“… the null hypothesis is never proved or established, but is possibly disproved, in the course of experimentation. Every experiment may be said to exist only to give the facts a chance of disproving the null hypothesis.” – R. A. Fisher

The null hypothesis takes a central role in the scientific process because it is impossible to reject or accept the experimental hypothesis. This is because there are an infinite number or variables that can have an effect on the difference caused between conditions, whereas the null hypothesis is simple.

Firstly, the null hypothesis states that there will be no differences between the conditions, or treatments. In other words, the samples being tested are from the same population and any differences that are found are very likely to be due to chance. The experimental hypothesis states that there will be a difference between conditions or treatment groups; this can be two-tailed (the direction of the difference is not stated) or one-tailed (the direction of the difference is stated, i.e. there will be an increase or decrease). For example, a researcher is studying the effects of smiling on dating behaviour. The null hypothesis would state that there is no difference between the amount of times a person smiles and the number of dates they go on. A one-tailed experimental hypothesis would state that there would be a difference between the amount of times a person smiles and the number or dates they go on, and the two-tailed could state that the more a person smiles the more number of dates they go on.

However, it is worth noting that some do argue that the null hypothesis is of no importance in the scientific world, such as Savage – “Null hypotheses of no difference are usually known to be false before the data are collected…when they are, their rejection or acceptance simply reflects the size of the sample and the power of the test, and this is not a contribution to science.”

To some extent this argument is true, as I mentioned in my previous blog entry a very large sample can make a small effect appear significant, when it is not, and research papers are normally not published if p = >.05 (the results do not show significance). Notwithstanding, the null hypothesis is still an important component of hypothesis testing, this is a method used in statistics where sample data is used to evaluate a hypothesis and make inferences about a population. Before selecting a sample the hypotheses (null and experimental) predict the characteristics that a sample should have. Once a random sample of the population, the researcher is interested in, has been gathered the obtained sample is compared with the prediction made in the hypothesis. There are 4 main steps to hypothesis testing:

  • State the hypotheses (null and experimental)
  • Chose an Alpha Level, this sets the criteria for making a decision about whether or not the null hypothesis is supported or should be rejected. 
  • Collect the data and compute the relevant sample statistics.
  •  Finally the researcher needs to make a decision about whether to reject the null, this happens when the sample is unlikely to occur if the null is true. So if the alpha level was set at .05 and the results showed p= <.05 this would be that there is a less than 5% chance of the null hypothesis being true therefore the null would be rejected. Or the researcher can fail to reject the null, this would be when the results show p= >.05 therefore falling out of the critical region and supporting the null. (After all the null hypothesis can never be proven because statistical tests deal with probability.)

Therefore the null hypothesis is crucial in the scientific process because it sets concrete boundaries allowing the researcher to know whether, or not, their results are significant or likely to have occured by chance.

Further reading of the null hypothesis can be found by visiting – http://www.null-hypothesis.co.uk/science//item/what_is_a_null_hypothesis