Data quality is an increasingly critical issue in the majority of information intensive businesses. Indeed, nowadays, in the information era, data quality is a crucial requirement for modern enterprises to obtain a competitive advantage.

In this thesis, a particular and widespread context characterization has been considered: multichannel cooperative information systems. It combines two adopted systems typical of the modern economy: multichannel information systems and cooperative information systems. The former classifies an organization considering its own architecture while the latter classifies the organization as part of a larger system considering its role in the interaction with other organizations. In such context, information quality plays a key role both inside organizations, as it reduces errors and increases process efficiency, and in inter-organizational relationships, as it represents a critical component of each company’s trustworthiness. Indeed, as regards multichannel information systems, each organization accesses its internal data and the primary goal of data quality assurance is the assessment and monitoring of data values and, possibly, their improvement. In cooperative information systems (CIS), data quality assurance is important in communication processes and data quality values should be exchanged along with data.

In order to assure high data quality levels, companies must design and implement a complete data quality management program. In the literature, models for data quality management have been proposed. A big limit in the application of the proposed methodologies is that they only provide guidelines without providing tools or specific algorithms that support users in the assessment, monitoring, and improvement phases. Focusing on the critical issues related to the assessment phase, the literature does not provide an exhaustive set of metrics that organizations can apply. Only a few algorithms have been developed for a subset of dimensions, such as accuracy [53], completeness [44], consistency [42], and timeliness [5]. Quality assurance is instead faced by the need for objective measures of quality, since most users cannot judge the quality of data and simply have to trust data sources.

Along these considerations, the first contribution of this thesis is the design of an architectural component, called Quality Factory that supports the quality assessment, monitoring, and improvement of internal
data (see Chapter 4). The proposal of an architecture that enables the adoption of a total data quality management aims at filling the lack of semi-automatic tools to evaluate and improve data quality. It is not possible to provide completely computer-based procedures for data quality management, but it is at least possible to limit the human interventions to control and monitoring activities. This is achieved, in the thesis, by applying a rule-based methodology that manages the activity and the interactions of the modules that compose the Quality Factory.

Specifically, the Quality Factory supports the users offering a complete set of tools for a comprehensive data quality program that combines some of the successful assessment techniques and algorithms presented in the literature with new algorithms and techniques. For example, a mathematical model to evaluate the impact of the architectural choices on data quality in multichannel environment has been developed. The results can be used as support of decisions related to the definition of design criteria for the choice of the most appropriate integration strategy depending on the level of data quality that the enterprise would achieve.

Furthermore, considering that the importance of exchanged data intensifies in a cooperative information systems, a fundamental and innovative functionality designed with the Quality Factory is the generation of a Quality Certificate that is associated with data when they satisfy quality requirements. Fundamentally, the user has the possibility to agree with the organization on quality levels specified through the list of acceptable values associated with the data quality dimensions. The organization undertakes to provide the specified quality level to the users and notifies it with the certification process. The certificate assumes great significance in the information exchange among different organizations and particularly between the organization and the data users.

Note that in order to evaluate with which degree data meet quality requirements, it would be necessary to assess data along the process in which they are involved and under the user’s perspective. In the thesis, a model in which the assessment algorithm varies on the basis of user profiling and personal requirements is also proposed.

Finally, to the best of our knowledge, in the context of multichannel cooperative information systems, the data quality problem has not been considered yet both from a methodological point of view and from an architectural perspective. This thesis aims at providing a first contribution in this specific research field.

Data quality is a critical success factor to many activities of the information age, including the development and operation of a data warehouse (Wixom and Watson 2001). The issue of data quality is recognized in the development of the warehouse; however, there is no formal methodological approach to dealing with the quality issues.

This research was targeted at the development of a framework and relational database tool to improve the process of making data quality decisions in the creation of a data warehouse. A framework for examining data quality issues, the Data Quality Analysis Framework (DQAF), is proposed. An associated tool for the collection of metadata, the Data Quality Knowledge Management (DQKM) tool, is also proposed. An exploratory study was conducted to determine if data gathered using the DQKM would support resource allocation decisions for data quality efforts.

