Also this week: another social science journal bans p-values, figure aspect ratio vs. your data, JBS Haldane vs. Hannibal Lecter, and more.
Want to know what new Am Nat EiC Dan Bolnick thinks about “gatekeepers” and scientific publishing in the digital age? Here you go! I think it’s great for an EiC to spell out his thinking at such length.
The journal Political Analysis is banning p-values. Their reasons for doing so are:
- an isolated p-value is not evidence for a model or the associated scientific hypothesis
- lots of people have written lots of papers criticizing p-values
- relying on p-values leads to publication bias because non-significant results go unpublished
- banning p-values will stop p-hacking
- many social scientists misunderstand p-values
All of those reasons are pretty debatable to my mind, especially in the absence of any discussion of the alternatives. For instance, if you think p-values are widely criticized and misunderstood, wait ’til you see how widely criticized and misunderstood Bayesian posterior probabilities are! No statistical procedure works at scale. And if you think there’s no Bayesian equivalent of p-hacking, think again.
Sticking with Political Analysis, they’re switching from double-blind review to single-blind, because
banning p-values wasn’t controversial enough on its own it’s taking too many editorial staff resources to check for and correct author-identifying information in submissions. To my mind, that’s an argument for Am Nat’s approach of just leaving author names and institutions off the title page. That’s arguably better than nothing; reviewers won’t always recognize the authors. And hopefully leaving off the author names and institutions reminds reviewers who do recognize the authors to focus on evaluating the science. Thereby hopefully preventing perverse outcomes that could arise from blinding being seen through non-randomly with respect to author attributes.
How Pittsburgh illustrates the best playbook to revive old industrial cities with good universities: build a tech cluster around the university, welcome immigrants, fund the arts, use redevelopment to create a few nice neighborhoods, and go from there. What say you, Cleveland? (home of Case Western Reserve)
How your figure’s aspect ratio can make it misleading.
JBS Haldane channels his inner Hannibal Lecter.
And finally, Stephen Heard imagines Valentine’s Day in the Ediacaran. 🙂
I’d just like to throw in a few words of caution about the Pittsburgh article. That article cites several trends that appear to support the idea that Pittsburgh’s solution to deindustrialization is working out well for everyone — the booming tech industry, increases in total real domestic product, a growing millennial population that is highly educated, and increasing average wages. Inferring that this is a model system for revamping the Rust Belt based on those trends is highly problematic, at least if one is concerned about a truly inclusive solution to deindustrialization.
A booming tech industry and an influx of highly educated workers are only good for a relatively small number of individuals — namely, those at the top of these industries and the small number of workers that are qualified for jobs these industries offer. However, the Rust Belt still has a huge population of workers that do not have the education and skills required for these jobs. So, not only do these individuals miss out on the economic gains, but the influx of higher income folks drives up the cost of living substantially, forcing people out of their communities. Juan Gonzalez’s excellent book Reclaiming Gotham describes the devastating impacts this has on communities of color.
-Here’s a brief bit on gentrification in Pittsburgh: https://www.wnyc.org/story/gentrification-pittsburgh-tale-two-cities/
Another side effect of increasing tech industry is that cities offer perverse tax incentives to these companies to set up shop in their downtown areas (look what is happening with the current bid war related to Amazon right now: https://www.seattletimes.com/business/amazon/this-city-hall-brought-to-you-by-amazon/). So, the argument that these companies will build up the tax base is extremely flawed — Pittsburgh is no exception (https://nonprofitquarterly.org/2017/09/12/pittsburgh-grapples-rise-nonprofit-industrial-complex/).
It shouldn’t come as a surprise to ecologists that interpretations based on averages (like increasing average wages) are highly problematic. Averages don’t describe the overall distribution. In this case, they paint a very misleading picture of the the economic gains Pittsburgh is experiencing. While mean wages have gone up, lower wages have been stagnant for decades. Unsurprisingly, poverty rates in Pittsburgh have increased from 10.6% in 2000 to 11.5% in 2007 to 12.3% in 2012. In Pittsburgh, as in most of the country, poverty disproportionately affects children, the elderly, people of color, female-headed households, and people with disabilities (https://www.urban.org/research/publication/poverty-and-income-insecurity-pittsburgh-metropolitan-area). In Pittsburgh, the median wage of whites has increased by 8.1%, while the median wage of African Americans has decreased by 19.6% (https://www.pghcitypaper.com/Blogh/archives/2017/05/11/is-pittsburghs-economy-growing-for-everyone). The end result is higher average wages but increasing income inequality in Pittsburgh. Similar problems exist with interpreting increases in total real domestic product.
