On Wednesday, February 14, 2024 at 4:57:38 PM UTC+11, Sylvia Else wrote:quote a number.
<https://www.abs.gov.au/statistics/health/causes-death/provisional-mortality-statistics/latest-release>
I just need a sanity check before I raise this with the relevant agency.
If you scroll down to "Age specific rates, 2023, 2022, Baseline", and
look at the three right hand columns.
How can the 2023 figures for each age group be less than the
corresponding baseline average, but the all-ages number be greater than
the baseline average?
If any 2023 age group were above the baseline average, then all-ages
number could go either way, because of different total populations in
each age group, but with all age groups being below the baseline
average, I just don't see it.
It's simple enough. The Covid-19 epidemic killed off quite a few vulnerable people, and the survivors represent samples from the more robust elements of the population.
Years ago it was claimed that males who survived past 80 in good health and women who survived past 85 in good health were predominantly drawn from that more robust population.
It's a statistician's play ground. In practical terms you can't work out precisely what population you belong to - you might just be a lucky snowflake.
An Australian male has an average life expectancy of 81.2 years, so at 81 and a couple of months I might be expected to drop dead soon. As an Australian male of 81, I've actually
got a life expectancy of 8.44 more years. In reality, as an 81-year-old Australian male who never smoked and managed to get a Ph.D. I'm a member of an even longer-lived cohort, but they don't split the life expectation table finely enough that I can
This seems to happen not just for both sexes, but for each sex individually.
Both sexes have been selected in much the same way.
--
Bill Sloman, Sydney
<https://www.abs.gov.au/statistics/health/causes-death/provisional-mortality-statistics/latest-release>
I just need a sanity check before I raise this with the relevant agency.
If you scroll down to "Age specific rates, 2023, 2022, Baseline", and
look at the three right hand columns.
How can the 2023 figures for each age group be less than the
corresponding baseline average, but the all-ages number be greater than
the baseline average?
If any 2023 age group were above the baseline average, then all-ages
number could go either way, because of different total populations in
each age group, but with all age groups being below the baseline
average, I just don't see it.
This seems to happen not just for both sexes, but for each sex individually.
Sylvia.
<https://www.abs.gov.au/statistics/health/causes-death/provisional-mortality-statistics/latest-release>
I just need a sanity check before I raise this with the relevant agency.
If you scroll down to "Age specific rates, 2023, 2022, Baseline", and
look at the three right hand columns.
How can the 2023 figures for each age group be less than the
corresponding baseline average, but the all-ages number be greater than
the baseline average?
If any 2023 age group were above the baseline average, then all-ages
number could go either way, because of different total populations in
each age group, but with all age groups being below the baseline
average, I just don't see it.
This seems to happen not just for both sexes, but for each sex
individually.
Sylvia.
On 14/02/2024 05:57, Sylvia Else wrote:
<https://www.abs.gov.au/statistics/health/causes-death/provisional-mortality-statistics/latest-release>
I just need a sanity check before I raise this with the relevant
agency. If you scroll down to "Age specific rates, 2023, 2022,
Baseline", and look at the three right hand columns.
I see what you mean. It doesn't seem to make any sense.
All the numbers for Jan-Sept 2023 are below long term average (because a proportion of those who would naturally have died in 2023 were killed permaturely by Covid). The column total is above the long term average.
You are right - this data makes no sense at all.
How can the 2023 figures for each age group be less than the
corresponding baseline average, but the all-ages number be greater
than the baseline average?
It can't even with really weird population ratings.
If any 2023 age group were above the baseline average, then all-ages
number could go either way, because of different total populations in
each age group, but with all age groups being below the baseline
average, I just don't see it.
This seems to happen not just for both sexes, but for each sex
individually.
Sylvia.
I'll hazard a guess that some of the numbers presented there are summed
over a different period so that the larger numbers are not correct.
Spreadsheets allow people to make very creative mistakes!
On 15-Feb-24 1:41 am, Martin Brown wrote:
On 14/02/2024 05:57, Sylvia Else wrote:
<https://www.abs.gov.au/statistics/health/causes-death/provisional-mortality-statistics/latest-release>
I just need a sanity check before I raise this with the relevant
agency. If you scroll down to "Age specific rates, 2023, 2022,
Baseline", and look at the three right hand columns.
I see what you mean. It doesn't seem to make any sense.
All the numbers for Jan-Sept 2023 are below long term average (because a
proportion of those who would naturally have died in 2023 were killed
permaturely by Covid). The column total is above the long term average.
You are right - this data makes no sense at all.
How can the 2023 figures for each age group be less than the
corresponding baseline average, but the all-ages number be greater
than the baseline average?
It can't even with really weird population ratings.
If any 2023 age group were above the baseline average, then all-ages
number could go either way, because of different total populations in
each age group, but with all age groups being below the baseline
average, I just don't see it.
This seems to happen not just for both sexes, but for each sex
individually.
Sylvia.
I'll hazard a guess that some of the numbers presented there are summed
over a different period so that the larger numbers are not correct.
Spreadsheets allow people to make very creative mistakes!
Thanks.
There's a update due in a couple of weeks. I'll wait to see whether that >suffers the same issue, and contact them if so. Otherwise I'll just
assume it was a one-off error.
Sylvia.
On Wed, 14 Feb 2024 16:57:30 +1100, Sylvia Else <sylvia@email.invalid>
wrote:
<https://www.abs.gov.au/statistics/health/causes-death/provisional-mortality-statistics/latest-release>
I just need a sanity check before I raise this with the relevant agency.
