Economic Growth
Case #9
1. Is it proper to multiply the average order size, $42.33, by the number of addresses (1,300,000) in the target mailing?
a. No, there is far too much variability in responses, including a massive outlier, to have any confidence in this average. The response rate is very low, one would be concerned as to why the rate of response was only 9.2%. The question would therefore be whether the remaining 90.8% will follow the same pattern or will they buy anything at all. There is also the question of whether the sampling frame is representative of the population in this case the target mailing list.
2. Is it better, as suggested, to multiply the endpoints of the confidence interval by the target mailing size?
a. No, it doesn’t show how the data is spread within the range, from the minimum to the maximum. Instead it only provides the two extremes. Even if the endpoints were within say a 95% Confidence Interval from the mean, the mere fact that we are using the target mailing population makes it improper as the sample mean does not seem to be representative of the population mean.
3. Would it be better to multiply by the size of the frame used to select the random sample?
a. It’s better but, there was not enough of a response to multiply against the entire 600,000 frame. Of that frame, less than a thousandth of a percent responded. 600 samples would have been a better number, unfortunately, the amount of response was less than 10% of the mailings, resulting in an even more miniscule amount of information to represent a large population. In the end, the response is simply not enough to ascertain any real information.
4. Should anything else trouble you about this situation?
a. The survey forces far too many assumptions on the analysts. There was a very low response rate for the questionnaire, causing the sample to be far too low. As for those who replied with an affirmative answer, the numbers are far
1. Is it proper to multiply the average order size, $42.33, by the number of addresses (1,300,000) in the target mailing?
a. No, there is far too much variability in responses, including a massive outlier, to have any confidence in this average. The response rate is very low, one would be concerned as to why the rate of response was only 9.2%. The question would therefore be whether the remaining 90.8% will follow the same pattern or will they buy anything at all. There is also the question of whether the sampling frame is representative of the population in this case the target mailing list.
2. Is it better, as suggested, to multiply the endpoints of the confidence interval by the target mailing size?
a. No, it doesn’t show how the data is spread within the range, from the minimum to the maximum. Instead it only provides the two extremes. Even if the endpoints were within say a 95% Confidence Interval from the mean, the mere fact that we are using the target mailing population makes it improper as the sample mean does not seem to be representative of the population mean.
3. Would it be better to multiply by the size of the frame used to select the random sample?
a. It’s better but, there was not enough of a response to multiply against the entire 600,000 frame. Of that frame, less than a thousandth of a percent responded. 600 samples would have been a better number, unfortunately, the amount of response was less than 10% of the mailings, resulting in an even more miniscule amount of information to represent a large population. In the end, the response is simply not enough to ascertain any real information.
4. Should anything else trouble you about this situation?
a. The survey forces far too many assumptions on the analysts. There was a very low response rate for the questionnaire, causing the sample to be far too low. As for those who replied with an affirmative answer, the numbers are far