The method of selecting a part of the aggregate statistical substance for finding out the information concerning the characteristics of the whole is known as sampling. (Das, 2000, p 160). In analyzing the data of the schools for studying the effectivity and successfulness of the new curriculum over the older one the stratified sampling strategy has been taken. In the total sample of 80 schools, 40 schools would be selected from 100 schools that took the new experimental curriculum and the other 40 schools would be taken from 300 schools that kept the old curriculum.
Regarding the sampling, the members of the school board have made three suggestions for increasing the effectivity of the analysis in support of the new experimental curriculum. All these three suggestions, however, if taken would dampen the quality of the experiment in many ways and ultimately the whole study would fail to produce the true picture about the successfulness of the new experimental curriculum.
According to the first suggestion, all the 80 schools should be taken from the best performing schools which have taken the new experimental curriculum. If this suggestion is taken for sampling then the test will be biased toward the success of the new experimental curriculum. Sampling error of the test will also be very high since this sampling procedure does not incorporate those schools which are still performing better with the old curriculum and those schools which have performed worse with the new curriculum.
Similarly if sampling is done with the second suggestion or the third suggestion, the test will also be biased and sampling error of the test will be high. Besides, if sampling is done with any of these three suggestions, the result of data analysis will not provide the clear picture. In that case by analyzing data we may infer that the new curriculum is better than the old curriculum or worse than the old curriculum though in real it is not the case.
Therefore we will not be able to justify whether or not the new curriculum should be roll out to all schools in the district. 2) We have analysed the data which are collected through stratified sampling. a) In our study when we compare the outcome of the analysis of the schools which took new curriculum with those which took old curriculum we find some interesting features which have been shown by the table 1 and table 2 as follows.