After translation equivalence has been established, what other two types of equivalence need to be established to ensure transferability? Also explain why these equivalences are important.

Question

What will be an ideal response?

Answer 1

Translation equivalence is a necessary, but not a sufficient, condition for ensuring transferability to another culture. The measure must also demonstrate two other types of equivalence. The first is conceptual equivalence--that is, the attribute being measured has a similar meaning across cultures (Brett, Tinsley, Janssens, Barsness, & Lytle, 1997). Measures must produce the same conceptual frame of reference in different cultures, which means that different cultures are defining an attribute in the same way (Cheung, 2008). The construct of leadership offers an example of nonequivalence, for it tends to refer to organizational leadership in the United States and to political leadership in Asian cultures (House, Wright, & Aditya, 1997). As another example, consider that the Western notion of “truth” is irrelevant in Confucian thinking (Adler, Campbell, & Laurent, 1989). Respondents also must interpret response options in the same way (Riordan & Vandenberg, 1994). For example, “neither disagree nor agree” may be interpreted as indifference in one culture and as slight agreement in another (Aguinis et al., 2001). Ensuring that constructs are comparable when testing for between-group differences is critical in cross-cultural research (Cascio, 2012b).
Finally, metric equivalence requires that statistical associations among dependent and independent variables remain relatively stable, regardless of whether a measure is used domestically or internationally (see Chapter 5). Correlation matrices and factor structures should also remain similar (Cheung, 2008; Salgado, Viswesvaran, & Ones, 2001). This was the approach taken by Tsaousis and Nikolaou (2001) in demonstrating the conceptual and metric equivalence of a measure of the five-factor model (FFM) of personality in the Greek language. With respect to integrity tests, Berry, Sackett, and Wiemann (2007) reported no major differences in scale means, standard deviations, criterion-related validities, or admissions of counterproductive work behaviors across five countries.