Question

Need the solution in 500 - 600 words

Answer 1

(a) A multi-stage cluster sampling is appropriate in this situation.

Reasons : For effective research to be conducted across multiple geographies, one needs to form clusters that can be achieved only using multiple-stage cluster sampling technique. Here the researcher plans to survey a number of families across the country. So he can divide the entire country’s population into cities (clusters) and further select cities with highest amount of Television watching time.(as single-stage sampling would be time and cost consuming). Hence, he can select number of families from each of the cities by SRS to carry out the survey.

In short :

1) Convenience

2) Takes less time and cost less

3) Easy to implement

4) Higher margin on data accuracy

If we are limited by budget and don’t have time to run around the country to conduct interviews, we can consider using cluster sampling.

(b) In cluster sampling, generally one variable is considered and studied. In this case as the clusters are formed based on highest amount of Television watching time, the variable of interest should be amount of debt. Because if the debt is more then we can conclude that the theory may be correct and exposing to TV commercials can result into high debts.

(c) Data type of the variable : If we consider debt as "high" debt, "low" debt, the variable can be ordinal but if the survey is strictly based on monetary terms, we can call it Quantitative variable, A broad category that includes any variable that can be counted, or has a numerical value associated with it.

(d) Problems usually confronted :

i) Non-response : people may not be interested in talking about their debts thereby resulting in non response.

ii) Overlapping : Sometimes, adequate number of cases from the stand point of increasing the precision of sample is not selected, an overlapping effect may take place.

iii) Problem due to Homogeneity : The groups or families must be as heterogeneous as possible, i.e each of the clusters should contain distinct and different sub-populations to represent the entire country. But this might not always be the case. Hence the problem.