## Biology Tutorial

#### Intro

In scientific experiments, scientists must be careful to eliminate bias from their studies. Bias is a preference for a result that may or may not be accurate. It may be intentional or unintentional. It can appear in the design of an experiment, the selection of a sample, the recording of the data, the analysis of the data, or in the formation of the conclusion. Peer review, where a study is examined by other scientists before its acceptance, is a measure against bias.

#### Sample Problem

A scientist is researching the various causes and risk factors of cancer. She intends to use this information to help the general population know if they are at a higher risk. She collects data from a cancer ward in a hospital in New York City. This data includes lifestyle questions about diet, smoking, and profession. She analyzes the data to find what factors are most frequent. One of her findings is that most of the patients had lived in a large city for most of their life. Why should she not conclude that living in a city can cause cancer?

#### Solution

While her sample size may have been large enough to be statistically meaningful, she only drew her sample from one hospital. This hospital is in New York City, a large city. It is logical that many of the patients were from the area and likely lived in the city.

This would be an unintentional case of bias caused by the sampling.

A stronger experiment would involve samples drawn from multiple hospitals in a variety of areas to gather more data. Cancer is not regional, so the risk factors should be consistent no matter where the people are.

There is another problem with her experiment. She did not have a control group.

This means that she could find that most of the cancer patients ate meat, and she might conclude that eating meat is a risk factor. However, if the percentage of cancer patients who eat meat is the same as the percentage of people who eat meat in the general population, there is no reason to think there is a connection.

A silly example to illustrate this: Most of the cancer patients speak English. Therefore, speaking English must be a risk factor for cancer. That is absurd. Comparing the percentage of people who speak English in both the control group (the general population) and the experimental group (the cancer patients) should have the percentages about equal. There is no significant difference, so speaking English is not proven to be a factor.

Be aware that having a control group may not be sufficient to protect an experiment from bias. The control group must match the experimental group. For example, if the hospital is in New York City but the control group is drawn from people in France, there would be a difference in the percentage of people who speak English. You must control your controlled variables between the control and the experimental group.

So, for a sample drawn from a hospital in a large city without a control group for comparison, our scientist cannot conclude that living in a city can cause cancer. She does not have the evidence due to her sample selection. If she makes the conclusion anyway, she has a biased conclusion.

# About The Author

 Erin W Biology Tutor With ACT Help, English Vocabulary An I'm a certified teacher with 2 years of experience teaching biology, a Master's degree in Education, and a Bachelor's degree in Conservation Biology. I was third in my high school graduating class. I believe that every student can learn any subject. Not everyone learns in the same way with the sam...
9 Subjects
KnowRo Tutor
5 Tutorials
Jackson, TN
Get Tutoring Info

^