I stumbled across this local Ann Arbor craigslist posting (since flagged for removal, and hence reproduced below) that advertises for a “ghost student” to take someone’s online master’s courses for them. Surprisingly, on the surface, the poster wants someone to earn a B average and occasionally fail classes. Why not pay for As? Potentially for a couple of reasons.

1. Maybe you’re getting compensated for doing this masters and want to drag it out. Seems unlikely, considering how much they are willing to pay
2. They realize the single biggest problem with faking academic credentials is getting caught, and if you want to avoid detection, it’s best to fly under the radar.

I strongly suspect that #2 is the major factor here. Carol Moseley Braun was found out within just days of misstating she had degrees from Harvard, when she merely held visiting fellow position at the school. Marilee Jones, on the other hand, claimed degrees from far less-renowned institutions, but was able to keep them on her resume for 28 years, because who is going to lie about having a degree from Union College? (You may remember Jones as the resume-padding fraud who, as dean of admissions for MIT, urged applicants not to pad their resume. Comedy gold.) Similarly, no one is going to check into your college records if you earn a B average and fail several classes. But if you are the shining star of your graduating cohort, somebody might just try to verify that you’re really taking your classes.

Full text of the ad:

What I need:
A U of M student
who is capable of good quality Masters level work
who is dependable
lives in the area
honest
will be around for 21 months or at least 9 months
knows APA formatting

To do what:
Take my online Masters Degree program for me as if they were me

For how long:
21 months

The game plan:
I need B level 3.0 work I need you to occasionally fail classes so the degree takes a little longer too

So that’s:
6 credits which is two 3 credit classes every 16 weeks, 16 weeks is one semester, one class at a time so one class every 8 weeks, total of 30 credits

When you fail:
You must participate weekly in the course discussions and submit assignments weekly and participate actively so you are not withdrawn from the course. You must write based on personal opinion and cite Wikipedia and do poor work so as to fail but actively participate the whole semester.

The pay:
$100 per week when you get the B’s. Paid to you the Monday after the Sunday night everything is due after you submit the work (I will be checking that you do the work). I can meet up with you for cash or we can arrange to set up a prepay VISA SERVE card or other such system for electronic transfer of funds or paypal so we don’t need to ever meet up and you just get paid online directly. When you fail you only get $50 a week because it’s less work. So $1,600 for the semester with B’s. $800 for the F’s or D’s.

Interested????
E-mail me with a little about yourself and why you would want to do this and how you would have the time to do the work. I will respond quickly. I need to pick someone by October 28. Please e-mail about yourself. If after a couple e-mails you sound like you’re the one. I would want to see some examples of your papers you’ve written for school to know you can do the work. Then meet up in person to give you the books and my passwords and everything.

START OCTOBER 28

Obligatory LOL at the requirement that the applicant be “honest.” But if people are going to cheat, I’d rather see them do it cleverly. Stupid cheaters are the most depressing thing in academia.

Planet Money recently ran an excellent story on the relationship between your college major and your eventual earnings. This is an important and generally under-appreciated aspect of the college decision. Broadly speaking, earnings are more strongly correlated with major than with school quality, but people spend a lot more time worrying about the latter. The basic relationship they are talking about is summarized in a short article entitled “The Most (And Least) Lucrative College Majors, in 1 Graph”:

I was careful to say “correlated” above because this graph highlights one of the key ways in which econometrics differs from (the rest of) statistics. Econometrics has an almost singular focus on isolating causal relationships in situations where simple correlations like the above are potentially misleading. That’s its single biggest strength, because humans are great at pattern-recognition and our tendency to jump to causality from observed correlations is hard to overcome. Think about describing this graph to a high school senior who is headed off to college. “People with engineering degrees tend to earn more than people with social work degrees” is something you can say without qualification. “Major in engineering because you will earn more money” is not.