The research presented in this paper reflects two processes: the development of the DQAF and DQKM, and the exploratory study of the tool. The process involved analyzing qualitative data from a series of interviews and quantitative data from the results of the exploratory studies. It involved the construction and population of the DQKM, based on data obtained from a case study conducted at the Center for Technology in Government. Thus, the work reflects a multi-method approach, providing a result that has practical value as well as academic rigor. Additionally, it lays the groundwork for a stream of research that is expected to last many years.

This research contributes to knowledge in three areas: 1) It addresses the need for a methodological approach for assessing data quality in the context of a data warehouse project; 2) It provides a tool for effectively capturing and managing metadata in the complex environment that results when integrating multiple data sources; 3) It builds on the concept of fitness for use, providing a mechanism for allocating resources for data quality projects based on the interaction of the data field, the data quality dimension, and the use of the data.

The dissertation investigates information quality in multi-actor service networks. A clear connection between the quality of information and success of business has been acknowledged, but tools for analysing information quality in network environments on the basis of qualitative data have been lacking. There is also a limited understanding of information processes of virtual networks of public and private service organizations in the literature.

In this dissertation, a novel framework for information quality analysis is developed and operationalized. It extends previously developed methods and provides a fundamentally different way to assess information quality, contrary to earlier quantitative studies. In addition to investigating and developing management of information quality, the dissertation focuses on collaboration and networking within the heterogeneous multi-actor service networks.

Operationalization of the newly developed framework for information quality analysis is undertaken in the virtual network environment of safety telephone services for ageing people. These services utilize rapidly developing well-being technology. The analysis is based on data from interviews with professionals working in several service networks of different types and sizes.

The results provide a detailed account of the state of information quality and network collaboration in the case networks. The results can be utilized as guidelines when planning information-related matters in the case networks in the future. Practical recommendations for the branch of safety telephone services are formulated. The dissertation also contains a thorough assessment of the usability of the framework of analysis. Suggestions concerning future use of the framework are formulated. The dissertation thus contributes to development of new tools for analysing information quality. It also suggests directions for future research.

Quality information is critical to organisations’ success in today’s highly competitive environment. Accounting information systems (AIS) as a discipline within information systems require high quality data. However, empirical evidence suggests that data quality is problematic in AIS. Therefore, knowledge of critical factors that are important in ensuring data quality in accounting information systems is desirable.

A literature review evaluates previous research work in quality management, data quality, and accounting information systems. It was found that there was a gap in the literature about critical success factors for data quality in accounting information systems. Based on this gap in the literature and the findings of the exploratory stage of the research, a preliminary research model for factors influence data quality in AIS was developed. A framework for understanding relationships between stakeholder groups and data quality in accounting information systems was also developed. The major stakeholders are information producers, information custodians, information managers, information users, and internal auditors.

Case study and survey methodology were adopted for this research. Case studies in seven Australian organisations were carried out, where four of them were large organisations and the other three are small to medium organisations (SMEs). Each case was examined as a whole to obtain an understanding of the opinions and perspectives of the respondents from each individual organisation as to what are considered to be the important factors in the case. Then, cross-case analysis was used to analyze the similarities and differences of the seven cases, which also include the variations between large organisations and small to medium organisations (SMEs). Furthermore, the variations between five different stakeholder groups were also examined. The results of the seven main case studies suggested 26 factors that may have impact on data quality in AIS.

Survey instrument was developed based on the findings from case studies. Two large-scale surveys were sent to selected members of Australian CPA, and Australian Computer Society to further develop and test the research framework. The major findings from the survey are: 1. respondents rated the importance of the factors II consistent higher than the actual performance of those factors. 2. There was only one factor, ‘audit and reviews’, that was found to be different between different sized organisations. 3. Four factors were found to be significantly different between different stakeholder groups: user focus, measurement and reporting, data supplier quality management and audit and reviews. 4. The top three critical factors for ensuring data quality in AIS were: top management commitment, education and training, and the nature of the accounting information systems.

