4th Do you have numerical data?

2020/11/12 Elhuyar Zientzia



If in an affirmation you can measure what you want to express, that is, if you can argue with numbers, it will be much more credible, more objective. What cannot be measured, that is, the ambiguous and the qualitative, can have many explanations and subjective interpretations.

We will explain it with a simple example: It is autumn and in this season time is very variable. That is, one day it can be hot and the next cold, one day it rains and the next it is sunny… You are at home and you have to decide what clothes you wear to go out to the street. You look out the window and see a neighbor sweating dressed in shorts and short-sleeved t-shirt. It's clear! It's hot out! You've had to get out of the house and quickly get back to changing clothes because you were cold.

What we see many times is not what we think and what the saying says can be fulfilled: “The empty half and the other half blew,” so you have to look for evidence to see if the beliefs are true. In this case, a critical thinker would analyze the information and try to find evidence to check what he thinks. That is, you would try to know the temperature of the street. It indicates that the mobile phone app performs 12º C on the street. Now yes, through measurement you have achieved an objective data and you have a strong argument to decide what clothes you wear.

Why was the neighbor sweating? Many reasons can explain what you've seen from the window: for example, it comes from sports. What is clear is that the initial observation (looking out the window) was not enough to draw conclusions about the temperature it makes and make decisions from those conclusions.

The example makes it clear that the arguments will be more consistent if the criteria for deciding something or analyzing the credibility of an information are measurable (or based on credible data).

All this does not mean that something immediate is false; undoubtedly, many qualitative issues emerge truths, but without relying on data it is more difficult to defend its credibility.


Importance and risks of statistics

We have seen that arguments, when completed with measurable parameter data, are more credible, but these data must also be credible, so the data used to argue any claims are also correct. Scientific data is useless without statistics and must meet minimum requirements to be acceptable.

The sample, for example, is very important for drawing conclusions based on data. For data based arguments to be consistent it is necessary to use statistically representative samples.

For example, suppose we want to know the opinion of the Basque population about bullfighting. We prepared a survey to analyze the data and asked 100 users of a nursing home in Donostia for their opinion on bullfighting. 75 people have responded that they like bullfighting, another 20 do not like bullfighting and the rest do not think about bullfighting. According to these data, we can say that 75% of the population of the Basque Country is attracted by bullfighting. Would you say the above statement is credible?

A critical thinker would be clear that the above statement is not correct. On the one hand, because the sample size is not adequate; it is very small to conclude the opinion of the Basque population on any subject. On the other hand, because only the inhabitants of Donostia have been asked, so it does not collect the opinion of all Euskal Herria. And finally, because it has been done to users of a nursing home; it does not reflect the opinion of people of all age stripes.

With the example above it is clear, therefore, that to believe the data you have to check the origin of the same and if they are correct.

Correlation is another statistical concept that must be used very carefully when formulating an argument, since correlation does not always mean causality. Causality refers to the cause and effect of a phenomenon, in which something directly causes a change of something else. Correlation is the comparison or description between two or more variables. Therefore, correlation does not always mean causality, that is, that two phenomena occur simultaneously does not mean that one has produced the other.

There is another associated fallacy, known as “Cum hoc ergo propter hoc”, which says that when we have two events that occur together, one is the cause of the other.

For example, let's look at the following argument: Many drug users have psychiatric problems and many with psychiatric problems use drugs. Therefore, drug use generates psychiatric problems. Although the conclusion may be true, the argument is false, since the correlation between drug use and psychiatric problems cannot guarantee the cause-effect relationship. Drug use can cause psychiatric problems, but it can also happen that psychiatric problems cause drug use, or that both are due to a third party, or that there is no relationship between both facts and is a coincidence.

When two events occur at once, it may be tempting to admit that one causes the other, but in addition to the statistical correlation, more information is needed to properly conclude that there is a causal relationship between one and another event.


Lichen measurement:

As discussed in the previous paragraphs, it is easier to argue for a measurable statement or information, but measurement should be objective. Therefore, the number of lichens of information received from social networks does not guarantee credibility, and the taste of many people does not mean that what is indicated is true.

If by reading an information people agree with what is said in it, they receive many lichens, but that does not. This is closely related to another fallacy called argumentum ex populo. Ex populo arguments are used in populist discourses, in politics, in the media and in everyday debates, using phrases such as: “And I don’t say it, everyone says it,” “Most people have the same opinion,” “Everyone knows that’s so,” etc.

This fallacy is based on the inadequate use of logic, since things are never true because someone knows them (or because everyone knows them), but because they coincide with the evidences shown. Therefore, be careful with the information that many Likes have or become trending topic, since the number of followers does not guarantee their credibility.


To download the document click here.