Americans have long been, in the words of historian Patricia Cline Cohen, "a calculating people." Consequently the sources available for doing quantitative American history are enormously rich and varied. They include census returns, birth and death records, tax lists, membership lists of clubs, churches, and other organizations, business records, social surveys, price lists, city directories, and loads of other quantifiable collections of information.

Yet by comparison to sociologists, psychologists, economists, and political scientists, historians also confront a distinct limitation when they utilize quantitative data. For the most part, historians study dead people and records left behind by dead people. We cannot go back and ask our subjects new questions if we do not like the questions that were asked, say, by a census taker in 1820—or if we cannot read the census taker’s handwriting. Likewise, we cannot design our own experiments comparing a test group to a control group to determine if factor A really made a historical difference. Instead we often must settle for data that were originally assembled by somebody with a different agenda than our own. But we can still be creative. For example, although we cannot ask voters why they preferred one presidential candidate to another in 1852, by using newspaper reports and government records, we can determine how people voted or at least how a group of people voted in a given electoral district. We can then compare voting patterns to patterns suggested by other available data, such as tax assessments, occupations, and/or the religious and ethnic composition of a particular group of voters. With the help of various statistical techniques, we can then make carefully limited inferences about why people voted they way they did, even though nobody asked them directly to explain their motivations.

So if you are given an assignment to use quantifiable sources, where should you begin? Basically, there are two different (but not mutually exclusive) starting points. You can either (1) begin with an existing data set and think about what question(s) you could answer using that data set; or (2) begin with a question (or set of questions) and look for data that would help you answer it (or them). Which starting point you choose will usually depend on the particular character of the assignment and your relative access to various data sets. You may be provided with a package of pre-assembled documents and data or—at the opposite end of the spectrum—you may be given "free rein" to choose your own topic and locate your own evidence. Between these extremes, you might be directed to use census data available on microfilm, on CD-ROM, or on the Internet. Or you might go to the local courthouse and gather information on the different kinds of criminal cases adjudicated during a given time period.

As you develop your project, keep in mind that not all data are equally reliable. Before you invest a lot of time and energy analyzing a particular set of data, you will want to have a general sense of how the data were collected, by whom, and for what purpose. For example, property assessments in tax lists may or may not represent the market value of taxpayers’ estates. Sometimes assessors applied formulas that consistently discounted market prices or left assessments unchanged when market conditions fluctuated. Undoubtedly the meaning and accuracy of tax lists could be affected by whether the assessor simply asked the taxpayer to estimate the value of his or her property or, alternatively, undertook an independent examination of the estate. Likewise, some kinds of property might be exempt from taxation altogether. The more you can find out about the way the data were originally compiled, the better. If you determine that the process was biased, you may still be able to use the data, by taking steps to correct for the distortion in the initial collection procedures.


The following three tables contain data on employees working in cigar and tobacco manufacturing in Tampa, Florida, from 1908-1910. Examine the tables and think about what kinds of historical questions this data can, and cannot, help you to answer.

Study of Employees in Cigar and Tobacco Manufacturing in Tampa Florida (1908-1910)

Table 1—Employees of each race for whom information was secured, by sex (Note: total number of employees in cigar and tobacco industry in Tampa is approximately 10,500)

General Nativity and Race

Number

Percent Distribution

Male

Female

Total

Male

Female

Total

Native born of native father
White	711	305	1,016	10.2	17.0	11.6
Negro	80	10	90	1.1	.6	1.0
Native born of foreign father, by country of birth of father
Canada	1	0	1	*	0	*
China	1	0	1	*	0	*
Cuba	195	0	195	2.8	0	2.2
England	6	0	6	.1	0	.1
France	3	0	3	*	0	*
Germany	11	0	11	.2	0	.1
Ireland	1	0	1	*	0	*
Italy	36	0	36	.5	0	.4
Mexico	2	0	2	*	0	*
Spain	50	0	50	.7	0	.6
West Indies (other than Cuba)	7	0	7	.1	0	.1
Foreign born (by race)
Canadian, French	0	1	1	0	.1	*
Canadian, other		2	2	0	.1	*
Cuban	3,013	532	3,545	43.3	29.7	40.5
English	15	3	18	.2	.2	.2
French	6	1	7	.1	.1	.1
German	15	0	15	.2	0	.2
Greek	0	1	1	0	.1	*
Hebrew, Russian	1	0	1	*	0	*
Hebrew, other	4	0	4	.1	0	*
Italian, North	13	12	25	.2	.7	.3
Italian, South	833	791	1,624	12.0	44.1	18.6
Magyar	1	0	1	*	0	*
Mexican	19	2	21	.3	.1	.2
Negro	45	4	49	.6	.2	.6
Scotch	0	1	1	0	.1	*
Spanish	1,881	127	2,008	27.0	7.1	22.9
West Indian (other than Cuba)	10	0	10	.1	0	.1
South American (race not specified)	1	0	1	*	0	*
Grand Total	6,961	1,793	8,754	100.0	100.0	100.0
*= less than0.005 percent

 

Table 2—Percent of foreign born male employees in each specified occupation before coming to the United States, by race

Race	Number Reporting Complete Data	Percent who were engaged in—
.	.	Manufac- turing of cigars and tobacco	Other Manufac- turing	Farming or farm labor	General labor	Hand trades	Trade	Other Occupa- tions
Cuban	2,342	87.3	0.0	7.0	0.3	1.5	2.7	1.3
Italian, South	472	25.0	1.1	39.2	1.7	18.2	3.8	11.0
Spanish	1,100	49.8	0.4	28.4	2.3	4.4	12.0	2.7
. .	.	.	.	.	.	.	.	.
Total	4,043	68.3	0.2	17.1	1.0	4.3	5.5	3.0

 

Table 3—Percent of foreign born female employees in each specified occupation before coming to the United States, by race

Race	Number Reporting Complete Data	Percent who were engaged in—
.	.	Manufac- turing of cigars and tobacco	Other Manufac- turing	Farming or farm labor	Domestic service	Sewing Embro- iderng & lace making	Teach- ing	Trade
Cuban	168	87.5	0.0	0.0	8.3	3.6	0.0	0.6
Italian, South	149	67.8	0.7	3.4	20.1	6.0	0.0	2.0
..	.	.	.	.	.	.	.	.
Total	361	75.9	0.3	2.8	13.9	5.0	0.3	1.9

Which conclusion can you draw from the data provided?

Women's potential for work in this field, based on previous work experience, attracted Italian immigrant families to Tampa. More Italian immigrant women in Tampa worked outside the home than either Cuban or Spanish women. Italian immigrant women made up the largest proportion of immigrant women working in cigar and tobacco manufacturing in Tampa.