Researchers are said to use the 0.05 level of significance to test for the null hypothesis most often. This shows that there is a 5% chance of rebutting a correct null hypothesis and this is usually known as type 1 error. It is termed as the conventional significance level of hypothesis testing used mostly in all social sciences. However, one will want to use a lower level of significance of 0.01 if there are multiple tests to be performed or if the chance of making the type 1 error occur is significantly high. This is done so that the error level might be less than p< 0.01. An example of situations that use a lower level of significance of 0.01 is experiments conducted in epidemiology or neuroscience where the error margin is very low (The Survey Systems Para. 5). Nonetheless, researchers also use a higher level of significance of up to 0.10 or 90% probability of an event occurring. This means that the chance of rejecting a null hypothesis is 10%. The use of a higher significance level is as a result of having a high chance of 10% of an event not taking place. In other words, a higher level of significance is chosen so as not to commit a type 2 error. This error is commonly termed as the error of omitting to disprove an incorrect null hypothesis. This is an indication that the chances of rejecting the hypothesis are high as compared to the testing of the hypothesis using a lower significance level of 0.01. An example of such a hypothesis test is conducted in a two tailed hypothesis test with high populations (European Social Survey Para. 2).
Works Cited
European Social Survey. Standard Error and Significance Level, 2013. Accessed on 21 March 2012 < http://essedunet.nsd.uib.no/cms/topics/regression/4/1.html>
The Survey Systems. Significance in Statistics & Surveys, 2012. Accessed on 21 March 2012
