The fundamental purpose of collected data is to provide evidence upon which you can base a conclusion. The purpose of a conclusion is to arrive at a clarifying generalization after a thorough evaluation of all of the collected information. Your collected data becomes the support for your conclusion.
Bias, in data collection, occurs when the questions used to gather data provide a fundamental preference toward specific answers, such as if a survey asked the question: "Do you like the taste of your current, highly metallic and bitter water?" The question suggests that the recipient should answer "No," making the question itself biased. Bias, in a conclusion, occurs when the conclusion you arrive at is either not supported by the collected data or not founded on the data analysis technique you used.
Collected data is always in one of three forms: qualitative (opinion, observation or experimental results), quantitative (statistical) or mixed (a combination of qualitative and quantitative). Each type of data requires a different form of analysis, but they are always in these forms. A conclusion is in an evaluator form, wherein the collected data is tested against a chosen analytical technique to arrive at an observation. For instance, if you compare the statistical percentage of people who believe that their tap water tastes metallic to your chosen standard, which is a base average percentage of people in the area who feel the same way, you can determine whether specific places in the area are receiving poorer water.
Data is collected through specific techniques, including surveys, experiments and observations. From this large collection of information, some pieces of collected data are discarded after you evaluate which information is valid and useful. As an example, once you locate the area where individuals state most often that their water tastes metallic, you can eliminate outlier areas and focus the rest of your research on the affected area. Conclusions are not gathered. Instead, you arrive at a conclusion through an analysis of your collected data, compared to your standard. The result is that a conclusion is unique and supported by your collected data.