description.qmd

---
title: "Upward Mobility from Poverty Metric Descriptions"
font: Lato
format:
  html:
    embed-resources: true
    toc: true
    mathjax: null
    css: www/web_report.css
    output_dir: "."
    include-in-header: analytics.html
editor_options:
  chunk_output_type: console
params:
  state_county: !expr c("42095", "42017", "48453", "09003")
  fake_labels: false
  state_title: false
execute:
  echo: false
  warning: false
  message: false
  error: true
  output: asis
number-sections: false
---

<!-- Google Tag Manager (noscript) -->

<noscript><iframe src="https://www.googletagmanager.com/ns.html?id=GTM-K2NMF8C"

height="0" width="0" style="display:none;visibility:hidden"></iframe></noscript>

<!-- End Google Tag Manager (noscript) -->

<br>

![](www/images/mobility-icons/mobility-infographic-icons-panel-text.png){width=8in fig-align="center"}

<br>

## <span style="color:#62A1C0">Pillar: Opportunity-Rich & Inclusive Neighborhoods</span>{#pillar-opportunity-rich-and-inclusive-neighborhoods} 

![](www/images/mobility-icons/mobility-infographic-house-icon.png){width=2in fig-align="center"}

##### [PREDICTOR: HOUSING AFFORDABILITY](https://upward-mobility.urban.org/housing-affordability)

**Metric: Ratio of affordable housing units to households with low, very low, and extremely low income levels** 

This metric reports the number of available housing units affordable for households with low (below 80 percent of area median income, or AMI), very low (below 50 percent of AMI), and extremely low (below 30 percent of AMI) incomes relative to every 100 households with these income levels. Housing units are defined as affordable if the monthly costs do not exceed 30 percent of a household’s income. For this metric, the stock of available housing units includes both vacant and occupied units and both rental and homeowner units. A unit is considered available for households at a given level of income if its monthly cost is affordable at that income level—regardless of the income of the current occupant. 

**Validity:** Affordable housing ratios of this type are widely applied in studies of local housing market conditions and trends. Both the income categories and the affordability standard are well established and accepted in both research and policy. 

**Availability:** This metric can be constructed using data from the US Census Bureau’s American Community Survey and income categories defined by the US Department of Housing and Urban Development, both of which are publicly available nationwide. 

**Frequency:** This metric can be updated annually. For less populous communities, it may be necessary to pool multiple years of data and report moving averages. 

**Geography:** This metric is available at the county and city level.

**Consistency:** Affordable housing ratios can be computed consistently for all counties and cities over time. Because the income categories are calculated relative to AMI, the affordability metric appropriately reflects local economic conditions.

**Subgroups:** Because these ratios focus on the characteristics of the housing stock, stratifying by demographic subgroups is not relevant. However, housing units in each affordability category can be stratified by size (number of bedrooms) and tenure (owned or rented). 

**Limitations:** This metric does not account for availability of the affordable units. Higher income households may occupy units that are otherwise deemed affordable to lower-income households, rendering these units unavailable. This metric also does not account for the quality of the affordable units, which may be of substandard quality or too small to meet the size of the occupying household. This metric is somewhat sensitive to patterns of residential mobility. For example, if the number of households with very low incomes were to decline (because of out-migration), this metric would show improvement even if no additional affordable units were produced. We calculate both the number of low-income households and the number of affordable houses by looking at the income threshold for a family of four, regardless of the size of the actual household in question. 

<br> 

##### [PREDICTOR: HOUSING STABILITY ](https://upward-mobility.urban.org/housing-stability) 

**Metric: Number and share of public school children who are ever homeless during the school year** 

This metric identifies the number of children (age 3 through 12th grade) who are enrolled in public schools and whose primary nighttime residence at any time during a school year was a shelter, transitional housing, or awaiting foster care placement; unsheltered (e.g., a car, park, campground, temporary trailer, or abandoned building); a hotel or motel because of the lack of alternative adequate accommodations; or in housing of other people because of loss of housing, economic hardship, or a similar reason.

**Validity:** These data are reported by school administrators and generally verified by local liaisons and state coordinators. This is a direct and well-established measure of homelessness for children that results from and reflects housing instability among families and unaccompanied children. The definition of homelessness used for this measure extends beyond literal homelessness to effectively include the full range of circumstances in which a family does not have a stable home of their own. 

**Availability:** The US Department of Education requires every local education agency to collect and report these data and data are made public nationwide. 

**Frequency:** This metric can be updated annually. 

**Geography:** This metric is available at the county and city level. The boundaries of local education agencies can be aggregated to other geographies.

**Consistency:** This measure is consistently defined, collected, and reported for all local education agencies nationwide. 

**Subgroups:** This metric can be disaggregated based on students’ disability status and whether they are enrolled in English-as–a-second-language courses. 

**Limitations:** Because this metric is limited to public-school children, it does not capture homeless adults who are childless, and it does not capture homelessness among children who do not enroll in public school. Further, it could show improvement if the families of homeless children move to a neighboring jurisdiction or if policies “push” them out. This metric is sensitive to patterns of residential mobility if large numbers of families with very low incomes flow into or out of a local education agency’s boundary. 

<br> 

##### [PREDICTOR: ECONOMIC INCLUSION](https://upward-mobility.urban.org/economic-inclusion) 

**Metric: Share of residents experiencing poverty who live in high-poverty neighborhoods** 

This metric measures the share of residents in a city or county who are experiencing poverty and who live in high-poverty neighborhoods (measured by census tract). People and families are classified as being in poverty if their income (before taxes and excluding capital gains or noncash benefits) is less than their poverty threshold, as defined by the US Census Bureau. Poverty thresholds vary by the size of the family and age of its members and are updated for inflation, but do not vary geographically. A high-poverty neighborhood is one in which over 40 percent of the residents are experiencing poverty.

**Validity:** Measures of poverty concentration have been widely used to measure the extent and severity of economic exclusion and isolation. The more concentrated and separate people in poverty are from better-resourced neighbors, the more isolated they are from the larger community and the social and economic resources and opportunities it can provide. Because this metric reflects the structural conditions facing a city or county’s residents, changes in the metric possibly caused by people moving into or out of a jurisdiction do represent changes to those structural conditions. 

**Availability:** The metric can be constructed using data from the US Census Bureau’s American Community Survey, which is publicly available nationwide. 

**Frequency:** This metric can be updated annually. 

**Geography:** This metric is available at the county and city level; however, aggregation to the city level is imprecise because census tracts and census places do not always perfectly align. 

**Consistency:** Poverty concentration can be consistently defined and calculated for all cities and counties over time. 

**Subgroups:** This metric can be disaggregated by the race and ethnicity of the head of household. 

**Limitations:** This measure can be sensitive to the overall poverty rate of a city or county. Therefore, changes in poverty concentrations need to be assessed with reference to the city or county’s overall poverty rate.

