SoylentNews is people
SoylentNews
from the probably-where-I-went-wrong dept.
Educators, policymakers, and parents have begun to focus more on children's math learning in the earliest years. Yet parents and teachers still find it challenging to know which kinds of early math skills merit attention in the classroom. Determining how to help children achieve in math is important, particularly for children from low-income families who often enter school with weaker math knowledge than their peers. A new longitudinal study conducted in Tennessee has found that low-income children's math knowledge in preschool was related to their later achievement—but not all types of math knowledge were related equally. The findings suggest that educators and school administrators may want to consider carefully which areas of math study they shift attention to as they develop curricula for the early years.
Conducted by researchers at Vanderbilt University, the study appears in the journal Child Development.
The study followed 517 low-income children from ages 4 to 11; the children were primarily Black and all qualified for free or reduced-price lunch, a measure of poverty. When the children were in the last year of preschool and near the end of first grade, researchers tested general skills (including self-regulated behavior, work-related skills, and reading) and six math skills (patterning, counting objects, comparing quantities, understanding written numbers, calculating, and understanding shapes). When the children were at the end of fifth grade, researchers tested a range of math knowledge, including knowledge about numbers, algebra, and geometry. The aim of the study was to determine whether children's math skills at ages 4 and 5 predicted their math achievement at age 11.
Preschool math skills supported first-grade math skills, which in turn supported fifth-grade math knowledge, according to the study. In preschool, children's skills in patterning, comparing quantities, and counting objects were stronger predictors of their math achievement in fifth grade than other skills, the study found. By first grade, patterning remained important, and understanding written numbers and calculating emerged as important predictors of later achievement.
(Score: 0) by Anonymous Coward on Wednesday December 07 2016, @05:06AM
This is nonsense. There are many biases, most notably the parents - parents who hassle their toddlers into learning how to count will hassle their teenagers into learning how to do calculus.
(Score: 0) by Anonymous Coward on Wednesday December 07 2016, @06:02AM
There should really be a 'social science' warning on articles like this.
(Score: 2) by tfried on Wednesday December 07 2016, @07:14AM
True, the study does nothing to establish causality. A "real" test would be to research the effect of (randomly assigned) programs. But that is simply a heck of a lot more difficult to accomplish. My personal guess on this is that they may be on to something about the importance of "comparing quantities" and "counting". Lack of a basic understanding of quantities may well be at the heart of Dyscalculia [wikipedia.org] for instance. Children lacking in that area will understandably have difficulty keeping up (while - notably - Dyscalculia does not necessarily imply that students will have difficulty with the "memorizable" parts of maths, including basic calculations). I don't see any particular reason to assume that this ability should not be accessible to early training.
On the other hand, "patterning" looks quite a bit like your regular (maths / logic) intelligence test to me. Quite obvious that this should affect maths achievement even years later. Quite dubious that training can achieve much beyond an initial "so that is how to approach these exercises"-effect.
Perhaps, what is a more important insight from the study, however, is what they did not find to be related to later achievement: understanding shapes. From my personal - hazy - recollection, that made up quite a big share of preschool "maths", and here's indication that this time might better be devoted to something else...
(Score: 2) by maxwell demon on Wednesday December 07 2016, @08:35AM
I think it depends a lot on how many patterns you already know and therefore can recognise immediately. A larger "library" of patterns also means you've got more material to recognise "meta-patterns" (those need not be at a level that you could explicitly tell them, it just matters that they are available to your "pattern matching engine") that help you recognise new pattern.
The Tao of math: The numbers you can count are not the real numbers.
(Score: 0) by Anonymous Coward on Wednesday December 07 2016, @12:30PM
> True, the study does nothing to establish causality.
But TFA heavily implies causality. Junk reporting (I didnt read the paper, I assume the authors did not imply causality but social scientists meh).
> A "real" test would be to research the effect of (randomly assigned) programs
I understand meta-studies have shown that no matter what program you put students into, it demonstrates achievement while the paper is being written. When those programs are rolled out, the improvement is much more muted. The interpretation is that just putting researchers into the school causes improvement irrespective of what the researchers are doing.
(Score: 0) by Anonymous Coward on Wednesday December 07 2016, @03:43PM
The same is true of classroom discipline studies. It doesn't matter what system you use to run the classroom, but you will see improvement over the control group, apparently because the research is being conducted.
When I was in ed school (thank God I got out of that shithole of insanity), I looked into one such study and found that the author of the study had dropped about half his initial data because it did not confirm that the system he was testing actually improved classroom discipline. It probably didn't help that the researcher was also selling the system as a set of books and seminars. Demonstrating that his system worked meant money in his pocket, so dropping data that showed no improvement over the control group meant bolstering the case for schools to hand him more money. I have no doubt that such methodology is at play in many education studies. There are more "methods" and seminars being sold in education than in used car sales, and they're of about the same quality. Education research will continue to suck as long as the researchers are seedy salesmen pushing their products rather than dispassionately measuring results.
(Score: 2) by tfried on Wednesday December 07 2016, @08:02PM
The interpretation is that just putting researchers into the school causes improvement irrespective of what the researchers are doing.
Well, yes and no. The takeaway is to never compare an intervention (program) to no intervention, but at least a credible dummy. Not entirely unlike giving a placebo to control subject in a medical study, instead of just nothing. Of course in order to be a credible comparison, you'll need to put in roughly the same amount of professional involvement, which means you've just doubled an already high cost.
That's not really an excuse for cutting the corners, but sometimes an unavoidable reason for doing it despite knowing better...
(Score: 0) by Anonymous Coward on Wednesday December 07 2016, @01:26PM
Yeah. Math is for pussies.