65
Table 8.
Global top 15
solar module manufacturers, 2013.
Rank
Manufacturer
Location of
Headquarters
Founded
year
Annual Report
(Form)
Number of
Samples
a
1
Yingli Green Energy
China
1998
2007-2015 (20-F)
9
2
Trina Solar
China
1997
2006-2015 (20-F)
10
3
Canadian Solar
Canada
2001
2006-2015 (20-F)
10
4
Sharp
Japan
1912
2006-2015 (other)
10
5
Jinko Solar
China
2006
2010-2015 (20-F)
6
6
First Solar
USA
1990
2006-2015 (10-K)
10
7
ReneSola
China
2005
2006-2015 (20-F)
10
8
Kyocera
Japan
1959
2006-2015 (20-F)
10
9
JA Solar
China
2005
2006-2015 (20-F)
10
10
Hanwha SolarOne
South Korea
1997
2007-2014 (20-F)
8
11
Sunpower
USA
1985
2006-2015 (10-K)
10
12
Suntech
China
2001
2006-2012 (20-K)
7
13
Solar Frontier
Japan
2006
2009-2015 (other)
6
14
REC Group
Norway
1996
2006-2014 (other)
8
15
(Hanwha)
Q-cell
Germany
(South Korea)
1999
(2015)
2015 (20-F)
1
Total
-
-
-
125
Note: Because an annual report describes the corporation’s status in the previous year, the
period of 2006-2015 corresponds with the time frame of this study, which is 2005-2014.
Method
Content analysis is defined as a technique for making inferences from texts based
on systematic coding.
103
It has been used for many organizational studies since it enables
exploring difficult-to-study issues in management fields.
104
It enables researchers to
103
Stemler, Steve. "An overview of content analysis."
Practical Assessment, Research &
Evaluation
7, no. 17 (2001): 137-146.
104
Levihn, Fabian, Cali Nuur, and Henrik Blomgren. "Corporate
response to climate
change mitigation: What can we learn from annual reports of European industries?"
International Journal of Industrial Engineering and Management
2, no. 3 (2011): 77-86.;
Waller, David S., and Roman Lanis. "Corporate social responsibility (CSR) disclosure of
advertising agencies: an exploratory analysis of six holding companies' annual reports."
Journal of Advertising
38, no. 1 (2009): 109-122.
66
access individual and collective values, intentions, cognitions and attitudes.
105
Content
analysis can be used in a flexible way by taking quantitative or interpretive approaches.
Frequency count of codes has commonly been used in many studies, and some studies
have analyzed qualitative terms and themes emerged from their investigation.
Both quantitative and qualitative approaches were taken
to analyze the changes in
the external environment of the solar module manufacturers and their reactions to it.
First, a structured content analysis with pre-determined codes was conducted to classify
the risks faced by the firms in recent ten years. Second, a qualitative approach was taken
for developing a description of the reactions of the firms to the changing environments.
For the structured content analysis, each risk factor in every annual report was
assessed and classified according to the pre-determined codes. The annual reports on
Form 10-K and 20-F include a specified format describing risks, which provides detailed
descriptions of each risk factor. The description of each risk
factor consists of a main
statement and detailed explanation. Both the main description and the details of the risk
factor were classified as one unit, which was the unit of analysis in this study. As a result,
the unit of risk factor consists of a few paragraphs. Although the annual reports written in
their own forms provide less standardized descriptions of risk factors, they also provide
separate sections of risks. This enabled coding of each risk factor. The analysis was
restricted to the risk section of each annual report and every text
of the risk section was
coded.
105
Duriau, Vincent J., Rhonda K. Reger, and Michael D. Pfarrer. "A content analysis of
the content analysis literature in organization studies: Research themes, data sources, and
methodological refinements."
Organizational Research Methods
10, no. 1 (2007): 5-34.
67
The codes for risks were pre-determined based on literature initially, and were
revised during the research process. To reduce rater bias, the Weber protocol for coding,
which has been widely referenced in
the content analysis studies, was adopted.
106
The
coding categories were defined, and were tested with samples, which were 10% of the
data. The results of the testing showed that the codes were not exhaustive and some codes
were not accurate enough. The codes were revised based on these testing results, and they
were tested again with the same 10% of the samples. All the data was coded with these
revised codes. After completing coding, the reliability and accuracy were assessed again.
Since no
issue was found, the coding was completed.
The risks are classified according to the source of risks:
internal environment
, and
external environment
(Table 9). The risks from external environments are classified again
into seven categories:
competition, demand, policy, resource, supplier, technology
, and
other
. The categories for the risks from external environment were initially determined
based on the study of Meijer, Hekkert, and Koppenjan, and they were revised
after testing
the codes.
107
106
Duriau, Vincent J., Rhonda K. Reger, and Michael D. Pfarrer. "A content analysis of
the content analysis literature in organization studies: Research themes, data sources, and
methodological refinements."
Organizational Research Methods
10, no. 1 (2007): 5-
34.;Weber, Robert Philip.
Basic content analysis
. (London: Sage, 1990).
107
Meijer, Ineke SM, Marko P. Hekkert, and Joop FM Koppenjan. "How perceived
uncertainties influence transitions; the case of micro-CHP in the Netherlands."
Technological Forecasting and Social Change
74, no. 4 (2007): 519-537.
