Qualitative research has been challenged with persistent criticisms for its unreliability, or that it is deficient relative to more quantitative approaches to social research. For research to be considered of value, it has too strongly fulfil the standards of research: validity, reliability, generalisibilty and objectivity. Concepts such as reliability and validity are generally associated with measuring in quantitative research approaches. Yet, many different opinions on reliability and validity are existent. From the 1980’s, qualitative research had been rejected on the grounds of reliability and validity, and has been criticised that concepts of validity are incompatible and therefore should be dismissed.
On the other hand, it has been argued that attempts should be made to ensure validity in qualitative research in order for results to become more credible (Chase 2001; Denzil 2000).
This paper will discuss the basis of these criticisms, and the types of standards against qualitative research are measured. Prior to that, the definitions of qualitative and quantitative research will be examined.
Thereafter, the uses of reliability and validity in quantitative research are discussed in order to provide a foundation of the meanings of these terms and to provide a discussion on how the qualitative research paradigm challenge these standards.
Traditionally, quantitative research is usually portrayed as demonstrating many hallmarks of natural science approach and is based on a positivism perspective and has an objectivist conception of social reality. Researchers utilise quantitative measures and experimental methods in order to test hypotheses and to determine the relationship between variables – they strive to obtain correlation, causality and relationships (Babbie 1995: Bryman 2001).
The quantitative research paradigm involves the accumulation of facts and causes of behaviour, emphasises quantification in the collection and analysis of data, information is gathered numerically and is quantified and summarised and the final results is processed statistically (Charles 1995: Golafshani 2003). Positivists portray the world as a formation of observable and measurable facts. Whereas qualitative research, is based on the interpretivist perspectives believe that reality is socially constructed, complex and constantly changing. Because the two paradigms hold different views on the nature of the world, therefore, different logic of research procedures; methods and instruments are required to obtain the desired results (Blaxter).
The process of measurement is vital in quantitative research. The role of ‘measurement’ is an operation undertaken by the observer on the physical world (Golafshani 2003). Measurement allows the researcher to delineate fine differences between people in terms of characteristics in question. It also provides a consistent device or yardstick for making distinctions and provides the basis for more precise estimates of the degree of relationships between concepts (Bryman 2001). Steven (1946) describes measurement as numbers being assigned to objects or events in accordance to results. Thus, measurement is associated with numbers, objective hard data, and can be statistically quantified (Babbie 1995: Golafshani 2003).
According to positivism, ‘reality’ can be captured via the use of research instruments i.e experiments and questionnaires (Blaxter 2006). The researcher will aim to break down phenomena by constructing an instrument that will allow the researcher to use measurable and common categories that can be applied to variety of subjects or situations. This is attained by the use of standardisation. Facts, attitudes and behaviour is measured through questions followed by a restricted number of predetermined response categories or numbers (Golafshani 2003: May 2001).
However, a crucial question is whether the devised instrument is measuring what it is intended to measure. Thus, when creating a test, validity is crucial in addition to ensuring that the results are reliable and repeatable. (Bryman 2001: Golafshani 2003).
Reliability is one of the primary standards of research that quantitative research are measured against. Reliability refers to the consistency of a concept. This occurs when repeated measures whether its done by someone else and at a different place, using the same measures produce identical or very similar results i.e. results can be directly comparable. As apposed to erratic, unstable or inconsistent results (Babbie 2004).
Kirk and Miller (1986) state that the following are three important factors involved when considering the reliability of a measure: the degree of consistency of measures (internal reliability), secondly the stability over time and lastly the similarity within a certain time period (Bryman 2001: Golafshani 2003).
Social researchers have developed several techniques for cross-checking the reliability of the measures they devise. The most obvious way of testing for the stability of a measure is the test-retest method. This involves administrating a test/measure i.e questionnaire to a group and then re-administrating it to the same group on another occasion. In order to insure stability, the results should have little variation over time (Bryman 2001). The higher the level of stability, denotes a high level of reliability therefore results are repeatable. However, this approach poses a few problems in evaluating reliability. Respondents answers may be influenced as this method may provoke the respondent to a subject matter. Secondly, events may intervene between the first and second occasions of administration of the tests resulting in a change in scores due to characteristics of the respondent. This results in errors of measurement and hence influences the degree of consistency (Bryman 2001: Golafshani 2003).
The researcher may be able to devise an instrument which is reliable if repeatability and internal consistency are apparent. However, the instrument itself may lack in validity (Golafshani 2003).
A further and in many ways the most important criteria of research is validity. Traditionally validity derives from the positivist perspective. Validity is a term describing a measure that accurately reflects the concept is in intended to measure.
Some writers advocate that the researcher should estimate the construct validity of a measure. This involves the deduction of hypotheses from theory that is applicable to the concept (Bryman 2001). The decision of the collection of data and the research strategy to be undertaken is determined by the construct which is the initial concept, notion, question or hypothesis. Wainer and Braun (1998) stress that in order to validate the investigation, quantitative researchers actively influence the interaction between the construct and data generally via the use of application of a test or other process. Thus, validity of a test will be considerably reduced by the involvement of the researcher on the research process.
