e population. One can easily imagine the fact that such a sample can never be drawn from the population despite all the techniques and knowledge applied on the subject. Put it straight, the survey research pro knows it very well that the findings from the sample won't be 'exactly' same if the same were to be extracted from the whole population. Thanks to statisticians, they have evolvedWhat is the importance of sampling in survey research? This question, if rephrased like this - can survey research be done without sampling? The answer will definitely be a big no. After all, what is a sample? A sample in survey research is nothing but a small portion of the population being put under the scanner for the study. Let's take a real life example. Suppose we want to buy five kilograms of mangoes. For the starters, mango quality is decided by its fragrance. What would a buyer do? Would she smell all the mangoes from the heap and then take out five kilograms or pick up few randomly and decide on the rest? Chances are bright that the buyer would smell 'select few' and ascertain the quality of the 'whole lot'. A survey research pro would call this 'select few' to be the sample and 'whole lot' as the population.

The same principle applies in survey research and survey analysis. One selects a sample, studies it and 'assumes' that the findings from this small group of persons is valid for the whole population. But is this the best method to be deployed in a survey? Surely this method can't be termed best, but the most optimum one. Imagine a survey research on fruit juice involving more than one million consumers. Is it possible to study all of them? Because it is not possible to study all of them (for want of time and resources), one deploys sampling methods to select a representative group from the population.

A perfect sample is called truly representative sample if it possesses all the characteristics found in the population. One can easily imagine the fact that such a sample can never be drawn from the population despite all the techniques and knowledge applied on the subject. Put it straight, the survey research pro knows it very well that the findings from the sample won't be 'exactly' same if the same were to be extracted from the whole population. Thanks to statisticians, they have evolved

kilograms of mangoes. For the starters, mango quality is decided by its fragrance. What would a buyer do? Would she smell all the mangoes from the heap and then take out five kilograms or pick up few randomly and decide on the rest? Chances are bright that the buyer would smell 'select few' and ascertain the quality of the 'whole lot'. A survey research pro would call this 'select few' to be the sample and 'whole lot' as the population.

The same principle applies in survey research and survey analysis. One selects a sample, studies it and 'assumes' that the findings from this small group of persons is valid for the whole population. But is this the best method to be deployed in a survey? Surely this method can't be termed best, but the most optimum one. Imagine a survey research on fruit juice involving more than one million consumers. Is it possible to study all of them? Because it is not possible to study all of them (for want of time and resources), one deploys sampling methods to select a representative group from the population.

A perfect sample is called truly representative sample if it possesses all the characteristics found in the population. One can easily imagine the fact that such a sample can never be drawn from the population despite all the techniques and knowledge applied on the subject. Put it straight, the survey research pro knows it very well that the findings from the sample won't be 'exactly' same if the same were to be extracted from the whole population. Thanks to statisticians, they have evolved

to be the sample and 'whole lot' as the population.

The same principle applies in survey research and survey analysis. One selects a sample, studies it and 'assumes' that the findings from this small group of persons is valid for the whole population. But is this the best method to be deployed in a survey? Surely this method can't be termed best, but the most optimum one. Imagine a survey research on fruit juice involving more than one million consumers. Is it possible to study all of them? Because it is not possible to study all of them (for want of time and resources), one deploys sampling methods to select a representative group from the population.

A perfect sample is called truly representative sample if it possesses all the characteristics found in the population. One can easily imagine the fact that such a sample can never be drawn from the population despite all the techniques and knowledge applied on the subject. Put it straight, the survey research pro knows it very well that the findings from the sample won't be 'exactly' same if the same were to be extracted from the whole population. Thanks to statisticians, they have evolved

a survey research on fruit juice involving more than one million consumers. Is it possible to study all of them? Because it is not possible to study all of them (for want of time and resources), one deploys sampling methods to select a representative group from the population.

A perfect sample is called truly representative sample if it possesses all the characteristics found in the population. One can easily imagine the fact that such a sample can never be drawn from the population despite all the techniques and knowledge applied on the subject. Put it straight, the survey research pro knows it very well that the findings from the sample won't be 'exactly' same if the same were to be extracted from the whole population. Thanks to statisticians, they have evolved

e population. One can easily imagine the fact that such a sample can never be drawn from the population despite all the techniques and knowledge applied on the subject. Put it straight, the survey research pro knows it very well that the findings from the sample won't be 'exactly' same if the same were to be extracted from the whole population. Thanks to statisticians, they have evolved many techniques to calculate these errors. Let me illustrate this with a simple example. If you were to forecast demand for a gadget, there could be two ways to present your findings

a. I am 99% confident that the number of units sold next year will be between 3 and 4 million b. I am one percent confident that the number of units sold next year will be between 3,245,998 Now which of the two is more practical? The first one gives a vague idea of sales figures with a high confidence level while the second is very accurate but with very low confidence. This confidence level (interval) theory helps a survey research pro select a sample and present the findings based on a confidence interval.