Do Smarter Workers Work Less? | Hacker News Reader

Do Smarter Workers Work Less?

economix.blogs.nytimes.com

55 pointsrglullis15 comments

Open articleView on HN

Incorrect title.

Correct title would have been: workers with a better education tend to works jobs with shorter hours.

But even that is possibly misleading. Best title:

Workers in states with a higher average level of education tend to work jobs with shorter hours.

(which says nothing about the workers themselves. We'd have to look at the data to know)

Indeed.

For example, if you just looked at the fact that higher-income states tend to lean Democratic politically, you might conclude that higher-income individuals lean Democratic. But nationwide, and in each state alone, higher-income individuals lean Republican.

These potentially-counterintuitive results are examples of Simpson's Paradox:

http://en.wikipedia.org/wiki/Simpson%27s_paradox

It's possible the same effect is present in hours worked: that even though high-income/high-education states work fewer hours on average, in each state individually higher-incomes/more-education is correlated with more hours worked. More detailed data would be required to know for sure.

If having an education causes women to enter the workforce more (either in order to pay off college loans or because they feel like they'd be wasting their degree if they stayed at home), then that could cause the correlation. Total hours per person would be more in the educated states, but hours per worker would be less. But we'd need more data to figure out if this is actually the case.

His result: Education seems to play a big role in how long a state’s average resident works, and for what wage.

This sentence is false based on the data given. The data in the article was a survey of employers. Thus it only measures how long employed workers work, not residents as a whole.

UPDATE: The full study is behind a paywall, but the abstract states a contradictory result from that of Florida's: "ws us to construct a longer time series. We find several interesting patterns. The married women with the largest increase in market hours are those with high-skilled husbands. When we compare households with different skill mixes, we also find dramatic differences in the time paths, with higher skill households having the largest increase in average hours over time." Source: http://www.emeraldinsight.com/Insight/viewContentItem.do;jse...

I like the plot of the immigration data. Immigrants are people with initiative, and they bring up the level of the people they live among.

It's good by they way to see these data presented as a scatterplot rather than just seeing the correlation number, which is much less informative.

http://en.wikipedia.org/wiki/Correlation_and_dependence

I'm dubious. There are a couple other factors:

1. Coastal cities have high costs of living; the top-right of the graph is dominated by coastal states with big cities; the bottom-left is dominated by landlocked, low cost-of-living places. 2. Wouldn't immigrants go places with high hourly wages? If you move in with your folks in New York or SF, why would you go to North Dakota?

It sounds like the cause and effect relationship goes both ways: immigrants move to high-wage places, and the presence of all those ambitious, extra-hard-working folks causes economic growth and higher wages. But it's conceivable that one of those relationships could reduce wages, instead; one could easily read this as a graph demonstrating that every time you get high wages, you have this almost Malthusian influx of immigrants.

Immigration has definitely raised my standard of living, though; the company I work for was founded by an immigrant.

This result is pretty much well known for ANY economist, even people with only econ 101. For instance, check the first graph of this link: http://www.statemaster.com/encyclopedia/Labour-market

The logic is the following:

1) More educated people earn higher wages (if not, then why spend so much time and money on education?).

2) Higher wages gives you more money for consumption

3) As people fill their consumption needs, they now want more leisure. For that, they work less hours.

> His result: Education seems to play a big role in how long a state’s average resident works, and for what wage.

There's nothing here that suggests the causality arrow runs in that direction. Here's another take "all workers prefer to work less; workers who have foresight and long time horizons get education so as to avoid working second hobs".

> This scatterplot suggests that state hourly earnings are positively associated with the percentage of immigrants (correlation of 0.64).

Sigh.

This could mean anything, from "having lots of cheap yard workers around makes it easier to put in weekend hours" to "high hourly rates draw in migrants", to just about anything else.

The states where workers work less seem to be on the colder climate zones; it could be that workers in the north just have less time to get things done climatically speaking; after all, if you work on the road, you have to take in account of day light and frosty conditions into effect.

As about the indoor ones, they probably want to get home early thanks to the shorter daylight?

Technically, the length of the days will average out. Longer in the spring and summer (when the weather is nicer for outdoor work), shorter in the winter.

However, the number of snow days won't average out, but I doubt that is being factored into the data.

Can someone with a stats background explain why the article claims there's a correlation? Those charts had R^2 values of .428, .344, and .403. I thought low R^2 values indicated no correlation.

I think of "correlation" as a measure of how much of the variation in one variable affects that in another. This explains why acceptable correlation levels vary between disciplines: one variable explaining 40% of the variation in another is pretty good in social sciences, where you expect all sorts of confounding variables, but quite bad in, say, (well designed) physics experiments.

Disclaimer: my statistics background consists entirely of one class (AP Statistics).

>one variable explaining 40% of the variation in another is pretty good in social sciences, where you expect all sorts of confounding variables, but quite bad in, say, (well designed) physics experiments.

I think this is part of the answer to your question, nhebb, but also I think we might have some confusion about r. In my understanding, (which is also limited to only AP Statistics) r is the value representing how well correlated the two categories are, and r squared represents how much of the correlation the categories you're using at that time accounts for. Does that sound right?

That makes sense, but I still have trouble believing the article's conclusion.

In my experience, people with advanced education work more hours. Doctors, lawyers, engineers, programmers, accountants,... the list of professions that require a lot of 40+ hour work week goes on. Most lower wage jobs, by contrast, work straight 40 hour work weeks

Also, is the data adjusted for cost of living? Why are they using average wages instead of median wages?

Overall, it just looks like a crappy study that doesn't lead to tangible insights.

How does hours per week top out at 37? 37? Thirty seven? Three-Seven? I can't really figure out how the average is that low except as an assumption that almost no one who works puts in overtime.

Now I feel like a chump for putting in so much overtime.

I don't see how it's very productive to work more than that, particularly in a creative discipline. Actually, I think I'm at my most effective when I "work" less than that - but I also spend idle time thinking things over in the background. That thinking builds up a head of steam for the next period of work, but after six or so hours of that at most, I'd have to be doing some kind of mindless grunt work to continue being effective.

Reply on news.ycombinator.com