The meanings of reliability and validity in quantitative research as described above are considered to be inadequate by qualitative researchers as these terms described in quantitative terms cannot be applied to the qualitative research paradigm. Qualitative researchers challenge these terms by focusing on credibility, transferability and precision for evaluating the results of qualitative research as apposed to replication.
Qualitative research is concerned with words as apposed to statistical procedures. Alternatively, qualitative research seeks to produce results from some phenomenon, real-world settings where the researchers endeavour is to avoid influencing the phenomena of interest and thus allowing it to progress naturally (Patton 2001: Golafshani 2003).
Qualitative research has rejected the practices and norms of the natural scientific model and particular positivism (Bryman 2001. Thus, each party supports apposite sides of the philosophical nature of the paradigm resulting in different type of knowledge produced between qualitative analysis and quantitative inquiry. It has been stated that during the research process, qualitative researchers try to avoid associating themselves to the highest extent yet qualitative researchers have started to accept their involvement and position in the research. This idea is supported by many writers who state that as the real world is subject to change, the immersion and researchers involvement plays a critical part as the presence of the researcher is vital in order to record changes. However, credibility is crucial to researchers in both quantitative and qualitative research. In quantitative research, the credibility is determined by the construction of instrument, however the researcher is the instrument in qualitative inquiry. Thus, validity and reliability in regards to quantitative research is to do with credibility whereas the credibility in qualitative research is dependant on the researcher skills (Patton 2002: Golashfani 2003). The concepts of reliability and validity are regarded separately in quantitative research. However, in qualitative research, they are considered together and expressed via other terminologies which incorporate both for example, credibility, trustworthiness.
Although, reliability and validity are important criteria in establishing and assessing the quality of research for the quantitative researcher, the idea can be applied to all research. If the notion of testing is perceived to be a way of obtaining information therefore for any qualitative study, quality is seen to be the fundamental test. Qualitative and quantitative research hold different judgments in evaluating the quality of studies which has resulted in the concept of reliability as a irrelevant matter. Stenbacka (2001) argues that the term reliability is irrelevant when judging the quality of qualitative. She further argues that if reliability is to be used as a criterion in qualitative research, subsequently the study is deficient.
Here enters a debate as to whether the criteria for assessing quantitative research can be applied to qualitative research. Three separate positions are apparent within the ongoing debate as according to Hope and Waterman (2003). Firstly, the notion that qualitative research is anti-realist, thus it is unable to apply assumptions from realist qualitative methods, secondly, the adoption of the positivists validity criteria, although slight alterations of terms maybe required and lastly, the formation of a separate criteria from those adopted in quantitative research.
Firstly, we will address the issue of philosophical belief in the value of qualitative inquiry. The controversy stems from the long-standing debate in science over how best to study and understand the world. This is known as the ‘the paradigms debate’ a paradigm in this case being a particular worldview where philosophy and methods intersect to determine what kind of evidence one finds as acceptable (Patten 2002). For example, this dichotomy is explained by Powers and Knapp (1990) as an orientation towards a realist view against an idealist view of the world. Realists believe that the world can be understood somewhat directly, they suggest a material world, that the world exists independent of us. Whereas idealists believe the perceptions of the world are mediated through a sequence of distorted lenses and can only be understood subjectively. Idealism argue that the world consists merely of representations and thus a creation of the mind. Thus, the attempt to apply criteria to assess qualitative research is doomed because the idea of criteria is incompatible with anti-realist assumptions thus validity and reliability are irrelevant.
LeCompte and Goetz (1982) write about reliability and validity in relation to qualitative research but invest in new concepts such as internal and external validity. According to LeCompte and Goetz, external reliability refers to the degree of replicability of a study. Lecompte and Goetz accept that it is unattainable to ‘freeze’ a social setting, thus replication is problematic. Therefore, it has been suggested that a similar role should be adapted to that of the original researcher. Internal reliability, has similarity to inter-observer consistency. This involves an agreement of results between members of the research team. Internal validity refers to whether a good match is apparent between researchers observations and the theoretical ideas they develop. External validity, refers to the degree of generalisibility of findings across social settings (Bryman 2002).
Since qualitative and quantitative research is rooted on completely different epistemological and ontological assumptions, thus many writers feel that the positivists, quantitative validity criteria is inappropriate. Leininger (1994) asserted that the application of quantitative validity criteria to qualitative are awkward, confounding and confusing (Chase 2001).
On the other hand, qualitative researchers challenge these claims. According to Patten (2001), while constructing a study, the factors of reliability and validity are of importance. Thus a second position in relation to validity and reliability in qualitative research can be discerned. Several qualitative researchers reject the validity criteria adopted by positivists. They argue against the assumption made by realists that a reality external to our perception is apparent. These qualitative researchers challenge the standards quantitative research are measured against by initiating different standards in order to judge the quality of research.
The incompatibility of the concepts of reliability and validity with the underlying assumptions and principles of qualitative research resulted in the generating new language and concepts of validity and reliability to reflect the interpretivists perspective and to distinguish quality in qualitative research (Chase 2001; Patten 2002; Seale 1999).