Section 22. Using Small Area Analysis to Uncover Disparities

• What do we mean by small area analysis?

• Why conduct small area analysis?

• When should you conduct small area analysis?

• Who should be involved in conducting small area analysis?

• How do you conduct small area analysis?

Community assessments of various kinds often require us to use statistics and other information relating to a certain area. Usually, that area is a city, a county, or even a state or province. That kind of information doesn’t always tell us what we need to know, however.

Suppose, for instance, you’re examining health disparities in a large city. You might have citywide statistics telling you about such things as the frequency of various diseases and medical conditions, emergency room visits, and health costs per resident. What these statistics won’t necessarily tell you is how conditions differ among neighborhoods or groups within the city’s population.

What population(s) bear a disproportionate burden of health problems? Is there a neighborhood where disparities are particularly prevalent? Without answers to these kinds of questions, you can’t focus prevention and care where they’re most needed or examine environmental, social, and other factors that may contribute to disparities in health among different groups within the population of the city as a whole. The same is true for issues other than health as well – violence, access to goods and services, traffic flow, and numerous others. If you don’t have information pinpointing where problems – and assets – exist, you can’t address or use them effectively.

This section deals with just that issue. By using small area analysis, you can better understand where needs and assets are, tailor problem solutions to the areas where they’re really needed, and allocate resources so that they will be as effective as possible. We’ll examine what small area analysis is, when it’s appropriate, and how to conduct it.

What do we mean by small area analysis?

The definition of a small area depends on your needs. It may be a geographic area, a political or administrative district, or even a particular group of people.

The point of small area analysis is to focus on specific areas or populations so that you can see differences among small areas within a larger statistical pattern. If, for instance, the smallest unit for which statistics exist is a county, there may be great differences among different towns or areas within that county when it comes to a particular issue, say the percentage of children with asthma. The statistics for the county as a whole may be average, but one or two small towns may have much higher rates than the rest of the county. Clearly, those differences deserve closer examination and may suggest where additional attention or resources are needed.

The difficulties in conducting small area analysis come in defining the areas you’re concerned with and in finding ways to obtain reliable information for those areas. We’ll discuss both these issues – defining small areas and gathering information – in more detail in the how-to part of this section.

As with most efforts described in the Community Tool Box, small area analysis can often work best when it’s possible to involve the community in planning the assessment. Those who live in the community and are affected by the issues that the assessment reveals may have good ideas about how to divide the larger community into small areas (i.e., towns, neighborhoods, housing developments, etc.), and about what to look for.

Why conduct small area analysis?

• Small area analysis can identify disparities in health and services. This can uncover lack of access, underserved areas, and populations disproportionately affected by adverse health, economic, and social conditions.
• It can uncover issues you wouldn’t otherwise see. A city or province may appear relatively problem-free in overall statistics, but small area analysis can identify geographic or population pockets within the larger area where problems are serious.
• It helps with deciding where to allocate resources. If you know where problems are most serious, you can target scarce resources appropriately.
• Small area analysis clarifies what problems, issues, and assets exist where.
• Small area analysis can help identify possible causes or contributing factors to a condition. By comparing the statistics and situations of a number of areas, you may be able to identify factors associated with why a condition is more common in one area than another. A nearby industrial facility, for example, might be one factor worth investigating when one area has higher childhood asthma rates than another. The geographic association alone, however, does not establish that the facility caused the difference.

The classic example is from 19th Century London. The physician John Snow mapped cholera cases and found a concentration around the Broad Street water pump. His investigation helped connect the outbreak to contaminated water, and the pump handle was removed.

When should you conduct small area analysis?

• When small areas are what you’re responsible for. If you’re part of a coalition representing a number of neighborhoods or several small rural towns, for instance, or if you’re part of or working with a public health agency, you’ll have to consider the needs of several small areas within a larger region.
• When you have to allocate limited resources. Small area analysis can inform your decisions about where and how to use your funding, personnel, and other resources most effectively and efficiently.
• When overall statistics don’t seem to be telling the story. You’re aware of a large and apparently growing number of diabetes cases in a clinic serving one part of the city, but the statistics for the city as a whole don’t seem to reflect this. Small area analysis can help you understand what may be happening.
• When you’re trying to pinpoint where conditions are concentrated and what may contribute to them. Where do most instances of youth violence take place, for example? Particular blocks or housing developments?

