Level of significance, Applied Statistics

Level of Significance: α

The main purpose of hypothesis testing is not to question the computed value of the sample statistic, but to make judgment about the difference between the sample statistic and a hypothesized population parameter. The next step after stating the Null and Alternative Hypotheses, is to decide what criterion to be used for deciding whether to accept or reject the null hypothesis.

When we choose 5% level of significance in a test procedure, there are about 5 cases in 100 that we would reject the hypothesis when it should be accepted, that is, we are about 95% confident that we have made the right decision. Similarly, if we choose 1% level of significance in testing a hypothesis, then there is only 1 case in 100 that we would reject the hypothesis when it should be accepted.

Suppose, that under a given hypothesis the sampling distribution of a statistic θ is approximately a normal distribution with mean

E (θ) and standard deviation (Standard Error) σθ


1879_level of significance.png


Then z = 2357_level of significance1.png

is called the standardized normal variable or z-score, and its distribution is the standardized normal distribution with mean 0 and standard deviation 1, the graph of which is shown above.

From the above figure, we see that if the test statistic z of a sample statistic  θ lies between -1.96 and 1.96, then we are 95% confident that the hypothesis is true [since the area under the normal curve between z = -1.96 and z  = 1.96 is 0.95 which is 95% of the total area].

But if for a simple random sample we find that the test statistic (or z-score) z lies outside the range -1.96 to 1.96, i.e. if z  > 1.96, we would say that such an event could happen with probability of only 0.05 (total shaded area in the above figure if the given hypothesis were true). In this case, we say that z-score differed significantly from the value expected under the hypothesis and hence, the hypothesis is to be rejected at 5% (or 0.05) level of significance. Here the total shaded area 0.05 in the above figure represents the probability of being wrong in rejecting the hypothesis. Thus if z  > 1.96, we say that the hypothesis is rejected at a 5% level of significance.

The set of z scores outside the range -1.96 and 1.96, constitutes the critical region or region of rejection of the hypothesis or the region of significance. Thus critical region is the area under the sampling distribution in which the test statistic value has to fall for the null hypothesis to be rejected. On the other hand, the set of z scores inside the range -1.96 to 1.96 is called theregion of acceptance of the hypothesis. The values -1.96 and 1.96 are called critical values at 5% level of significance.

From the above discussion we can formulate the following rule of decision:

Decision Rule (Two-Sided Tests)

Significant level

z Value






| z |  > 1.96

| z |  < 1.96

| z |  > 2.58

| z |  < 2.58






Posted Date: 9/15/2012 3:48:09 AM | Location : United States

Related Discussions:- Level of significance, Assignment Help, Ask Question on Level of significance, Get Answer, Expert's Help, Level of significance Discussions

Write discussion on Level of significance
Your posts are moderated
Related Questions
Primary and Secondary Data: Primary Data: These data are those are collected for the first time. Thus primary data are original in character and gathered   by actual observat

Advantages of Sampling Why should we settle on a sample instead of studying the entire population?  Sampling has the following advantages over a census (study of the entire pop

Lorenz Curve   It is a graphic method of measuring dispersion. This curve was devised by Dr. Max o Lorenz a famous statistician.  He used this technique for wealth it i

As we stated above, we start factor analysis with principal component analysis, but we quickly diverge as we apply the a priori knowledge we brought to the problem. This knowled

discuss the mathematical test of adequacy of index number of formulae. prove algebraically that the laspeyre, paasche and fisher price index formulae satisfies this test. What is

Cluster Sampling Here the population is divided into clusters or groups and then Random Sampling is done for each cluster. Cluster Sampling differs from Stratified Sampl

Correspondence Analysis (CA) is a generalization of PCA to contingency tables. The factors of correspondence analysis give an orthogonal decomposi:ion of the Chi- square associated

Purposive or Judgement Sampling Under this method of sampling, the choice  of selection of sample  items from the universe  depends exclusively on the judgement  of the investi

Regression line drawn as Y=C+1075x, when x was 2, and y was 239, given that y intercept was 11. calculate the residual

In simple regression the dependent variable Y was assumed to be linearly related to a single variable X. In real life, however, we often find that a dependent variable may depend o