Can You Lie With The Data?

Today more than ever the question, can you lie with the data? charges special value. The amount of fake news that is circulating in the media and social networks is causing havoc in many parts of the world. Movements such as the anti-vaccine and anti-mouth mask are examples of the effects of assuming data, which although true, are far from being understood in its proper dimension and within its entire context.

The biases in reading the data are also present in market research. Quantitative analysts with their experience are able to identify these areas of opportunity when presenting data. However, it is never necessary to bear in mind some examples of the incorrect use of data and information in order to base our decisions with objective and truthful inputs. Here is a list of the most common mistakes when communicating and understanding data:

  1. Decontextualize the data. In recent days we have witnessed headlines that state that most new cases of a certain disease occur among those vaccinated. This statement is only half true in countries with extensive immunization coverage. Suppose that in a population of 100 people, 95 are already vaccinated and 5 are not. In that same population today there are 20 cases of a disease: 16 among those vaccinated and 4 among those who do not. In this example, of the 20 patients, 80% are vaccinated people and 20% of those who are not. One media could affirm that most of the patients are vaccinated, but at the same time it can omit that it is also true that of the total inhabitants of this fictitious town, 80% have not become ill and that among all those not vaccinated, the percentage of patients also is 80%.
  2. Infer data without exhaustive knowledge of the universe. The design of a quantitative study that seeks to know the characteristics of a population will seek to obtain a representative sample of all the segments of the universe or, if possibilities allow it, will take a census of all its members. However, in some countries we have seen that the tests to detect a disease are limited to certain groups: mild and severe symptomatic. In this case, it is impossible to know information such as the general rate of positivity since the non-symptomatic and those who refuse to take a test are underestimated. This error is very similar to that of surveys carried out on social networks: only those who can view the invitation, who have an active account and who wish to participate will respond.
  3. Do not contrast the data. On many occasions, the study we analyze is the first exercise of its kind. However, in other cases it is possible to compare our data with those of other investigations in such a way that we can verify whether the findings are true or have some degree of error. Even if it is not possible to find similar data, it is possible to identify approximation indicators that allow us to know whether or not we can trust some figures. In the case of the health crisis that we are experiencing, the best example is the comparison of different databases to find out if the figures for the same indicator are within statistically valid expected ranges or there are wide discrepancies between different sources that make us dismiss certain data.
  4. Ignore the relativity of the numbers. Although it may seem obvious, it is not the same to work with absolute numbers as with relative ones. Some people support their actions by pointing out that the percentage of deaths worldwide from a certain disease is just over 0.05%. Initially, this percentage may be lower, but what is not stated at the same time is that 0.05% of a total of 7.8 billion people are almost 4 million. Another example of this type of imprecision is when speaking of minorities in India or China. In the case of India, a minority of 1% of the census population means 13.72 million people.

The execution of quantitative studies requires great attention to detail and unrestricted adherence to industry standards. Acertiva’s quantitative analysts are committed to the correct and truthful design, execution and presentation of research findings. We are attentive to your messages to know your needs in depth and to put our experience in market research at your disposal. Write us now.