First, you discuss Location – but the table and commentary do not match. In the table
First, you discuss Location – but the table and commentary do not match. In the table
you make it sound like there are 100 in the sample, when in reality there are only 50.
The bar graph is correct. (Same exact comments apply to the variable SIZE)
For the variable INCOME, you list a lot of descriptive statistics, and they are all correct.
But some are not explained – e.g. your statement
“kurtosis coefficient is negative implies that the height of the distribution is less than normal.”
is really unclear. If you do not understand kurtosis, it is better not to mention it (it will be hard
to explain to the client in any case). The SD and range are not measures of central tendency,
they are measures of variation – and the coefficient of determination is a good one to mention,
but it is NOT 30.17, it is 30.17%. And again you should not mention it at all if it is not explained.
(It is the ratio of the SD to the mean; so the SD is only about 3/10 the size of the mean).
For the first pairing, the graph used is not appropriate; a scatterplot should NOT be used when
one of the variables is categorical. And the concept of a “positive relationship” would not make
any sense in this context. Moreover, the statement
“I can conclude all locations are approximately expected to have almost same income level”
cannot be supported by your graph. What would have worked here would have been a simple bar
graph with a bar for each location and the height being either the median income or mean income.
Had you done this, you would see that there is indeed a significant difference in the average incomes.
For the other two pairings, the scatterplots are okay.
And again you should not be quoting statistics like “rho” if you do not know what they mean!
The Pearson product moment coefficient would probably be easy to explain to the clients; but
either you should actually explain it or leave it out.
