Identifying Anomalous Results

Anomalous results are odd results - those which are not in keeping with the rest of the results or which do not follow any correlation you have spotted. If you have repeated your experiment several times, you will usually find fewer anomalous results than if you carry it out only once.

There can be many reasons for anomalous results. You may not have noticed that one of the control variables wasn't fully under control for that particular result. You may have added a little too much of something. You could even have written down the result incorrectly or mis-read it when transferring it from your notes to your report.

When commenting on anomalous results, always try to offer some sort of explanation e.g. "this result is much higher than those either side. This might have happened if I didn't control the temperature properly and it went higher during this part of the experiment".

There are two things that you can do with anomalous results. Firstly, if you have the time, you can repeat that part of the experiment, taking great care with quantities and making sure that the control variables are fully under control! Hopefully, a repeat of the experiment will give you a figure that fits with the pattern of the other results. The second thing that you can do is to simply not use them when drawing your conclusion. You may think that not using some of your results that don't fit the pattern that you have spotted is cheating. It isn't - anomalous results can make your conclusion unreliable especially if you are using numbers such as a numerical correlation of the independent and dependent variables. Professional scientists do this all of the time.

When writing up your experiment, it is really important to mention the anomalous results even if they don't appear in your final results table or graph/chart. It shows that you have worked carefully and are aware that experiments are not perfect. If you have been very careful and repeated the experiment several times, you may not have any anomalous results. In this case, simply mention in your evaluation that you can't see any anomalous results. That will tell the person marking your work that you are aware of the concept of anomalous results even if you didn't have any to comment on.

See how much you know about identifying anomalous results and what to do about them by playing this quiz on the subject.