STAT3280FinalProject

Introduction

The world is increasingly becoming more secular, and no countries are spared from this shift. Recently I had the pleasure of taking a very insightful class at UVA where we discussed this shift: its historical context, key texts and theories about the shift, and budding counterculture resurgence movements against this shift. However, what no one can accurately explain is its cause. We discussed certain factors such as modernization, urbanization, and higher education and literacy levels, all of which have some merit. At the end of the class, it was nearly agreed that higher education levels had contributed the most to the resistance against organized religion. However, I’m still not so sure.

“Ten word answers can kill you…. they’re the tip of the sword.” - Josiah Bartlett (Martin Sheen), The West Wing

The world is too complicated for these vast changes to be explained away in a sentence, or a page, or even a whole book sometimes. Questions like these, ones with no real answer, always have fascinated me since I arrived at UVA, especially since there haven’t been too many asked in my Statistics courses. Great people like Max Weber and Peter Berger have spent their lives tackling issues like these, and they never got to use ggplot or build a shiny app to impose their worldviews. So, I thought I’d try to fill this gap as I continue my studies, in both Statistics and beyond.

Secularization in the Numbers

Let’s start by asking a few simple questions. What is secularization? Well, that’s easy – it’s the rate of people as a whole becoming less attached to organized religion and identifying as such. We will measure and test this trend throughout this project, to see where this trend is taking place. How do we measure this trend? The data featured here has been collected over these years and more by Pew Research Center to display the raw information that reflects societal shifts. It is the ultimate source of these figures, having collected from over 27,000 censuses, surveys, and more. The data only includes a few metrics – each nation’s population and religious makeup in 2010 and 2020, an RDI score calculation (more on this later), a ranking of this metric, and a regional average to view a larger outlook over time. It includes complete data from 201 nations across the world. As a reminder, this data only includes the years of 2010 and 2020. Different events, movements, leaders, and ideas have been born in between these years, and no country has followed the same path across this decade.

Let’s start by exploring the data a bit more. What does RDI mean? This metric can be confusing at first, but it’s very simple. We just take the number of followers of our 7 categories divided by the total population to give us a percentage of religious outlook. This can be seen in this bar graph here. By selecting any country, we see a stacked bar chart depicting religious outlook in each country surveyed. This can give us a greater vision of that our data contains, and this phenomenon.

Explore the data with this Interactive Shiny App

Take a look at some specific countries. What do you notice about countries where most people follow the same religion, like Afghanistan? What about a more even country, such as Singapore? You might notice a few patterns. In fact, this data shows us how so much can change in just 10 short years.

From the data, and this graph, a few trends are increasingly clear: . The world is becoming more secular in most areas and places. People are leaving all religions to become unaffiliated with any organized religion. . Countries with higher religious diversity experience more people that are unaffiliated with organized religion.

The measure of RDI

Let’s take a look at that second point. What is RDI, and how does it relate to the data? Well, think of it this way. If you live in a country where most people follow the same religion, it probably manifests itself more boldly in your everyday life. More women might walk around in niqabs, or you might see churches on every block. Businesses might be closed on Sundays. If you grow up with these ideas all around you, will you see this religion as a choice? Or might you see it as a tradition, a custom, a necessary part of life? These questions are far too complex for us to talk about further, but we can measure this trend worldwide.

We can see a direct inverse comparison between RDI score and “religious strength”, or share percentage of the largest religion in a country. This makes sense, as if one religion clearly dominates the share, then RDI will be quite low. This information is still very important for comparing RDI to other metrics, and whether the countries with low RDI and one dominant religion still experience the same secularization rate.

Note: there are several really interesting ideas here. Is religion tied to nationalism in countries where one religion dominates society? What does “dominating society” really mean? We won’t discuss these ideas, but if they sound interesting, I recommend you look at theories by Durkheim, Ammerman, or others.

Additionally, a simple boxplot of the data reveals that countries classified as High or Very High in Religious Diversity Index (RDI) have experienced greater rates of secularization. For RDI, scores globally range from 0.03 to above 9. Scores from 1 to 2 are classified as Low, and from 2 to 5.5 as moderate. Only scores above 7 score as very high, and there are very few of these. Now, let’s group each nation into these clusters, and measure their average rate of secularization from 2010 to 2020.

As we can see, it is quite clear that there is a correlation between having higher religious diversity and having a higher percentage of people who classify as Unaffiliated. There are several fascinating theories on this concept, none of which I can fully explain. However, think back to the point behind RDI. If you live in a city, or maybe a college town, where diversity of thought is more prevalent, the people next to you might not be as outwardly religious anymore (this is also studied and debated by sociologists) but instead keep their faith to themselves. In cultures where religions are celebrated and practiced more equally, then the social space is not dominated by any one. Thus, the factors we discussed beforehand do not apply as much. Religion seems more optional, more muted. Certain sociologists might claim that if you think religion is a choice, and not a tradition, then people are more often not choosing. However, once again, it is not so easily explained.