<br> 

##### [PREDICTOR: RACIAL DIVERSITY](https://upward-mobility.urban.org/racial-diversity) 

**Metric: Index of people’s exposure to neighbors of different races and ethnicities** 

This metric is constructed separately for each racial or ethnic group that is readily available and reliable, and reports the share of each group’s neighbors who are members of other racial or ethnic groups. The exposure index reflects the racial and ethnic diversity of neighborhoods. For example, the exposure index would report the share of people who are Hispanic; white, non-Hispanic; and other races and ethnicities in the census tract of the average Black, non-Hispanic person. Higher values of the index indicate more neighborhood diversity and more day-to-day exposure of people to neighbors of different races and ethnicities. 

**Validity:** The exposure index is one of several widely used measures of residential segregation or inclusion. It captures the multiracial or multiethnic diversity of American communities today and reflects the experience of individuals of different races and ethnicities, and it provides a robust picture of neighborhood racial and ethnic composition. Because this metric reflects the structural conditions facing a city or county’s residents, changes in the metric that may be caused by people moving into or out of a jurisdiction do represent changes to those structural conditions. 

**Availability:** The metric can be constructed using data from the US Census Bureau’s American Community Survey, which is publicly available nationwide. 

**Frequency:** This metric can be updated annually. 

**Geography:** This metric is available at the county and city level; however, aggregation to the city level is imprecise because census tracts and census places do not always perfectly align. 

**Consistency:** Exposure indexes can be consistently defined and calculated for all cities and counties over time. 

**Subgroups:** This metric is by definition disaggregated by race or ethnicity. 

**Limitations:** This measure can be sensitive to the overall racial or ethnic composition of a city or county. Therefore, changes in exposure indexes need to be assessed with reference to the city or county’s overall racial or ethnic composition. Further, although this index can be constructed annually, it may take many years to observe appreciable changes.

<br> 

##### [PREDICTOR: SOCIAL CAPITAL](https://upward-mobility.urban.org/social-capital) 

**Metric: Number of membership associations per 10,000 people** 

Based on the US Census Bureau’s County Business Patterns (CBP) data, which measure the total number and type of establishments for all counties in the US, this metric is a ratio of the number of membership associations (e.g., civic organizations, bowling centers, golf clubs, fitness centers, sports organizations, religious organizations, political organizations, labor organizations, business organizations, and professional organizations) per 10,000 people in a given community.

**Validity:** The CBP data from the US Census Bureau are well established and widely used by academics and other researchers across the country. Research supports its use as a measure for social trust because social trust is enhanced when people belong to voluntary groups and organizations. People who belong to such groups tend to trust others who belong to the same group. The more such groups per person, the more likely that individuals in those communities belong to one or more groups. 

**Availability:** The metric can be constructed using the CBP dataset, which is publicly available nationwide. 

**Frequency:** This metric can be updated annually. 

**Geography:** This metric is available at the county and city level. This metric is also available at the ZIP code level.

**Consistency:** This metric is clearly defined and has been consistently measured across time since 2012 and for the entire population of the US. CBP data are derived from the Business Register, maintained and updated by the Census Bureau to track all known single- and multi-establishment employer companies in the US.

**Structural equity\* and subgroups:** Data used to construct this metric can reflect racial or socioeconomic subgroups by matching it with demographic characteristics at the ZIP code level, such as racial or ethnic composition or concentrated poverty. 

**Structural relevance\*:** This metric reflects the availability of opportunities for engagement and relationship-building and important structural support for social capital. 

**Limitations:** This metric captures only a certain aspect of social capital. It is trying to measure social associations within a community and is likely the best measure within the context of business and professional organizations. Nevertheless, it cannot capture (a) social associations at the granular, individual level, or (b) smaller, more informal organizations that would not be in a position to self-report to the Business Register. 

\* These selection criteria on structural equity and structural relevance were explicitly added in a review of new metrics added to the Upward Mobility Framework in fall 2022. 

<br> 

##### [PREDICTOR: SOCIAL CAPITAL](https://upward-mobility.urban.org/social-capital) 

**Metric: Ratio of Facebook friends with higher socioeconomic status to Facebook friends with lower socioeconomic status**

“Economic connectedness” measures the extent to which low- and high-socioeconomic status individuals are friends with one another. Specifically, the economic connectedness metric is twice the average share of high-socioeconomic-status friends (individuals from households ranked in the top half of all income-earning households) among low-socioeconomic-status individuals (individuals from households ranked in the lower half of all US households based on income) in a given community. An economic connectedness measure of 1 represents a community that is perfectly integrated across socioeconomic status, with half of all low-socioeconomic status individuals’ friends being of high socioeconomic status. The metric is meant to be a measure of the interconnectivity, by location, between people from different economic backgrounds (i.e., with different levels of social capital and exposure).

**Validity:** This metric measures an important aspect of social capital: the extent to which members of a community associate with people with varying social statuses. The connections made through this type of social capital help facilitate and develop an individual’s power, autonomy, and sense of belonging in their community. These data come from research that has been peer reviewed and published. 

**Availability:** The metric is publicly available through Opportunity Insights’ Social Capital Atlas. 

**Frequency:** These data were released in summer 2022. Because this was an inaugural data release contingent on the publication of new research, it is unclear if it will be updated or with what frequency. 

**Geography:** This metric is available at the county and city level. These data are available nationally in a standardized format for all counties. The smallest geography for which these data are available is the Zip Code Tabulation Area (ZCTA) level, which can be approximately aggregated to other geographies. 

**Consistency:** This metric is clearly defined and is currently consistently measured across populations and geographies. We cannot know if it will be consistently measured across time, because it is new and has yet to be updated.

**Structural equity\* and subgroups:** Data used to construct this metric can reflect racial or socioeconomic subgroups by matching it with demographic characteristics at the ZIP code level, such as racial or ethnic composition or concentrated poverty. This metric will help identify structural equity in the community by identifying the relative abundance or lack of social cohesion between community members with different levels of socioeconomic status. 

**Structural relevance\*:** This metric is a systemic condition of economic mobility because it speaks to what level of socioeconomic intermingling is supported in the social environment of a given area. This metric, however, is also an outcome of greater economic mobility, since economic connectedness is positively correlated with increased mobility from poverty. 

**Limitations:** The biggest limitation of this metric is its novelty. Although it was developed by reputable scholars and peer reviewed, it lacks the established track record of other metrics. Moreover, it focuses on the financial aspects of social capital and does not capture other important elements like popularity and community ties. This metric may be sensitive to residential mobility into and out of a city or county, but not to an extent likely to affect its aggregate values. This metric also relies on the continued popularity and use of Facebook as a social media platform, the user base of which has been skewing older in recent years. Without continued engagement from the same user groups and the introduction of younger populations as they age, comparability and consistency over time may be compromised. This metric can only be calculated for ZIP codes containing at least 100 people. 

