It appears several papers are concluding it's fine to open school, and one of the pieces of evidence is a Swedish "study" from FHM which states that school teachers don't have a greater risk of illness than other occupations.

That's from this document

https://tinyurl.com/yyrpnn85 

That paper, however does so by referencing *another* FHM report, claiming an analysis there that showed school teachers and related staff aren't diagnosed with CV19 more than other work groups.

https://tinyurl.com/y45op9uf 

When you go to that paper though, it references a *third* paper, also by FHM.

13. Folkhälsomyndigheten. Arbetsmaterial covid-19 och yrkesgrupper per 2020-05-25

"Working document Covid-19 and Employment Groups per 2020-05-25"

When I tried to find that reference, I couldn't. I instead found a comment by TV4s Anneli Megner Arn on a web page about accessing Swedish documents. In it, she describes asking @Folkhalsomynd for this and being told it was covered by secrecy laws.

https://www.allmanhandling.se/2011/08/12/a-5-1-internt-arbetsmaterial/#comment-72829

So the actual research that I've seen some countries, and some organisations, use to justify opening schools is unpublished and unavailable.

However ... back in the first publication, there is a table, apparently based on this working paper -

The heading is "Relative risk per occupation within schools. # of cases and # active per occupation in the population. Median age per diagnosis. Relative risk (95% konfidensintervall)

Now, these are small and messy numbers. It's from tests sometime prior to May 25, when tests where only done if you were seriously ill. Others (like myself) were told to stay at home and never tested.

As I've mentioned elsewhere, FHM also seems to have made no attempt to control for age and sex, too of the biggest correlates of seriousness of CV19. Who is more likely to be sicker? Men, in particular older men.

According to OECD data, teachers are far more likely to be women than men.

https://stats.oecd.org/Index.aspx?DataSetCode=EAG_PERS_SHARE_AGE#

I've been unable to find the data, but I'm fairly certain that Taxi drivers, whom FHM claims is the most at risk group, are far more likely to be men

Given just this data, you would *expect* teachers to have significantly lower risk of serious covid-19 (and thus appearing in positive test statistics), than say taxi drivers.

But I realised something about the data in the table above. It allows us to do a "natural experiment".

The data for the teachers is for positive tests from sometime between the first cases towards the end of February, and when the "secret" working report was published May 25.

March 16, the Swedish government announced that high schools and universities would move to distance ed

So for a significant part of the data, we had teachers in schools (primary school/grundskollärare) and teachers not in school (high school/gymnasielärare). That's these two groups -

Unlike, apparently, FHM, I have some familiarity with basis tests of statistical significance, so I thought I'd pop these figures into a Chi Square analysis.

Primary school teachers were significantly more likely to test positive than high school teachers.

This is *despite* primary school teachers being overwhelmingly more female, and younger. Furthermore, the data includes the period when high school teachers were still at school

Now, can we say for sure this means closing high schools decreased the risk for teachers? Experience over the past 6 months indicates that's what the newspaper headlines would say!

It *probably* actually does mean this, but as I said, this data is messy. Lots of problems with it

What we *can* with 100% certainty say is that this apparently unpublished, secret report by Folkhälsomyndigheten should *absolutely* not be used as the basis for any decision making. It *cannot* be used to conclude opening schools is safe.

/end