1994) Research paradigms in education. In Husen, T. & Postlethwaite, N. (eds.). The
international encyclopaedia of education. London: Pergamon, pp.5051-5056.
Husen, T. & Postlethwaite, N. (eds.) (1994). The international encyclopaedia of education.
London. Pergamon.
Kasper, G. & Dahl, M. (1991). Research methods in interlanguage pragmatics. Studies in
Second Language Acquisition, vol. 13, no. 2, pp.215-247.
Kasper, G. & Rose, K. R. (1999). Pragmatics and SLA. ARAL, vol. 19, pp.81-104.
Keats, D. M. (1994). Interviewing for selection and clinical research. In Husen, T. &
Postlethwaite (eds.). The international encyclopaedia of education, London:
Pergamon, pp.3003-3007.
Kvale, S. (1996). Interviews: an introduction to qualitative research interviewing. Thousand
Oaks: Sage Publications.
Labov, W., (1966), The social stratification for English in New York city, Washington DC:
Centre for applied linguistics.
Labov, W. (1972). Some principles of linguistic methodology. Language in society, vol. 1,
pp.97-120.
Labov, W. (1981). Field methods used by the project on linguistic change and variation.
Sociolinguistic working paper 81. Austin, Texas: South western educational
development laboratory.
Levenston, E. & Blum, S. (1987) Discourse completion as a technique for studying lexical
features of interlanguage. Working Papers in Bilingualism, vol. 15, pp.13-21.
Maeshiba, N., Yoshinaga, N., Kasper, G., & Ross, S. (1996). Transfer and proficiency in
interlanguage apologizing. In Gass, S. M. & Neu, J. (eds.), Speech acts across
cultures, Berlin: Mouton De Gruyter, pp.155-187.
Merriam, S. M. (1988). Case study research in education. San Francisco: Jossey-Bass.
Milroy, L. (1987). Observing and analysing natural language. Oxford: Blackwell.
Olshtain, E. & Cohen, A. (1983). Apology, a speech act set. in Wolfson, N. & Judd, E., ( eds.),
(1983). Sociolinguistics and language acquisition. Rowley: Newbury House,
pp.18-35.
Patton, M. (1990). Qualitative evaluation and research methods. Newbury Park, Calif.,
London: Sage.
Rintell, E. & Mitchell, C. J. (1989). Studying requests and apologies: an inquiry into method In
Blum-Kulka, S., House, J., & Kasper, G. (eds.). Cross-cultural pragmatics.
Norwood NJ: Ablex, pp.248-272.
Robinson, M. A. (1992). Introspective methodology in interlanguage pragmatics research. In
Kasper, G. (ed.). Pragmatics of Japanese as native and target language.
Honolulu: HI University of Howaii at manoa, pp.27-82.
272
Rose, K. R. (1994) On the validity of discourse completion tests in non-western contexts.
Applied Linguistics, vol. 15, No. 1, pp.1-14.
Saville-Troike, M. (1997). The ethnographic analysis of communicative events. In Coupland, N.
& Jaworski, A. (eds.) (1997). Sociolinguistics: a reader and a coursebook.
Basingstoke: Macmillan Press.
Seale, C. (1999). The quality of qualitative research, London: Sage publications.
Seidman, I. E. (1991). Interviewing as qualitative research: a guide for researchers in
education and the social sciences. New York: Teachers College Press.
Stubbs, M. (1983). Discourse analysis. Oxford: Basil Blackwell.
Takahashi, T. & Beebe, L., (1987). Development of pragmatic competence by Japanese learners
of English. JALT Journal, vol. 8, pp.131-155.
Wolfson, N. (1983). Rules of Speaking. In Richards, J. & Smith, J. (eds.) (1983). Language
and Communication. London: Longman, pp.61-87.
Wolfson, N. (1989). Perspectives: Sociolinguistics and TESOL. Rowley: Newbury House.
Wolfson, N., Marmor, T., & Jones (1989). Problems in the comparison of Speech acts Across
Cultures. In Blum-Kulka, S., House, J. & Kasper, G. (eds.), (1989). Cross-
cultural pragmatics: Requests and Apologies. Norwood, New Jersey: Ablex
Publishing Corporation, pp.174-196.