Quantitative: Stratified Sampling

Sample test: Quantitative Method for Business Research
Note: This sample test is just a sample for you to see what a real exam looks like. That means that it is not necessarily same as a real exam.
1. Any measure to characterize a variable of a sample is called
a) a sample
b) a census
c) a statistics
d) a parameter

2. The score of student’s aptitude test is an example of
a) a categorical nominal data
b) a categorical ordinal data
c) a continuous interval numerical data
d) a discrete interval numerical data
e) a continuous ratio numerical data
f) a discrete ratio numerical data

3. Which one(s) could be used for analyzing nominal data? Circle all applied.
a) percentage
b) percentile
c) CV
d) mean
e) standard deviation

4. Which one could be used for analyzing ordinal data? Circle all applied.
a) percentage
b) percentile
c) CV
d) mean
e) standard deviation

5. If we like to check the reliability of an instrument using similar instrument which measures same concepts and/or constructs, it is
a) test-retest reliability
b) equivalence reliability
c) internal reliability
d) inter-rater reliability
e) intra-rater reliability

6. If your survey instrument is very sensitive to the “halo effect”, which one could have the most serious?
a) Mail survey
b) Phone survey
c) Face to face interview
d) Structured record review
e) Structured observation
7. Please write down four factors which affect sampling error and for each factor, write down the effect of sampling error (i.e., if a factor is up, then what happens to the sampling error)

8. Please write down four factors to decide sample size and for each factor, write down the effect of sample size (i.e., if a factor is up, then what happens to the sample size)

9. Please write down when the systematic sampling is better than simple random sampling.

10. Please write down why we do validity tests for a survey instrument.

Please write down why we do reliability tests for a survey instrument.

11. Please explain the concept of precision and accuracy with respect to reliability and validity.

12. CEO of an Auto maker company is concerned about the recent report of an economic journal. The journal says that complaints of SUV customers in US are increasing and varied. As such, she likes to know about satisfaction and demographic information about her customers who bought SUV which her company made recently. Let us assume that satisfaction is the main variable. The sampling design is following: 100 customers were selected from a list of customers who had applied for auto loan through her company. A sampling technique was the stratified sampling with the gender of customer as strata. From the analysis of loan applications, the following information was obtained: Gender (male =55% and female =45%), Age (Young=30%, Middle=50%. Old=20%), Residential (West=30%, Midwest=15%, East=35%, South=20%).
a) Please explain the procedure how to take 100 customers by the stratified sampling with gender as strata.
b) Is the stratified sampling better than the simple random sampling? Explain.

13. Continued from question #13: Even though the cost of inviting is large, she decided to invite them since she liked to have a precise data. The CEO had sent invitation letters to 100 customers selected by the stratified sampling and 60 of them accepted the invitation. The CEO interviewed them to collect the data and used the outputs of interview to do statistical analysis with 99% confidence. From the 60 customers interviewed, the following information is obtained: Gender (Male =30%, female=70%), Age (Young=10%, Middle=25%, Old=65%), Residential (West=25%, Midwest=20%, East=40%, South=15%).
a) Please identify population element, target population, sampled population, frame, and unit.

b) Please talk about four survey errors that the CEO might have in doing this research.

c) For each survey error presented, please discuss whether the error might be serious or not for this case.

d) Please talk about how to improve accuracy and precision of survey for this case.

e) Please explain that 100 customers are large enough for this case.