Professional Documents
Culture Documents
Case Study:: A) We Know That, Normal Distribution Is Distribution That Is Bell Shaped. It Is Symmetric
Case Study:: A) We Know That, Normal Distribution Is Distribution That Is Bell Shaped. It Is Symmetric
Geoff Brown is the manager for a small telemarketing firm and is evaluating the sales rate of
experienced workers in order to set minimum standards for new hires. During the past few
weeks, he has recorded the number of successful calls per hour for the staff. These data appear
below along with some summary statistics he worked out with a statistical software package.
Geoff has been a student at the local community college and has heard of many different kinds of
probability distributions (binomial, normal, hyper geometric, Poisson, etc.). Could you give
Geoff some advice on which distribution to use to fit these data as well as possible and how to
decide when a probationary employee should be accepted as having reached full production
status? This is important because it means a pay raise for the employee, and there have been
some probationary employees in the past who have quit because of discouragement that they
would never meet the standard.
Successful sales calls per hour during the week of August 14:
4 2 3 1 4 5 5 2 3 2 2 4 5 2 5 3 3 0
1 3 2 8 4 5 2 2 4 1 5 5 4 5 1 2 4
Descriptive statistics:
Which distribution do you think Geoff should use for his analysis? Support your
recommendations with your analysis. What standard should be used to determine if an employee
has reached “full production” status? Explain your recommendations.
Analysis:
a) We know that, Normal distribution is distribution that is bell shaped. It is symmetric
about mean.
That is normal distribution can be used. Since the number of sales calls is a discrete number so
we cannot use normal distribution directly. Here binomial distribution can be used with normal
approximation.
Hence, normal distribution can be used.
b) Which standard should be used to determine if an employee has reached full production
status?
Manager of the firm is evaluating for sales.
Let x = Number of sales per hour.
If we consider poisson random variable (x) then multiply discrete random variable with count
data.
i.e. x is the number of observations in given time interval.
We can say that,
x ~ Poisson (λ)
So, if we set particular standard of number of calls/hour required to reach full production status
then comparing sample mean (x̄) of each employee to rate of observations (λ) we can decide
whether employee have reached the “full production” status.