2 RESEARCH DESIGN
2.4 Data Collection and Data Analysis
In this research both parties of the technology partnership are considered equally important.
Therefore both small and large firms are studied. They are compared to each other and to other small or large firms. A dyad, i.e. a potential partnership of two, is analyzed. In order to research trust we need to focus on individuals, since only persons are able to trust, even though persons and organizations may be objects of trust. Individuals are interviewed and their reality is interpreted. The nested levels in the phenomenon are interpreted mainly through individuals, even though additional material is used to complement the interpretations. Thus the unit of analysis is the individual decision-maker and the potential partnership i.e. the exchange relationship (Möller and Wilson 1995d). Therefore, to understand and interpret the role and nature of trust in the asymmetric technology partnership formation the following different levels must be studied.
Environmental level technological
level Organizational
level
Individual level Small
firm
Large firm
Figure 5. Different Levels in Data Collection and Interpretation
The environmental context i.e. the nature of the high-velocity environment is important to understand. Also the organizational level and organizational factors are important as the partnering organizations’ propensity to the partner varies in accordance. In a large firm there are several decision-makers. During the research project it became clear that at least two levels of commitment were needed: strategic and operational. Individual decision-makers and the context have an important role in partnership formation. Thus the decision-makers (both at strategic and operative level), two firms (a dyad) and the context (e.g. environment and technological development) are studied. Asymmetric technology partnership formation is also impacted by the nature of technological knowledge. Because the nature and role of trust is under focus, the individual decision-makers’ ability to trust and be trusted must be studied.
Selection of Cases: Multiple Case Designs
In this study the case firms and interviewees have been chosen to illustrate and inform the researcher. The aim has been to learn as much as possible and therefore as informative cases as
possible have been chosen. Also Stake (1995) suggests maximization of learning as a first criterion for selecting the cases. Thus the basic idea has been to find either small or large ICT companies involved in asymmetric technology partnering, and suitable persons knowledgeable enough to be interviewed about the theme.
The cases were also selected so that pairs were formed (dyads). However, for reasons of confidentiality and trust, the names of the persons interviewed (and their companies) were deliberately not revealed to other informants. Rather, the informants in both case firms were allowed to discuss freely their relationships with any partner and tell stories1 of projects, negotiations or cases. This way of working emerged naturally2. However, the persons interviewed did refer to real cases and compared their experiences with several companies to illustrate the issues related to asymmetric technology partnership formation.
Due to the intimate nature of the key research issue (trust) anonymity and confidentiality allowed the respondents to talk freely, as they did not have to be afraid of the potential consequences to their existing or future business relationships. The researcher has sketched the inter-relationships of the interviewed companies herself (see Figure 6). This “anonymity” did not seem to create problems and it was possible to collect the relevant information on the managerial experience and perception on the issues of trust, technology and asymmetry in asymmetric technology partnership formation was possible to collect.
Evidence from multiple–case designs are seen to be more robust than single-case studies (Eisenhardt 1989a, Johnston et al. 1999). In this study new cases have been added, when additional information has been sought. E.g. when addressing trust, a case where trust and/or mistrust could be visible, was sought. A new case was investigated to see whether the expected issues on asymmetric technology partnership formation and evolution of trust would be pertinent or not. Thus the model on the role of trust was continuously “tested” against different cases to see how well the model described reality. The additional information from new cases has added to the knowledge base by either confirming or altering the a priori view of the researcher. In Figure 6 the case firms and their relationships are illustrated.
1 As a researcher I often felt that listening to these stories was exciting. My task was to solve a puzzle and I wanted to learn as much as possible to be able to interpret the respondents’ reality and answer my research questions. In many cases also the interviewed persons shared the excitement, e.g. two of the informants suggested that we should write a book together of their experiences as entrepreneurs.