Why not? Well, engineering degrees are hard, and you need to be very smart, diligent, and good at math to finish one. These abilities are things that, in part, you carry with you into college, and someone with that skill set will probably earn more money than someone who lacks it, even if they both choose to major in Studio Arts. The differences in the above graph will hence tend to overstate the benefits of majoring in engineering. Everyone (or at least everyone who has taken Stats 101 course) has heard the mantra that “correlation does not imply causation”.

But oftentimes correlations are causal. Econometric reasoning can help us see when they aren’t, and why. It also can give us a sense of the likely direction of the bias (upward, in this case – the effect of majoring in engineering on your wages looks bigger than it actually is) and what you might do to get better inferences (try to get your hands on pre-college measures of diligence and math ability). In its sophisticated extreme, econometrics is employed by (very clever) economists to isolate the actual causal relationship in question, say the impact of college major choice on eventual earnings.

I don’t know the right answer to the question of how much majoring in engineering will increase your earnings, but if you got a BA in counseling psychology, you can take some solace: that probably didn’t cost you a full $50,000 per year. However, it also wasn’t free, and college students should definitely pay more attention to that cost – and other associated costs, like higher unemployment rates.

The latest paper from our ongoing project on the potential of oral sex as a safer sex strategy in sub-Saharan African has been accepted for publication at the International Journal of Sexual Health. This paper, which I coauthored with Sallie Foley of the University of Michigan Graduate School of Social Work and my advisor (and study PI), economics Professor Rebecca Thornton, measures the how common oral sex is in Malawi and what factors are associated with it. Here’s the abstract:

Despite medical evidence that female-to-male oral sex (fellatio) carries a lower risk of HIV transmission than unprotected vaginal intercourse, little research exists on the practice of fellatio in Africa. We use two samples of men from Malawi, one rural and one urban, to examine the prevalence of oral sex. While 97 percent of the rural sample and 87 percent of the urban sample report having had vaginal sex, just 2 percent and 12 percent respectively say they had ever received oral sex. Only half of the rural sample, and less than three quarters of the urban sample, report having heard of oral sex. Education, exposure to newspapers and television, and condom use significantly predict oral sex knowledge after controlling for other confounding factors, while exposure to radio does not. The large gap between sexual activity and oral sex prevalence suggests that fellatio should be taken into consideration as a potential component of a HIV-prevention strategy, but further quantitative and qualitative research that includes women as well as men is needed to understand its potential benefits and drawbacks.

A copy of the final manuscript, not yet typeset, is available here. The full citation is: Kerwin, J. T., Thornton, R. L., & Foley, S. M. (in press). Prevalence of and Factors Associated with Oral Sex among Rural and Urban Malawian Men. International Journal of Sexual Health. doi:10.1080/19317611.2013.830671

We got accepted just before I left for my latest trip to Malawi (for another project), and turned in the proofs just as the fall semester was starting here at Michigan, so putting up a link to the article fell way down my to-do list. Today they sent me an email reminding me that not only is the manuscript already online, but that I also have a limited-use public link to it that people can use to access the PDF if they can’t get through the Taylor and Francis paywall. Contact me here or through my umich email address if you want the link.

The previous paper to come out of this project was called “Missing Safer Sex Strategies in HIV Prevention: A Call for Further Research”, and was published in African Population Studies. Our next step will likely be to exploit some novel data we have collected on people’s attitudes towards oral sex and sexual activity in general, which fills some of the gaps left by the just-accepted article. In particular, it includes women instead of just men, and has a qualitative component (we did a set of gender-segregated focus groups).

