Choice of Research Method

The research method defines the set of procedures and techniques for gathering and analyzing data Strauss & Corbin, 1998; p. 3). For our study, we choose a Delphi method utilizing mainly qualitative techniques. Although the study includes quantitative elements, we argue that the research should be classified as qualitative, as explained in the following.

3.2.3.1 The Delphi Method

This study uses the Delphi method to investigate the research question by collecting the opinions of procurement experts on the topic. The Delphi technique can be described as a method for structuring a group communication process so that the process is effective in allowing a group of individuals, as a whole, to deal with a complex problem Linstone & Turoff, 1975; p. 3). Skulmoski et al. explain the Delphi method as an iterative process to collect and distill the anonymous judgments of experts using a series of data collection and analysis techniques interspersed with feedback Skulmoski et al., ; p. . Both explanations emphasize the characteristic of the Delphi method to conflate the judgement of individual experts towards a profound and mutual understanding of the research problem. Typically, the research is designed in multiple stages where the subsequent research phase is built on the results of the previous phase, iterating until the research question is answered and the process stops (Skulmoski et al., 2007).

Suitability of the Delphi Method

By breaking up the research process into subsequent rounds, the Delphi method allows to develop the research design during the research process. This flexibility is particularly useful in research areas where only limited knowledge or empirical data exists (Skulmoski et al., 2007, Linstone &

Turoff, 1975), which is the case for our research topic. Blockchain technology is a very new technology that is rarely introduced to purchasing departments. Consequently, the understanding of the technology s impact on purchasing processes and transaction costs is very limited.

Furthermore, the Delphi method can be recommended for studies where subjective judgement instead of an accurate analysis is suitable (Linstone & Turoff, 1975). This study assesses the effects on perceived transaction costs , since the derivation of transaction costs from financial data is hardly possible (Grover & Malhotra, 2003). Consequently, the Delphi method seems a suitable technique to combine the perceptions of different experts. In addition, Linstone and Turoff (1975) name practical reasons to choose a Delphi method: For example, the participation of multiple respondents with diverse backgrounds and no history of communication amongst each other, the need for more participants than feasible to interact jointly in a group meeting, and the confidentiality or dominance of certain sub-groups calling for anonymity. The Delphi method

provides a solution for those cases by allowing for a participative research process among a diverse sample, while appreciating the participants anonymity. Apart from this, the high flexibility of the Delphi method facilitates the research process under resource and time constraints, which makes it highly suitable for graduate research (Skulmoski et al., 2007).

Number and Purpose of Delphi Rounds Number of Delphi rounds

There is no generally prescribed number of rounds a Delphi study should apply (White, 2017). The number of iterations required depends on the nature of the research question, the heterogeneity of the research sample as well as the level of agreement desired and reached (Skulmoski et al., 2007).

Since our research objective is of an exploratory nature and the study does not aim to find confirmation through high consensus, but rather pursues to discover information, and since the study chooses a homogenous sample, two Delphi rounds can be considered adequate (Skulmoski et al., 2007; White, 2017; Giunipero et al., 2012). With every subsequent round, the response rate can be expected to fall, which can hamper the research process (Skulmoski et al., 2007).

Purpose of Delphi Rounds

In line with the exploratory nature of our research question and the choice of an inductive approach, the purpose of choosing the Delphi method for this inquiry is to uncover and understand the impact of blockchain technology on transaction costs. It is important to underline that the purpose is neither to quantify the effects on transaction costs, nor to find absolute confirmation of the effects. Therefore, it is adequate to use exploratory interviews in the first round, which allow to openly gather the participant s ideas which thereafter serve as input for development of questions for the second round (Lazerfeld & Thielens (1958) in Strauss & Corbin, 1998; Skulmoski et al., 2007). The purpose of the second round is to suggest the ideas of each participant to the other participants in order to hear their feedback. The output is expected to be a list of factors or explanations that helps to understand how blockchain technology impacts transaction costs of procurement activities.

Details about the methods chosen in each round will be explained in the chapter 3.4 Delphi Design.

3.2.3.2 Qualitative Research

Qualitative research describes any type of research that produces findings not arrived at by statistical procedures or other means of quantification Strauss & Corbin, 1998, p. 10-11).

Qualitative studies focus on the causal relationships, situational environment and meanings behind what is investigated, while numerical measures recede into the background (Denzin & Lincoln,

1998). In comparison to quantitative research, the technique for collecting and analyzing data is not mathematical, but rather interpretative (Strauss & Corbin, 1998).

When to use qualitative research, methods depends on the nature of the research question (Strauss

& Corbin, 1998). Qualitative research methods can be advantageous when studying phenomena that are difficult to describe numerically, like processes, perceptions or emotions (Rasche, 2010;

Strauss & Corbin, 1998). Likewise, qualitative methods can be a good choice when investigating unknown, or sparsely known topic areas (Strauss & Corbin, 1998).

Willis et al. (2007) point out, that qualitative methods are often the preferred choice for interpretivist research, because the richness of detail they create allows to investigate the contextual setting. However, Willis et al. (2007) continue that interpretivists do not believe that there is any universal standard of research that prescribes the use of specific methods. For interpretivists, both quantitative and qualitative methods can constitute potentially lead to understanding, but none of them can be truly objective. What makes a research method suitable for interpretivist research is not what kind of data it creates, but that it allows to understand the context that the data accrues in.

Although both deductive and inductive research approaches can make use of either qualitative or quantitative methods (Willis et al., 2007), the flexibility of qualitative methods can accrue benefits if one aims to induce theory (Strauss & Corbin, 1998); rather than sticking to pre-defined categories, qualitative research facilitates an openness to investigated a research topic in great details, using a small sample, and including perceptions of different people as well as the situational background (Rasche, 2010).

However, qualitative and quantitative methods do not need to be treated strictly separated. As much as qualitative data can be quantified, qualitative research can apply quantitative elements while keeping a qualitative research focus. For example, when including background information like age of participants, sample size or response rate, this information is given numerically, but can be used in an interpretative manner (Strauss & Corbin, 1998). Strauss and Corbin (1998) point out that qualitative and quantitative research methods do not have to preclude one another. Instead, both methods can supplement, complement or even develop each other – a continuum of intermediary positions is possible. This interplay of methods should be considered wherever the research topic favors a combination of methods, also beyond the need for triangulation. For example, Lazersfeld and Thielens (1958) suggest that the development of survey questions should be based on exploratory interviews (in Strauss & Corbin, 1998). Strauss and Corbin further suggest

that research is a circular project p. , pointing out that suitability of methods often only becomes clear during the research project.

Since blockchain technology is a new topic to procurement literature and its implications are not much investigated, we find qualitative methods most useful for our study. We derive our findings through the interpretation and summarization of expert opinions. Qualitative methods facilitate to investigate details and context of documented opinions (Rasche, 2010). Acknowledging the subjective bias, we cannot avoid in our roles as observing researchers (Willis et al., 2007), we intend to move choices of prioritization to the research participants. For this purpose, we find the quantification of statements most useful, finding consensus between participants through selection of statements by quantity. Nevertheless, the bulk of our research is based on the interpretation of the experts expectations and of their context.

We want to reiterate that we believe that the research design should be reviewed throughout the development of the investigation (Strauss & Corbin, 1998). Particularly, because we investigate a topic about which little is known, we want to adjust the research method to the given situation, rather than binding ourselves to the use of pre-defined procedures.