Nations Report Card For Science

quakeaddict writes: "The US Dept of Education released the Nations Report Card for Science. Compared to 1996, it seems as if we as a nation are doing worse.The site also has cool tools to see how various states performed in other subjects.. There is also a very cool interactive tool that provides a wealth of information about how well kids did in all the subjects crossed with all sorts of questions (e.g. What was Johnnys score if he says he watched 1, 2 or 3 hours of TV a day etc...)."

Comments

The numbers look pretty constant to me

[hubie]

I have great concern regarding the emphasis and quality of science and mathematics teaching in the primary and secondary schools in the US. I believe much more can and should be done to improve the teaching, retention, and understanding for these subjects from K-12. Having said that, I looked at the stats on the science test results and things basically have not changed, so I think that, based on the test results, it is a bit misleading to say that we are doing worse as a nation.

The score changes from the 1996 test show only a statistical difference in the 12th grade results, and these differences are marginal at best. Even within the statistically different 12th grade results, the only statistical change was in the group that scored in the 50-percentile (a promising stat from that figure is that there was a statictically significant increase in the top performing 8th graders).

One graph that I found troubling was the one showing the numbers above and below basic proficiency levels for the 12th graders, where the numbers falling below understanding the basics increased. We can hope that this is a statistical fluctuation and not the start of a trend.

By the way, the web site is very impressive in how much information is presented from the test question to the error on the test results. My biggest beef with statistics reported in the media is that they either never give error bars, or they'll ignore the errors; they'll report political poll results as one candidate ahead in the polls even if that person is ahead by less than the margin of error (this leads to the whole topic of basic ignorance of relative risk and you don't want to get me going on that rant!).