Last week I received the most impressive spam I have ever seen. Here is the text, with identifying characteristics redacted:

Subject: GLOBAL WARMING
From: CLIMATE CHANGE AWARENESS

 Attn: Sir/Madam, View attached letter and contact Prof. XXXXXX

Thanks,
Ms. XXXXXXX

Attached as a PDF is a letter that begins:

This is to inform you that the United Nations Framework Convention on Climate Change (UNFCCC) in Collaboration with the Nelson Mandela Foundation and William J. Clinton Foundation has held an Internet Raffle Draw, and your Email Address was among the 18 Email Addresses that was picked through the computer ballot system. The United Nations Framework Convention on Climate Change/Nelson Mandela Foundation and William J. Clinton Foundation was conceived with the objective of human growth, educational and community development, and to create awareness to the dangers posed on our planet by climate change.

To celebrate the 18th anniversary of the United Nations Framework Convention on Climate Change (UNFCCC), the United Nations Framework Convention on Climate Change/Nelson Mandela Foundation and William J. Clinton Foundation in conjunction with the United Nations (UN) is giving out a yearly award of $900,000 (Nine Hundred Thousand United States Dollars Only) to 18 lucky recipients.

I’ve uploaded the letter to my personal webspace (after stripping identifiable information, which frankly is probably false anyway). The whole thing is worth reading.

What’s impressive about this is the way it plays on a prospective victim’s emotions, especially if the target is of a socially liberal persuasion. Nelson Mandela and Bill Clinton and the UN Framework Convention on Climate Change are involved. I’m being asked to help protect and save our planet! And I’ve won a grant – these big international organizations totally make grants, right? It couldn’t hurt to see where this is going. Also helping its credibility is the professional-looking email address for the “consultant” the target is referred to.

I now wonder whether there are other variants on these, targeted at different groups. For conservatives, for example, you would want to invoke religious groups, Ronald Reagan, and maybe Freedom House.

My mother passed on this excellent piece from The Week that discusses the psychology of lotteries. It’s well worth reading – it covers all of the ways that lotteries exploit our mental and emotional weaknesses. But the lead paragraph has one of the most confusing explanations of a lottery that I’ve ever seen. For context, the basic idea behind a lottery is to sell n tickets for $Z apiece, and award the winning ticket some amount $X that is much less that n times $Z. One person wins a big cash prize, and you make a tidy profit. That’s the underlying grift behind all the mind games lottery organizers play. But according to The Week (and the underlying Nautilus post) it’s not how the Powerball drawing on May 18th, 2013 worked:

To grasp how unlikely it was for Gloria C. MacKenzie, an 84-year-old Florida widow, to have won the $590 million Powerball lottery in May, Robert Williams, a professor who studies lotteries, offers this scenario: Head down to your local convenience store, slap $2 on the counter, and fill out a six-numbered Powerball ticket. It will take you about 10 seconds. To get your chance of winning down to a coin toss, or 50 percent, you will need to spend 12 hours a day, every day, filling out tickets for the next 55 years. It’s going be expensive. You will have to plunk down your $2 at least 86 million times.

Williams could have simply said the odds of winning the $590 million jackpot were 1 in 175 million. But that wouldn’t register. “People just aren’t able to grasp 1 in 175 million,” Williams says.

If the jackpot is $590 million, and the odds are 1 in 175 million, then each ticket has an expected value of $3.37. At $2 per ticket, you make a profit, on average, of $1.37 for each ticket you buy. Taken at face value, these numbers imply the folks running Powerball (a cabal with the lovely moniker of “Multi-State Lottery Association”, which is for some reason abbreviated MUSL) lost money on that drawing. There are tons of caveats here: to play the eponymous “powerball” costs an extra dollar, and taking the money all at once reduced MacKenzie’s payout to $370.8 million.* But if these details matter, then they should have been clarified. I’m assuming the real issue is an error in the probability of winning, but a lazy Google search was unable to confirm that.

Irrespective of the cause, this is an extremely telling error to make in an article about why we (foolishly) play the lottery. The author himself couldn’t do the math to see that the numbers implied that the expected profit from the example lottery was positive. This actually shows that Robert Williams is on the wrong track! It’s not just that people can’t understand small probabilities – in general, we can’t compute expected values either.

