Dangerous Statistics - also fake news

Warning: Math ahead.

I wanted to talk about this because recently, I witnessed first hand how statistics can be manipulated to say basically whatever you want it to.

A friend of mine and I were discussing abortion and his argument was that if abortion is made legal, abortion rates will go down. I provided him the statistics for abortion in the United States which shows that following the legalization of abortion, abortion rates have gone up and have only started decreasing after 17 years and are still not at the same rate as it was when abortion was illegal. It must also be noted that 3 states have ceased to provide their own numbers on abortions which could mean a higher number for the raw number of abortions occurring.

See: https://en.wikipedia.org/wiki/Abortion_statistics_in_the_United_States

We can see the abortion rates have a high of 364 and a low of 188 induced abortions for every 1,000 live births. This goes from around 15% to 26% of pregnancies ending in an induced abortion. Now, what shocked me was how my friend decided to use the numbers. He got the total number of abortions and divided it by the total number of women (pregnant or not).

Even at the peak in 1990 with 1.4M abortions but since he divided it by the total number of women, he comes up with a high of 1.04% abortion rate and a low in 1973 and 2015 at 0.39%. His argument goes that one must look at the total population (all women) to see if abortion is an issue that needs addressing. He further claimed that for the government to consider it a threat, it must at the very least have an incidence rate of at least 1% since abortion would be a medical procedure (this is erroneous but I went with it). I pointed out that in a 5-year window, you could easily come up to your 1% threshold to make it matter. Adding the abortions of 5 years divided by even the total population of the US (~380M) could easily break the 1% threshold he required.

He did something strange though. He said that to consider a 5 year window, you'd have to multiply the population by 5 as well. At this point, I knew he was either knowingly lying or had absolutely no idea how statistics work. And he even claimed to be a statistician! Anyone knows that in a 5 year period, your population doesn't increase to 5 times!

I questioned him why he would consider as the population, all the women in the US as opposed to just the women who were pregnant (i.e., live births when you removed further miscarriages). His reasoning was that since it's a women's issue, you have to take the total population.

The Guttenmacher Institute, which advocates for reproductive health and abortion rights, doesn't hide under that kind of faulty statistics (though I'll have to recheck that). Even when one considers only women, they say that 1 in 4 women will have an abortion before the age of 45. That's a long ways off from 0.39%! (See: https://www.guttmacher.org/united-states/abortion)

To challenge his reasoning, I countered that the total number of people diagnosed with HIV from 1986 to the present, in the Philippines is at around 40,000 (apparently it's at close to 47,000) and with a population of 108M, that would come out to an incidence rate of 0.0435%. I then asked him if he thought that HIV in the Philippines is an issue that does not need to be addressed because the incidence is so low?

This is where it gets scary. He then proceeded to say that for HIV prevalence, one would have to look at a specific population (sex workers, people who inject drugs, a specific age group, etc.) and you could arrive to a more significant number where you can't ignore it. I pointed out to him that where he would like to see it being an important issue, he reduces the population size and when he doesn't want it to be an important issue, he wants to increase the population size as much as possible.

What makes this scary is that this guy is educated. He has a master's degree and it bother me that educated people can use accurate data to come up with inaccurate conclusions. As they are educated, people will often take their word for it and proceed to spread this falsehood.

It's very easy to do but luckily, it's very easy to spot as well. People who manipulate data will try to hide their manipulation by citing accurate data. Whenever I see percentage rates thrown around, I know it has two components - percentage and base (remember your high school lessons?). It's usually in the base where they try to sneak in their deception.

Once you see certain qualifiers, which narrow down national issues or broaden more local issues, it may be prudent to look at the source. I'll demonstrate this with a popular statistic that Planned Parenthood, an abortion provider in the US, usually states.

They claim that abortion is only 3% of the services they provide. Why then are pro-life groups so keen on defunding Planned Parenthood when abortions are just 3% of their services? Notice that while pro-life groups are protesting against funding (money), Planned Parenthood responds with number of services.

When you take a closer look at the stats, Planned Parenthood includes as one of their services things like, pregnancy tests, issuing contraceptives, checking vital signs, all as individual services. So a $500 abortion procedure would be counted the same as a $10 pregnancy test. While the numbers are true, they're meant to mislead.

Be careful with statistics you see online. If you were to ask me for a rule of thumb, I would say that statistics that sound unbelievable are probably unbelievable. However, if you're able to find sources on both sides of an issue cite the same stats, you're most likely in the clear.