\* These selection criteria on structural equity and structural relevance were explicitly added in a review of new metrics added to the Upward Mobility Framework in fall 2022. 

<br> 

##### [PREDICTOR: TRANSPORTATION ACCESS](https://upward-mobility.urban.org/transportation-access)

**Metric: Transit trips index**

This metric reflects the number of public transit trips taken annually at the census tract level by a three-person single-parent family with income at 50 percent of the Area Median Income (AMI) for renters. This number is percentile ranked nationally into an index with values ranging from 0 to 100 for each census tract. Higher scores reflect better access to public transportation. 

**Validity:** This metric was designed in partnership with the US Department of Transportation and has been used by the US Department of Housing and Urban Development in community efforts to affirmatively further fair housing. Several scholars have also used this metric and data in published in peer-reviewed journals. Although other arrangements of family composition, income, and housing status are possible in constructing this index and are available in the data, these characteristics were intended to more closely characterize a lower-income household in the community and are the most validated of other household combinations.

**Availability:** The estimates come from the Location Affordability Index, which is publicly available. 

**Frequency:** This metric used to be released every three years. The Urban Institute is working to develop an updated version of the metric.

**Geography:** This can be measured at the census tract or neighborhood level. Values can be averaged at higher levels of geography. For example, one can calculate a population-weighted average value among all census tracts in a county to determine a county-level value.

**Consistency:** This metric can be calculated the same way over time. 

**Subgroups:** This metric is based on a lower-income population, notably single-parent families with two children earning half the local AMI among renters. 

**Limitations:** This metric cannot alone capture the concept of transportation access. This must be used in partnership with the transportation cost index to cover geographies that may not have an extensive public transportation system, such as rural areas. 

<br>

##### [PREDICTOR: TRANSPORTATION ACCESS](https://upward-mobility.urban.org/transportation-access)

**Metric: Transportation cost index**

This index reflects local transportation costs as a share of renters’ incomes. It accounts for both transit and cars. This index is based on estimates of transportation costs for a three-person, single-parent family with income at 50 percent of the median income for renters for the region (i.e., a core-based statistical area). Although other arrangements of family composition, income, and housing status are possible in constructing this index, these characteristics were intended to more closely characterize a lower-income household in the community. Values are inverted and percentile ranked nationally, with values ranging from 0 to 100. The higher the value, the lower the cost of transportation in that neighborhood.

**Validity:** This metric was designed in partnership with the US Department of Transportation and has been used by the US Department of Housing and Urban Development in community efforts to affirmatively further fair housing. Several scholars have also used this metric and data in articles in peer-reviewed journals.

**Availability:** The estimates come from the location affordability index, which are publicly available.

**Frequency:** The location affordability index data are updated every three years.

**Geography:** This can be measured at the census tract or neighborhood level. Values can be averaged at higher levels of geography. For example, one can calculate a population-weighted average value among all census tracts in a county to determine a county-level value.

**Consistency:** This metric can be calculated the same way over time. 

**Subgroups:** This metric is based on a lower-income population, notably single-parent families earning half the local area median income among renters. 

**Limitations:** Transportation costs may be low for a variety of reasons, including greater access to public transportation and the density of homes, services, and jobs in the neighborhood and surrounding community. It is important not to consider this metric alone but rather in combination with the transit trips index to more fully measure the concept of transportation access. 

<br>

## <span style="color:#F584BE">Pillar: High-Quality Education</span>{#pillar-high-quality-education} 

![](www/images/mobility-icons/mobility-infographic-book-icon.png){width=2in fig-align="center"}

##### [PREDICTOR: ACCESS TO PRESCHOOL](https://upward-mobility.urban.org/access-preschool) 

**Metric: Share of children enrolled in nursery school or preschool** 

This metric measures the share of 3- and 4-year-old children in a community who are enrolled in nursery school or preschool. 

**Validity:** Federal agencies such as the National Center for Education Statistics use household survey data to ascertain nursery and preschool enrollment.

**Availability:** The metric can be constructed using data from the US Census Bureau’s American Community Survey (ACS), which is publicly available nationwide. 

**Frequency:** This metric can be updated annually. Survey data are collected annually and used to create one- and five-year estimates. For less populous communities, it may be necessary to pool several years of data to obtain reliable estimates. 

**Geography:** This metric is available at the county and city level. 

**Consistency:** Information pertaining to nursery and preschool enrollment in the ACS is measured the same way across all geographies in the same year. Changes to the ACS in the future could influence comparisons over time.

**Subgroups:** The data can be disaggregated by race or ethnicity, gender, and other demographic factors. For less populous communities and for certain demographic groups, several years of data may need to be pooled to provide reliable estimates.

**Limitations:** This metric can change over time if fertility patterns change or if families with young children who move out of or into the community have very different propensities for enrolling their children in preschools than parents with young children who remain in the community. Because ACS data do not capture the quality of preschool, enrollment figures may overstate exposure to the kinds of programs most likely to improve short-term academic outcomes and long-term outcomes such as mobility from poverty. 

<br> 

##### [PREDICTOR: EFFECTIVE PUBLIC EDUCATION](https://upward-mobility.urban.org/effective-public-education)

**Metric: Average per-grade change in English language arts achievement between third and eighth grades** 

This metric reports the average annual improvement in English language arts (reading comprehension and written expression) observed between the third and eighth grades.

**Validity:** The state assessments are well defined and validated but vary by state. Because the Stanford Education Data Archive (SEDA) has standardized these to be nationally comparable, this metric has been widely used and a reliable measure of student achievement. Higher rates of improvement in English indicate more effective public education, which improves the upward mobility of children from less advantaged backgrounds.

**Availability:** The metric can be constructed using data from the SEDA, which is publicly available nationwide. 

**Frequency:** This metric can be updated annually.

**Geography:** This metric is available at the county, metropolitan area, and school district level. 

**Consistency:** Tests of student progress vary from state to state and can change over time if states modify their tests. The SEDA has standardized these to be nationally comparable.

**Subgroups:** The SEDA has adjusted scores by race or ethnicity, gender, and socioeconomic status. SEDA data comes from EDFacts, which reports proficiency levels based on raw data by race or ethnicity, gender, disability status, limited English proficiency status, homeless status, migrant status, and economically disadvantaged status.

**Limitations:** Literacy performance reported in “levels” is sensitive to movement in and out of a community over time.

<br> 

##### [PREDICTOR: SCHOOL ECONOMIC DIVERSITY](https://upward-mobility.urban.org/school-economic-diversity) 

**Metric: Share of students attending high-poverty schools, by student race or ethnicity** 

This set of metrics is constructed separately for each racial or ethnic group and reports the share of students attending schools in which over 20 percent of students come from households earning at or below 100 percent of the federal poverty level.

