The term stakeholder is in the context of open data often used when identifying “barriers” (Barry & Bannister, 2013; S. Martin, Foulonneau, Turki, & Ihadjadene, 2013) or “impediments” (Zuiderwijk, Janssen, et al., 2012) to open data. The literature about barriers and impediments often makes only implicit reference to stakeholder roles or treats them very broadly, e.g. as users. Furthermore, the identified barriers are to some extent as structural, legal and technical and not directly related to specific stakeholder roles. However, the literature is very extensive and detailed in identifying impediments which pose a challenge specifically for users of open data (see esp. Zuiderwijk, Janssen, et al., 2012). This also gives some indication for requirements users have. The table below gives a summarised overview of the barriers found in the literature, drawing upon a modified classification of Zuiderwijk et al. (2012).
|Category||Barrier, Impediment etc.||Sources|
|Availability andQuality||Completeness: Incomplete, partially available data; lacking data; unpublished data; non- original, processed data; insufficient meta data||Blakemore & Craglia, 2006; Zuiderwijk, Janssen, et al., 2012; Zuiderwijk, Jeffery, et al., 2012|
|Timeliness: outdated, non-updated data||M. Janssen et al., 2012; Lee & Kwak, 2012|
|Relevance: largely irrelevant, uninteresting data||Blakemore & Craglia, 2006; M. Janssen et al., 2012; Zuiderwijk, Janssen, et al., 2012|
|Quality: Bad, inaccurate or indeterminable data quality, ontologies and meta data||Conradie & Choenni, 2012; Huijboom & van den Broek, 2011; M. Janssen et al., 2012|
|Accessibility and Findability||Search Barriers: finding the proper dataset; no advanced search facilities; missing, incomplete or incorrect meta data||Conradie & Choenni, 2012; M. Janssen et al., 2012; Zuiderwijk, Janssen, et al., 2012|
|Access Barriers: Registration requirements;requests necessary; exclusive, restrictive access||Blakemore & Craglia, 2006; M. Janssen et al., 2012; Meijer & Thaens, 2009; Napoli & Karaganis, 2010; Zuiderwijk, Janssen, et al., 2012|
|Cost Barriers: fees due; inappropriate pricing||Huijboom & van den Broek, 2011; K. Janssen, 2011; S. Martin et al., 2013; Zuiderwijk, Janssen, et al., 2012|
|Fragmentation: fragmented sources; duplicated data||Conradie & Choenni, 2012; Vickery & Wunsch-Vincent, 2006|
|Usability||Licensing: Restrictive licenses; incoherent licenses; incomprehensible licenses||K. Janssen, 2011; S. Martin et al., 2013; Zuiderwijk, Janssen, et al., 2012|
|Machine-Readability: non-machine-readable formats; lack of good API||Zuiderwijk, Janssen, et al., 2012|
|Reliability: no reliable long term preservation of data; unclear provenance of data and trustworthiness of source||S. Martin et al., 2013; O’Riain, Curry, & Harth, 2012; Zuiderwijk, Janssen, et al., 2012|
|Manual Effort: various arbitrary data transformations necessary; data cleaning||Ding et al., 2011; M. Janssen, Charalabidis, & Zuiderwijk, 2012|
|Linking: difficulties to link data and meta data||M. Janssen et al., 2012; King, Liakata, Lu, Oliver, & Soldatova, 2011; Zuiderwijk, Janssen, et al., 2012|
|Interoperability: open data infrastructures not interoperable with other systems; fragmentation of software and applications||M. Janssen et al., 2012|
|Compatibility of Vocabulary and Structure: different definitions of data, terminologies; too much vocabularies; general lack of standards||Conradie & Choenni, 2012; Huijboom &van den Broek, 2011; M. Janssen et al., 2012; S. Martin et al., 2013; Zhang, Dawes, & Sarkis, 2005|
|Understand- ability||Meaning: domain knowledge necessary to understand data; jargon in data and meta data; meaning and meaningful interpretation of data are unclear; statistical expertise necessary||M. Janssen et al., 2012; King et al., 2011; S. Martin et al., 2013; Zuiderwijk, Jeffery, & Janssen, 2012|
|Validity: methods of data gathering unclear||Zuiderwijk & Janssen, 2012|
|Support: no expert advice available; lack of service by data providers to use raw data; lack of dialogue between data producers and consumers||M. Janssen et al., 2012; S. Martin et al., 2013|
|Visualisation: data and meta data are not visualised||Zuiderwijk, Janssen, et al., 2012|
The barriers listed in table 2.1 are largely expressed from a user point of view. However, they also give some indication of the multitude of other stakeholders involved in the more general decisions, e.g. regarding licensing, costs and fragmentation/consolidation of sources. As Huijboom and van den Broek (Huijboom & van den Broek, 2011) state, “whereas the drivers lie predominantly outside government, the barriers are within government organisations.” Whether or not this is entirely true, various barriers mentioned in the literature indirectly refer to the vertical and horizontal differentiation between different governmental levels. Especially regarding policy making in the context of open data, these levels’ roles are rarely appreciated. Accordingly, open data process models treat the public sector as a uniform entity. However, differences i.a. in regard to authority, amount of data, technical capabilities are notable among the local level, the regional level, the national level, the supranational level (e.g. the EU), the international level (e.g. the World Bank), and non-political levels (e.g. academic institutions). Predominantly the legal literature appreciates the role this administrative and political complexity plays (K. Janssen & Dumortier, 2003; K. Janssen, 2011).
One core element of open data is making data publicly available that is previously stored and kept internally by public agencies. Similarly, inter-agency information sharing involves making data available beyond the organisation’s boundaries, if only within the public administration. Here, previous research highlights how different values and cultures within public administration impact on information sharing practices (Yang & Maxwell, 2011). For example the kind of data that is collected and stored, how data are defined and how it shall be interpreted and used is affected by professional values (Dawes, Cresswell, & Pardo, 2009). Making data available thus means giving up the prerogative of definition and interpretation, because the data can be interpreted and used very differently in ways, what public agencies might view as detrimental (Pardo, Cresswell, Thompson, & Zhang, 2006). The completeness, accuracy of information and the misusing of the data by others which might even incur liabilities is considered as a further barrier for data providers (Yang, 2012). The lack of resources, especially staff shortage, is also mentioned as to inhibit data sharing. (Landsbergen Jr. & Wolken Jr., 2001; Yang, 2012). On the other hand trust has been shown to positively impact knowledge sharing practices (Willem & Buelens, 2006). These factors can be presumed to have a negative effect on open data as well, especially since most of them should be more favourable within the public sector then beyond sector boundaries. However, how these factors impact on implementing open data and what the results are, is so far barely understood.
References can be found here: OpenDataMonitor Project – Shared References