Introduction

In any data analysis process, the start of it is the data collection process. Data collection makes all the difference in the result of the analysis. The method of collecting data is also called Sampling Method. There are many methods present which make the process of data collection more efficient and effective.

The process is more difficult when we are considering the collection of data for the purpose of measuring the economic parameters like measuring the development of the country. We call it in other words, Population Survey.

There are basically two different methods of survey, Cross-Sectional and Longitudinal Surveys. But before, I start explaining these two methods, I want to explain the importance of the sampling.

Sampling the Population is a process of choosing particular collections of population from the total population. But this isn't as easy as it seems. It is one of the most difficult tasks of all.

Let’s say we have 10,000 people, now if we were to choose 50 people from this population for the purpose of studying the population, what would you do? Obviously, the main target would be to select those particular people who would represent the total population. It's the fundamental problem of sampling? How is that going to be possible, 50 people representing the 10,000 people? Wouldn't it be easy, if it was the whole population or even may be 5,000 of them?

But, sampling is done because surveying every individual of the population would not be practically feasible. So what are the methods of attaining the sample?

First step, is to understand the total population qualitative and start categorizing them, in terms of

1. Religion

2. Demography

3. Age

4. Gender

5. Literacy etc.

This kind of analysis gives you the variation in the population and lets you understand the population in qualitative manner. This is very important as it now narrows down the choices of the sample and the sampling method becomes more unbiased. Yes, that is the word, Unbiased Sampling or Random Sampling.

Lets take one example to illustrate the Random Sampling:

If the sample you take has 80% male and 20% female then you can say the sample is biased. The sample has more male population then female which makes the analysis more male based than female. Therefore, your sample must have no such trends in it that might create a doubt upon the integrity of sample.

So now let’s move on to the survey type.

## Types of Survey

### Cross-Sectional Survey

This is the kind of survey that is taken at a particular instance. This means, the data is collected in order to study the effect of change in one variable to the other variable. This doesn't look at the change with respect to the time. For example, the effect of new advertisement in the sales of that product can only be studied in a particular time.

### Longitudinal Survey

This is the survey to understand the trend in the population.

#### Trend Studies

In this kind of study, a particular sample is chosen from a particular population but may not be the same people. The survey is usually carried out over a long period of time. The target would be the variation of certain parameter in a population with respect to time, giving some historical significance.

#### Cohort Studies

This is similar to the Trend Studies, except that with the change in time, the sample is always taken from the same population as before. For example, if you sample a 4th grade students and asked them their favorite music genre and then asked the same population after 5 years the same question, when they are in 9th grade, then it will reveal their change in their choice. If instead a new sample of 4th grader were taken after 5 years, it would be Trends analysis.

#### Panel Studies

In this study, a sample is choosen and is continuingly queried with in a period of time, tracking them where ever they go. In this analysis, the person in the sample donot change.This method is usually difficult because continuos tracking of the people are necessary.

## Conclusion

So, far, we have understood the type of survey and the importance of unbiased sampling. Always remember that there is no absolute sample or 100% unbiased sample. There is always an approximation in the analysis. Therefore, the next time you see any result from the statistical survey try to understand how the population was sampled; this might help you to understand the analysis better.

The above discussion is the basic of any survey and applies to any kind of survey analysis.

## No comments:

## Post a Comment