# 35. Statistical manipulation - Biased Sampling - Selective bias ## 35.5. Methodology/Refinements/Sub-species ### 35.5.1. Self-selection This is the accidental or deliberate selection of certain individuals to participate in a survey or to act as subjects in an experimental study. This happens when individuals nominate themselves as participants in a study or survey or when a study designer or pollster nominates individuals using a limited selection process. The result is a biased sample. "Self selection" is often used to describe a situation where the characteristics of the selected people in the survey create skewed figures. For example, a right-wing newspaper survey of its readers will inevitably give rise to results which are biased by the self-selected participants, i.e. vocal and right-wing. In commercial polling circles, a poll suffering from this kind of bias is termed a self-selecting opinion poll or "SLOP" poll. It is a common form of manipulative mechanism and fairly easy to recognise and discredit by its victims. #### 35.5.1.1. Pre-screening Another hybrid of self-selection is "pre-screening", where the organisers of a poll or study deliberately or accidentally screen parts of the sample population in order to manipulate the likely outcome of the poll or study. Methods of polling already "accidentally" force certain pre-screening biases on the sample population. For instance, telephone polls already screen out members of the population without a telephone. #### 35.5.1.2. Conclusion Self-selection makes it very difficult for a statistician to determine causation, because it is impossible for a researcher or a manipulator's victim to know whether the conclusions of a study resulted from real causes or simply from the biases of self-selected samples. As a result of self-selection, there are often other factors affecting an outcome than just the cause itself. Self-selection bias causes problems when evaluating a product's usefulness for ordinary members of the public and researchers, and it obscures the results of market research.