Who should conduct small area analysis?

• Community members concerned with or affected by conditions that create disparities among groups in different areas or with different characteristics.
• Public health agencies, officials, and coalitions
• Other public agencies that provide services (e.g., public assistance, children’s services)
• NGOs (non-governmental organizations) and community-based human service organizations
• Community activists
• Police and fire departments, and municipal service departments
• Hospitals
• Community developers
• Community and regional planners

How do you conduct small area analysis?

Assessing and analyzing the needs and assets of small areas is similar to doing the same for larger areas, with one important difference: reliable data about small areas may be much harder to obtain. The availability of information depends on a number of factors: who is gathering it (the state, federal government, county, etc.), what it’s meant to be used for, the size of the population, privacy and confidentiality requirements, and how well-defined the areas are for which it’s collected. Choosing small areas for analysis may be a matter not only of what you want to learn, but also of which areas you can gather useful data for, where you’ll get it, and what you’ll do with it.

1. Ensure community participation. Involve community members at the very beginning of the process. They can help with all the steps that lead to a comprehensive assessment, they’re likely to know the small areas well, they are the ones likely to be most affected by your effort, and their involvement can provide information and context you otherwise might not have access to.

Enlist those you know to help you find the people and groups that should be involved. If you don’t have direct contact with key informants, leaders, or others with deep connections in the community, you might start with community meetings, information shared through schools and community organizations, and other public outreach strategies to assemble a core group. That group can then help to identify and recruit others whose participation is important to your effort.

2. Identify the outcomes you hope will result from small area analysis. What exactly are you trying to accomplish?

• Find and address disparities among areas or populations.
• Address issues in places or among populations where they are most serious.
• Identify potential environmental or social factors that may contribute to the disparities or conditions you’re concerned with, and plan how your effort will address them.
• Figure out how to most effectively allocate resources in order to have the greatest impact.

3. Define the small areas you’ll examine. In some cases, your choices might be obvious. If you’re responsible for a county, for instance, you might want to divide up the area by town.

In other cases, defining small areas may not be so simple. The real question is often whether it’s possible to get reliable information on the areas you’re interested in. Government departments and agencies, municipalities, businesses, and nonprofit organizations all may collect information in different ways and for different geographic and demographic units. Some data may be available for ZIP Code Tabulation Areas or other geographic units; some may refer to particular populations; some may only describe larger areas, such as counties or provinces. A few examples of small areas you might want to examine:

• City or town
• Neighborhood
• Rural village or group of villages
• City block
• Public housing complex or other specific housing development
• Favela or similar informal settlement
• Specific racial, ethnic, cultural, or faith group
• People with a particular health condition
• Public health or other government agency service area
• Service area of a hospital, human service organization, etc.
• County
• Water district
• Census tract

Defining small areas, therefore, may be a matter of balancing the availability and reliability of useful data with the areas you’re most concerned about.

4. Choose the information you’ll look for. What you’re seeking should tell you something about the issues you’re concerned with, and help you understand and address disparities among areas and population groups. There are several possibilities here, any or all of which may be important to your work.

• Demographics. This category includes:
  • Population size
  • Age
  • Income
  • Education
  • Race/ethnicity
  • Country of birth/citizenship
  • Employment
  • Marital status
  • Family size
  • Home ownership
• Environmental conditions that affect issues of concern. Some examples:
  • Land use policies
  • Location of industrial plants
  • Density of fast food and alcohol retailers
  • Access to services by location or public transportation
  • Access to safe areas to walk, jog, bicycle, etc.
• Social determinants related to disparities. These can include:
  • Income inequality
  • Racial or ethnic discrimination
  • Access to health care and health insurance
  • Social and cultural norms, practices, and expectations that may influence health-related behavior
  • Social capital. This is the web of social connections, trust, mutual support, and obligations that people develop through their relationships with others. Differences in social connection and support may be relevant to health and other community conditions.     
  • Food insecurity
• Context. This refers to the setting in which the small area and its residents live and function.
  • Area history. Are there groups with a history of working together on common issues or goals?
  • Relationships among groups and key individuals.
  • Is there new or expanded leadership, or a loss of/change in leadership?
  • Is there political commitment to change conditions?