*There are also federal taxes to consider, but these would just have lowered the expected profit to lottery players, and wouldn’t change the fact that MUSL lost money.

About halfway through my current 48-hour-long (!) voyage to Mulanje District in Southern Malawi, I am reminded of Chris Rock’s classic rant about how little he was taught about Africa in school, and how goddamned far away it is.

The only thing I know about Africa is it’s far – Africa is far, far away. Africa is like a 35-hour flight. So you know that boat ride was real long.

Here’s a link to the beginning of the whole bit about education, from his Bring the Pain standup act. It’s all great, but the part about Africa starts at 40:25. And Rock isn’t totally off-base. Sure, you can get to some parts of Africa from the US pretty quickly, but going to many parts of the continent takes a crazy long time.

Indeed, the most recent leg of my journey, from Atlanta’s scenic Hartsfield-Jackson International Airport to O.R. Tambo on the outskirts of Johannesburg, is the fourth-longest nonstop commercial flight in the world, and the longest operated by a US carrier. In all, I will spend some 27 hours on planes and another 14 or so in airports to get to my destination (plus several hours of other transit). This route, which gets me directly to Blantyre’s Chileka Airport, avoids up to a day of extra travel time and a long bus ride from Lilongwe down south.

A lot of this could clearly be done more quickly, in theory. For example, Johannesburg is an airline hub for Africa because it’s a big city that’s had a major airport for a long time, not because it’s a conveniently central location a la America’s Salt Lake City. And the effective remoteness of places like Southern Malawi has real consequences for people other than itinerant development economists: it raises prices of imports and weakens the region’s ability to export goods at a profit.

Sue Dynarski links to a typically awful news article on a correlational public health study, along the lines of the “chocolate prevents cancer!” garbage that inspired this blog’s name. This one is about a claimed link between coffee and early mortality in people under 55. What’s striking about this case is that I had already bookmarked an article by Alice Walton on the same study that does a great job of presenting the results for a mass audience.

Here’s what Alice gets right:

1) Emphasize the preliminarity and uncertainty of the results. Walton does this consistently, from the very beginning:

Those of us under 55 who drink a lot of coffee – more than four cups per day – may be at greater risk of an early death. [emphasis added]

2) Put the study in the context of the existing literature. She points out that these results are inconsistent with the “mishmosh of coffee studies all pointing at different outcomes”

But perhaps most uplifting of all is to remember that findings from a number of earlier studies contradict the new one and suggest that coffee is actually, at least on average, good for us. In fact, one recent study in The New England Journal of Medicine, following some 400,000 people, suggested that drinking up to six cups per day is actually linked to reduced mortality from all causes – 10% for men and 15% for women.

3) Talk about the theoretical reason for the reason – or lack thereof:

One problem is that no one really knows what mechanism/s could explain the coffee-death link. Some are candidates, however: There’s coffee’s ability to boost epinephrine (adrenalin) levels in the body, its inhibition of insulin function (though this is controversial), and the fact that it may raise blood pressure and homocysteine levels, which are both known to increase heart risk (though since heart disease was not increased in the study, these seem less likely).

4) Distinguish between simple correlations and actual measurement of causal relationships. Chip Lavie, the study’s author, is totally up front about this. He’s not really an outlier, either – most researchers who do this kind of work realize that their results are a guide to future research rather than established, incontrovertible facts. The real culprits in overselling these correlational results are university PR departments.

Also keep in mind, the current study only points to a correlation, not cause-and-effect. And it only measured coffee consumption at one time-point, not many throughout the years. There could be a lot of other things at play. “It is impossible to know if this association is causal or just an association,” says Lavie, “so one does not want to over-state or over-hype the dangers of drinking more than 28 cups per week, although I personally will make an effort to keep my cups at 3 or less most of the time.”

