Chapter 4
Obtaining Data
Section 4.1
1.
Operational definitions are used to define measurement procedures.
Benchmarks are objects or procedures
used to compare two or more products or processes.
2.
There would be many correct answers.
Here is one of them.
Numerous factors affect the fuel efficiency of a car.
These include the size and weight of the car, driving
style, road conditions, speed, type of gasoline, etc.
To simplify the problem, consider the task of determining the highway fuel efficiency of just one
make/year of car, a 1999 Ford Mustang.
Suppose you select 10 1999 Ford Mustang cars with similar options.
Each car is inspected to ensure it is in
good mechanical order.
The tire pressure is checked to ensure it is at the recommended level.
The oil is
changed and 10.0 gallons of gasoline (with an octane rating of 87) is filled into each vehicle.
A test section of a highway in the vicinity of, say, San Angelo, Texas is selected. A test date during which
time the weather is clear and the temperature remains between 70 and 75 degrees Fahrenheit is selected.
Each of 10 test drivers is employed to enter the highway, accelerate to 55 mph, and then drive their vehicles
(using standard rules of the highway) while maintaining 55 mph speed until such time as the vehicle runs
out of gasoline.
Upon completion of the trial, the distance each car traveled is recorded to the nearest one tenth of a mile.
Then the fuel efficiency for each car is computed by taking the distance traveled divided by 10.0 gallons,
producing a fuel efficiency value in miles/gallon.
Finally, the 10 cars’ miles/gallon fuel efficiency values
are averaged together and reported as the average highway fuel efficiency of a 1999 Ford Mustang under
test conditions.
This preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentChapter 4
Obtaining Data
2
3.
For example, you might use a definition such as:
Temperature at 2:00 pm in a fixed, unshaded area on top
of City Hall
.
4.
The published reference would more properly be considered a benchmark, since a benchmark is a well
defined object whose characteristics are already explicitly known, as are the first 200 digits of the number
π
.
Section 4.2
5.
An obvious method would be:
Obtain an exact count of the number N of words in the dictionary.
Use a
random number generator (e.g., in Excel or a statistical software package) to select a random sample of n
integers from the numbers 1, 2, 3, .
.., N
.
Although this sounds simple, finding N is usually not easy.
For
example, should you count singular and plural forms of a word as being the same or different words?
Some operational definition will have to be created so that you know exactly what words to count when
finding N, and even after doing so, finding N will be a huge task so this is generally not a practical method.
There are many methods that may sound good, but do not actually produce random samples.
For instance,
one such method is:
Pick a page at random, then pick a column at random (assuming there are 2 columns
per page), then pick a line at random from the column, then use the word (in bold) that is described by that
line
.
This is the end of the preview.
Sign up
to
access the rest of the document.
 Fall '08
 Staff
 researcher

Click to edit the document details