National Statistics and yearly population census minimum criteria

In 2015, for a National Statistics dataset to qualify, it must have a population census updated at least once a year.

I think only small countries may be able to conduct a national population census every year (or even the implied possibility of doing more than one census per year). The population census in Brazil is made by the National Institute of Geography and Statistic (IBGE) once every ten years. The last one has been in 2010. So, should National Statistics dataset be ruled out of the Open Data Index in all but the year a full census is actually performed? Or are future projections of population admissible for this dataset?


I would personally make the submission to the open data index and add the details in the comments section. I would prefer to have an entry that points me to some data rather than no entry at all.

@herrmann this is great feedback and it will help to refine the dataset definitions and the help text for the next census. Keep the questions coming :smile:


+1 to @Stephen on the comment. Just write everything in the comment sections.
@herrmann - where are all of these submissions for Brazil? I dont see them in the index and the deadline is Monday night.

Thanks for the tips, @Stephen and @Mor.

We find it much easier to research and write the information collectively in a shared document than doing each dataset individually. So we are just finishing writing the submission texts in Etherpad and will submit them in the proper form very soon, in time for the deadline.

1 Like

mmmm, i see your point that small countries should be more able to a population census each year but actually the Isle of Man does the same as the UK, full census in 2011 and every 10 years and interim half census in 2016 every 5 years inbetween the full census

i guess in small countries the net changes in population in absolute numbers are probably less important than in big countries – a million people would need extra budgets etc but a hundred people?

often there are sometimes much more accurate and lower cost proxy indicators anyway such as GP/doctor registrations and appointment in the previous 12 months

1 Like

Having yearly population estimates is a somewhat more plausible idea than doing a full census every year. However, the wording in the methodology (“Population updated at least once a year”) seems to imply that a full yearly census is a requirement for the dataset to be even considered.

I wonder how many countries (if any) would be able to fulfil that strict criteria.


agree with estimates, i dont think any countries could meet the stated actuals criteria

1 Like

I reached out to the elections unit to understand issues.
They explained that in smaller jurisdictions the total budget for venues/police presence/officials overtime or temp recruitment/election mailouts/advertising and likely availability of independent election monitors and willing returning officers (high paid lawyers from the law society volunteering for low expenses!) means that votes at polling stations are transferred to a single constituency counting station and that the process does not tag votes to a specific polling station just an aggregated constituency.
So what would seem like a minor process amendment to sub-total in polling station batches at the constituency counting house and then total the sub-totals is the major requirement to prioritise new legislation in the 5 year legislative programme to draft and pass new legislation (Representation of the People Act).