Introduction to Collecting Data: Experimental and Observational Studies

Topic: Intro to Collecting Data

Introduction to Data Collection Methods

Collecting data is a fundamental step in statistics, as it determines the validity and reliability of conclusions drawn from analysis. There are two main ways to collect data: experimental studies and observational studies. Understanding the differences between these methods is crucial for interpreting results and establishing causation.

• Experimental Study: Involves applying a change (treatment) and measuring its effects. This method allows researchers to assume causation if the experiment is well-designed.

• Observational Study: Involves observing and measuring variables without applying any changes. Researchers cannot assume causation from observational studies; they can only identify associations.

Key Definitions

• Experiment: A study in which the researcher actively applies a treatment or intervention to observe its effect on a variable.

• Observational Study: A study in which the researcher observes subjects without manipulating any variables.

• Causation: A relationship where one variable directly affects another.

• Association: A relationship where two variables are related, but one does not necessarily cause the other.

Examples and Applications

The following examples illustrate the distinction between experimental and observational studies, and whether causation can be inferred:

Scenario	Type of Study	Causation?
Testing a medication by giving 16 subjects a placebo, and 16 the medication	Experiment	Yes
Surveying 200 college students about their sleep habits	Observational Study	No
Rolling a die to decide who gets a temporary raise	Experiment	Yes

Application Questions

• Example 1: A manager wants to test whether increasing store hours increases profits. They randomly select half of the stores to increase hours and compare profits. This is an experiment because the manager applies a change (increasing hours). Causation can be inferred if the experiment is well-controlled.

• Example 2: A store surveys its target demographic and finds that 80% would purchase a product after heavy advertising. This is an observational study because no change is applied; causation cannot be assumed.

• Example 3: An office manager surveys employees about their feelings on personal growth and achievement. This is an observational study.

• Example 4: A software company tests a new app by randomly assigning users to use it and measuring results. This is an experiment.

Summary Table: Experimental vs. Observational Studies

Feature	Experimental Study	Observational Study
Researcher applies change?	Yes	No
Can infer causation?	Yes (if well-designed)	No
Example	Clinical drug trial	Survey of sleep habits

Key Formula

While no specific formula is required for distinguishing study types, understanding the concept of random assignment is important in experiments:

• Random Assignment: Each subject has an equal chance of being assigned to any treatment group. This helps eliminate bias and confounding variables.

Random assignment can be represented as:

$P(\text{assignment to group}) = \frac{1}{\text{number of groups}}$

Conclusion

Distinguishing between experimental and observational studies is essential for interpreting statistical results. Only experiments with proper randomization and control allow for causal conclusions, while observational studies can identify associations but not causation.