The key contribution of this thesis is the theoretical framework developed from the analysis of the findings of this research, which is the first such framework built upon empirical study that explored factors influencing data quality in AIS and their interrelationships with stakeholder groups and data quality outcomes. That is, it is now clear which factors impact on data quality in AIS, and which of those factors are critical success factors for ensuring high quality information outcomes. In addition, the performance level of factors was also incorporated into the research framework. Since the actual performance of factors has not been highlighted in other studies, this research adds new theoretical insights to the extant literature. In turn, this research confirms some of the factors mentioned in the literature and adds a few new factors. Moreover, stakeholder groups of data quality in AIS are important considerations and need more attention. The research framework of this research shows the relationship between stakeholder groups, important factors and data quality outcomes by highlighting stakeholder groups’ influence on identifying the important factors, as well as the evaluation of the importance and performance of the factors.

This research began a journey towards improved maturity around data quality management in New Zealand health care, where total data quality management is ‘business as usual’, institutionalised into the daily practices of all those who work in health care. The increasingly information intensive nature of health care demands a proactive and strategic approach to data quality to ensure the right information is available to the right person at the right time in the right format, all in consideration of the rights of the patient to have his/her health data protected and used in an ethical way. The work extends and tests principles to establish good practice and overcome practical barriers. This thesis explores the issues that define and control data quality in the national health data collections and the mechanisms and frameworks that can be developed to achieve and sustain good data quality.

The research is interpretive, studying meaning within a social setting. The research provides the structure for learning and potential change through the utilisation of action research. Grounded theory provides the structure for the analysis of qualitative data through inductive coding and constant comparison in the analysis phase of the action research iterative cycle. Participatory observation provided considerable rich data as the researcher was a member of staff within the organisation. Data were also collected at workshops, focus groups, structured meetings and interviews.

The development of a Data Quality Evaluation Framework and a national Data Quality Improvement Strategy provides clear direction for a holistic and ‘whole of health sector’ way of viewing data quality, with the ability for organisations to develop and implement local innovations through locally developed strategies and data quality improvement programmes. The researcher utilised the theory of appreciative enquiry (Fry, 2002) to positively encourage change, and to encourage the utilisation of existing organisational knowledge. Simple rules, such as the TDQM process and the data quality dimensions guided the change, leaving room for innovation. The theory of ‘complex systems of adjustment’ (Champagne, 2002; Stacey, 1993) can be instilled in the organisation to encourage change through the constant interaction of people throughout the organisation.

The relationship between information and decision-making is a complex one and has been the subject of extensive research spanning several decades, with roots in nineteenth and early twentieth century economic theories. More recently, researchers have suggested a relationship between the quality of information and the quality of decision-making, with a consequent relationship with organizational strategy; however, there has been very little research in which this relationship was investigated systematically. This research set forth contextual and conceptual models relating information quality to strategy and then provided an empirical investigation of the relationship between information quality and organizational outcomes, with information intensity hypothesized as a moderator of that relationship. Data for this study were collected through a Web-based survey of individuals associated with an industry consortium and were evaluated through a combination of multiple regression analysis, moderated regression analysis, and subgroup analysis. Data analysis revealed evidence that the relationship between the quality of information and organizational outcomes is systematically measurable, in that measurements of information quality can be used to predict organizational outcomes, and that this relationship is, for the most part, positive. An unexpected finding was that different regression models emerge when stakeholder roles in an information system are taken into consideration. Data analysis did not reveal support for the hypothesis that information intensity moderates the relationship between information quality and organizational outcomes.

The recent progress in information and communication technologies is encouraging diversity in the utilization of existing data. Data are now routinely pooled from multiple systems and physical locations, and integrated in creative ways for various decision-making purposes. From a managerial perspective, however, there are growing concerns in regard to the quality of the output information, and the economic justification of costly investments in new technologies.

