Social science research is carried out based on the measurements, due to the negligence of the researchers, such as inappropriate wording in questionnaires, improper investigation forms, the poor survey environment, or respondents' personality, mood, level of education, age, gender and other individual psychological factors, all these will cause bias in the results. Thus, after the data is collected and analyzed, reliability and validity test should be conducted to ensure the correctness of the results, and then to improve and correct the results. As an excellent and rigorous empirical research, it must based on good measurement, and only good research instrumentation can bring high reliability and validity. Reliability and validity is the criteria for assessing the quality of academic research (Barnes & Vidgen, 2001). In this essay, the author aims to explain the importance reliability and validity in the choice of research designs and methods.
The concept of reliability and validity mainly comes from the tradition of positivism quantitative research, because the purpose of quantitative research is to identify the universalism of social phenomena or social behavior, therefore, the objectivity and credibility of the research methods become very important assessment indicators (Baumgarten, 2012). Morse et al (2002) believe that reliability's role is to ensure that research results can be replicated. First, for quantitative research, replicability contains at least three dimensions: First is the results' stability, which means after many tests, the results obtained are quite close. The second is internal reliability. internal reliability is applicable to the research of multiple indicators. For example, the researches taking questionnaire as the main data collection tool will use the inherent reliability to measure the degree of fit of the questions in the questionnaire with the corresponding concepts (Bryman, 2008). The third is inter-observer consistency and intra-observer consistency. Luoma (2004) argues that the former assesses the consistency of inter-observer to explain the same phenomenon; the latter tests it the consistency of the same observer to explain the same phenomenon in different circumstances. When the explanation of different observers for the observed phenomenon is similar, which means the observer consistency is high. Similarly, when the same observer obtains similar explanation for the same phenomenon in each situation, it passes the intra-observer reliability test.
The reliability of quantitative research is tested by statistical calculations, in contrary, qualitative research does not use number information or use non-systematic data collection and analysis, so that their reliability is questioned by quantitative research academics, therefore, Kirk & Miller (1986) advocate the qualitative researchers document data collection and analysis methods clearly, LeCompte & Preissle (1993) recommend for the replicability of qualitative research results, they argue that if want to replicate the qualitative research results, the status of the researcher, the choice of participants, social context, data collection and analysis methods should remain same.
Validity is defined as the extent to which the instrument measures what it aims to measure (Miller, nd), it is the benchmark used to measure whether the research methods are suitable for research purposes, that is, to assess whether the results reach the research purposes. According to Rubin & Babbie (2009), in quantitative research, there are mainly several kinds of validity: the first focuses on measuring the applicability of the data collection method, commonly referred as measurement validity, it covers all activities related to measurement in the research. The second is internal validity. What internal validity concerns is how to prove the correlation...