Henry Brownstein writes in The Rise and Fall of a Violent Crime Wave:
While increases [in violent crime] in small cities may have been as large or larger than the
increases in New York, the phenomenon of the contemporary violent crime wave is
perhaps best demonstrated by the pattern of the violent crime rate in New York City.
While the increase in New York may not have been the greatest among U.S. cities, the
decline to date has been the most impressive.(1)
Here are (a portion) of the data he presents to make his point.
Year
New York
Montgomery,
Alabama
1980
2125.7
520.8
1981
2219.8
475.6
1982
2028.3
372.1
1983
1868.3
289.0
1984
1845.8
238.5
1985
1881.3
292.9
1986
1995.2
210.2
1987
2036.1
168.1
1988
2217.6
234.5
1989
2299.9
356.9
1990
2383.6
516.8
1991
2318.2
903.6
1992
2163.7
775.4
1993
2089.8
871.9
1994
1860.9
964.3
Find the percent and log percent (L%) changes in the rate of violent crime for the two
cities.
Measure:
New York
Montgomery
Percent Change
Total log percent change
1980-1994
Log percent change
1980-1990
Log percent change
1990-1994
You will have to use a calculator or the "calculator" or "mathematical expressions"
in MINITAB and the Student Version of MINITAB, respectively.
Make sure that the log percent changes for the two time periods sum to the total.
Use 1994 as the base year or Y in the standard formula.
Now enter the raw data into MINITAB. Since there are only three columns, it shouldn't
take very long.
Attach carefully labeled time series plots of the raw data: New York series versus year and
Montgomery versus year.
Go to Graph, then select Plot from the menu.
Important:
In Version 11 (the full version) do this now: after opening the plot dialogue
box, go immediately to Regions, then to Multiple graphs. Check the
Overlay option so that all the plots appear on one graph or page.
In the student version do this now: open the plot dialogue box, then click
Frames, and go to Multiple graphs. Check the Overlay option so that all
the plots appear on one graph or page.
Then return to the dialogue box, select variables (i.e., series for New York is the
first Y, year is X; then the series for Montgomery is the second Y, and year is
again X).
The plot should have a title and proper labels for each axis. You, of course, supply
this information in the Annotation and Frame options. Also, use the connect with
symbols.
Now obtain the natural logs of the two series. Depending on whether you have the full or
student version, use the Calculator or Mathematical expressions options under the Calc
menu. Label and store these new variables in two available columns.
Attach a carefully labeled time series plot of the two transformed variables plotted against
year.
Again, remember to use the multiple graphs option so that both graphs appear on
the same page.
Properly label the graph and attach.
In the space below briefly summarize your conclusions:
Now try this. If you have an inordinate amount of trouble, just skip the problem and tell
me what you did.
I want you to attempt to "download" data from the internet and use them in
MINITAB. (You can use SPSS but will have to adjust the instructions.)
These data are "Number of persons arrested for violent crimes."
Both the raw numbers and the rates per 100,000 population are provided.
The file is in a so-called "lotus" format. (Lotus is a spreadsheet program like Excel
and somewhat similar to MINITAB.) Lots of data are available in Lotus format so
it's worth becoming familiar with them.
Before starting look over the instructions because it's easy to make mistakes. Pay
attention to names, spelling, columns numbers, and other details.
Click on Index of available files, then select Crime and arrest data.
Select Number of arrests for violent crime...spreadsheet.
When you click on the file you should see or be given an opportunity to
Save File...
Save the file on your diskette or on your hard drive using (for simplicity)
the Bureau's file name; i.e., violarr.wk1.
The data consist of several columns of text data that must be converted and
cleaned before processing by MINITAB or SPSS.
The variables are
Year
Total number of arrests
Number arrested under 15 years of age
Number arrested 15 to 17 years.
Number arrested 18 to 20 years.
Number arrested 21 to 24 years.
Number arrested 25 years and older.
Year (again)
Total arrest rate: total number arrested per 100,000 population
Rate per 100,000 aged less than 15 years.
Rate per 100,000 aged 15 to 17
Rate per 100,000 aged 18 to 20
Rate per 100,000 aged 21 to 24
Rate per 100,000 aged 25 years or more.
Note the file contains some unused columns
STUDENT VERSION OF MINITAB
You have to clean these data. Fortunately, the Student Version makes the
process fairly easy.
First, start MINITAB and choose Open worksheet.
Click Lotus 1-2-3 and
then Select File.
In the Open Worksheet dialogue box find and select the file, which should
be on your diskette or hard drive under the name violarr.wk1 or something
similar.
You can now get rid of blank rows (rows with stars or "*" in them)
manually. Or use Manip and Delete rows.
In the delete columns box delete rows 1:14 20 26 32.
Note the colon (:) when specifying a range of numbers.
You will also have to delete a couple of rows in Column 1. Just
search for rows with asterisks in them.
Label the columns in the usual way such as Year for column 1, Total for
column 2, and so forth.
Name the second Year column Year2 or simply delete it.
Once you have only numbers in the spread sheet you should be able to plot
the various time series.
FULL VERSION (Version 11)
Surprisingly importation of a lotus spreadsheet is a bit more complicated
when using version 11.
First, start MINITAB and choose Open worksheet.
In the Open Worksheet dialogue box click the arrow in the Files of Type:
window and choose 1-2-2 (*wk?)
Look for your file, which should have the extension wk1, and open it.
I suggest that you label each column with a name such as Year, Total, and
so forth.
Then go to Manipulate and Delete rows to delete rows with text in them.
Or, you can do so manually by using the mouse and pointer to highlight
columns and pressing the delete key.
Most of columns will have a "T" in them to indicate that the data are
textual. These columns have to be converted to numeric.
Go to Manip and choose Change Data Type.
Then select Text to numeric
Pick a variable such as c1 or Year, then type the same column in
the Store in window.
That is, just store the "converted" variable in the same
location with the same name.
Once all of the columns have been converted to numeric data and the stray
text and asterisks removed, you can start plotting.
Attach a carefully labeled time series plot of the number,
not rate, of arrests of
the five age groups and the total.
There will be six plots on one graph.
Next transform these five series to logarithms, using natural logs or "loge" and
plot the transformed series.
These commands are found under Calc
and Calculator or Mathematical
expressions.
Attach these plots to this assignment.
If you have trouble, e-mail me or wait for the next class.
1. Henry H. Brownstein,
The Rise and Fall of A Violent Crime Wave: Crack Cocaine and the Social
Construction of a Crime Problem (Harrow and Heston, 1996) pp. 10-12.