Trustworthiness in Research

ㅡ Validity

Validity is a fundamental concept that pertains to the accuracy of the research tools employed, ensuring they truly measure what researchers intend to assess (Bryman and Bell, 2011; Li et al., 2006; Dunn et al., 1994). It encompasses four distinct dimensions:

1. Content Validity: Content validity underscores the necessity for a measurement to accurately represent the entirety of the construct under examination (Bryman and Bell, 2011; Li et al., 2006; Dunn et al., 1994). This means that the measurement should encompass all relevant aspects of the concept being studied. However, it's important to note that content validity may not always be entirely objective, as it often depends on the discernment of the researchers themselves (Churchill, 1992).

2. Construct Validity: Unlike content validity, construct validity is concerned with how well a theoretical framework is translated into the actual measurement items (Hair et al., 2014; Dunn et al., 1994). It is considered valid and robust when both convergent and discriminant validity are present (Hair et al., 2014; Dunn et al., 1994). Convergent validity is typically established through statistical techniques such as confirmatory factor analysis, which helps confirm that related constructs are correlated (Churchill, 1979), while distinctive constructs remain separate (Anderson and Gerbing, 1988).

3. Convergent Validity: Convergent validity is achieved when the standardized coefficients in a measurement model demonstrate statistical significance (Demo et al., 2012). This involves relying on techniques like confirmatory factor analysis to ensure that related constructs exhibit correlation (Churchill, 1979), while distinct constructs remain distinct (Anderson and Gerbing, 1988).

4. Unidimensionality: Unidimensionality asserts that each measurement should be influenced by a single underlying construct (Anderson et al., 1987; Dunn et al., 1994; Garvin, 1987; Gerbing and Anderson, 1988). This aspect is vital in ensuring construct validity and relies on the idea that a measurement item should not be a mixture of multiple unrelated concepts.

ㅡ Reliability

Reliability in research is all about the consistency, stability, and internal coherence of a measurement tool (Mentzer and Flint, 1997; Rattray and Jones, 2007). It assesses whether a tool consistently produces the same results when the same research is conducted repeatedly (Campbell and Fiske, 1959). In simpler terms, reliability emphasizes the degree of agreement between multiple attempts to measure the same characteristic using similar methods.

Various methods, such as test-retest, split-half, and Cronbach's Alpha, are employed to assess reliability (Bagozzi, 1984; Mentzer and Flint, 1997; Hair et al., 2014). Cronbach's Alpha, although widely used, has faced criticism because it can increase the number of measurement items (Hair et al., 2014; Malhotra et al., 2017). In response to this, alternative approaches like Average Variance Extracted (AVE) and Composite Reliability (CR) have emerged from Confirmatory Factor Analysis, offering researchers a more comprehensive way to evaluate reliability (Hair et al., 2010).

In summary, validity ensures that research tools accurately measure what they are intended to measure, while reliability ensures that these measurements are consistent and dependable over repeated research efforts. Researchers must carefully consider and address these factors to ensure the trustworthiness of their research findings.