Discussion Topics

Topic 1 - Data collection in relation to timeliness of output (in COVID times).

Can purposive samples be quickly set up?

Can already available data sources, even if there is some selectivity, be used?

Can new data sources be combined with existing data sources?

If timeliness is so crucial, are greater margins of error acceptable?

What other options are available to increase timeliness: are there institutes using now casting?

Topic 2. Handling selectivity in available data sets/use of non-probability samples.

Examples of such selective sets are web-scraping/online data.

Are there institutes using opt-in web surveys of businesses?

When non-probability samples are linked to a business register (if possible) selectivity in the data sets can be identified. What would be the best ways to deal with such selectivity in the case of business statistics?

Can these data sources be combined with probability samples? Are background variables available to correct for this selectivity?

Topic 3: Regionalisation of statistics.

Data on business statistics are often measured using enterprise (group) data, what if figures partitioned to regions are required?

Can they be measured at the establishment level?

Are model-based estimates being used?

How can they be verified?


Sponsored by: