Research Project

Introduction This research adopted the data which originally are suicide total number. However, if we use the suicide total number to analyze, our results may be wrong because suicide total number is affected by population. Model one and Model two is a contrast with two plots made by different data. The number of suicide in Asia is the largest one, but the highest suicide rate happened in Europe. My second result is that the worldwide suicide rate is decreasing as a whole. In addition, I connect it with the per capita GDP. It is recognized that the worldwide suicide rate is decreasing while the per capita GDP is increasing.

Research Questions

1. The relationship between suicide rate and GDP per capita
2. Hypothesis: suicide rates are negatively correlated with per capita GDP
3. Whether the hypothesis could be proved across the world, or across continents, or in developed or developing countries

Data Sets

The project uses the Data Science Labs version of gapminder, with the following additional data sets from Gapminder.org:

1. Suicide, total deaths

All data sets are available from Gapminder.org under a CC-BY 4.0 license.

Principal investigator

This project, Suicide Rate Research Project , was submitted on 21 June 2021 by Liyi Chan, ID: 006, in partial fulfillment of the requirements for ENG 3208A: Telling stories with Data, Shantou University, Spring Semester 2021.

Results Analyses

Suicide Number and Suicide Rate: 1990s

These two plots show us that the variable population will affect the result of my project. In addition, we found out that Europe is the continent which holds the highest suicide rate among the five continents. So next, we will grasp this continent to find more details. But now, Let’s back to the analysis of the world.

Suicide Rate: 1990-2015

This is a plot about the suicide rate around the whole world. We can see that the rate was decreasing from 1995.

Per Cap GDP: 1990-2011

This is a plot about the per cap gdp of the whole world. We can see that it was increasing from 1990 but decreasing from 2009.

Result Analyses

Suicide Rate ~ Per Cap GDP

According to first model, we gain that the suicide rate of Europe is negatively related to the per cap gdp. But they are roughly negative correlations, which are actually not particularly accurate.

EU Rates: 1990-2015

The second model tells us that suicide rates in Europe, over time, have been relatively low and concentrated in a big range of rates. But in the 2000s, there were more extremely high suicide rates. This proves that GDP per capita is not the main factor that affects whether the suicide rate is high or low, but there is a link.

EU Suicide Rates

The third data sheet tells us a frightening fact, that is, among European countries, Russia’s suicide rate has been high for a long time.We therefore examine the impact of GDP per capita in Russia as a representative of a European country with a high suicide rate.

Russia: Rates ~ Per Cap GDP

The fourth model makes it more clear that the negative correlation between GDP per capita and suicide rates is not very strong. Obviously, the change in the suicide rate in Russia was a sharp change between 1990 and 2000, and the change was on the rise. Its suicide rate has since declined. A review of other sources suggests that Russia experienced a severe economic recession in the 1990s after the collapse of the Soviet Union, which may have contributed to the rise in suicides. Unfortunately, our data can not confirm this result.

---

Result Analyses

Suicide Map 1995 & 2005

In this part, we will analyze suicide rates and per capita GDP in some developed countries. At the same time, we’re still going to apply the hypothesis to it. First, let’s look at 1995, which was the year with the highest suicide rate in the world since 1990. Russia had the highest suicide rate during the year. However, most of the countries with high suicide rates are in Europe and Asia. Data are missing for a few countries.

Developed Nations Contrast

First, we select several developed countries with high level of development, which are the United States, the United Kingdom, Japan, South Korea and Australia. Suicide rates in the five developed countries vary widely as a whole. Or even in two parts. On one hand, Japan and South Korea have high suicide rates; on the other hand, only South Korea has seen a straight rise in suicide rates. South Korea’s suicide rate is on the rise even as its per capita gdp is increasing.

Another Contrast

We then looked at the comparison of GDP per capita across countries by swapping the positions of the variables. We found only two countries that may fit our hypothesis. Japan has a much higher suicide rate than the United States even though its per capita gdp is higher. South Korea seems to have experienced a serious setback during this period, and this setback will have a long-term impact on the country. In short, our hypothesis cannot be tested in data from all countries. In other words, suicide rates may be linked to different main causes in different countries, and the causes may vary from country to country.

Result Analyses

Brics Nations Contrast and Another Contrast

1. Finally, we look at the five BRICS countries. Two of the countries that stand out are Russia, which has a particularly high suicide rate, and Brazil, which has a particularly low suicide rate. From the second model, we can see that Brazil’s GDP per capita has been high, and the suicide rate has changed very little. The closest countries to our hypothesis may be China and India.
2. If we link GDP with the level of economic development and suicide rate with mental stress, we can know from the above cases that a high economic level does not necessarily reduce mental stress, and maybe there is no positive correlation between economic level and happiness.

Overall View

In Overall View, we have ten nations here. By gathering them, we have a more clear comparison. There may be six countries consistent with our hypothesis. But I am not sure. So last part, we conduct a linear model.

The Reason I use Two Contrast Plots

The difference between the Brics Contrast (p1) and Another Contrast (p2) is the position of the variables, which was mentioned in Developed Nations. But why do I do that? If you look closely, in p1, we can clearly see the comparison of suicide rates in different countries from year to year. But we use a “round ball” to represent GDP per capita, which only gives us a rough approximation of the amount of GDP per capita. Thus, I made p2, where we can clearly compare the GDP per capita of each country and correlate it with the suicide rate.

---

Analyses

We further verify the above results through linear model.

1. Rates Summary: The summary shows us the average, minimum and maximum suicide rates for ten countries between 1990 and 2011. As we can see, Russia is the country with the highest average suicide rate, which also confirms the results of our previous data. What is more, Russia’s minimum suicide rate is even higher than some countries’ maximum suicide rates.
2. Model: We can see that the Brazil data is completely accidental. India, Russia, South Africa, South Korea, United Kingdom are 99.9% Confidence Level (CL). Highly significant (in more cases). Less than 0.1% of result occurring by chance. China is 99% CL. Less than 1% result occurring by chance. Japan and United States are 95 CL. Significant at this widely-used standard. Less than 5% result occurring by chance.
3. Next, we arrange the model coefficients in order. That could build up a visualization of Confidence Level of the ten nations. The lower (more unlikely) the p-value is, the stronger case we have for rejecting the NULL Hypothesis. Look at the model coefficients, we can see the p-values of United States, Japan, Brazil are higher than 0.05, which means there is no significant difference. For this, We do NOT reject the NULL Hypothesis. But for other countries, the p-value is low, so that we REJECT the NULL Hypothesis.
4. CI Range: If the CI contains zero, then the data have no statistical significance. The countries such as United States, Japan, Brazil, have no statistical significance.

Conclusion

Let’s make a conclusion. Through all the data analysis, I always reminded the original hypothesis that suicide rates are negatively correlated with GDP per capita. We looked at the distribution of suicide rates around the world, and then we took what we found – that Europe has the highest suicide rate of any continent – and put it to the test. It turned out that Russia was a major factor, with one of the highest suicide rates in the world.

Next, we analyzed data from five developed countries and five BRICS countries. The final conclusion is that our hypothesis is not necessarily valid, and some countries do not conform to this situation. Even in countries that fit our hypothesis, there may be other factors at play. As for the suicide rate, we should look for the reasons for the high or low suicide rate in a specific country, which will help us better understand the psychological situation of the country.