Statistical models and techniques applied to the International Manufacturing Strategy Survey (IMSS) database

Vundla, S

Abstract

The research explores statistical models and techniques which 'have been or might be used'
to investigate relationships within the international manufacturing strategy survey (IMSS)
database and other similar databases. The sample data for the research was taken from the
fourth round of the IMSS (2005/06) database, which was the latest available release at the
time when the main report was compiled. The data set contained in that release was the one
analysed throughout this thesis.
The research had two broad aims:
(a) To contribute towards management insight, by uncovering those factors that affect
business performance, as evidenced from the IMSS data set. The IMSS data set is
an important source of information about manufacturing companies.
(b) To provide a detailed discussion of statistical models which can be used to analyse
ordinal data, and employ data reduction techniques to generate generic measures.
The models are investigated and tested on the IMSS data set. It is hoped that this
can create a useful body of knowledge for similar studies, based on this data set or
other similar data sets.
The statistical models which were examined include latent variable models (latent class
analysis and latent trait models), ordinal logistic regression, multiple linear regression and
structural equation modelling. Thus, for example, with regards to the latent variable
models, the latent trait and latent class analysis models were fitted to the IMSS database in
an attempt to establish whether the latent variable is better modelled as a continuous or
discrete variable. It was concluded that the latent trait model provides a marginally better fit
to the data compared to the latent class analysis model, which suggests that the latent
variable(s) is better modelled as a continuous variable.
A further aim of the research work was to contribute to the existing literature on
manufacturing strategy. Through examining and linking operations strategy, such as manufacturing and competitive strategies, to business performance, the research has
uncovered a better understanding and insight into the various relationships. It has thus
contributed towards managerial knowledge in this domain. A primary intention was to
demonstrate that to achieve and maintain core competencies, and also as a pre-requisite to
being able to formulate viable long-term strategies, managers should better understand the
links between market characteristics, competitive strategies, manufacturing strategies,
manufacturing performance and business performance.
The results of the data analysis suggest that more market characteristics variables are highly
significant to business performance indicators compared to the competitive strategy
variables. The fact that competitive strategy variables have a less dominant influence on
business performance could be due to the fact that the manufacturing firms who responded
to the survey are driven by market conditions and customer needs. The results imply that
companies are using competitive strategies to compete against their rivals, while their
performances relative to the past are more affected by market conditions. Furthermore, two
main explanatory variables have been identified as consistently significant across all the
performance measures, and these are 'market dynamics' and 'product design and quality'.
These results affirm the importance of quality (as represented by product design and
quality) among manufacturing firms, and the importance of responding (by firms) to the
dynamics of their environment. The 'market dynamics' in which a firm operates have an
important role to play on performance attainment and even survival of a business.
This thesis could be regarded as "offering rich insight and the drawing of specific
implications". Therefore, the main contribution of the research is testing and explaining the
relationships among manufacturing strategy, competitive strategy, market characteristics
and business performance. Since the thesis is based on the analysis of a large database, we
argue that the results from this research provide a basis for further studies in the area and
note that these results are easily generalisable to other similar studies.
However, owing to the fact that the questionnaires which were used to conduct the surveys
differed from one collection phase to another, this study is a cross-sectional rather than a
longitudinal one. Because some of the relationships investigated were concerned with strategies for improving performance the cross-sectional approach rules out the possibility
of unearthing actual improvements in respondents' performance over time. Such issues are
thus limitations of the research.
Finally, we note that studying both market characteristics and competitive strategies and
their relative impact on overall business performance is a new feature of this research and
this adds to the literature on the use and analysis of the IMSS data. Relationship studies
based on this data set have, in the past, tended to concentrate on the impact on
manufacturing performance only.

Citation

Vundla, S. Statistical models and techniques applied to the International Manufacturing Strategy Survey (IMSS) database. (Thesis). University of Salford

• Thesis Type: Thesis
• Deposit Date: Aug 12, 2021
• Award Date: Nov 1, 2012