Is scientific research well done?

Tabby Cat
Oliver-Bonjoch, CC BY-SA 3.0

Sometimes, while reading items published by journals such as Science News, it looks like some research currently being carried out is platitudinous. Either it leads to the discovery of things everyone knows, or time and efforts are spent to investigate in fields that no one cares about. We know that many researchers are anxious to publish, and they must justify somewhat the funds they receive, but up to that point?

Let’s look at a very recent news (April 2019):

Cats recognize their name. A study suggests our feline friends can tell the familiar sound of their name from other words. A paragraph of this news adds: As for whether or not a cat understands what a name is, well, only the cat knows that.

Anyone who has had a cat (I had one half a century ago) knows that cats recognize their name. Was it necessary to do a research about this, probably spending public money, to discover something that everyone knows?

Zero risk does not exist

No entry: radiation risk

We would like to live in a world where we run no risks, but that is impossible. Whenever we get into a car, cross the street, turn on the gas, or play sports, we run a risk. The most elementary acts of our life entail a risk: breathing polluted air; getting exposed to the natural radioactivity in buildings; passing under a roof just when a tile is falling down... We have always known that life is synonymous with danger, and we have adapted to that. In our time, however, it seems that the threshold of risk we are willing to tolerate has fallen down. In other words: we are now more cowardly.

The media are largely to blame. Trying to attract readers and increase their profits, they often encourage states of opinion close to panic. We can see it in the way many news are presented, especially those affecting health (mad cow syndrome, bird flu, SARS, type A influenza, whatever...); the viability of human life on Earth (global warming, collision with an asteroid); or the economy (times of crisis). Many of these threats are real, but they are systematically exaggerated.

Consequences of statistical ignorance

In a study performed a few years ago [1], the following problem was proposed to 18 expert consultants for AIDS patients:

Helen has tested positive for AIDS. How likely do you think she is really an AIDS patient? What would you advise her?

The input data are:

1.      The probability of having AIDS, when one belongs to a population without special risk, is 1 in 10,000.

2.      The sensitivity of the AIDS test is 99.9%. In other words, the probability of a false negative is 0.1%.

3.      The specificity of the AIDS test is 99.99%. In other words, the probability of a false positive is 0.01%.

The result of the study was as follows:

·         The 18 experts agreed that the probability that Helen is an AIDS patient is greater than 90%. Most thought that the probability is greater than 99%. Some even claimed that is greater than 99.9%.

·         All experts said that they would advise Helen to inform her family, make everybody test for AIDS, and start taking medication.

The slaughter of the innocents

This image shows a possible result of the tests
described in the post. Black figures represent
pregnant women who tested negative.
Red figures are false positives.
Only the white figures are true positives.
Above-left are the results
for the Down syndrome

A study performed a few years ago [1] describes a delicate situation regarding certain widely used medical tests. The triple test has been performed on many pregnant women in order to detect whether the fetus will have the Down syndrome and other deficiencies. When the test is positive, gynecologists often recommend an amniocentesis or a chorionic villus sampling (biopsy), but as these tests involve some risk, many pregnant women do not want to do it, and some may decide to have an abortion based only on the results of the triple test.

The Down syndrome affects approximately 0.13% of fetuses. The problem is that the triple test used to detect it has a 70% sensitivity (or what is the same, the probability of a false negative is 30%) and a 91% specificity, which means that the probability of a false positive is 9%.

How should we interpret these numbers?

Suppose the triple test is performed on 1000 pregnant women. By applying the above data we conclude that:

•         The likely number of fetuses affected by Down syndrome will be 1 or 2.

•         The probability of detecting them with the test is 70%.

•         90 women (9% in 1000) will test positive, even if their child won’t be affected by the Down syndrome.

On-line bullying

In an article by the press agency Europa Press published on June 28 2012, which refers to a study performed by Microsoft among youths in the age range 8 to 17, it is stated that 37% of Spanish youths suffer on-line bullying through the Internet. This looks like a high figure, but it may depend on how bullying is defined.

Reading the article, it appears that 17 per cent of the polled declares having been addressed in an unfriendly way, 13 per cent have been targets of mockery and 19 per cent have felt insulted. Also, 24% of the youth confess that they bully other people.

Neither in the Europa Press article, nor in the summary of the Microsoft study, is there a definition of unfriendly behavior and the other forms of bullying. It appears that the youths who answered the poll just considered it thus.