2 For confidentiality reasons the names of the other case companies or persons interviewed were not revealed to the interviewees. This seemed natural, as it would not have been possible to “go around” and interview people about the experienced trust or mistrust in ongoing business relationships. Also, because in 1999 on I was also working for Sonera Corporation, it was important to let the interviewed persons to decide how open they wanted to be.
Small Software Company Zeta 2 MD interviews Internet Technology Net
2 MD interviews 1 technological expert interview
participant observation
Small Software Company Alpha 2 MD interviews INCUMBENT
ICT COMPANY B 3 interviews:
Partner director, director, vendor manager INCUMBENT ICT COMPANY A
2 interviews:
Director, manager
INCUMBENT ICT COMPANY E
1 manager INCUMBENT ICT COMPANY D
3 interviews:
Director, dev.director, middle manager Small Software Company Strada
1 MD interview participant observation
INCUMBENT ICT COMPANY F
3 interviews:
Partner director, manager, partner manager Small Software Company Next
2 MD interviews participant observation
Small Software Company Mia 2 MD interviews, 1 internal workshop
participant observation Small Software Company One
2 MD interviews
Technology Venture True 2 Partner dev. mgr interviews
Manager interview
Figure 6. Interviewed Small and Large Technology Firms and their Relationships
Data Collection: Interviews and Participant Observation
In addition to being elusive and multi-dimensional, trust is also a complex and sensitive issue, and thus close and open discussions with key informants were an important means to gather information (more on the use of the qualitative method, see Eisenhardt 1989a, Johnston et al.
1999, Yin 1989). Yin (1989) also notes that case studies are good in measuring and recording behavior, whereas a survey may only focus on verbal information. In addition to in-depth interviews also the participant observation method (Yin 1989) has been explored in both company-internal and inter-organizational contexts (negotiations and workshops). In this study the role of this method was more confirmatory than revealing. Also introspection has been used as the researcher has investigated the role of trust in inter-personal relationships in the asymmetric technology partnership formation context. According to Giddens (1993, 24) “it is the very ontological condition of human life in society as such” and self-understanding is connected integrally to the understanding of others.
The interviewees were chosen on the basis of the researcher’s knowledge and all the interviews were conducted personally. The interviewees were chosen so that they had direct knowledge and experience on asymmetric technology partnerships. If the person had also experience of both small and large firms, he was supposed to have relatively more personal experience and understanding on the issue of asymmetry.
Table 1. Interviewed Persons
Company Person´s position Age and sex Working experience
Large ICT Company B Vendor manager 35-40, male 10 yrs
Large ICT Company D Manager 35-40, female 10 yrs
Large ICT Company D Director 35-45, male 14 yrs
Large ICT Company D Account manager 35-45, male 15 yrs x
Large ICT Company E Project manager 25-35, male 5 yrs
Large ICT Company E Director 35-45, male > 15 yrs
Large ICT Company F Partner program manager 40-50, male > 15 yrs
Large ICT Company F Partner manager 25-35, male 5 yrs
Large ICT Company F Partner expert 25-35, male 5 yrs
Large ICT Company F Manager 35-45, female > 10 yrs
Technology Venture True Business Dev. manager 45-55, male > 15 yrs x Technology Venture True Partner Dev. manager 25-35, male 3 yrs x Internet Technology Next MD/ chief technologist 25-35, male 5-10 yrs x Small Software Company Mia Managing director 35-45, male > 10 yrs Small Software Company Zeta Managing director 25-35, male 10 yrs Internet Technology Net Managing director 25-35, male > 5 yrs
Internet Technology Net Manager 25-35, male > 5 yrs
Internet Technology Net Technological expert 25-35, male > 5 yrs Small Software Company Strada Managing director 35-45, male 15 yrs x Small Software Company Alpha Managing director 25-35, male 10 yrs Small Software Company One Technological director 25-35, male 10 yrs x
The interviewed persons were very interested in the research since they found that it had
managerial importance. The interviewees from small software suppliers were mainly managing directors (decision-makers) or technology experts with customer/partner responsibility. The interviewees from large ICT firms were managers responsible for the relationships with small suppliers. Also all small firm managers had personal experience on cooperation with large technology firms. The themes were defined in advance, but the questions and their order varied by interviewee and situation. The interviews were semi-structured in a sense that there was a question list prepared in advance but it was used freely (see Appendix I for the research questions). The interviewer guided the discussion and returned later with some interesting questions or asked additional ones as the research project advanced.