5) Take the effect size with the crazy huge magnitude out of the lede, if you mention it at all. Dynarski points out that the 20%-50% range cited by the US News article on the coffee study is completely implausible:

Here is a sniff test of the magnitude of this estimate: a similar, correlational analysis showed that light smoking (less than half a pack of day) is associated with an increase in all-cause mortality of 30%. Heavy smoking (more than half a pack a day), an increase of 80%.  These magnitudes are in the same ballpark as the coffee study, which immediately suggests to me that the coffee estimates are absurd.

To Walton’s credit, she puts that number deep in the article’s text and caveats it heavily. The news media needs more science writers like her, and much less of just copy-pasting official press releases and adding in some hyperbole at the top.

Eric Chyn passes along this Marginal Revolution post in which Tyler Cowen asks “to what extent is the choice of venue for study due to what I will call ‘social science infrastructure’?” By “social science infrastructure” Cowen means having a pool of experienced field workers, plus a population that is accustomed to being studied and other less-tangible factors:

I don’t mean roads and bridges. I mean having trained armies of local assistants, data gathering and processing facilities, populations which are used to signing informed consent forms, medical clinics which understand how to work with social scientists and register data, and other less visible assets.

This list hits on some important factors, but it is missing at least of key items:

1. Existing data, in large quantities and available to the public. I have a colleague who works mostly with secondary data and was born and raised in Haiti but doesn’t do research on the country because there’s nothing to work with. Less data means fewer existing papers, a smaller existing literature available for informing your research (and selling it as publication-worthy!), and it makes running power calculations harder. Running an experiment on a population that has never been studied is an exercise in wasting your time and money.
2. Experience in the country, both on an individual basis and on the part of one’s colleagues and advisor. It’s much easier to do research in a place where you know the language, have friends, understand aspects of the culture, can get around, and so forth. A very close proxy for having these things personally is knowing some
3. Ongoing projects to work on. These are a great way to get experience in the country, to pilot survey questions, and even to run mini-experiments. This isn’t only for grad students – faculty tack their questions onto surveys and jump into collaborations as well.

Cowen essentially answers his own question by noting that a large number of field RCTs are set in Western Kenya. Malawi is another development research hotspot, and Northern Uganda seems to be growing as one as well (and I have worked in both). Social science infrastructure is essentially the whole reason why there is so much geographic concentration of development research.

But he missed out on asking the broader question: why are there such huge gaps in social science infrastructure across countries? The answer is suggested by the three components I listed above. What ties them all together is path dependence. At some point fairly long ago, people decided to start doing research in country X. The reasons for this are some mix of the obvious practical issues noted by Cowen (security, language, etc.) and totally idiosyncratic things. This then makes the next set of projects drastically easier: all of a sudden people know someone working in a country, they have a contacts to ask about who to hire as employees, they can pilot their work while interning on another study, and so forth. Then those people have colleagues and students, and the cycle continues.

In the case of Malawi, one of the key early events was the beginning of the MDICP, a longitudinal study of contraceptive use and sexual health behaviors. A tremendous share of the foreign social science researchers that work in Malawi have close connections to that project (including myself – my advisor worked on the 2004 wave, and used it as a platform for her job market experiment). The same goes for the human capital needed to do research – to pick one example, IKI, emerged from the local MDICP research staff.

It’s hard to overstate the strength of the path dependence effect in development research. The difficulty of blazing your own trail is a huge barrier to getting work done. From my perspective, the marginal cost of doing a project in Malawi is a tiny fraction of doing the same thing in Haiti. On the benefit side of the ledger, I’d agree that more people should study Haiti, but economics in particular puts strikingly little weight on the geographic origin of a given dataset. To first order, the economics profession thinks American data is all you need to study anyway; getting people to take any developing-country data seriously can be a challenge, so it’s not surprising that there’s a limited payoff to doing research in a new or under-studied locale.

