Tuesday, October 28, 2014

Use Caution When Reading Research Results

Yet again I came across research results that come from a study design that has at least one major flaw. 

Calves were assigned to either a control group or one of four treatments. On one hand, since the calves were born over six days the group assignment made sure calves in each group represented the same profile of birth dates. 

On the other hand, no mention was made of making each group equally representative by either birth weight or immunity status (i.e., blood serum total protein levels). Just saying, "... and randomly assigned to one of five treatments," in my mind leaves too much possible bias in group populations given that they were measuring health and rate of growth of the calves in response to the experimental treatment.

Thus, when presented with study results I recommend asking about the methods used to assign the calves to the various treatments in the research.

No comments: