Positivism offers the best approach to research in academia because it demands large data samples. Positivists argue this because they believe that more data yields more representative and accurate results. The positivist approach is grounded in data, they rely on it to test their hypothesis. The use of a large data set produces more representative and accurate results, thus creating greater certainty and credibility to the research and its conclusions. Large empirically testable data is used to generate theories of social behaviour which are grounded in hard evidence, allowing theories to be applied and replicated in other social contexts.  Data which relies on small samples is subject to error because each piece of data has the potential to alter the results of the sample. There is emphasis on large data samples because to use a small data sample introduces bias to the research, something positivists reject. A small sample size collects data from a small subsection of the total population and will only reveal the views of that subsection, not the entire population. This creates bias because it presents the evidence of a select section of the population. If for example polling stations only collected the opinions of men, or of people in cities, the results would be biased toward their opinion rather than the whole population’s opinion. A positivist approach to sociology and other social sciences look for four things to test their hypothesis: validity, reliability, accuracy, and representativeness. These are achieved through large data sets. By using a large number of sources, the positivists can make a decisive contribution to their field. One source alone can be ambiguous and lead to false conclusions, however using a variety of source types the researchers can formulate accurate theories and reach almost foolproof conclusions that cannot be easily discredited.
The principle drawback of positivism's reliance on large data sets for the bulk of their evidence is their unquestioning approach to the results. Because positivism assumes data, statists, and numbers to be objective truths, they fail to account for the possibility of skewed results. Data can be inaccurate if collected badly. Especially with large data samples which rely on large bodies of people providing data, there is a risk of unauthentic and constrained responses. Unlike qualitative methods, there is no direct supervision over the responses and nothing to guarantee that respondents are answering truthfully and authentically. A respondent may simply choose to randomly tick boxes to save time. At the same time, respondents who do seek to answer truthfully may be constrained by the options provided. If the participant's honest and authentic response is not among one of the provided responses, the data suffers because it does not capture the true response. Large data sets are only as good as the the methods through which the data is collected. Therefore, an over-reliance on data as objective truth fails to account for human error in collection.
Rejecting the premises