International and national media have reported that the number of victims of sexual abuse in religious contexts in Spain is 440,000, with 200,000 directly abused by priests. In this short article, we want to provide a statistical reflection with the sole purpose of contributing to open paths to dialogue that will facilitate that people of different options can use the maximum of our energies in the solutions that prevent and overcome the sexual abuse of children. Of course, we know that everything we say can be completed, corrected, and even refuted by new contributions; we will be glad if that happens because that is how science advances to improve society. In any case, we also recognize that what we explain is a simplification that we commit ourselves to enrich with new statistical knowledge (confidence interval, random sampling error, and other errors, infinite population, etc.), if this article generates debate or questions.
There is an open debate about whether or not the 1.13% of people abused in the religious sphere in the survey sample accompanying the Spanish Ombudsman’s report can be extrapolated to the population as a whole in the way that, according to some media reports, would give 440,000 people. When we take a sample of a population, we are always making a mistake.
To calculate the error, there is a formula, but its automatic calculation can also be found directly on the Internet. For a total population close to 39,000,000, the sample achieved by the report’s survey (8,013 people) generates an error that the report says is at most 1.1 (for a 95% confidence interval). This would mean that the 1.13% of people abused could be 1.1 higher or 1.1 lower, i.e., 2.23% of people abused or 0.03% of people abused.
In other words, the percentage of people who have been abused in religious settings is exactly 1.13% of the sample, but we do not know what the proportion of the total population is, although we do know that it is between 2.23% and 0.03%. In absolute numbers, the possible range would be between more than 800,000 and less than 12,000 abused persons. Those who claim that the number is 440,000 do not take into account the error, they take 1.13% as fixed and multiply it by the total population considered, which gives about 440,000.
In reality, that 1.1 random sampling error is the maximum. I am sure that in the completed survey it must have been less, but today we do not know it concretely, with the information that has been made public so far. In any case, both science and all the people of different options who work in the service of the victims and survivors will continue to focus our energies on applying the scientifically proven solutions that have long been published by the international scientific community.