Determine the types of biases that could compromise your research. Also take into consideration your own personal beliefs. While there are various types of bias you need to watch out for, understanding any influences your research is susceptible to helps fend off a particularly egregious case of bias. For example, when writing a research paper covering the highly controversial issue of abortion, be aware when your own opinions start to take control of the research. On the other hand, a research paper on quantum physics is less susceptible to emotion. Quantitative bias is a more likely culprit.
Acknowledge the design bias in your research. First, try to include as many variables as possible to lessen the effects of design bias. Second, understand that it is nearly impossible to create the perfect, unbiased research paper no matter how hard you try. Lessen the effects of design bias by acknowledging the shortcoming of the experimentation in the research paper. This gives additional credibility to your paper.
Include large numbers of samples to avoid sampling bias. Sampling bias occurs when a researcher omits or over-includes one type of variable. This sways the results. Larger and more varied samples reduce omissions and over-inclusion biases.
Read any interview questions you have with an independent party to analyze interview bias. The language in your questions can steer responses in a particular direction or prompt a particular answer. It's difficult for the question-drafter to see this bias, so another person -- preferably someone without a stake in the research -- can look over your questions and look for biased phrasing.
Give outlying results the appropriate attention. Some research inevitably produces one or two results that do not fit in with the rest of the data. These are called outliers. These outliers should not be overemphasized as this produces what is called a false positive, a common type of bias. Outliers should be duly noted and analyzed, but never portrayed as significant.
Control the manner in which data is collected to avoid measurement bias. Measurement bias can compromise quantitative scientific research through a poor measurement scale. This, in turn, produces bad instrument measurements. For qualitative research papers, consider that test subjects also have their own biases. You can effectively protect your paper from a test subject's bias if you can accurately forecast what that bias or biases may be.