'Twas the night before census
April 2, 2022 11:48 AM Subscribe
Delayed a year because of Covid, Ireland will be enumerating the people tonight with a Census form [PDF]. It usually happens every 5 years and is substantively the same each go-round - so that comparisons can be more easily made. For the first time, households are being offered the option to send a message to the future; all of which will be sealed for 100 years. These Time Capsules are anything that can be written / drawn in a rectangle 17cm x 12.5cm.
Another progressive first: Q12 changes the default first option for "Religion" from "Roman Catholic" to "No religion".
On the other hand, Q2 What is your sex? Offers only 1 [__] Male 2 [__] Female. The organisers are anticipating some push back on this. Solution: forms which go non-binary will be randomly assigned M or F.
On the time capsule front, the censuses for 1901 and 1911 have been digitised for searching
Another progressive first: Q12 changes the default first option for "Religion" from "Roman Catholic" to "No religion".
On the other hand, Q2 What is your sex? Offers only 1 [__] Male 2 [__] Female. The organisers are anticipating some push back on this. Solution: forms which go non-binary will be randomly assigned M or F.
On the time capsule front, the censuses for 1901 and 1911 have been digitised for searching
Solution: forms which go non-binary will be randomly assigned M or F.
I didn't learn a lot in my statistics classes but doesn't this mess up the math given that you're introducing a bunch of junk data? If I'm AFAB but don't want to check a box, no matter which choice I get randomly assigned, they're both incorrect. So the official numbers will always be off by an unmeasurable number. Is there a math reason why they couldn't just let you leave it blank if neither option is right?
posted by bleep at 12:18 PM on April 2, 2022 [2 favorites]
I didn't learn a lot in my statistics classes but doesn't this mess up the math given that you're introducing a bunch of junk data? If I'm AFAB but don't want to check a box, no matter which choice I get randomly assigned, they're both incorrect. So the official numbers will always be off by an unmeasurable number. Is there a math reason why they couldn't just let you leave it blank if neither option is right?
posted by bleep at 12:18 PM on April 2, 2022 [2 favorites]
Maybe the software requires answers for all questions, and with no third option, it decides to introduce random noise that should filter out to being 50/50 on either side? That's my best guess.
posted by hippybear at 12:46 PM on April 2, 2022 [1 favorite]
posted by hippybear at 12:46 PM on April 2, 2022 [1 favorite]
the censuses for 1901 and 1911 have been digitised for searching
The same goes for the historical Births, Marriages and Deaths registries, which are available at Irish Genealogy. The Catholic parish registers are also available at the National Library of Ireland, but while they’re scanned, they’re not searchable.
It can be fun (or frustrating) trying to match people to their records in the various registries. You’ve both the problem of people seeming to age at different rates, and the fact that naming patterns were pretty unoriginal…
posted by scorbet at 1:22 PM on April 2, 2022 [1 favorite]
The same goes for the historical Births, Marriages and Deaths registries, which are available at Irish Genealogy. The Catholic parish registers are also available at the National Library of Ireland, but while they’re scanned, they’re not searchable.
It can be fun (or frustrating) trying to match people to their records in the various registries. You’ve both the problem of people seeming to age at different rates, and the fact that naming patterns were pretty unoriginal…
posted by scorbet at 1:22 PM on April 2, 2022 [1 favorite]
Thanks for reminding me but it's tomorrow (3rd) not today!!
If you want I will write something in the time capsule bit about metafilter in there
posted by BigCalm at 3:17 PM on April 2, 2022 [6 favorites]
If you want I will write something in the time capsule bit about metafilter in there
posted by BigCalm at 3:17 PM on April 2, 2022 [6 favorites]
The time capsule idea is amazing. It turns data collection into something tangible for families. How cool! Any background on how they came up with the idea?
posted by haptic_avenger at 7:16 PM on April 2, 2022
posted by haptic_avenger at 7:16 PM on April 2, 2022
I would describe the lates technologies available to me, guess on future ones and describe my hopes and dreams.
posted by St. Peepsburg at 8:51 PM on April 2, 2022
posted by St. Peepsburg at 8:51 PM on April 2, 2022
> doesn't this mess up the math given that you're introducing a bunch of junk data
This is an extremely standard approach.
I would consider "all nonbinary people are male" to be an approach that would generate junk data. Using a uniform 50/50 is the most principled choice possible (if you have to sign people M or F) from a statistical perspective.
The Law of Large Numbers tells us that it will tend towards a 50/50 split.
At a high level it will work pretty well, effectively just biasing the estimate towards 50/50.
There are two issues:
-> if NB is a large fraction of the data, this will work actually very well; literally nothing will be correlated with gender
-> for small slices of the data, (e.g. a small town with a large NB population), you could see a lot of variability.
If NB is 5% of the data, I would consider that to have minimal impact on the validity of the data, and adjust the tolerances a little bit for slice analysis.
posted by constraint at 9:18 PM on April 2, 2022 [1 favorite]
This is an extremely standard approach.
I would consider "all nonbinary people are male" to be an approach that would generate junk data. Using a uniform 50/50 is the most principled choice possible (if you have to sign people M or F) from a statistical perspective.
The Law of Large Numbers tells us that it will tend towards a 50/50 split.
At a high level it will work pretty well, effectively just biasing the estimate towards 50/50.
There are two issues:
-> if NB is a large fraction of the data, this will work actually very well; literally nothing will be correlated with gender
-> for small slices of the data, (e.g. a small town with a large NB population), you could see a lot of variability.