**Validity:** This set of metrics captures the interaction of economic and racial segregation of schools and therefore reveals whether (and to what degree) students of color are more likely than white students to attend schools with large concentrations of classmates experiencing poverty. Higher concentrations of students experiencing poverty are associated with worse achievement for all the students in a school.

**Availability:** The metric can be constructed using data from the National Center for Education Statistics Common Core of Data and the Urban Institute's Modeled Estimates of Students in Poverty, which are publicly available nationwide. 

**Frequency:** This metric can be updated annually. 

**Geography:** This metric is available at the county and city level. These data exist at the school district, city, and county levels. Because this metric reflects the structural conditions facing a city or county’s students, changes in the metric that may result from people moving into or out of a community may represent changes to those structural conditions.

**Consistency:** This metric can be consistently defined and calculated for cities and counties. 

**Subgroups:** This metric is by definition disaggregated by race or ethnicity. 

**Limitations:** Because traditional proxies for school poverty (i.e., the share of free- and reduced-price-meal students or the share of students directly certified for free meals) have grown inconsistent across time and states, this metric uses the Urban Institute's [Model Estimates of Poverty in Schools](https://www.urban.org/sites/default/files/2022-06/Model%20Estimates%20of%20Poverty%20in%20Schools.pdf) to identify school poverty levels. 

<br> 

##### [PREDICTOR: PREPARATION FOR COLLEGE](https://upward-mobility.urban.org/preparation-college) 

**Metric: Share of 19- and 20-year-olds with a high school degree** 

This metric is the share of 19- and 20-year-olds in a community who have a high school degree. 

**Validity:** Earning a high school degree is an important prerequisite for pursuing additional schooling, and although not all high school graduates are ready to enroll in college, high school completion is a well-understood and widely used measure of educational attainment. Data on educational attainment are collected in a variety of federal surveys.

**Availability:** The metric can be constructed using data from the US Census Bureau’s American Community Survey (ACS), which is publicly available nationwide. 

**Frequency:** This metric can be updated annually. Survey data are collected annually and used to create one- and five-year estimates. For less populous communities, it may be necessary to pool several years of data to obtain reliable estimates. 

**Geography:** This metric is available at the county and city level. 

**Consistency:** Information pertaining to educational attainment in general and on high school graduation in particular is measured the same way across all geographies in the same year in the ACS. Changes to the ACS in the future could influence comparisons over time.

**Subgroups:** This metric is disaggregated by race and ethnicity. Data used to construct this metric can be disaggregated by race or ethnicity, gender, and other demographic factors. For less populous communities and for certain demographic groups, several years of data may need to be pooled to provide reliable estimates.

**Limitations:** Young adults moving in and out of an area can influence this measure. This measure also does not capture the quality of schooling received.

<br> 

##### [PREDICTOR: DIGITAL ACCESS](https://upward-mobility.urban.org/digital-access) 

**Metric: Share of people in households with broadband access in the home** 

This is a measure of the share of people in households in a population that have a broadband internet subscription of any type (e.g., DSL, cable modem, fiber-optic cell phone, or satellite) at their place of residence.

**Validity:** The US Census Bureau uses a series of questions to measure aspects of digital access across the nation. Existing literature makes extensive use of these measures of digital access. 

**Availability:** Data on broadband access are available annually from the US Census Bureau’s American Community Survey (ACS), which is publicly available nationwide.

**Frequency:** This metric can be updated annually. 

**Geography:** This metric is available at the county and city level. 

**Consistency:** The metric can be measured in the same way across geographies and over time, but some changes were made in 2016 to the survey questions required to construct this metric, so values before and after 2016 should not be compared. 

**Structural equity\* and subgroups:** The digital divide is about the degree of inequity in digital access by demographics (including race and ethnicity) and by geography (e.g., urban versus rural). Because this metric can be disaggregated along those dimensions, it can provide insights into structural equity at the community level. The data can be disaggregated by household income and by demographics of the head of household, such as race, ethnicity, and gender.

**Structural relevance\*:** This metric measures whether individual households have broadband connections to the internet and thus reflect individual choices as well as more structural factors such as affordability and the availability of broadband services 

**Limitations:** Having broadband internet at home is not useful without a device from which to access it. Access to computing devices is an important component of digital access. Moreover, measuring broadband access alone does not account for other types of digital access (such as access to cellular data), but existing scholarship supports a measure of broadband access as it relates to closing the digital divide. There is no universally available measure of digital access that includes both broadband access and access to computing devices. Those who would like to further investigate digital access in their communities could also look at the data available through the ACS on access to computing devices. Finally, not all broadband is fast enough to meet the needs of all households. Adequate minimum speeds to effectively access all content types (e.g., streaming videos or virtual classrooms) may vary by household. 

\* These selection criteria on structural equity and structural relevance were explicitly added in a review of new metrics added to the Upward Mobility Framework updated in fall 2022. 

<br> 

## <span style="color:#8ECF86">Pillar: Rewarding Work</span>{#pillar-rewarding-work} 

![](www/images/mobility-icons/mobility-infographic-money-icon.png){width=2in fig-align="center"}

##### [PREDICTOR: EMPLOYMENT OPPORTUNITIES](https://upward-mobility.urban.org/employment-opportunities) 

**Metric: Employment-to-population ratio for adults ages 25 to 54** 

This metric is the share of adults ages 25 to 54 in a given community who are employed.

**Validity:** An employment ratio captures what share of adults in a community are engaging in work for pay. The prime-age employment-to-population ratio focuses on age groups that are traditionally older than college age and younger than retirement age. The prime-age employment-to-population ratio is a standard labor market indicator, which can be calculated using the Current Population Survey for states and the American Community Survey (ACS) for sub-state areas.

**Availability:** The metric can be constructed using data from the ACS, which is publicly available nationwide. 

**Frequency:** This metric can be updated annually. Survey data are collected annually and used to create one- and five-year estimates. For less populous communities, it may be necessary to pool several years of data to obtain reliable estimates. 

**Geography:** This metric is available at the county and city level. 

**Consistency:** Information on employment and age is measured the same way across all geographies in the same year in the ACS. It is highly unlikely that the ACS will change how it captures employment in the future.

**Subgroups:** This metric is disaggregated by race and ethnicity. Data used to construct this metric can be disaggregated by race or ethnicity, gender, and other demographic factors. For less populous communities and for certain demographic groups, several years of data may need to be pooled to provide reliable estimates.

**Limitations:** The Bureau of Labor Statistics (BLS) reports the official employment-to-population ratio monthly for those ages 16 to 19 and those age 20 and up. As such, the BLS-reported measure could be lower for communities that have many young adults attending college (and not working). Consequently, for our purposes, we recommend computing the employment-to-population ratio for adults ages 25 to 54 using data from the ACS rather than relying on BLS reports. Even when using ACS data, the employment-to-population ratio can drop if people who aren't working leave an area or if working people move in. 