Access and contacting was relatively easy despite of the sensitive issues. The interviews were conducted in very open and trusting terms1. Also the interviews started with more general topics and moved slowly to more sensitive issues. Interviewer stated open questions, and allowed interviewees great freedom to discuss issues of interest. Most interviews were done in the companies, but in a peaceful room where the interviewees were able to focus on the discussion and talk confidentially. The interviewees were promised the right to “wipe over” any part of the discussion later and read the discussion, as well as full confidentiality when the interviews were used in written research. There were issues in almost all the interviews, where the interviewees referred to strict confidentiality, “this is out of tape-recording.” Usually they referred to some especially delicate issues and experiences. However, the general part of the interview was usually illustrative enough and this did not cause any problem to data analysis. Often the interview was continued when the planned time ran out. Open questions were used in addition to in-depth cases when the concept of trust was explored. Trust as a concept was not discussed in the first interviews, but the researcher “fished around the theme” to avoid leading questions. Special care was taken not to impact the interviewed persons’ perceptions and ideas to give socially acceptable answers (see also Blois 1999 on sensitive handling of trust due to the fragility and complexity of the issue, and Schein 1992 on the basic principle of interviewing on values and assumptions). After saturation was reached the later interviews were more for exploration of raised new issues and for validatory purposes.
The interviews lasted some 1.5 – 2.5 hours. Some additional interviews were shorter, and there were numerous brief telephone conversations related to partnerships and to common interests due to my work tasks at that period. The interviews were recorded on tape and transcribed. All the expressions, “eh, well” and such were not written, but the tone and feeling were tried to keep as original as possible. No one refused to be interviewed. All the interviewed participants were sent the chosen citations from the interview for comments and acceptance.
Whilst working for a large technology firm I had a chance for participant observation in internal workshops, development groups and also in some negotiations for asymmetric technology partnership formation. In two of the cases I interviewed the boundary spanners2 of small and large technology firms both before and after their partnership negotiations and observed the actual meeting. It was useful merely to observe actors behavior and learn of their interpretation of he other actors’ behavior.
1 Trust may have partly developed through earlier contacts in almost all of the cases. Because of the working experience in the field I had many suitable contacts available.
2 A boundary spanner is an individual champion active in the inter-organizational interface. See e.g. Davenport and Prusak 1998 and Tushman and Scanlan 1981.
In this research several respondents and cases have been used to avoid bias. The respondents have also answered by practical examples about their relationships with each other, thus the data is supposed to be richer. The researcher made a second round when specific or paradoxical information had been given. In summer 2001 the preliminary results were discussed and confirmed with 3 small firm managers and 3 large firm managers. In addition some written data (business plans, economic figures, journal articles and www-pages) was used. Also some experts (corporate analysts, venture capitalists, consults and customers) were interviewed for a more general discussion on asymmetric technology partnerships.
Data Reduction and Analysis of Multiple Cases
The basic idea behind cross-case searching tactics is to force the investigators to go beyond initial impressions to structured and diverse lenses on the data. New cases either confirmed the emerging theory or disconfirmed it. When the evidence from another data source corroborates a pattern of a previous one, the finding is stronger and better grounded. However, disconfirming cases are valuable in providing an opportunity to refine or extend the theory (Eisenhardt 1989a, 541-544). In this research the found key characteristics of asymmetric technology partnership formation as well as conceptualization of trust were such that the issues came out several times.