5. Identify potential data sources for the small area(s) you’re concerned with. Some possibilities:

• The census and related national statistical surveys can provide a wide range of information. In the U.S., the decennial census counts the population and gathers basic demographic and housing information, while the American Community Survey (ACS) continuously collects more detailed social, economic, housing, and demographic information from a sample of households and releases estimates every year. You can search Census Bureau data in a number of ways:
  • By state or province, county, sub-county, city or town, parish, etc. Some Census Bureau data are broken out by administrative or statistical areas such as states, counties, places, and ZIP Code Tabulation Areas.
  • Census tract. In the U.S. Census, census tracts are relatively permanent small statistical subdivisions of a county or equivalent area. Standard census tracts generally have between 1,200 and 8,000 people, with an optimum size of about 4,000.
  •  Census block. Census tracts are further broken down into block groups and blocks. Census blocks are the smallest geographic areas for which decennial census data are tabulated and may cover large geographic areas with few or no people or very small areas in densely populated communities. Detailed ACS estimates are available down to the block-group level rather than the individual block level.
• Survey records. 
  • BRFSS (Behavioral Risk Factor Surveillance System) provides state-level health estimates, with selected local estimates available through the SMART BRFSS project for eligible metropolitan and micropolitan statistical areas and some counties. Local coverage is not available for every community.
  • YRBS (Youth Risk Behavior Survey) provides national and participating state, territorial, tribal, and local school district data about high school students. Local data are available only for participating jurisdictions with appropriate samples; YRBS results are not generally available by individual school, ZIP code, or census tract.
• Organizational and institutional files often contain valuable information about health, living conditions, income, etc. for very specific small areas and populations. Some organizations and institutions that might be helpful include:
  • Hospitals
  • Health and mental health clinics
  • Human service organizations
  • Schools and colleges
  • Economic development organizations

Most of these protect the privacy of individuals, but may have aggregated figures for the areas or populations they serve. Small-area data may sometimes be suppressed or combined when reporting them could risk identifying individuals or produce unreliable estimates.

• Municipal records – the records kept by cities and towns and their departments and boards – can often be helpful. Among these might be:
  • Police records
  • Planning department files (city/town/county)
  • GIS and other maps
  • Vital statistics (births, marriages, deaths)
  • Municipal boards (Board of Health, Zoning Board, etc.)
• State and federal government agency files may be public records or may be available – again, with individual identities protected – for research purposes. Some that could prove useful:
  • CMS databases
  • Department of Education
  • PRAMS (CDC Pregnancy Risk Assessment Monitoring System)
• Direct, hands-on information gathering can be effective when the areas you’re concerned with are small enough and you have the people – often volunteers – willing to go out and get the data. Some methods of direct data gathering:
  • Surveys, oral or written.
  • Individual and group interviews
  • Observation – watching or participating in the activities of people, places, conditions, etc. where they exist.

Be aware that if you decide to gather your own data, volunteers or staff members will probably need some training, especially if they have no real experience asking people for information or conducting research. How to find survey and interview participants, how and when to make observations, effective recording techniques, privacy and consent considerations, and the use of any special equipment – all of these and more may be part of a training for data collectors.

6. Determine how you’ll analyze the information once you have it. This is where “analysis” comes in. How will you extract what you need from the data you have? That depends on at least three factors:

• The quality of the information. Where did it come from, how was it gathered, how large was the sample or number of events, and how accurate and reliable is it?
• The timeliness of the information. Older data may not reflect current conditions, although how quickly data become outdated depends on what you’re measuring and how rapidly the community is changing. If there are data from several years, they may help identify trends, but those trends should not automatically be assumed to have continued beyond the period measured. If only one set of older data exists, be clear about its date and limitations when using it.

