Types and

Sources of Data

UAPP 702
Research Methods for Urban & Public Policy
Based on notes by Steven W. Peuquet, Ph.D.

1
 Distinction between
DATA and INFORMATION

 DATA is the raw material of information.

 may be adequate to provide the information
needed, or it may be partially adequate, or may be
useless.

 INFORMATION is very useful and contributes

in significant ways to answering one or more
of a researcher’s research questions.

2
 Types of Data

Cross-Sectional Data

Longitudinal data
 Trenddata (time series)
 Cohort data

 Panel data

3
 Primary Data

 data that was collected for the

expressed purpose of answering a
specific research question or set of
related research questions.

4
 Secondary Data
 was not collected for the expressed purpose of
answering a current specific research

 probably collected by persons other than those

involved in the current research project

 probably has been archived for some period of time

in some form

5
 Types of Data
Cross-sectional data
 the measurement of variables for a sample drawn from a
selected universe of units of observation (people, objects,
social artifacts, etc.)

 obtained at one point in time (or during a relatively short

period of time)
 a “snapshot” of that universe

Examples:
 US. Census of Population & Housing done in 2000
 Survey we took in class
 Poll taken before an election

6
 Types of Data (2)
Longitudinal: trend data (time series)
 consists of two or more cross-sectional “snapshots” taken at
different times

 definition of the universe of units of observation and the

variables measured remain the same…but…

 the specific units of observation included in each sample differ

 due to the use of random selection and
 changes in the specific units that meet the definition for being members of
the universe.

Examples
 US. Census of Population & Housing done in 1970, 1980, 1990, 2000.
 Multiple surveys in class done every week concerning whether the U.S. should enact
universal health care
 Poll taken once a week every week during the two months leading up to an election

7
 Types of Data (3)
Longitudinal: cohort data (time series)
 consists of two or more cross-sectional “snapshots” of a specific
subpopulation (cohort) taken at different times

 where the actual units of observation making up the

subpopulation (cohort) and the variables measured remain the
same, but…

 the specific units of observation actually included in each

sample differ due to the use of random selection.
 members of a subpopulation (cohort) are “followed” as a group over time

Examples
 Data on how attitudes about war changed over time for soldiers who are Viet Nam
veterans.
 Data on changes in math proficiency of children who began taking a new math
curriculum in the Middleburg School District in the 3 rd grade.

8
 Types of Data (4)
Longitudinal: panel data

 Consists of two or more measurements taken

 for the exact same units of observation and variables
 at two or more different points in time

Examples
 Study by the U.S. Bureau of the Census on the impact of welfare reform on specific
welfare mothers and their children over a period of 5 years

 Study of how the attitudes of specific students in UAPP 800 changed during the Fall
2002 semester regarding their views about the U.S. going to war against Iraq

 Study of how potential voters’ preferences for different candidates changed over a 6
month period leading up to the election for governor.

9
 Secondary Data Sources

 Raw, tabulated, or aggregated statistical data from

government, industry or special interest groups (e.g.,
U.S. Bureau of the Census)
 Journals and books
 Magazines and newspapers
 Internet and the World Wide Web
 Administrative or other records

10
 Problem of Fit
 Must always be concerned about a good fit between secondary
data and the research question…

 because the data was not developed and structured for the
specific purpose of answering the current research question(s)

 So…two options…

 Find the data to answer the question(s), which might require that primary data be
collected also, or

 Change the research question(s) so that it (they) can be answered by the data you
have.

11
 Problem of FIT
(continued)

“Story of the Man

Who Lost his Keys”

12
 Benefit of Secondary Data
 determining what secondary data is available helps
the researcher determine what primary data needs
to be collected.

 maximum use of secondary data minimizes

expenditure of time, effort and money.

 general rule of thumb is:

 least expensive: secondary data (cross-sectional or longitudinal)

 more expensive: primary cross-sectional data

 most expensive: primary longitudinal data

13
 Disadvantages of Secondary Data

 Data may be old

 Concepts not defined and operationalized in an ideal

manner
 e.g., definition of what constitutes "drug abuse”

 Not best unit of observation

 e.g., use of "household" instead of "family”

 Data has been aggregated, hence, data records for

each observed unit not available

14
 Evaluating Secondary Sources

 data collection is usually "purposeful,"

 and the purpose for why the data was collected can
influence

 the data collection procedure employed

 the definition of terms and categories
 the quality of the information
 the conclusions drawn

15
 Evaluating Secondary Sources (2)

Key questions to ask about a given secondary

data source . . .

 What was the purpose of the study?

 Why was the information collected?

 Who was responsible for collecting the information?

 What qualifications, resources, and potential biases are
represented in the conduct of the study?

 What information was actually collected?

 How were units and concepts defined?
 How direct were the measures used?
 How complete was the information?

16
 Evaluating Secondary Sources (3)

More key questions . . .

 When was the data collected? Is the information still

current, or have events made the data obsolete?

 Were there specific events occurring at the time the data

were collected that may have produced the specific results
obtained?

 How was the data obtained? What was the methodology

employed?

 How consistent is the data obtained from one source with

information obtained from other sources?

17
 Secondary Data Resources
There are numerous sources of secondary data.
 U.S. Bureau of the Census (http://www.census.gov)

 U.S. Census “American FactFinder” (http://factfinder.census.gov/servlet/BasicFactsServlet)

 U.S. Federal Government (http://www.first.gov)

 University of Delaware Library Data Bases (http://www.lib.udel.edu/db/)

 Center for Applied Demography and Survey Research, University of Delaware (http://www.cadsr.udel.edu)

 State of Delaware (http://delaware.gov)

 Inter-University Consortium for Political and Social Research (http://www.icpsr.umich.edu)

 Social Sciences Data Services, Northwestern University Library

(http://www.library.northwestern.edu/data)

18
 Secondary Data & Literature Review
 A literature review for a research project can be
thought of as “secondary” in nature…

 because the literature was not done for purposes of

supporting your current research project.

 literature is critical to applied social and policy research

because it is…
 critical to the formulation and refinement of the research
question(s) by helping to identify gaps in knowledge

 provides context for new research

19