If no contradictory evidence or explanation was found the emerging issues, themes or patterns were taken as tentative results (list of critical issues). The iteration between existing data, theoretical literature and additional interviews was used to confirm the early ideas to a point when a coherent and logical explanation was reached. That means, if potentially critical issues were logically sound and supported by other issues, they were accepted. If incongruent or surprising issues came out, they were analyzed and an explanatory framework was searched. E.g.
the emerged theme of individual-based fast trust led to in-depth analysis of the managerial context (see outer context: dynamic environment in chapter seven). Questions like why? In what conditions? Where is this valid? Have been continuously asked.
The data should be investigated and preferably coded in order to find themes and patterns (Glesne and Peshkin 1992, 132). Data reduction is part of analysis and…” a form of analysis that sharpens, sorts, focuses, discards, and organizes data in such a way that “final” conclusions can be drawn and verified” (Miles and Huberman 1984, 21). According to Eisenhardt (1989a, 541), during the process tentative themes, concepts and possibly even relationships between different variables should emerge. Evidence from each case should be constantly compared with the frame iterating towards a theory that closely fits the data. Eisenhardt (1989a, 541) further notes that a close fit is important because it takes advantage of new insights from the data and yields an empirically valid theory.
In this research project the data reduction process was as follows: A majority of the data was collected as 1) in-depth interviews, which were tape-recorded and 2) transcribed to text. The interviews made some 15-35 pages of written text each. The researcher then 3) read this text, and relevant parts (related to interview questions and research problems) were 4) translated into English by the researcher. Important issues were 5) written above the translated text in block letters (coding). This format enabled simple and concise overview of the interviews. Patterns, themes and concepts emerged from the data (for an example of the transcribed, translated and coded interview data, see Appendix IX).
The final important part of data analysis is conclusion drawing or verification, where regularities,
patterns, explanations, possible configurations, causal flows and propositions may be drawn.
Data analysis is interactive cyclical, and iterative (Miles and Huberman 1984, 22). Also Eisenhardt (1989a) notes the frequent overlap of data analysis with data collection. In this research the empirical analysis was done throughout the process and the various phases of the process were certainly not easy to separate in practice.
Data reduction
Conclusions:
drawing/ verifying Data
collection Data
display
Figure 7. Components of Data Analysis: Interactive Model (Miles and Huberman 1984, 23)
Analyzing and reaching traceable conclusions on qualitative data is challenging. Eisenhardt (1989a, 539) notes that analyzing qualitative data is both the most difficult and least codified part of the process. The data should be looked at in many divergent ways (Eisenhardt 1989a, 540).
Yin (1989) suggests either a “pattern matching” logic where the actual and expected patterns are compared, or “explanation building” where theoretical propositions are developed. In this thesis the explanation building approach has been used as theoretical propositions have been developed and validated through the research process.
“Data analysis… is the effort of researchers to manage and make sense of their data, to transform it from its acquired form… into a form that communicates the promise of a study’s findings.” (Glesne and Peshkin 1992, 145)
“Data analysis involves organizing what you have seen, heard, and read so that you can make sense of what you have learned. Working with the data, you create explanations, pose hypotheses, develop theories and link your story to other stories. To do so, you must categorize, synthesize, search for patterns and interpret the data you have collected.” (Glesne and Peshkin 1992, 127).
Because of the subjective nature of human knowledge it is to the researcher to try to understand why are they saying this? In what context have they experienced this? Even though challenging, it is important for the researcher to try to understand the respondent’s organizational context and basic assumptions. Even when wanting to be honest and frank, the interviewed persons are usually able to give only a partial and possibly biased view due to some previous incident not known to the researcher. The subjectiveness is accepted, but in order to get a more complete picture of the phenomena multiple sources of evidence have been used. For example several respondents have described the same business case and so multiple analyses have been melted to a more comprehensive view on the case. Several persons have been interviewed in order to get a
more valid view at the organizational level. In this thesis a rather dense description and large portion of context-based illustrations are used to elaborate the respondents’ worldview and actual context (see Syrjälä et al. 1996).