Researcher bias, also called "experimenter bias," is what happens when someone conducting a study intentionally or unintentionally influences the results based on their own expected outcome. An example of an intentional researcher bias might be creating a survey which asks the question, "What is your favorite food," and offers the possible choices of cake or dirt. Naturally, any respondents would feel that they had to answer "cake," because dirt is not even a food item. Of course, there are many other foods that might be a person's favorite. Here, the researcher has intentionally created a survey where there is really only one reasonable answer.
Unintentional researcher bias often stems from poor research design or a simple lack of experience and understanding. Let's imagine a developmental psychologist is interested in learning about how children like to spend their free time. To learn more, they have filled a room with toys, art supplies, and a television, and ask a number of children to go in and spend some time. While toys, art supplies, and television are likely to be things children do enjoy, their choices are limited here. The psychologist isn't really going to learn anything about how children like to spend their time unless it involves one of those three hobbies. This is an example of poor experiment design because it is too limiting, and it may very well stem from a lack of experience with children outside of a research setting.
In these two scenarios, it could be easy to understand how the researchers were biased in their studies. In the first, the researcher clearly wanted people to say that cake was their favorite food. In the second, the psychologist failed to account for other activities children might enjoy. We should also take into account the sampling methods used. Did the first researcher ask only pastry chefs to participate in their study? Did the psychologist ask their friends to bring their children in for the study?
Regardless of motivations, researcher bias does occur and takes persistence to overcome. The best ways to eliminate researcher bias is by framing open-ended questions and collaborating or otherwise sharing work so that it can be examined for biased elements.