The major part of this dissertation addresses these concerns through formal studies on the quality and value of information, based on information economics (IE) theory. The quality and value of information integration is studied from a standpoint that recognizes the fundamental role of information integration in information systems. The objective of this study is to create a domain-independent theoretical framework that can facilitate decision making on information integration. The framework classifies information integration situations using two information quality characteristics: informativeness and dependence; and links different conditions in terms of these characteristics with different predictions on the value of integration.

A second, related study centers on the questions of whether improving the accuracy of the input of an information system guarantees higher accuracy and economic value of its output, especially higher accuracy and economic value of forecasts. The study offers sufficient conditions under which the answer to these questions is positive, and also presents counter examples that suggest conditions under which the answer is negative. The results point to a contextual factor that can affect accuracy both ways: positive or negative, which has been ignored by data quality theory. This factor is dependence between errors.

Duplicate detection consists in detecting multiple representations of a same real-world object, and that for every object represented in a data source. Duplicate detection is relevant in data cleaning and data integration applications and has been studied extensively for relational data describing a single type of object in a single table. Our research focuses on iterative duplicate detection in XML data. We consider detecting duplicates in multiple types of objects related to each other and devise methods adapted to semi-structured XML data. Relationships between different types of objects either form a hierarchical structure (given by the hierarchical XML structure) or a graph structure (e.g.,given by referential constraints).

Iterative duplicate detection require a similarity measure to compare pairs of object representations, called candidates, based on descriptive information of a candidate. The distinction between candidates and their description is not straightforward in XML, but we show that we can semi-automatically determine these descriptions using heuristics and conditions.

Second, we define a similarity measure that is suited for XML data and that considers relationships between candidates. It considers data comparability, data relevance, data similarity, and distinguishes between missing and contradictory data. Experimental evaluation shows that our similarity measure outperforms existing similarity measures in terms of effectiveness.

To avoid pairwise comparisons and thereby improve efficiency, but without compromising effectiveness, we propose three comparison strategies: The topdown algorithm is suited for hierarchically related candidates where nesting represents 1:N relationships, whereas the bottom-up algorithm is suited when nested candidates actually exist in anM:N relationships in the real world. When candidate relationships form a graph, the Reconsidering Algorithm re-compares candidate pairs to improve effectiveness. Using a comparison order that reduces the number of re-comparisons nevertheless allows efficient and effective duplicate detection.

To scale to large amounts of data, we propose methods that interact with a database to handle retrieval, classification, and update of iteratively compared candidate pairs. Empirical evaluation shows that these methods scale linearly in time with the number of candidates, the connectivity of the graph, and the fraction of duplicates among all candidates. We are the first to obtain linear behavior for all these parameters, so, in summary, this thesis presents novel methods for effective, efficient, and scalable duplicate detection in graph, and in particular in XML data.

Finally, we present XClean, the first system for declarative XML data cleaning, for which we defined operators and a specification language that is compiled to XQuery.

Pervasive environments generate large quantities of data, originating from backend servers, portable devices and wireless mobile sensors. Pervasive sensing devices that monitor properties of the environment (including human beings) can produce a large data source. The unprocessed datasets may include data that is faulty and irrelevant, and data that is important and useful. If not managed correctly the large amount of data from a data-rich pervasive environment may result in information overload or delivery of incorrect information.

Context-sensitive quality data management aims to gather, verify, process, and manage the multiple data sources in a pervasive environment in order to deliver high quality, relevant information to the end user. Managing the quality of data from different sources, correlating related data and making use of context are all essential in providing end users with accurate and meaningful data in real-time. This requirement is especially true for critical applications such as in a medical environment.

This thesis presents the Data Management System (DMS) architecture. It is designed to deliver a quality of data service to its users. The DMS architecture employs an agent based middleware to intelligently and effectively manage all pervasive data sources, and to make use of context to deliver relevant information to the end-user. DMS components have been designed to manage: data validation; data consistency; context-based data delivery; knowledge management and distributed processing. The DMS components have been rigorously evaluated using various medical-based test cases.

This thesis demonstrates a careful, precise approach to data based on the quality of the data and the context of its use. It emphasises the DMS architecture and the role of software agents in providing quality data management.

