Evaluating experiments 

In a nutshell

The purpose of evaluating experiments is to determine if the original question has been answered by your results and to determine how the experiment can be improved.

Critiquing experiments

Looking critically at your experiment can help discover flaws which may or may not be present in your procedure. Have all the variables in your experiment been controlled? If not then the experiment isn't valid as certain uncontrolled variables may have an effect on the outcome.

Accuracy and precision

Accuracy and precision do not mean the same thing. Accurate results are results which are very close to the true answer. Whereas precise results mean results which are measured using sensitive instruments.

A Accurate results B Precise results

Example

Using a mass balance which can measure to the nearest $$\pm\ 0.001g$$ would mean precise results. However, the mass balance has some dirt on the surface which gives incorrect results. This is an example of precise but inaccurate results.​

Uncertainty

Definition

Uncertainty is the range of values around a measurement within which the true value is expected to lie.

Combining measurements

When combining measurements the uncertainties will also have to be combined.

Example

When combining two volumes of liquid each with an uncertainty of $$\pm 0.5\ ml$$ the total uncertainty in the combined solution will be $$\pm 1\ ml$$.​

Percentage uncertainty

Percentage uncertainty can be calculated with the following formula:

​$$percentage\ uncertainty\ (\%)\ =\ \frac{uncertainty}{reading}\ \times\ 100$$​​

Minimising percentage uncertainty

There are a few ways to reduce errors in measurements.

• Firstly, using the most precise equipment available is an easy way to decrease percentage uncertainty in measurements.
• Another way to decrease percentage uncertainty is to increase the amount of what you're measuring. 

Example

A mass balance has an uncertainty of $$\pm \ 0.01g$$. Since the uncertainty is $$\pm\ 0.01$$, there are two sources of error therefore the total uncertainty is: 

$$\pm\ 0.01\ g\ \times\ 2\ =\ \underline{0.02}\ g$$​​

Two readings are taken from this mass balance, $$5.2\ g$$​ and $$11.6\ g$$

$$\frac{0.2\ g}{5.2\ g}\ \times 100\ =\ \underline{3.85}\%$$

​​

$$\frac{0.2\ g}{11.6\ g}\ \times 100\ =\ \underline{1.72}\%$$

As we can see the larger readings have a smaller percentage uncertainty. Generally, the larger the reading the smaller the percentage uncertainty.

​​

Systematic and random errors

There are two main types of errors, random and systematic.

Random errors are errors which can slightly change the results every time the experiment is repeated. 

Example

When titrations are conducted, the volume is noted down and the actual volume may be slightly different. This can cause random errors in the titration volume. 

Systematic errors are errors which are the same every time the experiment is repeated. This is usually caused by an error in setup and or equipment used. 

Example

A mass balance is used to weigh a specific amount of solute. However, the mass balance is on an uneven surface and therefore gives incorrect readings. 

How can the experiment be improved?

Once the experiment is completed it is common to consider what can be done differently to either improve the precision or the accuracy of your results. 

• Will the data collected answer the question and were all the variables controlled? 

• How could the accuracy of the results be improved? Was the apparatus which was used of appropriate accuracy? Could more precise equipment be used to reduce percentage uncertainty?

• How many times was the experiment repeated? Were there many outliers and is it possible that systematic errors were present during the data collection stage?