In: Statistics and Probability
SCENARIO 2:
I need a final, complete sample of 200 respondents for my survey. My client provided my with 100,000 telephone numbers of their current customers to use as my sampling frame. Historically, we can expect about 10% of these numbers are no longer in use by the customer.
I am using a market research company that specializes in live telephone polling to collect the data. They have informed me that, even with a valid phone number, only 30% of people ever pick up and answer. Further, once people pick up, about half immediately refuse to participate in a study.
Fortunately, the only criteria for qualification is that the person is a "current customer" of the company, so anyone who is willing to take the study also qualifies for the study. The telephone polling company told me that I can expect about 25% of the people who start the survey will hang up or quite before the entire study is done.
Valid Sample Size Required (this is the total number of valid, complete responses you desire to analyze) | |
__% | BCI (% of people in sample estimated to have bad contact information, meaning you never reach them) |
__% | NR (% of people who you have correct information for, but you simply never make contact with after multiple attempts) |
__% | REF (% of people who you do contact, but simply refuse to participate in your study) |
__% | NQ (% of people willing to particiapte but cannot because they don't actually qualify for the study) |
__% | INC (% of people who qualify and start the the study, but fail to complete important sections of the study meaning their response is essentially an incomplete) |
Total Sampling Elements (this is how many people you'll have to try to contact in order to reach your final valid sample size) |
1. Total Sampling Elements = [answer here ]
Id also like to estimate the total cost
associated with my data collection activities (since I will have to
bill the client for it!). After talking with the telephone polling
company, here is the cost breakdown:
(1) There is a "setup fee" of $500.
(2) For a completed, valid response, it will cost
$12 per response.
(3) For people who start the study but don't
complete the study, it will cost $2 per person (we still have to
pay the hourly rate of the people conducting the interview)
2. Estimated Data Collection Costs = [answer
here ]
Hint: The TSE formula will be helpful for calculating the
costs, but you're going to need to do a little work on your own to
figure out the total data collection cost.
1) Total sampling elements = 1976
2. Estimated Data Collection Costs = $3034
Calculations:
From given data, we have following information:
Valid Sample Size Required = 200 respondents
BCI = 10%
NR = 70%
REF = 50%
NQ = 0%
INC = 25%
So, we have the following formula for valid sample size and Total sampling elements:
TSE*(1-BCI)*(1-NR)*(1-REF)*(1-NQ)*(1-INC) = Valid sample size
TSE = 1975.30 1976 contacts
2. Estimated Data Collection Costs
Out of total 1976 total sampling elements,
10% are BCI,
30% pick up the call (NR) and
50% refuse (REF)
So people who start the study = 1976 *(1-0.10)*(0.30)*(0.50) = 267
So you are likely to get 267 people who start the study.
Among them, you will get 200 completed valid response.
And 67 will not complete the study.
Estimated data collection costs = Setup Fee + Cost of incomplete response + Cost of completed response
Estimated data collection costs = 500 + 2*67 + 200*12 = $3034