If you scroll down to "Age specific rates, 2023, 2022, Baseline", and
look at the three right hand columns.
How can the 2023 figures for each age group be less than the
corresponding baseline average, but the all-ages number be greater than
the baseline average?
Age specific rates, 2023, 2022, Baseline
I read all rates higher than baseline except in 85+.
. . . . including all-age.
Monthly rates also higher all months but one.
If any 2023 age group were above the baseline average, then all-ages
number could go either way, because of different total populations in
each age group, but with all age groups being below the baseline
average, I just don't see it.
This seems to happen not just for both sexes, but for each sex individually. >>
Sylvia.
On Thu, 15 Feb 2024 16:22:27 +1100, Sylvia Else <sylvia@email.invalid>
wrote:
On 15-Feb-24 1:41 am, Martin Brown wrote:
On 14/02/2024 05:57, Sylvia Else wrote:
<https://www.abs.gov.au/statistics/health/causes-death/provisional-mortality-statistics/latest-release>
I just need a sanity check before I raise this with the relevant
agency. If you scroll down to "Age specific rates, 2023, 2022,
Baseline", and look at the three right hand columns.
I see what you mean. It doesn't seem to make any sense.
All the numbers for Jan-Sept 2023 are below long term average (because a >>> proportion of those who would naturally have died in 2023 were killed
permaturely by Covid). The column total is above the long term average.
You are right - this data makes no sense at all.
How can the 2023 figures for each age group be less than the
corresponding baseline average, but the all-ages number be greater
than the baseline average?
It can't even with really weird population ratings.
If any 2023 age group were above the baseline average, then all-ages
number could go either way, because of different total populations in
each age group, but with all age groups being below the baseline
average, I just don't see it.
This seems to happen not just for both sexes, but for each sex
individually.
Sylvia.
I'll hazard a guess that some of the numbers presented there are summed
over a different period so that the larger numbers are not correct.
Spreadsheets allow people to make very creative mistakes!
Thanks.
There's a update due in a couple of weeks. I'll wait to see whether that
suffers the same issue, and contact them if so. Otherwise I'll just
assume it was a one-off error.
Sylvia.
Worried about the Covid jabs you had then, Sylvia?
<https://www.abs.gov.au/statistics/health/causes-death/provisional-mortality-statistics/latest-release>
I just need a sanity check before I raise this with the relevant agency.
If you scroll down to "Age specific rates, 2023, 2022, Baseline", and
look at the three right hand columns.
How can the 2023 figures for each age group be less than the
corresponding baseline average, but the all-ages number be greater than
the baseline average?
If any 2023 age group were above the baseline average, then all-ages
number could go either way, because of different total populations in
each age group, but with all age groups being below the baseline
average, I just don't see it.
This seems to happen not just for both sexes, but for each sex
individually.
Sylvia.
On 14-Feb-24 4:57 pm, Sylvia Else wrote:
<https://www.abs.gov.au/statistics/health/causes-death/provisional-mortality-statistics/latest-release>
I just need a sanity check before I raise this with the relevant agency.
If you scroll down to "Age specific rates, 2023, 2022, Baseline", and
look at the three right hand columns.
How can the 2023 figures for each age group be less than the
corresponding baseline average, but the all-ages number be greater than
the baseline average?
If any 2023 age group were above the baseline average, then all-ages
number could go either way, because of different total populations in
each age group, but with all age groups being below the baseline
average, I just don't see it.
This seems to happen not just for both sexes, but for each sex
individually.
Sylvia.
I should just add that I thought I'd try to prove this impossible using algebra, and failed totally.
So then I used Excel, and was able to construct a counter example. So my intuition about this was wrong. Of particular significance appears to be
that the population numbers for the base line age groups will not
generally be the same as the population numbers for the current data age groups.
Oh well. Thanks to those who looked at this.
Sylvia.
On 2024-02-20, Sylvia Else <sylvia@email.invalid> wrote:
On 14-Feb-24 4:57 pm, Sylvia Else wrote:
<https://www.abs.gov.au/statistics/health/causes-death/provisional-mortality-statistics/latest-release>
I just need a sanity check before I raise this with the relevant agency. >>> If you scroll down to "Age specific rates, 2023, 2022, Baseline", and
look at the three right hand columns.
How can the 2023 figures for each age group be less than the
corresponding baseline average, but the all-ages number be greater than
the baseline average?
If any 2023 age group were above the baseline average, then all-ages
number could go either way, because of different total populations in
each age group, but with all age groups being below the baseline
average, I just don't see it.
This seems to happen not just for both sexes, but for each sex
individually.
Sylvia.
I should just add that I thought I'd try to prove this impossible using
algebra, and failed totally.
So then I used Excel, and was able to construct a counter example. So my
intuition about this was wrong. Of particular significance appears to be
that the population numbers for the base line age groups will not
generally be the same as the population numbers for the current data age
groups.
Oh well. Thanks to those who looked at this.
Sylvia.
It should be possible to get the numbers that went into this report
under "freedom of information".
Sysop: | Keyop |
---|---|
Location: | Huddersfield, West Yorkshire, UK |
Users: | 415 |
Nodes: | 16 (2 / 14) |
Uptime: | 108:49:16 |
Calls: | 8,692 |
Calls today: | 1 |
Files: | 13,257 |
Messages: | 5,948,428 |