Question

Selection biases in marketing

Answer 1

What is selection bias? Definition
What is selection bias?
Selection bias: examples
Selection biases in marketing

---

Selection biases in marketing

Almost all the decisions we make are influenced by prejudices, largely unconscious. One of these cognitive errors is selection bias, also known as sampling bias or selection effect. It is a cognitive bias that leads us to select elements in the wrong way, for example, sample data for a study, which produces results that do not correspond to reality. This effect can have a significant impact on market research and should therefore be considered by all companies..

Index

1. What is selection bias? Definition
2. What is selection bias?
3. Selection bias: examples
4. Selection biases in marketing

What is selection bias? Definition

Definition

Selection bias: Due to this phenomenon, also called the selection effect, all people have unconscious biases that prevent them from selecting samples for studies impartially, which leads to distortion of statistical data.

Selection biases greatly influence the selection of information, not only in market research or scientific research , but also in everyday situations. Only by being aware of our vulnerability to cognitive biases of this type can we take advantage of its effects in a positive way. Realizing that data has been incorrectly evaluated can be critical for companies and bring them many benefits: By deliberately counteracting selection bias when conducting a study, you get more reliable results..

What is selection bias?

Selection bias is a statistical error that appears when sample units are selected for a study. To get really valid data and results, try to limit its effects by all means. In marketing, for example, selection bias affects the objectivity of customer surveys and other market research tools. The reasons for this bias are varied and can be found both on the side of the study participants and the researchers selecting the sample units. One aspect that must always be taken into account is people's predisposition to participate , in case it is not by chance . Some of the more well-known selection biases are, for example, non-response bias, self-selection bias, or survival bias. The latter is observed when, in a study on the success or failure of the participants, the results only reflect the data of those who have been successful or of the "survivors", always involuntarily.

The precautions and measures necessary to reduce or avoid the effect of selection bias are relatively complex. In this sense, statistical techniques, such as the Heckman correction , are often used in order to obtain correct results in empirical social studies or in market research..

Selection bias: examples

The effects of selection bias have been considered for a long time, and not just in research. Also in business and everyday life we selectively select information and expose ourselves to distorted data . Cognitive biases contribute greatly to making mistakes in selection processes and, consequently, to obtaining erroneous results.

People are victims of selection bias all the time, which shows that we are not impartial , but that we must try very hard to at least aspire to a state of neutrality. The following examples reveal the profound implications of this bias.

Suppose we want to conduct a general brand awareness survey of a healthy dietary supplement. If we carry out the survey in gyms, health food stores or organic supermarkets, we will only ask the target audiences of said product. This can be useful, but the results should be evaluated with caution, because they will already be conditioned by the selection bias: in general, customers of gyms, health food stores or organic supermarkets are more receptive to the effectiveness and usefulness of healthy products. Therefore, it can be assumed that these groups will know the brand better than the rest of the people and that, consequently, the data will not have been evaluated in a neutral way.

Our second example of selection bias reveals the serious consequences of not choosing the sample subjects truly randomly. Let's suppose that a research team wants to carry out a survey on the economic situation of a country that represents as much as possible all the companies. However, to select the data, it uses the commercial register, so it only chooses the corporations and commercial companies that appear in that list. In this case, the selection bias is much greater than it may appear at first glance: due to its effect, the study not only excludes small businesses, but also the many freelancers, such as lawyers, doctors, architects or artists. , and to professionals who work part-time in all sectors.

This mistake is very obvious, and more experienced researchers probably won't make it. However, selection bias can appear in studies in a much more subtle way, distorting something as important as a country's economic forecast.

Selection biases in marketing

Selection bias is a challenge especially for market research and not so much for marketing activities. Already when evaluating the success of advertising campaigns , which is ultimately also a type of market research, possible sample bias must be taken into account - for example, to achieve effective campaign results.

Self-selection bias, a variant of selection bias, plays a very important role in customer and user surveys . This bias appears whenever the participants can decide whether they want to take the survey. If people who do not want to participate have a significantly different opinion (for example, dissatisfaction with the company) than those who do (for example, satisfaction with the company), the sampling bias could cause customer satisfaction to be overestimated. In this case, the selection bias can and should be reduced using statistical weighting techniques .

Selection bias is a particularly complex challenge for marketing because it often appears in combination with other cognitive biases. To counter this adequately in statistical terms, all biases that may affect the study must first be detected. For example, it is possible that selection bias appears at the same time as publication bias (when only positive results are published) or self-selection bias , which we have already described (when only certain groups participate).

The smaller the number of samples and / or data that is extracted from them, the more likely it is that selection bias will prevent the development of marketing campaigns based on real data. Undetected errors can distort the results of any study or survey, to the point of causing them to be totally arbitrary and ultimately lead us to make wrong marketing decisions with disastrous consequences . Rather, a well-designed studio will ensure that unconscious effects, unwanted distortions, and potential manipulations are avoided.

Note

To collect and evaluate data correctly, other phenomena must be considered in addition to selection bias, such as confirmation bias, retrospective bias or halo effect, which can also be used to make campaigns profitable. In addition, in marketing, some cognitive biases can be used to increase brand equity and sales: primarily, loss aversion and the related endowment effect.

Please, take into account the legal notice related to this article.

---