Nancy Marker passes on this NPR piece about residents of Kenya’s Mathare slum using satellite images and GPS to put their community on the map and lay claim to their property. This is a hugely encouraging development even if this is the last step: activists are able to use these maps to shame officials into fixing problems, and to provide assurance to homeowners that if they allow pipes to be laid

Even better, it could be one of the first steps toward what I think might be the next stunning technological leap for sub-Saharan Africa (and much of the developing world). Fifteen years ago, “stunning technological leap” and “Africa” rarely appeared in a single paragraph. Then cell phones happened:

Mobile phones allowed Africans to work around the perennial problems of poor infrastructure, badly-regulated utilities, and bureacratic gridlock that had kept virtually all of then from having a telephone. The continent essentially skipped over home telephones entirely, moving directly to mobile phones. Africa now has more mobile subscribers than the US, and Kenya is the world leader in mobile phone payments.* Ever since the advent of Africa’s cell phone miracle, I had been pondering what made it happen and where the next huge breakthrough will happen.

The key element seems to be pent-up demand for a given service that is constrained (maybe by infrastructure costs, or by badly-functioning bureacracies). When a technology appears that allows the constraint to be bypassed, a huge boom occurs. This fits the adoption of mobile phones to bypass landlines, and the adoption of M-Pesa to bypass formal banking (since banks are hard to reach and expensive to use). It also matches the emerging spread of mobile-phone internet access in Africa: people want to use the internet for a variety of reasons but physical computers require large fixed costs and a decent amount of infrastructure, hence there is a move directly to phone-based internet.

I see telling hints of a similar pattern in terms of maps and street addresses. Anybody who has spent time in the developing world has seen how common it is for streets to have no name, or the name to be unknown to most people. Even if some kind of neighborhood or street name exists, many buildings have no numbers and the house and building numbers that do exist are very poorly documented. This makes finding places you haven’t visited before, or receiving mail or other shipments, a nightmare. The cost of fixing this system would be fairly exorbitant – you would need to unify all the disparate patterns and names already under use, and try to prevent new residences from popping up without acquiring appropriate numbers. In places like Mathare, you’d have new structures built before you even finished mapping and naming the community.

So this appears to be a case of high infrastructure costs preventing people from obtaining usable addresses. But the very system of addresses is a strange, pre-modern relic. Plug an address into a web mapping service, and it has to guess what the actual coordinates of that address are so it can give you a map. For my childhood home, Google Maps currently points to the wrong driveway. Moreover, adding new addresses can be problematic: what if there aren’t enough spaces between numbers? How to indicate multiple units in one complex? In the future, I suspect that we will hand out the GPS coordinates of our homes as often as we give people our street numbers (and by hand out, I mean send directly using a smartphone app).

Africans, many of whom lack addresses and need to walk all the way into the center of town to do things like receive mail, have an opportunity to get there first – skipping over street addresses entirely. High-quality GPS receivers already have a 3-meter resolution, enough for all but the most densely-packed slums, and advances in the system will allow ever-greater precision. If ground-based enhancements are added then even current technology allows centimeter-level precision. I’m not yet sure where the money is in promoting this innovation, but if I were an entrepreneur looking at growth areas in African markets, I’d begin with GPS coordinates.

*For my technologically-deprived US readers, I should clarify that in Kenya you can use you phone to pay for things and transfer money, and this has been possible since 2007. For anybody reading this outside the US, I should note that we still use paper cheques to send money and pay for things; it’s like time-traveling back to the 1930s.

With modest embellishments, the movie’s depiction of commodities trading is accurate and realistic, according to the latest Planet Money episode. I was surprised.

The end of the episode also has a neat little explanation of how Billy Ray and Winthorpe’s short-selling scheme works. Also described in this article. The key detail is that it is possible to sell a promise to sell FCOJ (or anything else being traded) at some future date and price. If the price is expected to go higher than the agreed-on price, people will jump at the opportunity – and if it falls, you can make a tidy profit. The protagonists, knowing that the rally in orange juice is based on faulty information, win big and simultaneously bankrupt Duke & Duke. Awesome.