Bias

Bias results from collecting the data in such a way that one answer to the research question is given undue favor over another.  Since bias has to do with the evidence you collect to answer your question, it is a concern during the sample selection and data collection phases of the research process.

All researchers have biased views of their own questions.  You know what you want the results to be and, if you are not careful, you will unconsciously sway the study in that direction.  Therefore, when you can take precautions to maintain objectivity you should do so.  For example, if you have a choice between randomly selecting your subjects and subjectively picking them out of a group, you  should choose random selection.

Bias during the data collection means that the researcher either is influencing the responses of the subjects or is selectively recording observations according to conscious or unconscious predispositions.  Both of these problems are present in exploratory studies, especially those where the researcher is involved in the setting with the subjects.  If data collection is unstructured and the researcher interacts with the subjects to any great extent, it is virtually impossible to eliminate bias from the study.  It is possible, however, to plan for maximum objectivity by establishing some checks and balances.