Total Process Error: An Approach for Assessing and Monitoring the Quality of Multisource Processes

Rocci, Fabiana; Varriale, Roberta; Luzi, Orietta

doi:10.2478/jos-2022-0025

Cited by 5 publications

(1 citation statement)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In recent years, National Statistical Institutes (NSIs) are progressively moving to the production of official statistics based on the combination of data from different sources, with the aim to reduce costs and response burden while delivering detailed and high-quality information [13]. In this view, new strategies in producing the required outputs need to be developed exploiting as far as possible the integrated use of different data sources, which include data from survey data and administrative sources.…”

Section: Introductionmentioning

confidence: 99%

Multi-source statistics on employment status in Italy, a machine learning approach

Varriale

2023

METRON

Self Cite

View full text Add to dashboard Cite

In recent decades, National Statistical Institutes have started to produce official statistics by exploiting multiple sources of information (multi-source statistics) rather than a single source, usually a statistical survey. In this context, one of the research projects addressed by the Italian National Statistical Institute (Istat) concerned methods for producing estimates on employment in Italy using survey data and administrative sources. The former are drawn from the Labour Force survey conducted by Istat, the latter from several administrative sources that Istat regularly acquires from external bodies. We use machine learning methods to predict the individual employment status. This approach is based on the application of decision tree and random forest techniques, that are frequently used to classify large amounts of data. We show how to construct a “new” response variable denoting agreement of the data sources: this approach is shown to maximise the information we may derive by machine learning approach in some problematic cases. The methods have been applied using the R software.

show abstract

Section: Introductionmentioning

confidence: 99%

Multi-source statistics on employment status in Italy, a machine learning approach

Varriale

2023

METRON

Self Cite

View full text Add to dashboard Cite

show abstract

Augmenting business statistics information by combining traditional data with textual data: a composite indicator approach

Salvatore,

Biffignandi,

Bianchi

2024

METRON

View full text Add to dashboard Cite

Combining traditional and digital trace data is an emerging trend in statistics. In this respect, new data sources represent the basis for multi-purpose extraction of different statistical indicators, which contribute to augmenting the statistical information, for feeding smart statistics. The production of business statistics can benefit from the use of unstructured data, especially to study novel aspects which are not covered by traditional data sources. This paper proposes a methodological general framework for augmenting information by combining data, both structured and non structured. The statistical challenges of using unstructured data and their integration with traditional data are discussed. The methodological general framework is applied to the construction of smart composite indicators using social media data and their metadata. An empirical exercise illustrates how to apply the methodology in practice.

show abstract

John Dunne and Li-Chun Zhang's reply to the Discussion of ‘A system of population estimates compiled from administrative data only'

Dunne,

Zhang

2023

Journal of the Royal Statistical Society Series A: Statistics in Society

View full text Add to dashboard Cite

We are grateful to all those who submitted comments on the paper. Their comments give us an opportunity to revisit some of the key points in our paper, as well as to elaborate on some important issues which we could not cover in sufficient detail in the paper. Specific assumptions and general applicabilityThe three assumptions related to the potential DSE bias are concerned with linkage errors, erroneous records, and heterogeneous list capture probabilities, as commented by Baffour, Chai, Eleanor, Shlomo and Smith. While the first two types of error seem to have their own distinct causes, heterogeneous capture probabilities may be confounded with list dependence. Nevertheless, since the consequence of any list dependence is just heterogeneous capture probabilities, one only needs to focus on whether the capture probability of at least one list is sufficiently constant given the relevant covariates, such as by post-stratification, rather than deliberating whether independence can or cannot be assumed for different register sources. Diagnostics and tests for homogeneous B-list capture probability are useful tools in practice.In reality, where question marks cannot be entirely eradicated for any of the error sources, there are only two logical steps one can take.The first step is to define a strategy, by which some types of error can be reduced to a negligible extent compared to the remaining ones. In the PECADO system, decisions are made to make A-list undercounting the predominant problem that remains, which can be solved by the traditional wisdom of DSE implementation. Whereas the system described by Bernardi et al. sets up EPR overcounting as the main issue to be solved. This shows how different practical approaches are possible following the same strategy of 'reduce-and-conquer', given the data situations in different countries.The second step is to develop adequate methods to support the chosen strategy. For instance, trimming source-registers instead of individual records is found to be useful for the PECADO system. But it may not be feasible if the statistical agency has only limited access to various signs-of-life data sources. Chipperfield et al. ( 2023) investigate capture-recapture estimation given both non-negligible linkage errors and erroneous records. In particular, a record-level score for trimming can be modelled on linking census and administrative spine, such that the trimmed spine containing negligible erroneous records can be used as a list for capture-recapture estimation. Methodological extensionsSeveral needs and possibilities for methodological extensions are mentioned by Abbott, Allin, Eleanor, Henry and O'Farrell, Pfeffermann and Shlomo, which pertain to potential uses of more sophisticated models, disaggregating statistics to more detailed levels including special subpopulations, and producing attribute statistics by similar methods.

show abstract

Total Process Error: An Approach for Assessing and Monitoring the Quality of Multisource Processes

Cited by 5 publications

References 8 publications

Multi-source statistics on employment status in Italy, a machine learning approach

Multi-source statistics on employment status in Italy, a machine learning approach

Augmenting business statistics information by combining traditional data with textual data: a composite indicator approach

John Dunne and Li-Chun Zhang's reply to the Discussion of ‘A system of population estimates compiled from administrative data only'

Contact Info

Product

Resources

About