Experimental Design: the Randomized Blinded Study as an Instrument to Reduce Bias
From the previous chapter it should be clear that in most cases the before-after comparison in a single group of patients is an inadequate experimental design, as it fails to achieve comparisons free from bias. “Before” is not a good control for “after”, since the effects of many factors are mixed with the effect of the treatment, introducing all kinds of systematic errors. Generally, in this type of experimental design, bias has the effect of simulating or exaggerating the effect of the treatment.
KeywordsRandomization List Assessment Bias Allocation Ratio Simple Randomization Stratify Randomization
Unable to display preview. Download preview PDF.