Designing

Research design includes, for instance, defining unit of analysis, selecting cases which will be utilized in the study, identifying case studies and testing research design against different criteria which provide validity and reliability for the re-search (Yin 2014). The rere-search design is the logic that links the data to be col-lected to the initial questions of the study (Yin 2014). Creating a research design

is a crucial step in the study, because it commonly includes several main choices which the researcher has to make during the research process.

This study was aiming for a holistic understanding of crowdsourcing and its utili-zation in an organiutili-zation’s innovation activities, and included one main research question and three sub-research questions. Innovation activities are part of a mod-ern organization’s management activities and should cover several areas of man-agement which the research questions of this study also highlight. During the de-sign of the research, the initial framework was implemented, as illustrated in Fig-ure 11.

Figure 11. The initial framework of the study

The initial framework of the study included strategic management, innovation management and project management areas which utilize crowdsourcing for crea-tion of innovacrea-tions which have demand in the markets. The assumpcrea-tions of the study are that the selected management areas have connections to each other and that crowdsourcing is utilized in selected management areas. The empirical data collection, literature and selected case organizations were based on this initial framework. Therefore, the units of analysis in this study are management areas and crowdsourcing, as illustrated in Figure 11. The development of the GCM for the holistic innovation management model was based on the initial framework.

Case organizations were required to meet four criteria for crowdsourcing and its utilization in management areas and innovation-related activities. Moreover, management should have a clear focus in innovation-related matters. A case or-ganization should fulfill the following criteria for participating in the study:

1. A case organization should have management for strategies, innovations and projects.

2. A case organization should have experience of implementation of innova-tions within its own industry.

3. A case organization should be customer-oriented.

4. A case organization should have experience of crowdsourcing and its uti-lization in an organization, and especially in innovation-related activities.

The case organizations’ industries or size were not limiting factors in this study.

Therefore, this research was implemented in a wide range of industry sectors. The amount of case organizations was not defined in the design stage. Collection of case data continued until a saturation point was reached. The saturation point is reached when no additional data are being found and full understanding about the object under study is reached (Glaser and Strauss 2009; Holzemer 2010; Legard et al. 2003). Case organizations were selected based on criteria, and the saturation point was the point at which the collection of the empirical data was concluded.

Crowdsourcing is a relatively new concept for science and businesses which can also be seen in case organizations. Qualitative and quantitative approaches were selected for forming a holistic understanding of crowdsourcing and its utilization in an organization’s activities. The main case data of the study were collected in one cycle. Interviews were carried out as semi-structured interviews where man-agement areas and crowdsourcing were discussed. The second data collection cycle was a broad verification of the developed model. These interviews were carried out as structured interviews.

Unstructured interviews provide qualitative depth by allowing interviewees to talk about the subject in terms of their own frames of reference and, thereby, it provides greater understanding of research topic (Arksey and Knight 1999; May 1997). Semi-structured interviews have same elements than unstructured inter-views but includes list of questions that needs to be covered (Bernard 2006).

Structured interviews are the opposite of unstructured interviews and are associat-ed with survey research and includes standardizassociat-ed set of questions (May 1997).

This study included four research questions which the research should gain the answer to achieve the research objective. Some interview questions were defined beforehand but interviews were guided by the situation in case organization. This research was utilizing semi-structured interviews. A presentation was created to support the semi-structured interviews, and is illustrated in Appendix 1.

Collecting the quantitative data was the supporting data collection method in this study. Collection of quantitative data was implemented with the Webropol online survey system. The survey included closed- and open-ended questions for provid-ing more in-depth information about a case organization’s crowdsourcprovid-ing utiliza-tion in selected management areas. The aim of the survey was to support the qual-itative data and provide more points of view which semi-structured interviews cannot provide. Moreover, gaining access to all interesting case organizations is not always possible, because organizations might see it, for example, as time con-suming. Utilizing a survey in the research can provide some information relating to these cases as well. Therefore, it can be useful to have several methods of col-lecting the case data. Semi-structured interviews and the survey included partly the same questions, because the research topic and aim are the same. The survey included the official research introduction letter which is presented in Appendix 2.

The final version of the survey included 109 questions in total, and included as well closed-ended six-point Likert scale questions. The Likert scale can provide the attitudes of a respondent towards a particular subject (Black 1999; Wilson 2010). The six-point Likert scale was selected to avoid central tendency bias er-ror. Central tendency bias error occurs when respondents are reluctant to rate at-tributes at the extremes of the scale and tend to rate most atat-tributes in the middle of the scale (Smith and Roodt 2003). Appendix 3 presents the online survey form questions as used in this study.

This study aimed for the development of the GCM for holistic innovation man-agement which is a theoretical model. Development of the model was based on empirical results gained from case organizations. A second round of empirical data collection was implemented through structured phone interviews. A presen-tation was created to support the structured interviews. This presenpresen-tation is illus-trated in Appendix 4.