Introduction to Experimental Design
In AQA AS Biology, practical skills are just as important as knowing the facts about cells or DNA. At least 15% of your total marks come from questions about how experiments are designed, how data is collected, and how reliable the results are. Don't worry if you find the "maths" side of biology a bit daunting—experimental design is mostly about logic and fairness.
Think of an experiment like a fair race: you need to make sure everyone starts at the same line, runs the same distance, and uses the same timing equipment. If you don't, you can't really say who the fastest runner is!
1. Identifying Variables
A variable is anything that can change or be changed in an experiment. To make an experiment a "fair test," we categorize these into three main types:
The Independent Variable
This is the factor that you decide to change. It is the "cause."
Memory Trick: Independent starts with "I"—it is the one I change.
The Dependent Variable
This is the factor that you measure. It is the "effect" or the result.
Memory Trick: Dependent starts with "D"—it is the Data you collect.
Control Variables
These are all the other factors that could affect your results. You must keep these constant (the same) so they don't interfere with the relationship between your independent and dependent variables.
Example: If you are investigating how temperature affects enzyme rate (Required Practical 1), you must keep the pH and the enzyme concentration the same.
Quick Review: If you change more than one thing at a time, you won't know which change caused your result!
2. The Importance of "Controls"
A control experiment (or a control group) is different from a control variable. A control experiment is a set-up where the independent variable is removed or set to a "standard" level.
Why do we need it? To prove that the independent variable is actually responsible for the results. For example, if you are testing an antibiotic on bacteria, you would also grow bacteria with just sterile water. If the bacteria die in the water too, you know your antibiotic isn't the reason they are dying!
3. Accuracy, Precision, and Reliability
Students often get these terms mixed up, but they have very specific meanings in Biology:
Accuracy
How close a measurement is to the true value. If the "true" temperature of a liquid is \( 37.0 ^\circ C \) and your thermometer reads \( 37.0 ^\circ C \), it is accurate. If it reads \( 35.5 ^\circ C \), it is inaccurate.
Precision
How close repeated measurements are to each other. If you measure the same thing three times and get \( 20.1, 20.2, \) and \( 20.1 \), your results are precise. If you get \( 15.0, 25.0, \) and \( 20.0 \), they are not precise.
Repeatability and Reproducibility
Repeatability: Can you do the experiment again and get the same results?
Reproducibility: Can someone else follow your method and get the same results?
Key Takeaway: To improve the reliability of your mean (average), you should always carry out repeats. This helps you identify anomalies (results that don't fit the pattern) and ignore them when calculating your mean.
4. Uncertainty and Errors
No measurement is perfect. Every piece of equipment has a limit to how "fine" it can measure. This is called uncertainty.
Understanding Margins of Error
When you use a ruler, a syringe, or a balance, there is always a small "plus or minus" (\( \pm \)) value. For example, if a scale is accurate to \( \pm 0.01g \), a reading of \( 5.00g \) could actually be anything between \( 4.99g \) and \( 5.01g \).
How to reduce uncertainty:
1. Use equipment with a higher resolution (e.g., a gas syringe that measures in \( 0.1 cm^3 \) rather than a measuring cylinder that measures in \( 1.0 cm^3 \)).
2. Measure larger volumes or masses. A \( 0.5 cm^3 \) error is a huge problem if you are only measuring \( 2.0 cm^3 \), but it is a tiny problem if you are measuring \( 100 cm^3 \).
5. Evaluating Evidence: Correlation vs. Causation
In Paper 2, you are often asked to look at data (like lung disease or heart disease rates) and evaluate a conclusion.
Correlation
A correlation is a link or relationship between two variables. If Variable A goes up and Variable B goes up, they are correlated. This does not mean A caused B.
Causal Relationship
A causal relationship is when one variable directly causes the change in the other. To prove causation, scientists need to find a biological mechanism (a "how" and "why").
Example: There is a correlation between smoking and lung cancer. Scientists proved causation by showing that chemicals in cigarette smoke damage DNA and cause mutations in lung cells (see Section 3.3.2).
Common Exam Trap: If a question asks you to "evaluate" a claim based on a graph showing a link, always check for:
- Small sample sizes (not representative).
- Lack of control variables (other factors might be involved).
- No statistical tests mentioned (though you don't have to do the tests, you should look for "Standard Deviation" bars).
Summary Checklist
- Have I identified the Independent, Dependent, and Control variables?
- Is there a control experiment to compare against?
- Are there enough repeats to identify anomalies and calculate a reliable mean?
- Is the equipment precise enough to keep uncertainty low?
- Does the data show a simple correlation, or is there evidence of causation?