I’d also note that the article (and the one it cites concerning the influx of immigrants) doesn’t say anything about what type of economic situation immigrants are living in in Pittsburgh. This post is already too long, but I’d just note that a high proportion of the immigrants to Pittsburgh are Latinx (https://www.pghcitypaper.com/Blogh/archives/2017/06/07/immigrants-are-holding-up-the-pittsburgh-metro-area-population), and that Hispanics generally are not doing well economically in Pittsburgh (https://www.urban.org/research/publication/poverty-and-income-insecurity-pittsburgh-metropolitan-area). Related to this, a booming tech industry often results in an increasing tendency towards a “gig economy” (the article cites Uber setting up shop in Pittsburgh), which offers no benefits to workers, while union jobs in Pittsburgh continue to decline.
Finally, I’d just note that the economic impact of these large industrial cities was not restricted to the city limits. These cities have historically supported dozens of nearby communities (like the one I grew up in near Detroit) with good paying union jobs that were the foundation for the middle class in the Rust Belt. I’m going to go out on a limb and speculate that Google, Uber, Amazon, and the robotics industry will not be having similar effects.
All fair points.
I guess my question in reply would be what’s the alternative? What’s a better playbook, and where are the cities and towns successfully implementing that playbook? Honest question–I’m not an urban economist, I don’t have any expertise here, I only know what I happen to read online.
In passing, wanted to note that Noah Smith doesn’t suggest that playbook solely based on the example of Pittsburgh. Here’s a piece he did on Pikeville, KY: https://www.bloomberg.com/view/articles/2017-11-14/small-colleges-can-save-towns-in-middle-america He’s relying a lot on Enrico Moretti’s The New Geography of Jobs (which I haven’t read; it’s on my list) and Richard Florida’s work.
You mentioned Detroit; I’d be interested in pointers to pieces on the situation there and Wayne State’s role.
I think that’s the million dollar question. I’m not sure I can point to better case studies to be fair, and can’t comment much on Wayne State’s role in efforts to rebuild Detroit. I absolutely think that universities could play an important role in the future of these deindustrialized cites, but I think that will only happen if we critically evaluate the situation now. I do think that any solution is going to take a lot of political change at the national level (e.g., tying the minimum wage to inflation and gross productivity, revamping of unions, universal healthcare, tighter banking regulations, free or at least affordable public college, implementing a more sane tax system, etc.).
I think that conversations like this are important for getting us closer to a solution that works for more folks. Speaking of which, I just want to thank you for being willing to have such a conversation — I’m rather disturbed by how much hostility a lot of folks in academia have towards these ideas. Most of my motivation for commenting on this post was due to having heard all these trends cited before as evidence that the economy is doing great, while seeing firsthand that folks are still really struggling.
Here’s a link to a follow-up from Noah Smith that addresses some of the issues you raise, using the examples of Raleigh and Charlotte:
Replying to “A Graduate student…”
“tying the minimum wage to inflation and gross productivity, revamping of unions, universal healthcare, tighter banking regulations, free or at least affordable public college, implementing a more sane tax system, etc.).
Many of these suggestions will take you in exactly the opposite direction of higher wages for the middle class.
The recent studies in Seattle show clearly that raising the min wage reduces employment, just as a simple analysis of supply and demand predicts. In general, rasing the min wage beyond the market wage creates an inefficiency that drives up the cost of labor, so more societal wealth is expended for the same amont of product.
The same would be true of “free” (to the individual) higher education. When the price of something falls, more of it gets wasted. In the context of education, that likely means the price to the individual will fall but the *cost* to the public as a whole will rise, because more lower quality students will attend and these students will need even more support services. What happens? taxes rise. Who pays the taxes? Middle and low skill workers.
It’s not even clear that it’s desirable to have more educated workers: If the supply of educated workers rises, what happens to the price those workers can demand for their services? it falls. Right?
I can’t speak for Pittsburg, but here in the Pacific Northwest environmental regulations and a very strong hostility toward industry are a direct hit to middle and low skill workers and surely a signifcant driver of the homelessness that has turned seattle into an urban trash heap. The governer here is trying to impose a carbon tax. Who will pay that tax? Mostly low skill workers, who spend a disproportionate amount of their income on transportation and mostly drive older and cheaper cars. From pipelines to coal export terminals to logging and drilling, every reg that kills these jobs puts downward pressure on wages for med/low skill workers and makes more people homeless.