Brendan Nyhan links to this hilariously bad graph from the Wall Street Journal:
It’s cute how they scale the black line to go right between the red and blue lines, huh? I’m not quite sure how $7.25 can be 39% of something, while $5.15 is 10%, but I’m sure there’s a perfectly good explanation . . .
Follow the above link for more details. As Brendan notes, the graph says essentially nothing about the relation between minimum wage laws and unemployment (“Any variable that trended in one direction during the current economic downturn will be correlated with the unemployment rate among teens or any other group.”) and he also helpfully graphs the unemployment trends among the general population, which has a similar upward trend.
This is not to say that increases in the minimum wage are necessarily a good idea–that’s not my area of expertise. I’m talkin here about a horrible graph–all the worse, I fear, because of its professionalism. The above graph looks legit–it has many of the visual signifiers of seriousness, looking similar to a newsy graph you might see in the Economist, rather than like a joke graph of the sort identified with USA Today and parodied so well by the Onion.
P.S. I have no problem with the use of a crisp graph to make a political point; see for example here or here.
is there another y axis on the right hand side for the minimal wages? 0.7 dollar on this axis corresponds to 10% on the left hand axis
The graph is an optical illusion.
The strongly increasing minimum wage line creates the initial impression that unemployment has been increasing steadily also.
However, black teen unemployment is steady as wages go from $5.15 to $6.55; total teen unemployment is steady until $6.55. Maybe a complete financial collapse has more to do with unemployment than minimum wage?
See the second graph on this post for a slightly better representation — the difference between overall and teen unemployment versus minimum wage. I'm sure it wouldn't be difficult to extend the chart to include previous recessions and other minimum wage levels.
Also, Andrew, are you complaining about all graphs with two y-axes? Your comment that "It's cute how they scale the black line to go right between the red and blue lines, huh?" could be applied to every single chart that puts two separate measurements in the same space.
Ed: Indeed, this is the point that Brendan was making in the linked blog.
Fu: Yes, I'm complaining about all graphs with two y-axes. Many experts agree that this sort of graph is almost always a bad idea. I believe that Howard Wainer, for instance, has written a bit on this. The above graph is particularly bad because not only are the y-axes on different scales, but there's also no common baseline of zero. The visual effect is extremely misleading.
the thing they do not point out is that this period of rising teen unemployment was also a period of a recession.
Since 1960 there have been 18 increases in the minimum wage. Nine of them appeared to have no impact on teen employment and nine of them were accompanied by a rise tin the teen unemployment rate.
but of the nine time the teen unemployment rate rose the economy was also in a recession. In the nine where the teen unemployment rate was unaffected the economy was in an expansion.
This generates a rule that is 100% accurate. A rise in the minimum wage accompanied by a recession is accompanied by a rise in the teen unemployment rate but a rise in the minimum wage during an economic expansion has no impact on the teen unemployment rate.
This rule or experience clearly raises serious questions of whether the recent rise in the teen unemployment rate was due to the recession or the rise in the unemployment rate.
Probably some of each, but anyone saying it was just the minimum wage would seem to be an ideologue pushing a biased point of view and incapable of serious economic analyisis.
On this very question, the current TAS (The American Statistician, February 2010) has an article by Stanley, Jarrel and Doucouliagos arguing:
1. The connection between minimum wage and employment is an excellent example of publication bias rather than a true collection
and, most interestingly from a stats point of view
2. Often a meta analysis would be better if we threw out 90% of the studies and only kept the 10% most precise.
I'd be interested in comments by others on point #2.
zbicyclist – for 2 if you change "precise" to "relevant and methodologically sound" and think of projecting even higher – you get Don Rubin's response surface model for meta-analysis.
This paper provides some dicussion of scoring "methodologically sound" and suggests some partial pooling of observed quality effects to an overall modeling quality effect (weighting – especial 0 verus 1 weighting is often a bad idea).
Greenland S, O’Rourke K. On the bias produced by quality scores in meta-analysis, and a hierarchical view of proposed solutions. Biostatistics. 2001 463-471
The main challenge though – discussed in a previous post – may be to get statisticians to realize they can't just deal the study that happened to be brought to thier door/consulting practice.
K