QMBE 2786
Intermediate Statistics for Business and Economics
A BusinessRelated Review of
Elementary Statistics ( UNO Math 2314)
(A previous course in statistics is normally considered a mandatory prerequisite.)
1. Data categories, etc.
Data Categories
elements  entities on which data is collected
variable  an element characteristic of interest
observation  a set of measurements for a particular element
(see table 1.1, p. 5)
Measurement scales
nominal  variable is a label or name
ordinal  variable has nominal properties, but rank is meaningful
interval  rank is meaningful and the interval between values can
be expressed in a unit of measure. Interval data is always
numeric.
ratio  interval data for which the ratio of two values is meaningful.
Always numeric. Must have a zero value.
Qualitative data
is either nominal or ordinal and may be numeric or
nonnumeric.
Quantitative data
may use either an interval or ratio scale of measurement
and is always numeric.
Whether a particular kind of statistical analysis is appropriate will
depend on the type of variable(s) involved.
Exercises:
1. A possible source of data is company employee records. Some of the available data may be:
Name, address, s.s.#, number of sick days, salary, minority status, etc. Discuss these sorts of
data in light of the above categories.
2. Into what category of measurement scale does the variable " temperature Celsius" fall?
3. American universities are often rated by their "party school" status. Last year Tulane ranked
10th. What kind of measurement scale is this?
4. A questionnaire asks your religion. What kind of data is this?
5. Give an example of data that is nominal but not ordinal, ordinal but not interval, etc.
2. Descriptive Statistics.
Given a set of numbers, whether we regard it
as a
population
or as
a
sample
selected from a population, there are certain commonsensical descriptive procedures
( summaries) we can apply.
So what, by way of description, can you do with a set of numbers?
1. Improve the coherence of the data by organizing it
into frequency and relative frequency ( also
percent frequency) distribution tables. The mechanics of this procedure:
a) Pick the number of frequency classes so that coherence is actually
improved. This depends on the size and the nature of the data set, but
This preview has intentionally blurred sections. Sign up to view the full version.
View Full Documenta rough rule is to use 5 to 20 classes.
Review
p. 2
b) Choose the classes, that is, set numerical values ( for quantitative data)
that define the
classes. These are called the class limits. Do this in such a way that each item falls into
one and only
one class and, preferably, all the classes are the same width.
2. Make pictures or graphs that at a glance reveal the trends within
the data set. There are many possibilities: bar graphs, histograms ( a special
sort of bar graph), ogives, frequency polygons, pie charts, etc.
If you don't have a better hobby, you
can spend hours with the Excel Chart Wizard. From a purely mathematical point of view, the most
important of these is the relative frequency histogram.
3. Compute various items of summary arithmetic. These are also called
This is the end of the preview.
Sign up
to
access the rest of the document.
 Spring '08
 Easly
 Normal Distribution, Standard Deviation, Review p.

Click to edit the document details