observation as in case research
TRANSCRIPT
-
7/30/2019 Observation as in Case Research
1/2
Observation as in case research. Thus, in this section, we focus our comments on issuess
pertaining to survey research.
Primary data developed through a survey methodology have one very important
advantage over almost all secondary data. That edvantage is that the instrument is developed
specifically to adress and research question. Many marketing strategy question cannot be
adressed with any secondary data source. For example, there are no analogues to market
orientation which has been an important research topic recently (e.g Kohli and Jaworksi,
1990; Narver and Slater, 1990; Deshpande et al., 1993). Although Snow and Hambrick
(1980,p., 535) argued that selft-typing by managers was inferior to using archival data for
identifying current or historical startegic events, subsequent research has shown congruence
between archval measures and managers assessment of marketing conditions ( Keats and
Hitt, 1988), product market strategy (Shortell and Zajac, 1990) and performance (Dess and
Robinson, 1984). Thus, survey research the most appropriate, if not only, approach for
addressing some questions.
However, there are several decision that influence whether the survey research will be
externally and/or internally valid. These decision include : (1) selection of the sample of
appropriate breadth, (2) selection of the sample of appropriate size, (3) determination of the
appropriate level in the organization for data collection and analysis, and (4) development on
valid measures.
Sampling issues
Sample breadth
A substantial portion of marketing strategy research has been concerned with testing
general theories that apply across market bounderies. Exsample include the relationship
between pionering, quality, market share or market orientation and some measure of
performance. Establishing a general relationship between some element of marketing strategy
and performance implies that the relationship can be generalized over a variety of market
settings. This could lead to miximize its external validity.
A prerequisite for axternal validty is internal validity; does the research design control
for the effects of other factors that may effect or obscure the relationship under study? Unless
the research design has adequate controls, there can be little confidence in the finding that
result from the study. Internal validity is challengged by the relativity of strategy (Snow
and Hambrick, 1980, p., 531) which says that strategy conctruct have meaning within an
industry and competitive enviroment rather than across enviroments. For example, the values
-
7/30/2019 Observation as in Case Research
2/2
and practices taht constitute a strong market orientation are likely to be quite different in the
costumers financial services industry from those in the forest product industry. Since strategic
constructs are retaive phenomena that are largely influenced by industry condition and
competitors actions, it is difficult to control for industry-specific effect on strategy in board
samples and consequently to achieve internal validity.
The type of research that has the greatest internal validity is the laboratory experiment
where confouding influences are eliminated. While this type of research has been used to
study some strategic decision making issues (e.g Sawyer et al., 1979), it has little relevance
for studying the types of relationsship illustrated in Fing. 1. A reasonable alternative may be
to focus the early stages of research in a single corporation. Business units managers in a
decentralized corporation are likely to have substantial decision making autonomy
(Varadarajan and Clark, 1944) Thus