Reviewing the Scientific Method & “The Null Hypothesis”
Discussion Questions
1. Why is it important to be objective and methodical?
2. Why are sample size and replication important when collecting data?
3. Why is it important for scientists to share their findings?
4. How are controlled and natural experiments different? Why do we need each type?
Background
Scientific research involves asking questions about nature and collecting data or evidence that may lead to the answers. Investigators often frame their questions in the form of testable hypotheses, which are either accepted or rejected on the basis of the observational or experimental data compiled. This general process is termed the hypothetico-deductive method. The method consists of falsifying hypotheses. The process begins when initial observations or ideas are used as a basis to formulate a null hypothesis (H0).
The null hypothesis states that two or more data sets are no more different that one would expect by chance. The investigator then collects appropriate data and accepts or rejects the null hypothesis on the basis of that evidence. For example, initial observations may have led you to suspect that Shaw students are more likely to be found at the Sunoco station than at the Gas USA station. Your null hypothesis, then, could be framed as follows:
H0: Outside of Shaw, there is no significant difference between the number of students at the Sunoco station than at the Gas USA station.
Your next step would be to make observations about Shaw students and which stations they go to, by visiting the stations, giving surveys, or asking the owners. Naturally, the greater your sample size (the more observations you make), the better your chances are of determining the correct answer. At some point, though, you decide you have collected enough data to accept or reject the null hypothesis with a reasonable degree of certainty.
In this case, accepting the null hypothesis means either that there is no significant difference between the number of Shaw students at the two stations, or that your data were insufficient to determine a difference. Rejecting the null hypothesis means that a difference likely does exist. Rejection then leads to an alternate hypothesis:
H1: The difference between the number of students who go to the Sunoco station and the Gas USA station is too great to be ascribed to mere chance.
Having determined that a structural difference in form or pattern indeed does exist, the foundation is laid for a functional study to find the cause of that pattern. This functional phase begins with the phrasing of a second null hypothesis directed at cause-effect relationships. For example, you might say that there is no difference in the prices of snacks between the two stations and then test that hypothesis.
Situation
Two neighboring houses in East Cleveland are almost identical in appearance. Both are one-family homes, having a square footage of less than 1200 square feet, and both are surrounded by identical vegetation. Investigators have found one house to be susceptible to flooding, while the other appears to be in no immediate flood damage; yet the only significant difference between the two homes is their elevation.
· Home A is perched up above the water table. Also, the foundation of the house was recently waterproofed.
· Home B is on a lower elevation, with its basement sunk underneath the water table. The foundation of the house has never been waterproofed.
Frame a null hypothesis and outline a hypothetical strategy to investigate precisely why Home A is less susceptible to flooding than Home B. Include:
1. A general statement about the observed physical differences between the two homes
2. A null hypothesis formulated to test this difference
3. A suggested testing procedure
4. A statement rejecting the null hypothesis (assume your data suggest a difference)
5. An alternate hypothesis
6. An explanation of the possible cause of the difference
7. A null hypothesis of this cause-effect relationship
8. A suggested test strategy to evaluate this relationship.