Question

What are the critical aspects of designing a good experiment? Discuss briefly.

a) If a data set outcome of an experiment does not support the theoretical model, can we say the theoretical model is wrong? Discuss what would you double check in the design of the experiment?

a) If a data set outcome of an experiment supports the theoretical model, can we say the theoretical model is confirmed? Discuss what would you double check in the design of the experiment.

Answer 1

There are three main aspects of designing an experiment. One needs to be very cautious about meeting all the aspects while designing the experiment. Otherwise the experiment can be labelled faulty. The important aspects are as follows:

• The inputs or factors are very important in design of an experiment. Finding the suitable factors or inputs is utmost important task while designing the experiment. These factors can be classified as controllable factors and uncontrollable factors. Controllable factors are those on which can be regulated. It includes product, price, place, promotion. On the other hand uncontrollable factors can not be regulated easily. These are not in control of any individual. Example of uncontrollable factors are environmental condition, political condition, technological condition, legal conditin etc.
• Levels or settings of each factor makes a big roll in the study. With the change in level the condition of the factors often changes.
• Like input or factors, output or response of the experiment also matters. In designing a good experiment one need to be clear about the requirement of its output.

a) If data set outcome of an experiment does not support the theoretical model, we can not surely say that the theoretical model is wrong.

In this circumstances we need to evaluate different controllable and uncontrollable factors on which the experiment is made. First we need to evaluate the sample on which the whole experiment is made. We need to investigate whether the sample properly ,atches the entire population i.e. whether all the different clusters of the population are taken while representing the sample. Then we need to look at the place, inputs that are taken while making the experiment.

Now coming to the uncontrollable factors we know that those cannot be regulated by any individual. Hence the environmental condition, legal condition, political condition need not be same all the time in the place the experiment is made. We must know thae with the change in these condition the outcome of the experiment can differ to a large extent.

At the end of all we need to closely look at whether proper methods and calculations are done while doing the experiment. A small error in the method or a mistake in calculation may differ the outcome. Hence it is always recommended to use different data set and make the experiment twice or thrice before coming to a conclution about whether the data experiment matches the theoretical model and whether the theoretical model is consistent or not.

b) If a data set outcome of an experiment supports the theoretical model, it is not necesssary that the theoretical model is confirmed.

Again in this case we need to evaluate different controllable and uncontrollable factors. We need to check the sample taken while making the data set. Whether the sample properly represents the original population. If the sample is faulty and is not consistent with the population, we cannot say that it properly represents the theoretical model. The place of the experiment, inputs taken in the experiment are also important while doing bringing the data outcome.

The uncontrollable factors also change with the course of time. Environmental, legal, political condition don't remain same in a place for long. Hence the data outcome must be taken twice or thrice for the experiment in order to see whether it properly supports the theoretical model.

And at the last the method of the experiment and the calculation should also be rechecked eatch time as any small error or mistake can bring out a faulty outcome.