Opiate or Antidepressant

"Religion is the opiate of the masses" may be Karl Marx's well known saying, but the InterAmerican Development Bank (IADB, BID in Spanish), in a study released today, is putting a new spin on it.

According to them, religion is actually an antidepressant.

And Honduras, the authors say, is one of the least depressed countries in their international sample.

This was reported in La Tribuna as "IADB: religiosity reduces the depression of Hondurans", citing the EFE original.

Sound good? well, not so fast: take a closer look at how the study was conducted, and the conclusion seems a little, well, dubious.

The 25 page study, which is part of the IADB's sponsored research on the "Quality of Life", is the work of Uruguayan economists Natalia Melgar and Máximo Rossi. It began with the observation that past studies have indicated some correlation between personal life characteristics and risk factors for depression and sought to extend previous findings by exploring the relationship between specific environmental factors and depression.

To summarize, the study found the depression is

positively related to being a woman, adulthood, divorce, widowhood, unemployment, and low income.

Not surprisingly the study found depression correlated with living situations where there are high levels of inequality, especially in urban areas. But the study didn't actually find as strong a relationship as the authors expected between poverty and depression. And that's where the findings about Honduras that La Tribuna chose to emphasize come in.

The study's authors found that Honduras had one of the economically most uneven distributions of wealth, concluding that "the three least equitable countries are Bolivia, Brazil and Honduras". Yet Honduras had a lower level of depression that would be expected, which they argued "may be explained by the very high percentage of people with religious affiliation".

The wording here is significant: "high percentage of people with religious affiliation". So what did the study actually find?

The study relies on data from a CID Gallup poll in 2007 to assess depression. The question Gallup asked to assess depression was "Did you experience the following feelings during a lot of the day yesterday? How about depression?" The possible answers were "Yes", "No", "Don't know", and "Refuse". Approximately 80,000 people in 93 different countries answered either "Yes" or "No" to the question. Overall, 14.63% of the international sample answered "Yes" to the question, with 85.37% answering "No".

The most depressed country in the sample? Ethiopia, where more than 51% of the respondents reported being depressed.

Religiosity was one of the individual (personal) factors the study's authors tested against depression. Being religious was measured by response to a question asked in the 2007 CID Gallup poll about having attended a place of worship within the last seven days. If a person answered "Yes" to that question, they were considered "religious". They found that attending religious services had no significant effect on depression.

So, in fact, the authors conclude that religiosity-- a personal characteristic-- is not related to depression.

The study's authors, though, went further. They were interested in how what they call macro level factors-- characteristics not of individual people, but of countries-- affected depression.

So they used data on religion: the percentage of the population that was Catholic, Muslim, or Protestant. Each of these had a significant negative correlation with depression: the higher the percentage of the population affiliated with one of these religions, whether or not they attended church services, the less likely the country's populations was to be depressed.

Nine of 14 countries with the highest income inequality that unexpectedly had the lowest probability of depression also had a high proportion of the population that belonged to one of the major religions.

The list includes Honduras, Panama, Niger, Senegal, Jamaica, Uganda, Brazil and Mozambique.

But there is a problem here. For unexplained reasons, the study's authors used a 1980 source for the proportion of the population belonging to the major religions. So somehow, the proportion of the population belonging to a major religion in 1980 is correlated with the degree of depression in 2007.

Never mind that the US State Department 2008 report on Religious Freedom in Honduras states that

there are no reliable statistics on religious affiliation.

The State Department cites the same 2007 CID-Gallup poll used as the source for self-reported levels of depression in the IADB study as the best source for estimating religious affiliation in Honduras.

According to the State Department, the 2007 CID Gallup poll says that 47% of the Honduran population self-report as Catholic, with a further 36% self-reporting as Evangelical Protestant.

Melgar and Rossi do not state where their 1980 data come from, or why they used those data in place of the more recent results from the same poll that provided their data on depression.

From a purely social-science perspective, the IADB study is at best reporting a correlation, which does not say one factor causes the other. Both may be effects of some third factor.

Using data from two completely different surveys separated by 27 years means that the populations sampled were not the same. To see these as cause and effect would require a mechanism through which the religious affiliation of Honduras in 1980 led to a situation in 2007 that caused less depression among a later population. The mechanism involved is, shall we say, hard to imagine.

The fact that actual practice of a religion in 2007 did not correlate with levels of depression in 2007 is a hint that perhaps the 1980/2007 relationship is not real.

According to the 2007 Gallup poll, 11.96% of Hondurans reported being depressed. No comment on what the numbers would be in 2010.