False. Inferential statistics involves using a
sample to draw a conclusion about a corresponding population.
Page 6, #10
The
data set consisting of the annual salary for each employee at a company is a
population, since no employee salary is excluded from the data.
Page 6, #14
Population: Income of all home owners in Ohio
Sample: Income of home owners in Ohio with mortgages
Page 7, #18
Population: All American vacationers
Sample: 872 American vacationers surveyed
Page 7, #22
The
numerical value is a parameter, since “all new magazines” represents a
population.
False. For data at the interval level, you can
calculate meaningful differences between data entries.
Page 13, #16
Ratio. Note here that there is an inherent
zero. Also, it is possible to determine
length ratios with this data.
Page 20, #4
True
Page
20, #6
Sampling. A simulation is not relevant in this
case. Assuming the product is
distributed nationally, or even internationally, it would be prohibitively
expensive to do a census, so sampling makes the most sense. An excellent answer given was, “Assuming the
test marketing of the product with and without the label (experiment) is not
possible because of consumer laws requiring the warning, the best option is a
sampling of the population of consumers who would be questioned on how or
whether the warning would effect their purchase.”
A
number of people wrote that an experiment would be the best way to collect data. Reading the question very carefully,
you will see that it specifies that the study is to determine whether or not
consumers still buy the product.
Therefore, experimentation is not relevant in this case. The warning label change has already gone
into effect. Therefore, we cannot
create an experiment in which we distribute the product with the warning label
to some people, and without the warning label to others. We can only take a sample of our target
population (people who might have reason to use this product) and ask questions
regarding their purchases of this product and whether the warning label has had
any effect on their purchasing decisions.
Page 20, #12
Cluster sampling. Note the phrase “every occupied house in the grid is interviewed.” A census is taken of each of the 30 out of 200 grids that were chosen. This, along with the choice of some, but not all, of the grids identifies the sampling technique as cluster sampling. One potential source for bias is that only the occupied households were sampled. There were lots of other good answers on this question that addressed potential sources of bias.
Page 21, #16
Convenience
sampling. It was “convenient” for the
researcher to question the teachers as they left the lounge. However, the researcher sampled only those
teachers who visit the faculty lounge, or who happened to visit the faculty
lounge on the days or days that the researcher was there. This may be reason for bias, if, for
example, teachers who visit the faculty lounge all have similar teaching styles
and grading methods.
The question is biased, because it implies that drivers that change lanes several times are dangerous. A better question might be, “How would you characterize drivers who change lanes several times: dangerous or not dangerous?” Or, “Are drivers who change lanes several times dangerous?”