Talk on WeChat
Scientific Skills: Accuracy, Validity and Reliability
8 Mins Read | Written by Sara Lowe
Key Points Summary
- Accuracy: to how well the results of your experiment reflect the expected outcome
- Validity: how well you have controlled your experimental variables in order to maintain a fair test
- Reliability: how many times you repeat the experiment and come to similar results
Content
Accuracy
Accuracy refers to how well the results of your experiment reflect the expected outcome. A lot of the time this has to do with your equipment and how well it works
Lets take the example of weighing a cat. If I place a cat onto a scale and the scale reads 4kg I can assume that this is accurate as this is roughly what I expect my cat to weigh. However, if I put my cat onto a scale at the scale reads 100kg then I can assume that my scale is inaccurate because this result is drastically different to what is reasonable to expect.
Validity
Validity is about how well you have controlled your experimental variables in order to maintain a fair test. In science experiments, we are concerned with 3 variables, our independent, dependent and control variables.
Dymocks Tip: Think Validity has an V for Variables
To explore this concept, let’s consider the following example:
I am interested in measuring the effect of sunlight exposure on plant growth. I have three plants and I place one outside directly in the sun, one in a room with moderate sunlight and the other in a cupboard with no sunlight.
| Variable | Definition/ Description | Relation to Experiment Example |
| Independent Variable | This is the things that you change in your experiment and are interested in how this change affects results. In a valid experiment there is only one independent variable. | For example, in my experiment, the thing I am changing is how much light each plant is exposed to. |
| Dependent Variable | This is the thing that you are measuring | For example, in my plant experiment, I am measuring how tall my plants grow in centimetres. |
| Control Variable | This is everything else that must be kept the same in order to maintain a fair test. This is important so that we can conclude that the effect we are observing is because of our independent variable and not some other factor. | For example, in my plant experiment I should water all the plants the same, use the same soil to plant them in, use the same plant seed etc.That way I can say that any changes in plant growth can only be due to their different sunlight exposure as all else in the experiment was kept the same. |
Lets take a look at a sample NESA HSC question testing validity:
Answer: C
The experiment is invalid because there is more than one independent variable. For a valid test we need to have only 1 thing we change (i.e., only 1 independent variable), which in this case would be the amount of pollutant exposure. Everything else must be kept the same (i.e. our control variables). The amount of water, for example, is a control variable and needs to be kept the same.
Reliability
Reliability is how many times you repeat the experiment and come to similar results. If results of an experiment are consistent across many repetitions, then the experiment is deemed reliable.
For example, if I repeat an experiment 10 times and get similar results each time, then I can be confident that my results are reliable.
In scientific literature, the best way to determine the reliability of results to undergo a peer review process. This is when someone else repeats your experiment and gets the same results.
Dymocks Tip: Think Reliability has an R for Repeats
Conclusion
If you found this helpful and would like some help developing your scientific skills, come and book a free trial with us!