The post test analysis model involves some of the following factors in order to be applicable on various statistical conditions.

1. Two groups

2. Work on post test measures

3. Includes two distributional measures, each of which consider mean and average

4. Evaluating treatment effect that is the difference between the groups.

In the above factors the meaning o f the term difference is widely expressed which not only involve only the difference. Because it is actually affected by the variance of the group that is the difference which is evaluated by considering the low variability groups in which minimum overlapping is observed between the two groups. The consideration of difference as low variability case is important because when the variability is low the difference between the mean values is high which results in clear and absolute values and is easy to calculate. It is also stated that if the signal is high and the noise is low then the differences observed are apparent. [sky]

Signal/ Noise = difference between the group means / dispersion of groups

= mean (X_{T}) – mean (X_{C})/ SE [mean (X_{T}) – mean (X_{C})]

= value of T

In the formula, the numerator involves the actual values of the mean between the two groups that are the control group and treatment group where as the denominator involves the spread or variance of the two groups. In the post test only analysis, the evaluation of standard errors is made which means the possibilities of errors between the two groups. The standard errors in reality involve the information about the standard deviation between the various groups. The above stated formula is the ratio of T value that describes the difference of mean value and the variability concerns.

The post test analysis actually involves the three methods for the evaluation and three of the methods provide the same results. Following are the three methods;

• The T Test analysis

• One way analysis of variance

• Regression analysis

One method that is the T test analysis is described above. But most appropriately and general method used for the post test evaluation is regression analysis. The equation for the evaluation of regression analysis is

Y_{i}= B_{0}+B_{1}Z_{i}+e_{i}

In the stated regression analysis equation, Yi is the ith unit of the concluded value of the group; B0 is the coefficient of the intercept, B1 is the coefficient of the slope, the value of Zi varies for both the groups like for treatment group its value is 1 and 0 for control group and ei is the residual value for the ith unit of groups. As in the statistical point of view the values for all the terms in the equation that are replacement of b0 and b1 with B0 and B1 that are the intercept and the slope. The value of b1 is the value that is calculated as the difference of average on y values of two groups and it has the same values that are subtracted from the mean.

## References

• DeWitt, J., (2008), “Policy analysis and program evaluation”. Randomize post test only.

• Moore, David S. (1995), The Basic Practice of Statistics, New York: W. H. Freeman.