Regional Differences

Let’s explore this idea a bit more. Generally, countries resemble more the ones they border than the ones further away. We can easily identify regions from this idea, and Pew indicates that most countries (at least the ones studied) fall into one of 6. From the map of the world above, you may have noticed how regions tend to be more similar in change in unaffiliation from 2010 to 2020, but let’s look a bit more in depth at this idea.

We can take the trends we’ve already gleaned and start to apply them to other theories. If higher RDI can be correlated with lower religious domination, can we say simply that the rise in unaffiliation is tied to that? Not quite. Unsurprisingly, the world is too vast and different for that to be the end all be all. Let’s focus on regions – 6 is much less than 200+, after all, and this can lead us to our next theory.

Here, we have an interactive app that displays this trend a bit more in depth. Try this out for yourself: start by filtering RDI as much as you’d like. Which countries grey out from the map? Are these countries in the same region,or more spread out? Check the summary table as well – does the average secularization rate increase when you narrow the filter, or decrease? What about from the other side of the filter? Do you see the same effect?

Shiny interactive Map

Of course, it’s obvious some regions have much higher RDI than others – this is confirmed by this graph here. There is a large discrepancy between regional RDI scores - this must be a factor.

Measuring the Impact of Education

I’d like to come back to my original point. My classmates and I theorized that higher education rates contribute most to this trend of secularization. But is this actually true? Of course, it’s impossible to tell from one simple dataset or graph. However, the World Bank attempted to compile this information. They created a dataset measuring the ratio of people in each country who have completed secondary and tertiary education (some form of high school and university, or equivalent). This metric is one of the best ways to measure education levels, and thus we will use to measure against religious shift.

We will test the correlation between the change in education rate from 2010 to 2020 and the secularization rate from that same timeframe. If the correlation is closer to 1, then the data would suggest there is a relationship between the two, and my class has some standing.

Before we start comparing, let’s take a look at our education data. Is it increasing on average over time? The data is less complete than our standard religion data, but let’s verify anyway through a neat animation:

Looks good! We the world map get more orange over time, which is a positive indicator. The data includes every year in which the education survey was completed for each country. If a country was missing values for 2010 or 2020, I replaced them with the closest year possible for accuracy. The data has many missing values, so only 136 of 201 possible countries have acceptable data. However, this is as good as the data can get. So, what do you think the correlation will be between these metrics?

Unfortunately, it doesn’t seem the data supports our hypothesis. Coming into this project and this idea, I actually thought the correlation would indicate a trend. This point was going to be the main focus of my project. However, the more digging I did, and the more patterns I measured, it seemed as if the opposite was true: we cannot attribute this shift to one cause. It is an invisible force, slowly moving, yet seemingly unstoppable over the last 40 years or so. I only can measure the last 10, but this is true nonetheless. Not even one region can claim that its secularization is tied to its change in education rate.

The outliers to this theory

If education doesn’t explain the secularization trend, what countries are throwing it off? We can measure the residuals to our prediction line – the line on the graph above is what the expected secularization rate would be if education was a better predictor. Residuals are the distance from the point on the graph itself to that line. The further away from the line, the more unfit our guess was. So, with that said: which countries threw us off?

Some countries, like Australia and Chile, are way ahead of the curve on secularization. Others, like Ghana or Mozambique, are still holding on to their religious tendencies.

Can we extrapolate this by region? From the ideas we discussed earlier, it’s possible that regional differences make up the difference here. If the residuals are clustered by region, this could explain some of the difference.

It doesn’t look like it – the regions are following the same trends as we discussed above, where Africa lags behind with its lower RDI and higher religious domination. …

So if education doesn’t explain this momentum on its own, what does? I could keep making correlation plots to find the answer, but the keys that unlock the world won’t be found in a spreadsheet. Despite how many people throw themselves against the wall to find it, sometimes the answer really just is that there is no answer. Data can be a valuable tool to research areas like sociology, but there is no replacement for mere observation. What does the world look like, and how do we make sense of it when the data can’t make sense for us?

It looks like my class was wrong. Or were we? Again, we can’t say definitively either way. The world keeps getting smarter, and less religious, slowly and slowly over time. That can’t be ignored.

Conclusion

What was the point of this entire project? We didn’t prove anything, didn’t show any correlation, didn’t find the reason behind a phenomena. To me, that’s the key part: these ideas are complicated.If anyone claims they know the meaning behind secularization, or how to win an election, or how to invent the perfect hypothesis test, in just a few short graphs, don’t believe them. A semester-long discussion, backed by multiple talented, dedicated authors, professors, and students got it completely wrong. It will take time, and resources, to explain the reasons behind the world. It’s what makes the world worth it.