Definition of Null Hypothesis Testing

In many cases, accuracy is a matter of life and death. For example, automakers must know whether a vehicle will protect passengers during an automotive accident. Research can sometimes produce false conclusions due to errors in the research design, the instruments, compounding variables or many other factors. Therefore, researchers seek to minimize inaccuracies by testing the null hypothesis.
  1. Definition

    • A null hypothesis is one where the researchers try to disprove a hypothesis. The hypothesis is a claim or statement about a specific property of something. This is in contrast to the alternate hypothesis, which the experimenter seeks to prove. A hypothesis is a prediction that an experimenter makes based on the collected data. Null hypothesis is usually a commonly held view. An alternate hypothesis might state "trees grow only in sunlight" and a null hypothesis would state, "trees do not grow only in sunlight."

    Disproving a Hypothesis

    • Scientists try to disprove a hypothesis that they think is untrue. This can happen when they have a hypothesis that contradicts other scientific data or when they simply want to have accurate data. When researchers perform significance tests, if the test shows that there is a 95 percent or greater likelihood the null hypothesis is not true, the scientists reject the null hypothesis. Otherwise, the null hypothesis remains a hypothesis. Scientists can still later test and disprove the hypothesis.

    Significance Test

    • A significance test involves researchers comparing observed results to theoretical results. Using the tree example, researchers would observe a tree to see whether it grows in darkness. When measuring the tree, if the scientists notice an increase in the tree's size, then the hypothesis that the tree only grows in sunlight is invalid. The significance test is crucial, since data that seems to verify an alternate hypothesis might have arisen by chance. The null hypothesis gives scientists a chance to rule out a seemingly valid hypothesis. Scientists put the probability of the null hypothesis being accurate between zero and one, with zero indicating the hypothesis is false and one indicating that it is true. Scientists do not use percentages, since they are difficult to work with mathematically.

    Rejecting the Null Hypothesis

    • There is a risk of failing to reject a null hypothesis that is really not true. This risk is called the error of the second kind. Smaller discrepancies are difficult to detect by scientists, but large discrepancies can become obvious. Scientists sometimes conduct experiments and get results that completely contradict what they believe and what seems most likely, given other research. As a result, scientists often attempt to reject the null hypothesis by performing null hypothesis testing.

Learnify Hub © www.0685.com All Rights Reserved