Archive for the ‘Numbers and Statistics’ Category.

Even an Influential Chart Can Be A Graphics Fail

Presumably most of you have seen this chart frm a study that says that not only do Americans not know where the Ukraine is, but that desire for US intervention there is correlated with such knowledge or lack thereof (the less people understand where it is, the more they support intervention).

click to enlarge

I find the study results both depressing and unsurprising, so I won't comment on them per se.   Though I suppose if you confuse the Ukraine with the Yukon (as a number of respondents seem to), interventionism might make more sense.  My only question is:  where were such studies of domain knowledge vs. policy recommendations in the health care or minimum wage debate?

However much impact this chart has had, though, it is still a graphics fail in my mind.  Why?  Because the author attempts to portray a second variable by the dot color.  But the variable he or she chooses to portray is the distance of the point from the correct location (red being more correct, blue less).  But that is easy to see without the variation in color.  It is redundant information.

A much better chart would have been to color code each dot with that respondent's Ukraine prescription, from blue = intervention to red = non-intervention.  This way the chart would have supported the full findings of the study (link between geographical knowledge and policy prescription) rather than just one aspect (quality of geographic knowledge).

Update:  If so many people got the Ukraine and the Yukon confused, God help us if the next Russian crisis is in Georgia.

Why Does The US Appear to Have Higher Infant Mortality?

I am sure you have seen various rankings where the US falls way behind other western nations in terms of infant mortality.  This stat is jumped on by the left as justification for just how cold and heartless America is, and just how enlightened socialized medicine must be.  However, no one seems to bother to check the statistic itself (certainly the media is too incompetent to do so, particularly when it fits their narrative).  Statistics like this that are measured across nations are notoriously unreliable, as individual nations may have different definitions or methods for gathering the data.

And, in fact, this turns out to be the case with infant mortality, a fact I first reported here (related post on medical definitions driving national statistics here).  This week, Mark Perry links to an article further illuminating the issue:

The main
factors affecting early infant survival are birth weight and
prematurity. The way that these factors are reported "” and how such
babies are treated statistically "” tells a different story than what
the numbers reveal.  Low
birth weight infants are not counted against the "live birth"
statistics for many countries reporting low infant mortality rates.

According
to the way statistics are calculated in Canada, Germany, and Austria, a
premature baby weighing less than 500 kg is not considered a living
child.

But
in the U.S., such very low birth weight babies are considered live
births. The mortality rate of such babies "” considered "unsalvageable"
outside of the U.S. and therefore never alive
"” is extraordinarily
high; up to 869 per 1,000 in the first month of life alone. This skews
U.S. infant mortality statistics.Norway
boasts one of the lowest infant mortality rates in the world. But when
the main determinant of mortality "” weight at birth "” is factored in,
Norway has no better survival rates than the United States....

In the United States, all infants who show signs of life at birth
(take a breath, move voluntarily, have a heartbeat) are considered
alive.

If a child in Hong Kong or Japan is born alive but dies within the
first 24 hours of birth, he or she is reported as a "miscarriage" and
does not affect the country's reported infant mortality rates....

Efforts to salvage these tiny babies reflect this classification. Since
2000, 42 of the world's 52 surviving babies weighing less than 400g
(0.9 lbs.) were born in the United States.

Hmm, so in the US we actually try to save low-birthweight babies rather than label them unsalvageable.  Wow, we sure have a cold and heartless system here.  [disclosure:  My nephew was a very pre-mature, very low-birthweight baby who could have fit in the palm of your hand at birth and survived by the full application of American medical technology.  He is doing great today]