# By the Numbers

###### 3 comments on “By the Numbers”
1. Romy Joseph says:

Regarding population growth: Malthus says that population will grow exponentially when UNCHECKED. I used statistical linear regression to check the linearity of the the graph of population vs. time (in years). The correlation coefficient is almost 1.0, which means that population growth is linear, not exponential. Simply put, Philippine population growth is already CHECKED as per Malthus, a debacle to those who claim population is growing in a fast pace.

By the way, my friend Abe Llera, requested me to post this for we debated in facebook against pro-RH.

Hope this helps.

2. Romy Joseph says:

On GDP vs Population, even if there is a relationship, statistical science tells us that correlation is not causation. Which means the GDP growth can’t be attributed to population growth only. In short, GDP = f(population) can’t be established.

As an example of correlation not causation: statement 1 – my son’s height increases yearly. statement 2 – population increases yearly. There is correlation; therefore, population increase is caused by my son’s height.

Hope this helps too.

3. timothy2011 says:

Thanks, Romy, your explanations help a lot in providing the appropriate context for the numbers provided.

Yes, in science, a correlation alone certainly never implies causation. Any scientist knows this. For causation to be established, there are other conditions that MUST be satisfied together with correlation, such as: temporal precedence and the ruling-out of competing explanations for the association between variables, in this case, GDP and population growth. Typically, scientists establish casual relationships using the experimental method wherein the necessary conditions are set up and the hypothesized cause is manipulated to produce the intended effect.

When the whole population of a country and its economic performance are involved, it is difficult to test for causation for obvious ethical and practical reasons. In such cases, another way to test a causal hypothesis is to examine a causal model using a non-experimental design that approximates the conditions for establishing causality, such as using longitudinal data and statistically testing which variable — GDP or PGR — appears to be the better cause of the other, while statistically controlling for the effect of other variables (i.e., competing explanations) that might be related to the outcome.

In the case of GDP and PGR, such causal models have been tested by economists using data points from many countries spanning 40 to 50 years worth of GDP and PGR data in Asia and South America. Not surprisingly, the findings of these studies suggest that there is no clear-cut relationship between GDP and PGR, much less a causal relationship.

timothy2011

Top Posts & Pages