There is more to this narrow minded view of the testing system
than is obvious.
It makes a conclusion and presents a solution without any data that
supports the conclusion or solution.
They assert that a girls only class is the solution and the problem is boys
dominating the class.
They do not address the possibility that educators could simply be biased
and the same individual educators in a special class would imprint that same bias without
change. The result would then be identical.
Not addressed in this data is the assumption that systematic issues in education
are the reason girls do not invest themselves in STEM anything. Society also
adds to this...
Personal antidotal bias is that the smartest math and science person in my k-12 education
decided to pursue her dream as an artists. As a spouse in a traditional marriage
she could do this without concern for the finances of it. She was not alone although
another gal almost as smart ended up as a NASA outreach educator.
The other very important issue is that girls and boys do not develop on the same
schedule biologically. Moving the girls and boys into their own classes without
adjusting the schedule will also get the same result.
If educators are going to be honest they need to design an education programs that
allows biology and maturation shape the schedule of boys and girls class content.
In a K-12 school the age differences where one child can be 364 days older than another
must also be considered important.
One real issue is that standard tests are anchored on birthdays and on a calendar.
Adjusting the time for one groups tests vs. another would be seen as very unfair
yet it may be more so.
If we cold take sex out of this and substitute cognitive maturation we might get
better outcomes from the child's point of view. One clumsy attempt on this
is homogeneous grouping. Assignment into a group might be because an individual
was slow to grasp or simply unable to grasp the material.
I think the school has a bias and thinks they have a solution. Then they found numbers to support it.
There are girl and boy only schools where supporting data might live. One might be
a serious review of the famous schools and their curriculum as framed before standard testing.
These schools (perhaps 1780-1920) and their syllabus (teachers notes) may still exist
and may prove interesting.
Little of this matters -- TV sitcom and even cartoons have very rigid rolls for the sexes