If there is no up-to-date information, you have some choices as to how to proceed:

• You can use the existing data as a historical baseline, clearly identifying when it was collected and considering what may have changed since then.
• You can examine trends in data from several periods to develop a sense of the direction of change, while avoiding presenting a projection as if it were a current measured value.
• You can compare the available small-area data with information for the larger area and use other current information to help determine whether the earlier differences still appear to be relevant.

None of these is a perfect solution, so any conclusions based on older or estimated information should be described with appropriate caution.

• The geographic and demographic areas the information describes. The ideal, of course, is finding information that relates specifically to the areas you’re concerned with – urban neighborhoods, rural farming villages, townships, housing developments, or particular populations. You may be able to get useful information through other geographic units – ZIP Code Tabulation Areas, census tracts, or service areas, for instance. On the other hand, you may not be able to get information specific to the areas you want, and may have to settle for information relating to somewhat larger or geographically different areas. Be clear about those differences when interpreting the results.

Using childhood asthma as an example, we’ll list some questions to consider as you analyze the data:

• Do one or more neighborhoods within a city have a higher rate of asthma among children than the city as a whole?
• What are the differences in the rates of childhood asthma between the neighborhoods?
• Do children of different racial or ethnic groups in a particular neighborhood or housing development have different rates of asthma? What other environmental, social, economic, or access-related factors might help explain those differences?
• Are there factors in the physical environment that might contribute to or help explain differences? In the case of our example, these might include industrial facilities, nearby busy roads, rodents, indoor chemicals, or waste sites that could contribute to pollution or particles in the air associated with asthma. Volatile organic compounds, or VOCs, for instance, are found in many paints, solvents, glues, and other substances used in indoor and outdoor construction processes or materials, and some can irritate the respiratory system.
• Are there social, political, and/or economic factors (social determinants of health or SDOH) that might contribute to or help explain differences? Poverty, discrimination, housing conditions, environmental exposures, and differences in access to health care or other resources may all be relevant.
• Are there cultural or social factors that may influence health behaviors, access to information or care, or people’s experiences with the health system?
• Are there trends that show the issue increasing or decreasing in particular small areas or among particular population groups?
• Are numbers dependent on when measurements were taken?
• Do you know something the numbers don’t? The recent start of a new initiative or program, a recent shift in population, new or lost leadership, new funding – any of these and numerous other factors can have a positive or negative impact on the conditions you’re examining. If you’re aware of events or conditions that have arisen since your information was collected, you should take them into account.

When working with very small areas or populations, also pay attention to the number of observations behind a rate or percentage. A few additional cases can sometimes cause a large-looking percentage change when the underlying numbers are small. Data sources may suppress small counts or report wide margins of error for this reason. Avoid treating small differences as meaningful unless the data support that conclusion.

7. Evaluate your small area analysis effort. Small area analysis is often ongoing. Even if your effort is not, the chances are that you’ll want to do a small area analysis again. It’s important to know what worked well, what kind of information was most valuable, whether your understanding of contributing factors and identification of trends were accurate, and whether the information you gathered was in fact useful in addressing the issues and disparities you were concerned with.

If you collect and analyze the data carefully, small area analysis can help you develop a better assessment of the larger community, and allow you to direct your effort where it will do the most good.

In Summary

Small area analysis is the practice of separating out the data for small geographic areas and/or populations – neighborhoods, rural villages, housing developments, etc. – in order to identify and understand disparities among them and between them and the larger statistical area. This makes it possible to concentrate efforts and resources where they are most needed and where they may have the greatest impact.

Conducting small area analysis is a process that includes:

• Encouraging community participation
• Identifying the outcomes you hope to gain from conducting small area analysis
• Defining the small areas you’ll examine
• Choosing the information you’ll look for
• Determining where to find the information
• Deciding how to analyze the information once you have it
• Evaluating your small area analysis effort

Small area analysis can be tremendously helpful in understanding the real needs and assets of your community, and in directing efforts and resources accordingly.

We encourage the reproduction of this material, but ask that you credit the Community Tool Box: /.