If NB is 5% of the data, I would consider that to have minimal impact on the validity of the data, and adjust the tolerances a little bit for slice analysis.
posted by constraint at 9:18 PM on April 2, 2022 [1 favorite]
I hope someone takes the opportunity to draw a little comic strip in the Time Capsule box.
posted by Paul Slade at 11:49 PM on April 2, 2022
posted by Paul Slade at 11:49 PM on April 2, 2022
... or perhaps to write a full short story in very small writing. Why not plant the clues for a puzzle to be solved in 100 years time? Or offer your own lyrics for a century-spanning collaboration with a musician of the future? So many creative possibilities ....
posted by Paul Slade at 12:09 AM on April 3, 2022
posted by Paul Slade at 12:09 AM on April 3, 2022
Paul Slade: Why not plant the clues for a puzzle to be solved in 100 years time?
On it! My plan is to bury a Centenary Kist containing useful things from today [honey bee, teaspoon, wood-screws] which may not {have a purpose | be available} in 2122 and use the space on the census form to describe the location. GPS may not exist in 2022 per Ask, but Lat and Long will and a hyper-local map will be required for the details. Living rural in the Era of Backhoe it is non-trivial to pick a location confident in its unchanginghood.
posted by BobTheScientist at 1:06 AM on April 3, 2022 [2 favorites]
On it! My plan is to bury a Centenary Kist containing useful things from today [honey bee, teaspoon, wood-screws] which may not {have a purpose | be available} in 2122 and use the space on the census form to describe the location. GPS may not exist in 2022 per Ask, but Lat and Long will and a hyper-local map will be required for the details. Living rural in the Era of Backhoe it is non-trivial to pick a location confident in its unchanginghood.
posted by BobTheScientist at 1:06 AM on April 3, 2022 [2 favorites]
I'm going to write, "I bet generalized AI is only 5 years away".
posted by night_train at 2:09 AM on April 3, 2022 [3 favorites]
posted by night_train at 2:09 AM on April 3, 2022 [3 favorites]
Is there a math reason why they couldn't just let you leave it blank if neither option is right?
There’s a legal reason, allegedly. Per Report on the public consultation on content of Census 2021 and the Census Pilot Survey 2018 , Appendix 4,
There’s a legal reason, allegedly. Per Report on the public consultation on content of Census 2021 and the Census Pilot Survey 2018 , Appendix 4,
Several submissions were received during public consultation requesting the inclusion of questions on both gender identity and sexual orientation. These are complex emerging topics in the field of statistical data collection and as yet there is no international consensus on whether they should be incorporated in a census, how the questions should be framed and how classifications could be affected by the mode of data collection. Questions on both topics may also give rise to statistical disclosure control issues. These often arise for National Statistical Institutes when considering publishing data on relatively small population cohorts at small area levels, a defining geography for census data. Similar confidentiality issues may also arise when preparing cross tabular outputs.posted by zamboni at 6:36 AM on April 3, 2022 [1 favorite]
With regards to gender identity, the United Nations Economic Commission for Europe (UNECE) in its publication Recommendations for the 2020 Censuses of Population and Housing5 discusses the potential impact of a gender identity question on the responses to the sex question and advocates a rigorous testing programme before including the topic in a census. From a legal perspective, there is an EU requirement on CSO to collect census data on sex with a binary classification scheme but no requirement as yet to collect data on either gender identity or sexual orientation.
These considerations were raised by CSO and discussed at the Census Advisory Group meetings and neither topic has been recommended for inclusion on Census 2021. However, CSO recognises the importance of developing its capacity to collect and produce data on both topics and it is committed to doing so. The office introduced new questions on both gender identity and sexual orientation in its General Household Survey in the first quarter of 2019. The first release from this iteration of the survey was published in July 2019. The CSO will monitor the collection and production of data on both gender identity and sexual orientation with a view to developing questions which are well understood and acceptable to users, stakeholders and respondents and produce confidential, robust and consistent data. When the questions are suitably developed, they will be included in CSO’s larger household surveys and ultimately will be considered for Pilot testing for a full census. The inclusion of both topics in the household surveys will also facilitate the production of data at a national and regional level.
Richard Chambers collected some of the things that people were planning to write on twitter here. (There's a typical Irish mix of thoughtful ideas, complaints about the government, and Father Ted quotes.)
TwistedDoodles also had a comic for the occasion.
posted by scorbet at 2:23 AM on April 4, 2022
TwistedDoodles also had a comic for the occasion.
posted by scorbet at 2:23 AM on April 4, 2022
Solution: forms which go non-binary will be randomly assigned M or F.
To whoever set this up so that this is the option: I cordially invite you to FOAD. I am nonbinary. That doesn't mean that I'm male, or female, or that you can choose whichever of those options you consider more convenient. For me, it means NOT.
posted by Lexica at 3:26 PM on April 4, 2022 [2 favorites]
To whoever set this up so that this is the option: I cordially invite you to FOAD. I am nonbinary. That doesn't mean that I'm male, or female, or that you can choose whichever of those options you consider more convenient. For me, it means NOT.
posted by Lexica at 3:26 PM on April 4, 2022 [2 favorites]
« Older THE HEART COMES WITH CALF-HEARTS FOR A REASON. | The new e-Sport? Newer »
This thread has been archived and is closed to new comments
posted by GallonOfAlan at 12:06 PM on April 2, 2022 [4 favorites]