Experimental design and identifying variables
In a nutshell
Before performing experiments, it is important you plan it fully so your results can be used and trusted by other scientists. To do this, you must choose appropriate techniques, equipment and apparatus.
Results
When performing experiments, it is important that you produce results that are reliable, repeatable, reproducible, valid and accurate.
Feature | Description |
Reliable | Precise, reliable results do not vary much from the mean. However, reliability of results is affected by random error. |
Reproducible | Results that are reproducible can be recreated when performed by another person with a slightly different method or equipment. |
Repeatable | Repeatable results can be recreated by the same person, performing the same experiment with the same equipment. |
Valid | Valid results are results that answer the original question posed before performing the experiment. To do this, all variables must be controlled. |
Accurate | Results that are accurate are really close to the expected answer. |
Experiments should be repeated at least three times and a mean should be calculated where applicable. This will reduce the effect of random error on you results and make your results more reliable. Equally, doing many repeats and getting similar results makes your data repeatable which also increases the likelihood of reproducibility.
Variables
Definition
Variables are quantities with the potential to change.
Example
Temperature is a variable.
Normally in experiments, only one variable should be changed and this is called the independent variable. The variable that is being measured is called the dependent variable.
Other variables should be controlled to ensure only your independent variable is affecting the dependent variable. Negative controls are controls used to check only the independent variable is affecting the results.
Example
Experiment
| Testing samples with varying concentrations of glucose using quantitative Benedict's reagent. |
Independent variable | The concentration of glucose in the solution is changed between samples. |
Dependent variable | You are measuring the amount of reducing sugar in the samples. |
Negative control | Quantitative Benedict's reagent should also be added to a sample of water with the same volume. This should produce a negative result as glucose does not contain any reducing sugar. |
Apparatus, equipment and techniques
Before performing your experiment, you need to decide what you're going to measure and how you will do this. You also need to select the most appropriate equipment and apparatus to measure and perform the experiment.
Feature | Reason | Example |
Sensitive apparatus | So you can accurately measure changes. | If you require a 1 cm3 sample of water, it would not be a good idea to use a 50 cm3 measuring cylinder. |
Appropriate equipment | You must choose equipment that is appropriate for its function. | To measure absorbance of your glucose and quantitative Benedict's samples, you must use a colorimeter. You cannot use a colour scale as this will not be accurate. |
Appropriate technique | Your technique should be appropriate for your experiment. | When performing a serial dilution, it is important you add the correct amount of water to each test tube or each sample won't be diluted enough. |
Note: You must make sure you're measuring things in appropriate units.
Risk assessments
Before carrying out experiments, you must identify all of the risks and ethical issues to ensure the experiment can be carried out safely.
Three things that must be identified are:
- All of the dangers and potential hazards,
- Who is at risk of these hazards,
- How you can reduce the risk of these hazards.
Recording data
A results table is usually the best way to present your raw data from an experiment. You should make sure that your table has appropriate headers with units in the table headings.
Presenting your data in a table makes it easier to spot anomalous results. Anomalous results are results that do not fit well with the general trends of the rest of the data. These must be investigated so you can identify potential sources of error and ensure the errors are avoided in the future.
Note: You cannot delete anomalous data or ignore it, you should include it in your evaluation and conclusions.
