13th European Conference on eGovernment – ECEG 2013 1 | Page 631

2. Methodology

Wolfgang Palka et al.

To develop the classification schemes, we follow a qualitative, exploratory two‐step research approach. In step 1, we examine the similarities and differences as well as typical instances and their characteristics of OGD best practices. To do this, we conducted a multiple case study analysis as proposed by Eisenhardt( 1989) and Yin( 2009). We analyzed documents and portals of OGD initiatives in the USA, Great Britain and New Zealand. In step 2, we explore and refine the case study results by findings from relevant literature. By adding characteristics that are common from a theoretical perspective, we were able to develop classification schemes with a strategic and operative dimension that have a high generalizability for the OGD provision.

2.1 Data collection

We selected case studies from countries which can be described as first movers and best practitioners in this field. They have implemented OGD portals successfully several years ago – the USA, Great Britain and New Zealand. More so, these countries are working already on a large scale diffusion of OGD. We identified two areas as data sources:( 1) central documents from the selected countries, such as the " Memorandum on Transparency and Open Government”, legislative texts such as the " Freedom of Information Act " in the UK or coalition programs including " The Coalition: our program for government ", and( 2) concrete implementations of OGD initiatives and their functionality. In total, we identified 17 OGD portals, which consist of a huge number of sub‐portals, e. g., the idea portal on Data. gov. uk( Table 1).

Table 1: Identified OGD portals for data analysis Country OGD portal( URL)

USA

UK

NZ

2.2 Data analysis

After the identification of the relevant data sources, we performed a qualitative data analysis by examining the documents and implementations as mentioned above( Miles & Huberman 1994). More so, to identify the different instances and characteristics, the data analysis adopted open and axial coding set out by Strauss and Corbin( 1998). Two experienced coders were involved in the coding process. After the first step of iterative data analysis, known as open coding, emergent concepts and categories suggested by the data were selected and named. In the next step, known as axial coding, we put those data back together in new ways by making connections between the categories and concepts. For example, we identified concepts that answer the question“ Which political level is responsible for the OGD portal?”( i. e., we named the category“ Level of responsibility”): for the USA, we identified the White House and Department of State that are responsible for OGD portals; for the UK, we found the Chief Secretary to the Treasury; and for New Zealand we discovered the Department of Internal Affairs. As a consequence, we were able to name two instances: federal( White House), federal and ministry( Department of State, Chief Secretary to the Treasury, and Department of Internal Affairs). In the next step, we enriched these findings with existing knowledge from literature and connected these concepts / instances with the appropriate category.

3. Results

Based on the multiple case study analysis and its findings, we developed three classification schemes for the OGD provision via web portals. Their setting bases upon the creative technique of morphological boxes developed by Zwicky( 1966). This technique allows the structuring and investigating of a set of relations contained in multi‐dimensional problems. As a consequence, we derived the content of our classification schemes from the findings of the case studies.

609