The absence of data screening

How might the absence of data screening affect a researcher's data quality, their interpretations of the data, and thereby their interpretations of their study's findings? What quantitative rule may be used to determine univariate outliers, and are there situations in which deleting a case/participant may be justified? Explain.