Page 121 - Comparing Political Communication Theories, Cases, and Challenge

P1: JZZ/KAA P2: KaF
0521828317c05.xml CY425/Esser 0521828317 May 22, 2004 11:55

Designs and Methods of Research

the amount of foreign affairs coverage in one variable, by measuring the
length of newspaper articles. If, however, newspaper articles in country A
are generally longer than they are in country B, irrespective of their topic,
the result of a sum or mean index of foreign affairs coverage would almost
inevitably lead to the conclusion that the amount of foreign affairs cover-
age in country A is higher than in country B. This outcome would hardly
be surprising and would fail to answer the research question, because the
countries’ average amount of foreign affairs coverage is not related to the
national average length of articles. To get comparable results, that is, to
avoid cultural bias, the results must be standardized or weighted, for ex-
ample, by the mean article length. Analogous to this standardization and
weighting of the article length, Kolb et al. (2001) propose the weighting
of newspaper articles using the estimated amount of recipients, which
may improve inference of “public opinion” from the articles’ contents.
In the following three sections of the chapter, we will illustrate different
sources of bias on the three levels and try to show ways of avoiding the
trap of cultural bias.

Construct Equivalence and Construct Bias
Construct bias (van de Vijver and Tanzer 1997, 264) can be found
whenever the construct of interest is not equivalent in all the cultures
included in the study. To find out whether the construct is equivalent
in every country in question, the researcher will generally require ex-
ternal data and rather complex procedures of culture-specific construct
validation(s). Ideally, this includes analyses of the external structure, that
is, theoretical references to other constructs, as well as an examination
of the latent or internal structure.The internal structure consists of the
relationships between the construct’s subdimensions. It can be tested
using confirmatory factor analyses, multidimensional scaling, or item
analyses (Hui and Triandis 1985, 141–3; Bentler 1990; Pedhazur and
Schmelkin 1991, 60; Byrne and Campbell 1999; Caprara et al. 2000).
Equivalence can be assumed if the construct validation for every culture
has been successful and if the internal and external structures are identi-
cal in every country. As a less complicated but expensive alternative, one
could use triangulation, that is, a combination of different methods or
methodologies to measure the construct in each country (Denzin 1978;
Abell 1990). To reduce cost and complexity, the researchers could under-
take a survey of local experts or conduct focus-group discussions before
embarking on the main part of the study. It is obvious that a very good
knowledge of every country or culture included in the study is absolutely

101