Correct Response: B. Removing data because it differs from expected results would most strongly bias the researcher's analysis. Noting flaws in the research design that may have generated faulty data (A) does not introduce bias into the researcher's analysis of the data, but does suggest there are possible problems with the data set. Changing the way the data is presented graphically to improve readability (C) or how the data is summarized and averaged (D) does not bias the analysis as long as graphics are clearly labeled and averaged values are clearly described as averages.