Comment by jacobolus | Hacker News Reader

Comment by jacobolus | Hacker News Reader

Parent Full thread

Looks to me mostly like a cluster of kids with very high “calculation skills” and relatively low “empathizing quotient” is skewing their result. If you remove that group, the remaining data looks pretty uncorrelated at a glance: http://www.nature.com/articles/srep23011/figures/1

“Math Fluency is a measure of speeded application of arithmetic procedures. Problems included a mix of addition, subtraction, and multiplication with operands up to 10. Children were given 3 minutes to complete as many as possible. We computed a Calculation Skills composite measure, which combines the Calculation and Math Fluency subtests, in order to have a single measure capturing arithmetic ability.

“The primary guardian of each child completed the Combined Empathy Quotient-Child (EQ-C) and Systemizing Quotient-Child (SQ-C). The questionnaire was designed to be parent-report in order to avoid variance associated with children’s reading and comprehension abilities.”

Alternative summary: “Out of 112 children, among the bottom 105 children when ranked by arithmetic calculation speed, there’s no statistically significant correlation between arithmetic calculation speed and parental assessment of child empathy. The parents of the top 7 performers by arithmetic calculation speed uniformly think their kids have below-average empathy.”

Even if you remove those, look at the bottom of the data, across the horizontal axis. There is still a downward trend, although that is an eye judgement as well. That is, if you bin the kids according to empathy (x-axis), the lowest math achiever in each bin is worse as emphathy increases. The max in each bin doesn't seem to get better, however. The slope would definitely be milder if those offending points were removed.

Moral of the story: higher N is always better.

> Moral of the story: higher N is always better.

This is true in a way, but finding and studying larger effects is often a better strategy. To some extent, why talk about something that needs huge N to get above the noise?

There's also a cluster if kids with very low “calculation skills” (<85) and relatively high “empathizing quotient" (>30)

those combine into a trend. Look at the gap in the lower-right corner.

Of course, it could be bias where parents feel a need to pick a strength for their low-performing child.

>could be bias where parents feel a need to pick a strength for their low-performing child

Agreed. Even a small skew in that direction could cause most of this effect.

Yeah, but why would you do that? What would warrant considering that it would just be pure coincidence that the 7 best performers happen to have below average empathy scores? Of course there will be outliers in data. You must consider that there are also outliers on the bottom end of the math scale who seem to score above average on empathy.

What if the math group would have been up in the upper right hand corner? Would you then have made the same observation?

I don't really buy the "my children aren't good at everything but not bad at everything either so lets pick one", which was mentioned earlier. The questionnaire considered both SQ-C and EQ-C, and correlation was found between only one of them. You'd think both would have then been found to correlate.

"If we remove the samples that we don't like, we get the result we want!"

Note: I'm not a fan of reading too much into studies, but come on.

@jacobolus is correct IMO, according to these data, empathy can't be described as predictor of math ability.

See http://www.nature.com/articles/srep23011/figures/1

An outlier group of mathy kids turn out to have poor (but not the poorest) empathy skills. The rest of the cohort isn't skewed, at all.

I am glad it was noted. I don't really have time to read this (edit: and the comment by jacobolus saves me the effort). I have good empathy and social skills and innate math ability. My ex husband and our oldest son both suck at all three of those.

I think it is entirely fair to say that some oddball subgroup is skewing the results. It is no different from discussing mean, median and mode in real estate data.

Ever heard of the book "How to lie with statistics"?

I don't see how you can discredit the study by citing three data points. The plural of anecdote is not data.

You are completely misinterpreting my remark. I in no way was discrediting the study based on three data points. I was interested in reading it because of three data points, but I don't really have the time for it. Fortunately, someone else read it and found a glaring flaw and that is sufficient for my personal level of interest at this time, thus I won't be wasting my limited time on this today.

"thus I won't be wasting my limited time on this today"

That seems debatable as you have taken the time to comment on this story multiple times. I guess we all have opinions though.

My opinion about how I choose to use my time and what I have time for is the only one that counts. It isn't open to debate. You are basically engaging in a personal attack, for no apparent reason. Attacking me personally in no way adds to the discussion here. Do you have something to say about whether or not the original article has merit/is worth actually reading/etc?

If not, it would be extremely nice if more people followed not only the letter of the guidelines here but also the spirit and didn't engage in ad hominems or other smear campaigns.

Reply on news.ycombinator.com