Unintended consequences are rarely a part of the conventional wisdom.
This is, of course, is what makes conventional wisdom so dangerous; it thrives on the “obvious” and erases “what if” from the conversation; it replaces evidence with bold assumptions and words like “of course!” and “naturally!;” it kicks out critical thought and opens the door to groupthink.
In development – any field, really – it’s easy to get caught up in conventional wisdom. Naturally, giving out free shoes to kids is a great thing; of course voluntourism isa great way to harmlessly give back! Sites like WhyDev, A View From The Cave, and many others push back on it, but it’s ever-present.
In a way, mHealth, or mobile health, is in this stage right now. It’s one of the most exciting and innovative areas in global health; the ubiquity of mobile phones in sub-Saharan Africa and elsewhere has led many to pin their hopes on the technology to improve health outcomes, educate children and adults about the scourges of HIV and malaria, and connect individuals with virtual health care providers.
In developing countries, the data has so far been a bit thin on its broad-scope efficacy, but of course only good can come from it; what’s the harm? A recently-released paper by Julian Jamison, Dean Karlan, and Pia Raffler of Innovations for Poverty Action (IPA) highlights why an uncritical eye can be unwise, even harmful.
IPA, founded by Karlan a decade ago, is known for conducting rigorous randomized controlled evaluations in order to find out what works in development, then scaling promising interventions. It’s also known for an irreverent, skeptical attitude towards the conventional wisdom; for always asking, “is this really the best way?” and “how can we do better?”
Here’s the study: it’s Uganda in 2009, and Google, the Grameen Foundation, and MTN (a Ugandan cell network analogous in its ubiquity to Verizon or AT&T in America) have just worked together to create a health information “bot” that automatically answers questions about “sexual health, family planning, and local health services” via free text messages. The multi-lingual service allows users to ask questions in English or Luganda (an official language widely spoken in Uganda) by texting a number, and an algorithm guesses the information they’re looking for. Like that, a teen with a question about sexually transmitted diseases or condom use has the information he or she is looking for.
IPA thought this would make for an effective trial, so it worked with the group to randomize sixty villages in central Uganda; some received a “high-intensity marketing campaign by a professional marketing firm” about the service, and others didn’t (though the service was technically available to them as well).
The conventional wisdom – and the hypotheses tested – holds that having access to private, objective, and reliable sexual and reproductive health information should increase health knowledge, which would lead to a drop in risky sexual behavior, change attitudes towards safe sex practices, and reduce infidelity. Or, at the very least, it wouldn’t do any harm – how could it?
But it could, and it did; Jamison, Karlan, and Raffler found:
…no increase in health knowledge regarding HIV transmission or contraception methods, and no change in attitudes. Rather than seeing reductions in risky sexual behavior, we actually find higher incidence of risky sexual behavior, and more infidelity, although more abstinence as well. Overall, individuals perceive their behavior as being riskier, which could be an indication of better probabilistic assessments but is also likely a result of the riskier (self-reported) behavior and possibly a desire to answer the surveyor in a particular way.
So, what’s going on? Why is there a higher incidence of risky sexual behavior and more infidelity among users of the service? The authors don’t have conclusive data, but suggest that “sexual sorting” may have occurred; women, having learned about the risks of unsafe sex, demanded safe sex (and, failing that, denied sex altogether). Men, who perhaps didn’t want safe sex, found another girl to have risky sex with, cheating in the process.
That’s a pretty significant unintended consequence for an intervention designed to reduce risky sex and infidelity.
But, wait: that wasn’t part of the conventional wisdom! Those results are the antithesis of conventional: how many would have expected them? In retrospect, it makes a certain kind of sense, but would be difficult to guess in advance.
Jamison, Karlan, and Raffler have shown why it’s necessary to continually challenge the conventional wisdom with skeptical minds and rigorous evaluations; let’s all take their example to heart, and continue to critically look at our projects and other interventions hailed as the “obvious” solutions to the most vexing issues.
 For those both young and old enough to have grown up using AOL Instant Messenger, it was kind of like SmarterChild – though smarter, I hope