Page 118 - Comparing Political Communication Theories, Cases, and Challenge

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

Werner Wirth and Steffen Kolb

of turned-around quasi-experimental logic, studies following this
design focus on similarities between cultures, even though these
differ in the greatest possible way. One example might be the
search for correspondence in the yellow press in many different
countries.

The most different systems design has often been favored due to the higher
number of possible cases (i.e., “most different” countries) and due to
overly specialized research on single phenomena (similarly van de Vijver
and Leung 1997, 29–30). Hartmann (1995, 31–3) has pointed out that,
for the sake of a higher number of cases, the impact of context variables
hasbeenunderestimated.Interculturalcomparisonsthatvaluetheirdata
profiles higher than history, religion, tradition, and so forth, risk ending
up with trivial results. Berg-Schlosser (1997) suggests a parallel analysis
of the most different systems with the same outcome (MDSO) and the most
similar systems with different outcomes (MSDO) to improve the results.
Ragin (1989) prefers a medium number of cases to combine intensive
and extensive research strategies. The intensive strategy will produce in-
depth knowledge of the research objects, but it is only suitable for a small
number of cases due to the effort involved. For a large number of cases,
the extensive strategy would be preferable, but it remains a superficial
way to analyze cultures. In addition to these approaches, there are sev-
eral criteria for the selection of countries or cultures. For example, the
research objects could be chosen following a theoretically built factor
matrix or following some previously fixed guidelines for decision mak-
ing. In most cases, “modified most similar systems designs” seem to be
“seen as the optimum” (Niedermayer 1997, 97).
Whenever any kind of representativity is needed, random selection
will be advantageous. Random samples of countries, however, are rather
problematic from a statistical point of view, because the number of states
in the world is limited. Thus, a normal distribution for the social factors
under examination, that is, the precondition of random sampling, can-
not be assumed, and some parametric analytic methods may hereby be
renderedinapplicable.Moreover,manystatisticalmethodsfaceproblems
when applied under the condition of a low number of cases (Ragin 1989;
Hartmann 1995, 34; Aarebrot and Bakka 1997; Widmaier 1997). These
problems of low numbers of cases are, of course, irrelevant for stud-
ies that are measuring on an individual level, that is, analyzing cultural
groups by survey or experimental research or examining international
mediaproducts.Randomsamplingisappliedfrequentlyinpsychological

98