Evaluating data
The quality of any data should be evaluated before making any conclusions.
Precision, repeatability and reproducibility
| Term | Meaning |
| Precision | Measurements are in close agreement |
| Repeatable | Measurements are very similar when repeated by the same person or group, using the same equipment and method |
| Reproducible | Measurements are very similar when repeated by a different person or group, using different equipment and/or methods |
| Term | Precision |
|---|---|
| Meaning | Measurements are in close agreement |
| Term | Repeatable |
|---|---|
| Meaning | Measurements are very similar when repeated by the same person or group, using the same equipment and method |
| Term | Reproducible |
|---|---|
| Meaning | Measurements are very similar when repeated by a different person or group, using different equipment and/or methods |
Precision and repeatability can be seen easily from a table of results containing repeat measurement. If the repeat measurements are close together, the data is precise and repeatable.
Accuracy
Evaluation of the data should also consider How close a numerical result is to the true value.. A measurement is accurate if it is close to the The actual value that a measurement should be..
To ensure the data is as accurate as possible, work out the best estimate of the true value.
Identify any A measurement that appears very different to other repeat measurements. It should be included in the data unless a reason is found to explain it. (anomalous results) in the data. These are results that are very different to the others. For example, 2.2 and 0.1 are outliers:
| 0.9 | 1.0 | 1.2 | 0.9 | 1.1 | 2.2 | 1.1 | 0.8 | 0.1 | 1.2 |
| 0.9 |
| 1.0 |
| 1.2 |
| 0.9 |
| 1.1 |
| 2.2 |
| 1.1 |
| 0.8 |
| 0.1 |
| 1.2 |
Try to explain why the outlier is different. An outlier may be removed if there is a good reason to do so. For example, if there is a measurement or recording error.
Find the The mean is calculated by adding all of the data and dividing by the number of items of data. of the remaining results. To find the mean add together the results and divide by the number of measurements.
Sometimes, outliers are not obvious until a graph is plotted.
If a data point does not fit into a trend then it is an outlier. The result should be circled in the graph and labelled as an outlier.