The thesis is an inductive investigation of how the users of information make value-judgments about the content they encounter and retrieve from the Web. Specifically, it examines perceptions of IQ from the perspective of eighty "academic" high-end users, who regularly engage the Web and its search engines to search for and retrieve high-quality information related to their research, teaching and learning.

User results are explored and discussed in the context of a novel investigative framework, the "Combined Conceptual Life-Cycle" (CC/LC) model of IQ, which enables researchers to develop a more accurate research lens through which to examine user/information interaction and perceptions of IQ.

The findings associated with the thesis are also consistent with the proposal of a new "Ongoing Technology Acceptance Model" (OTAM), which facilitates the measurement of users' perception of the predictability of their technology interactions, and has the capacity to more accurately investigate user individual differences, and feedback mechanisms involved with user/system interaction.

Modern experimental science, or e-science, increasingly relies upon the use of information integration and analysis techniques to achieve its results. A central requirement in e-science is that the necessary data and service resources be contributed by many parties within a scientific community, trascending the boundaries of individual labs. This scenario bears the promise of reducing the overall cost of science by encouraging the reuse of scientific information on a large scale. At the same time, however, there is a risk that data of poor quality, resulting for example from inaccurate experiments, may propagate out of control and contaminate other experiments. To compound the problem, the fast-paced evolution of the experimental techniques is making it difficult to investigate and standardise methods of quality control. Quality assurance for e-science information is therefore an important and largely open problem.

In this thesis we argue that user scientists should play a central role in ensuring that the third-party information they wish to use is of acceptable quality. This is difficult, however, because users, who are not part of the information production process, are often left to estimate quality of data using empirical rules that are based on limited and indirect evidence of correctness. This results in implicit quality control rules being applied in a bespoke and intuitive fashion, if at all. Our research hypothesis is that these quality rules for data acceptability are a form of latent knowledge, for which we have coined the term quality knowledge, where objective measures overlap with the scientists’ subjective propensity to the risk of using errouneous data. In the thesis we investigate ways to make such quality knowledge explicit, and to exploit it in order to make e-science experiments quality aware in a principled way. Our main result is a model and architecture for Quality Views, i.e., quality processes that embody the user scientists’ personal criteria for data acceptability. We show that, with appropriate support from software tools, Quality Views can become reusable quality components that are easily integrated into e-science experiments.

Similar to a supply chain, an information supply chain is a dynamic environment where networks of information-sharing agents gather data from many sources and utilize the same data for different tasks. Unfortunately, raw data arriving from a variety of sources is often plagued by errors (Ballou et al. 1998), which can lead to poor decision making. Supporting decision making in this challenging environment demands a proactive approach to data quality management, since the decision maker has no control over these data sources (Shankaranarayan et al. 2003).

This is true in health care, and in particular in health planning, where health care resource allocation is often based on summarized data from a myriad of sources such as hospital admissions, vital statistic records, and specific disease registries. These data are utilized to justify investments in services, reduce inequities in treatment, and rank health care problems to support policy formulation (Berndt et al. 2003). This research targets the development of metrics that communicate data quality information at decision time that can be calculated when the final information product is created. To design and evaluate the result-driven data quality metrics this thesis utilizes the design science paradigm (Simon 1996; Hevner, March et al. 2004). A field study conducted in a Florida Health Planning Agency helps identify potential data quality measures and biases in the health planning context. The data quality metrics are designed and implemented with simple Online Analytical Processing interfaces in order to present these metrics to decision makers. The metrics were evaluated using two types of focus groups: exploratory and confirmatory.

The unallocated data incompleteness (UDI) metric considers the effects of null values in any of the grouping or filtering variables, providing an operational definition for aspects of incompleteness. Information volatility metric measures the rate of change in the values of stored data and is a measure of reliability. The final metric is not a calculation, but rather investigates drawing the attention of a decision maker to aggregated data based on small sample sizes to mitigate the bias of insensitivity to sample size.