Proposal Description

The development in technology and methods, and the use of new data sources put new demands and expectations on producer of official statistics regarding transparency and scientific integrity. These new sources and new expectations have led to prominent controversies in numerous countries, regarding, e.g., trade-offs among privacy, accuracy and granularity; the reproducibility and replicability of results; and assertions regarding causality based on limited observational data. These controversies in turn have led to a deep reconsideration of the best ways in which to ensure transparency and scientific integrity in the production, dissemination and use of official statistics. This session provides an in-depth review of those issues, with emphasis on practical approaches. In addition, the session highlights several ways in in which the development of artificial intelligence and machine learning have led to special challenges and opportunities in work with transparency and scientific integrity.


Paper 1 directs special attention to the alignment of transparent reporting with specific stakeholder priorities for enhancement of scientific integrity. Different priorities naturally lead to emphasis on different levels of the design of production, dissemination, and usage processes, which in turn correspond to differing degrees of emphasis on multiple criteria for transparency and scientific integrity.

Paper 2 explores Statistics Canada’s evolution in transparency considering both a priori and ex post facto approaches to transparent communication. The presentation will explain the Canadian context of social acceptability evolution and describe the related frameworks adopted. It will also illustrate their use by means of concrete examples.

Paper 3 discusses navigation of ethical considerations in data collection, in ways that balance technological advancements, respondent trust and data regulations. In particular, this paper aims to explore the potential impact of AI and the General Data Protection Regulation (GDPR) on individuals' willingness to participate in surveys, highlighting both positive and negative aspects.

Paper 4 highlights the use of technology to improve access to official statistics, while maintaining transparency and scientific integrity. Special emphasis is placed on work to develop an AI powered data access tool; and the importance of leveraging the power of the tool and underlying information to ensure that the user obtains the correct data and metadata to support their specific use.