<br> 

##### [PREDICTOR: JOBS PAYING A LIVING WAGE](https://upward-mobility.urban.org/access-jobs-paying-living-wage) 

**Metric: Ratio of pay on an average job to the cost of living** 

This metric shows average pay relative to the cost of living in a particular area. The metric is computed by dividing the average earnings for a job in an area by the cost of meeting a family of three’s basic expenses in that area.

**Validity:** Communities in which typical jobs pay a greater share of the local cost of living provide more opportunities for their residents to move out of poverty. Employer-reported data on wages paid are a reliable indicator of what jobs pay, and our metric is based on data collected and disseminated by the Bureau of Labor Statistics (BLS). Data on what it costs to meet basic expenses requires detailed studies of the cost of food, clothing, shelter, health care, and work-related expenses for each community. We rely on the work of well-regarded scholars at the Massachusetts Institute of Technology (MIT) to obtain estimates of the local cost of living.

**Availability:** The metric can be constructed using data from the BLS Quarterly Census of Employment and Wages, and data on living wage from the MIT, which are publicly available nationwide.

**Frequency:** This metric can be updated annually. 

**Geography:** This metric is available at the county and metro levels. Data on wages are available for the 365 largest counties in the US. About 75 percent of the US population live in the 365 largest counties. 

**Consistency:** Information on weekly wages is collected in a consistent fashion by the BLS. MIT uses a consistent methodology to compute living wages by county.

**Subgroups:** This metric cannot be disaggregated into subgroups because these data describe wages rather than the people earning those wages. 

**Limitations:** The measure cannot be disaggregated into subgroups. The measure relies on MIT’s computations of “living wages.” The living wage data is not available for every year.

<br> 

##### [PREDICTOR: OPPORTUNITIES FOR INCOME](https://upward-mobility.urban.org/opportunities-income) 

**Metric: Household income at the 20th, 50th, and 80th percentiles** 

Household income is a standard measure of financial well-being. The three levels help a community to track how and for whom incomes are changing in a given place as well as whether incomes are rising across the board or more so for those with higher incomes. To identify income percentiles, all households are ranked by income from lowest to highest. The income level threshold for the poorest 20 percent of households is the value at the 20th percentile. The 50th percentile income threshold indicates the median, with half of households earning less and half earning more. The income level threshold for the richest 20 percent of households is the value at the 80th percentile. 

**Validity:** These are well-established and frequently-used measures to assess the financial well-being of families by several federal agencies and many scholars. 

**Availability:** The metric can be constructed using data from the US Census Bureau’s American Community Survey, which is publicly available nationwide. 

**Frequency:** This metric can be updated annually. Survey data are collected annually and used to create one- and five-year estimates.

**Geography:** This metric is available at the county and city level. 

**Consistency:** Income data in the ACS are measured the same way across all geographies in the same year. The measure is fairly consistent over time but there could be changes in the phrasing and sequence of income source questions that may affect comparisons over time. When such changes have occurred in other federal surveys, such as the Current Population Survey, the Census Bureau provides bridge year data so users can assess the effects of survey changes. 

**Subgroups:** This metric is disaggregated by race and ethnicity. Data used to construct this metric can be disaggregated by race or ethnicity, gender, and other demographic factors. For less populous communities and for certain demographic groups, several years of data may need to be pooled to provide reliable estimates.

**Limitations:** The purchasing power of any particular level of income will vary based on the local cost of living. Also, because household sizes differ, the same income may be stretched across larger average households in some places relative to others. Like all metrics based on the characteristics of people living in an area, it can change because of residential mobility.

<br> 

##### [PREDICTOR: FINANCIAL SECURITY](https://upward-mobility.urban.org/financial-security-and-wealth-building-opportunities) 

**Metric: Share with debt in collections** 

This measure indicates the share of people in an area with a credit bureau record with debt that has progressed from being past due to being in collections. Debt in collections includes past-due credit lines that have been closed and charged off on the creditor’s books as well as unpaid bills reported to the credit bureaus that the creditor is attempting to collect. For example, credit card accounts enter collections once they are 180 days past due. The city-level measure captures the share of people in an area with a credit bureau record with any derogatory debt, which is primarily debt in collections. 

**Validity:** Consumers without other accessible financial resources may need to take out debt not only to pay for housing or education but also to pay for daily necessities such as food or utilities. The inability to pay back debts can signal current or near-term financial insecurity, particularly for families with lower incomes. Those households likely have few assets and as such may have negative wealth.

**Availability:** Drawn directly from credit reports, the credit bureau data are nationally representative and uniform across the country. The data are restricted and are not accessible directly from credit bureaus but are made available in aggregate form on the Urban Institute's [Debt in America feature](https://apps.urban.org/features/debt-interactive-map/?type=overall&variable=pct_debt_collections) and on the Urban Institute’s [Financial Health & Wealth Dashboard](https://apps.urban.org/features/financial-health-wealth-dashboard/).

**Frequency:** This metric can be updated annually.

**Geography:** This metric is available at the ZIP code level which can be aggregated to the county and city level. 

**Consistency:** The share of consumers with debt in collections can be measured consistently for all geographies. The measure is likely to remain consistent over time, unless the credit bureaus change the way overdue debt is captured in credit reporting. However, Urban Institute features like Debt in America and the Financial Health and Wealth Dashboard have different methodologies for measuring debt in collections. 

**Subgroups:** This metric can reflect racial or socioeconomic subgroups by matching it with demographic characteristics at the ZIP code level, such as racial or ethnic composition or concentrated poverty. 

**Limitations:** Aside from the limitations related to geography, subgroup analysis, and the definition of debt in collections, these data do not capture “credit invisible” households without a credit record. And as a measure of financial well-being, even if few consumers have no debt in collections, many may still have too little wealth or savings to be primed for upward mobility. This measure is somewhat sensitive to resident turnover. If many residents without overdue debt move into a county or ZIP code, or if many residents with overdue debt move out, this measure could shift without any in-household change in debt management. The measure from the Urban Institute’s [Financial Health & Wealth Dashboard](https://apps.urban.org/features/financial-health-wealth-dashboard/) captures all derogatory debt, which is primarily debt in collections, while the measure from the Urban Institute’s [Debt in America feature](https://apps.urban.org/features/debt-interactive-map/?type=overall&variable=pct_debt_collections) includes debt in collections only.

