Ideally, to test a hypothesis you would use an entire population as your sample. This would ensure the most accurate results. However this is not feasible in most cases, as participation from every individual is very unlikely to happen. This is why we use samples, as it is a less time-consuming, more cost-effective way of collecting data.
However despite the manual and economical benefits, this process has many weaknesses. Quite often convenience sampling occurs, where students offer themselves to participate. An example of this is how we personally offer our own participation for the third year projects on SONA. Although this is a quick and easy way to gather participants, it is certainly not scientifically accurate, as people who proactively offer to participate in studies may have certain characteristics that do not represent those of the whole population.
A different approach is random sampling, a procedure where each member of the population has an equal chance of being chosen. Although this prevents sample bias, it is found that quite often a study will want to look at the diversity in a population, and therefore maximum variation sampling is used which finds unusual and extreme participants.
The vast range of sampling methods makes it difficult for researchers to find which technique will give them the results than can be most accurately generalised to the rest of society. One major issue is sampling bias, where a specific section of the population is over-represented due to its dominance (e.g. an age group or race). As anticipated, a larger sample prevents this.
Piaget’s influential work on children’s cognitive development highlights some of the sampling issues. His theory was primarily based on the development of his own three children. He used this minute sample to generalise his theory to children across the world, and although evidence has shown his timeline for cognitive development occurs in industrialised societies (Goodnow, 1969), in other countries where education is poor the children reach each stage later than Piaget suggested (Dasen, 1975).
It is not to say however that the bigger the sample the ‘better’ the findings. Although large samples do increase the statistical strength of your hypothesis, studies with small samples (e.g. case studies) can provide us with valuable insight into psychological conditions and ideas. Although case studies such as Freud’s do not provide us with theories to be generalised to the wider population, the extensive detail produced opens doors for future research on a larger scale. Funding will always be more accessible to those who have prior research showing their ideas producing results.
Inevitably, the generalisation of a sample’s result is continuously done in psychology. It allows us to make predictions about the wider society that one simply doesn’t have the time and money to do accurately. As long as one recognises the issues of over-generalisation, and perhaps sometimes accepts that not all samples can represent all of the population and restricts conclusions because of this, accepted and universal theories can still be created.