DC and Forum: Please forgive me and correct me if I am misinterpreting the information presented, or merely oversimplifying it. Also, I may be ignorant of essential characteristics, and thus off the beam.
Apparently 1, there are measurable differences between the results of the two studies; 2, observer bias may or may not be equal and in the same particulars in each of the studies; 3, there is "certainty" that observer bias affects the usefulness of one or both studies; 4, observer bias is the only possible defect in the data. I wonder about so-called "natural" variations, and whether the studies were long enough (even though two years [?] seems long by current standards) to account for the non-observer noise of such fluctuations? That is, is it actually possible to parse or tease out the observer bias at all? It seems that the answer might lie in the basis for the "certainty" that observer bias exists (it possibly does), but that information is not given. WT At 07:30 AM 1/11/2007, D.C. McEwen wrote: >Greetings, > >Does anyone know of methods to account for observer bias in an intervention >analysis context? We are looking at anthropogenic impacts on invertebrate >community structure and production. We were commissioned to do a >traditional BACI based on our work (AFTER DATA = 2003 through 2005) and >another researcher's (BEFORE DATA = 1984 through 1986). Even though we >replicated their study as closely as possible, I am certain that much of the >difference we will find can be attributed to observer bias (e.g., skill >level in identification, time used to pick invertebrates from samples, >equipment differences, etc.). If there is anyway to attempt to partition >out the related variance, I would like to do so. > > Thanks. > > >Daniel C. McEwen, Ph.D. Candidate >North Dakota State University >Department of Biological Sciences >Stevens Hall 214 >Fargo, ND 58105-5517 >Phone: 701-231-6561 >Fax: 701-231-7149 >URL: http://www.ndsu.edu/ndsu/biosciences/students/dmcewen/ >Advisor: Malcolm G. Butler >http://www.ndsu.nodak.edu/ndsu/mbutler/