<br> 

##### [PREDICTOR: WEALTH-BUILDING OPPORTUNITIES](https://upward-mobility.urban.org/financial-security-and-wealth-building-opportunities) 

**Metric: Ratio of the share of a community’s housing wealth held by a racial or ethnic group to the share of households of the same group** 

This metric highlights racial and ethnic disparities in an important source of wealth—disparities that likely reflect structural inequities in wealth-building opportunities. The metric compares a racial or ethnic group’s (measured by the head of household) share of primary-residence housing wealth (measured using self-reported house values) in a community to its share of the total number of households. For example, if Black homeowners have 15 percent of the community’s primary-residence housing wealth but make up 45 percent of the community’s heads of households, then the metric is a ratio of 15%:45%. The greater the gap between these percentages, the more inequity in housing wealth in the community.

**Validity:** Although this ratio is not commonly used, each piece of the ratio is. The calculation of primary-residence housing wealth is consistent with the literature. The share of racial and ethnic groups among the household population is commonly used. The juxtaposition of these two shares has been used to highlight housing wealth equity and homeownership wealth gaps. 

**Availability:** The metric can be constructed using data from the US Census Bureau’s American Community Survey, which is publicly available nationwide. 

**Frequency:** This metric can be updated annually. For less populous communities, it may be necessary to pool several years of data to obtain reliable estimates. 

**Geography:** This metric is available at the county and city level. 

**Consistency:** This metric is defined consistently across race and ethnic groups, is consistently measured over time, and is comparable across geography.

**Structural equity\* and subgroups:** This metric captures the racial disparities apparent in the ways wealth is shared across a given population. Rather than focusing on an aggregate measure of total homeownership, this helps elucidate any racial or ethnic disparities in comparative housing wealth and how it has been accessed or distributed in that community. These disparities can signal instances of racism as a hindrance to achieving or benefiting from homeownership.

**Structural relevance\*:** This ratio characterizes the distribution of aggregate housing wealth to describe a community-level condition rather than an individual outcome such as the average value of household wealth. 

**Limitations:** This metric only captures relative wealth-building opportunities and does not capture absolute wealth-building opportunities. It is possible that a community has an equitable distribution of housing wealth in a community with very little housing wealth. Such a community would rate well under this metric. This metric will also be imprecise or suppressed for communities that lack racial or ethnic diversity. Although we refer to this metric as housing wealth, the data reflect homeowners’ self-assessments of the value of their homes and does not account for mortgage debt. Black and Latino households, on average, buy their homes with more debt, so the racial housing wealth disparities are likely to widen if mortgage debt is incorporated. Also, this metric does not account for other financial costs and benefits of homeownership that could affect wealth building, nor does it account for other important differences such as the average age of people in different racial and ethnic groups. One would expect older people to have higher-value homes than younger people, so some racial and ethnic disparities could be exaggerated by age differences. Therefore, this metric may not fully reflect the size of the actual housing wealth gap and could be misleading without a deeper understanding of homeownership and demographic circumstances in a community. Further, this metric focuses on only one form of wealth based on homeownership, and does not consider other forms of wealth like owning a business, stocks, or bonds. The metric only considers the race of the head of household and does not account for household size. 

\* These selection criteria on structural equity and structural relevance were explicitly added in a review of new metrics added to the Upward Mobility Framework in fall 2022. 

<br> 

## <span style="color:#8A9191">Pillar: Healthy Environment and Access to Good Healthcare</span>{#pillar-healthy-environment-and-access-to-good-health-care} 

![](www/images/mobility-icons/mobility-infographic-health-icon.png){width=2in fig-align="center"}

##### [PREDICTOR: ACCESS TO HEALTH SERVICES](https://upward-mobility.urban.org/access-health-services) 

**Metric: Ratio of population per primary care physician** 

This ratio represents the number of residents per primary care physician in a community. Primary care physicians include practicing non-federal physicians (MDs and DOs) under age 75 specializing in general practice medicine, family medicine, internal medicine, and pediatrics. If a community has a population of 50,000 and has 20 primary care physicians, for example, the ratio would be: 2,500:1. If a community has no primary care physicians, the ratio would be the population size to zero.

**Validity:** This metric is defined and established by the US Department of Health and Human Services, Health Resources and Services Administration, Bureau of Health Workforce. 

**Availability:** The metric can be constructed using data from the US Department of Health and Human Services’ Area Health Resource File, which is publicly available nationwide. 

**Frequency:** This metric can be updated annually. 

**Geography:** This metric is available at the county level. 

**Consistency:** This metric can be measured in the same way across geographies and over time, but the definition of a primary care physician changed in 2013, so values before and after 2013 should not be compared.

**Structural equity\* and subgroups:** These data cannot be disaggregated by demographic subgroups because we are unable to connect physicians to their patients.

**Structural relevance\*:** This metric measures the extent to which community residents have access to primary care physicians based on the presence of physicians in the community. As such, it is a structural feature of the community. 

**Limitations:** Because of financial and insurance constraints, the presence of physicians in an area does not mean that all local residents can access their services or that the care they access is good quality. Conversely, physicians in one county may provide services to residents of another county, and physicians are not the only type of primary care provider available to patients. Nurse practitioners, physician assistants, or other practitioners can also provide primary care services. This metric is not available by demographic subgroups and therefore cannot speak to differences in access by race or ethnicity. 

\* These selection criteria on structural equity and structural relevance were explicitly added in a review of new metrics added to the Upward Mobility Framework updated in fall 2022. 

<br> 

##### [PREDICTOR: NEONATAL HEALTH](https://upward-mobility.urban.org/neonatal-health)

**Metric: Share with low birth weight** 

A child born weighing less than 5 pounds 8 ounces (<2,500 grams) is considered to have a low birthweight. Children born below that threshold are at elevated risk for health conditions and infant mortality. This metric looks at the share of low-birthweight babies out of all births with available birthweight information.

**Validity:** This metric is the standard currently used by the Centers for Disease Control and Prevention as part of their national assessment on health among infants. 

**Availability:** The metric can be constructed using data from the Centers for Disease Control and Prevention’s National Center for Health Statistics, which is publicly available nationwide.

**Frequency:** This metric can be updated annually.

**Geography:** This metric is available at the county level. County-level estimates are available through public-use microdata files provided by the National Center for Health Statistics as well as through other data collection efforts, such as the Kids Count Data Center or the CDC WONDER system. 

**Consistency:** Medical advances have improved the outcomes for low-birthweight babies, implying this metric may change in the future. However, this has been consistently used over decades as a metric for neonatal health.

**Subgroups:** This metric is disaggregated by race and ethnicity. The share of children born with low birthweights can be disaggregated by the race or ethnicity of the mother, as well as by the mother’s age.

**Limitations:** These data are not readily available at lower levels of geography, such as neighborhoods, where disparities by race and socioeconomic status within a county are most notable. Large movements of women with risky pregnancies moving in or out of a community could influence this metric. 

<br> 

##### [PREDICTOR: ENVIRONMENTAL QUALITY](https://upward-mobility.urban.org/environmental-quality) 

**Metric: Air quality index** 

The air quality index is an index that summarizes potential exposure to harmful toxins at the neighborhood level. The index is a linear combination of standardized US Environmental Protection Agency (EPA) estimates of air quality carcinogenic, respiratory, and neurological hazards at the census tract level. Values are inverted and then percentile ranked nationally. Values per census tract range from 0 to 100. The higher the index value, the less exposure to toxins harmful to human health. 

**Validity:** EPA scientists and researchers link air pollutants to health effects that can manifest within a few hours or days after breathing polluted air.

**Consistency:** Air pollutants can be consistently measured across time and geographies. However, EPA standards may change in the coming years and should be reevaluated for consistency moving forward.

**Availability:** The metric can be constructed using data from the EPA’s AirToxScreen data, which are publicly available nationwide.

**Frequency:** Air quality information from the AirToxScreen are updated every three years.

**Geography:** This metric is available at the census tract level which can be aggregated to the county and city level.

**Subgroups:** This metric can reflect racial or socioeconomic subgroups by matching it with demographic characteristics of residents at the census tract level, such as racial or ethnic composition or concentrated poverty.

**Limitations:** These data may not be updated with enough frequency for some communities. Alternative data sources can offer information annually, or even daily, but they are only available at higher levels of geography and for a limited set of places and cannot be disaggregated by subgroups.

<br> 

##### [PREDICTOR: SAFETY FROM TRAUMA](https://upward-mobility.urban.org/safety-trauma) 

**Metric: Deaths due to injury per 100,000 people** 

This metric represents the number of deaths of community residents, both from intentional injuries such as homicide or suicide and unintentional injuries such as motor vehicle deaths, per 100,000 people in the community.

**Validity:** These data are collected by the National Center for Health Statistics and the Centers for Disease Control and Prevention. Injury can be traumatic, and people living in communities with a high incidence of injury can experience both the direct trauma from injury and vicarious trauma from injuries sustained by others, which can lead to psychological distress, increased rates of aggression, and diminished physical health. High rates of injuries that lead to death in a community, such as opioid overdose, suicide, traffic fatalities, and homicide, can lead to community-level trauma. The County Health Rankings and Roadmaps uses a five-year average to increase data reliability. 

**Availability:** The metric can be constructed using data from the National Center for Health Statistics Mortality Files and Centers for Disease Control and Prevention’s Wide-Ranging Online Data for Epidemiologic Research, which are publicly available nationwide. The metric is available publicly for counties through the [County Health Rankings & Roadmaps](https://www.countyhealthrankings.org/) website.

**Frequency:** This metric can be updated annually. 

**Geography:** This metric is available at the county level. 

**Consistency:** The metric can be measured in the same way across geographies and over time.

**Structural equity\* and subgroups:** This metric is not disaggregated by subgroups, but the underlying data can be disaggregated by race and ethnicity, age, gender, and education level. Because this metric can be disaggregated by race and ethnicity, it can be used to see how much exposure to trauma varies between racial and ethnic groups within a community.

**Structural relevance\*:** This metric is concerned with individual deaths, but deaths caused by injury can be reflective of both individual-level factors and structural factors such as neighborhood design, crime rates, and access to mental health services. 

**Limitations:** The metric captures only one aspect of exposure to trauma. Injury more generally (i.e., injuries that don’t lead to deaths) may still cause trauma, but data on that are not nationally available. Data are not available at the city level. 

\* These selection criteria on structural equity and structural relevance were explicitly added in a review of new metrics added to the Upward Mobility Framework updated in fall 2022. 

<br> 

## <span style="color:#FFD175">Pillar: Responsible and Just Governance</span>{#pillar-responsive-and-just-governance}

![](www/images/mobility-icons/mobility-infographic-govt-icon.png){width=2in fig-align="center"}

##### [PREDICTOR: POLITICAL PARTICIPATION](https://upward-mobility.urban.org/political-participation) 

**Metric: Share of the voting-age population who turn out to vote** 

This metric measures the share of the voting-age population that voted in the most recent presidential election.

**Validity:** This metric is well established. Scholars of political science have used this metric in articles published in peer-reviewed journals. 

**Availability:** The metric can be constructed using data reported by local governments, which is publicly available nationwide. 

**Frequency:** This metric can be updated after every presidential election (every four years).

**Geography:** This metric is available at the county and city level; however, aggregation to the city level can be imprecise. These data are broadly available at the electoral district level and can be approximately aggregated to other geographies.

**Consistency:** Voter turnout is measured consistently over time and geography, but the values can be volatile from year to year, with higher turnouts in years involving a presidential election, so comparison over time should account for that.

**Subgroups:** This metric cannot be directly disaggregated into subgroups because these data do not include voter demographics.

**Limitations:** Residential mobility can impact this metric, so it is important to interpret changes in voter turnout in the context of demographic shifts in the community. In local communities with higher rates of non-citizens, voter turnout can inaccurately reflect a community’s political participation. Communities with a population of non-citizens may consider additional local data to better assess political participation and civic engagement. Census data only identifies the citizen voting-age population and does not account for residents who are ineligible for reasons like incarceration or disenfranchisement.

<br> 

##### [PREDICTOR: DESCRIPTIVE REPRESENTATION](https://upward-mobility.urban.org/descriptive-representation) 

**Metric: Ratio of the share of local elected officials of a racial or ethnic group to the share of residents of the same racial or ethnic group** 

This metric measures the ratio of the share of the city council or county board from specific racial and ethnic groups to the share of city or county residents from those racial or ethnic groups. For example, if a community has 10 elected officials; 9 are white, non-Hispanic; and the community’s population is half white, non-Hispanic, the metric will read as “90.0%:50.0%.” If the share of local officials is higher than the share of people in the community, then this group is overrepresented. If the share of local officials is lower than the share of people in the community, then this group is underrepresented.

**Validity:** Scholars of political science have used this metric in articles published in peer-reviewed journals. 

**Availability:** Data on the racial or ethnic characteristics of city council or county boards can be collected locally. Communities should ask their local elected officials to self-report their racial or ethnic identity (see suggestions for collecting this information in the Boosting Upward Mobility [Planning Guide](https://upward-mobility.urban.org/boosting-upward-mobility-planning-guide-local-action) (Fedorowicz et al. 2022, pg. 27)). The racial and ethnic composition of residents in those communities can be calculated using data from the US Census Bureau’s American Community Survey, which is publicly available nationwide.

**Frequency:** This metric can be updated as frequently as elections occur. The population figure can be updated annually. 

**Geography:** This metric is available at the county and city level; however, aggregation to the city level is imprecise because census tracts and census places do not always perfectly align. 

**Consistency:** This metric can be calculated the same way over time and across geographies.

**Subgroups:** This metric accounts for race within its definition, but it may also be calculated for other subgroups.

**Limitations:** Although the movement of people in and out of the community can influence this metric, it is likely to be far more sensitive to shifts in the composition of elected officials in the short term. The number and race of public officials must be collected locally, and collecting information on the demographic characteristics of a local official may be challenging if they do not reveal this information publicly and are unwilling to report it to local data collectors. 

<br> 

##### [PREDICTOR: SAFETY FROM CRIME](https://upward-mobility.urban.org/safety-crime) 

**Metric: Reported property crimes per 100,000 people and reported violent crimes per 100,000 people** 

The Federal Bureau of Investigation’s (FBI) National Incident-Based Reporting System (NIBRS) provides a standard, well-defined measure of crime. Reported crimes are captured as crimes against persons (e.g., assault, homicide, human trafficking, kidnapping or abduction, and sex offenses) and crimes against property (e.g., arson, bribery, burglary or breaking and entering, counterfeiting or forgery, destruction/damage/vandalism of property, embezzlement, extortion or blackmail, larceny or theft offenses, motor vehicle theft, robbery, and stolen property offenses). Crimes against society and arrest-only offenses are also captured in NIBRS but are not used in our data. This metric shows crime rates as the number of reported crimes per 100,000 people.

**Validity:** NIBRS is the most widely used source to measure and compare reported crime across the country. The FBI provides definitions of each variable collected and provides technical specifications and a user manual. 

**Availability:** The metric can be constructed using data from NIBRS, which is gaining increased participation from agencies across the country. As of 2021, agencies reporting to NIBRS covered 66 percent of the US population. If a community is not included in NIBRS, the relevant and comparable data can be requested by the public directly from their local law enforcement agencies or may be posted on their local law enforcement website. 

**Frequency:** This metric can be updated annually. 

**Geography:** This metric is available at the county and city level. The NIBRS data are reported at the agency level. Information about the cities and counties that the agency has jurisdiction over are available in the NIBRS data. Accordingly, the data can be aggregated across all agencies to other geographies. 

**Consistency:** The NIBRS data are consistent across the communities that provide data to the FBI, because the FBI establishes the definitions of the crimes included. Data may be accumulated and compiled differently at the local level. Prior to January 1, 2021, the FBI’s Uniform Crime Reporting (UCR) Program provided a standard, well-defined measure of crime. Reported crimes were captured for four “index” violent felonies (murder or nonnegligent manslaughter, rape, robbery, and aggravated assault) and four index property felonies (burglary, larceny-theft, motor vehicle theft, and arson). The UCR program was retired on January 1, 2021, transitioning the national standard for crime statistics to NIBRS to improve crime measures nationally. NIBRS offers greater specificity in reporting offenses, includes more detailed information, and provides context to specific crime problems. 

**Subgroups:** The NIBRS data include demographic information about the age, race, and gender of victims and person suspected of committing the offense. However, demographics in these data are unreliable. Additionally, those not directly impacted by a crime can still be negatively impacted by general exposure to crime. Looking at differences in crime rates by neighborhood demographics can illustrate whether there are racial disparities in exposure to crime. NIBRS data does not include neighborhood level information necessary for this analysis, but local law enforcement agencies have anonymized geocoded incident-level data and some agencies share this data publicly on government websites. The Urban Institute’s [Spatial Equity Data Tool](https://apps.urban.org/features/equity-data-tool/) can be used for this analysis. 

**Limitations:** Reporting to NIBRS is not mandatory, and although most communities provide data, NIBRS does not capture the universe of reported crimes across the US. NIBRS measures crime reported to the police, so unreported crime is not captured in these data. An FBI analysis estimated that up to half of violent crime goes unreported to the local police, and research finds that some neighborhoods are less likely to report violent crime, especially where trust of police is low. As a place-based measure, reported crime is affected by mobility in and out of the community. Crime rates are based on the number of incidents per 100,000 residents. If the number of residents increases and crime remains constant, the crime rate could go down without any change in the number of reported incidents. Also, since crime tends to be concentrated in certain areas, if new residents are moving to places where crime rates were already low, the populations and areas experiencing the most crime may not see any change even if city-wide rates decrease. Relatedly, NIBRS does not provide data on crime at the neighborhood level, so it cannot track changes in crime or compare different places within a community. NIBRS reports data at the agency level and some agencies may have jurisdiction in multiple counties and cities. Similarly, multiple law enforcement agencies (e.g., state, county, city, university, tribal) can fall within a single county. Because of this, at the county level, crime counts may be underestimated in counties where some agency data is missing. 

<br> 

##### [PREDICTOR: JUST POLICING](https://upward-mobility.urban.org/just-policing) 

**Metric: Juvenile arrests per 100,000 juveniles** 

The Federal Bureau of Investigation’s (FBI) National Incident-Based Reporting System (NIBRS) provides information about arrests of people under the age of 18. Because individuals can be arrested multiple times, the NIBRS data reports the number of arrests and not the number of individuals arrested. The metric is for juvenile (defined as under 18 years of age and over 9 years of age) arrest for any crime or status offense, but the data can be broken down by offense type. Arrest rates can be calculated using population data from the American Community Survey. 

**Validity:** Although arrest behavior of the total population may be confounded by many factors, arrests among juvenile offenders can be more closely tied to overly punitive policing behavior. Research finds that juveniles are more likely to be arrested than adult suspects, after controlling for suspect race, gender, seriousness of offense, and amount of evidence. Research also finds large and disruptive impacts of juvenile justice system involvement on adult outcomes; juvenile detention is associated with lower educational attainment, lower rates of employment, and higher rates of criminal offending and incarceration as an adult.

**Availability:** The metric can be constructed using data from NIBRS, which is gaining increased participation from agencies across the country. As of 2021, agencies reporting to NIBRS covered 66 percent of the US population. If a community is not included in NIBRS, the relevant and comparable data can be requested by the public directly from their local law enforcement agencies or may be posted on their local law enforcement website. 

**Frequency:** This metric can be updated annually. Juvenile arrest data is available annually through the FBI. Arrest data before 2014 can be found on the Bureau of Justice Statistics Arrest Data tool.

**Geography:** This metric is available at the county and city level. The NIBRS data are available at the agency level, which can be aggregated to other geographies.

**Consistency:** These data are consistent across the communities that provide data to the FBI, because the FBI establishes the definitions of the crimes included in the index and of juvenile as between 10 and 17 years of age regardless of state definition.

**Subgroups:** This metric necessarily measures people within a particular age group but also provides arrest-level data including age as well as race or ethnicity and gender. Ethnicity data is inconsistently collected and frequently missing.

**Limitations:** Reporting to NIBRS is not mandatory, and although most communities are covered by NIBRS data, NIBRS does not capture the universe of reported crimes across the US. As a place-based measure, levels of arrests are affected by mobility in and out of the community, and because the measure is a rate, large increases or declines in the number of youth under age 18 in an area could also affect the metric.