News

Extended methodology: How we ranked the neighborhoods

The Kansas City Star set out to measure the quality of life of city neighborhoods as places to live.

We modeled our analysis on a similar project The Star did last year, when suburbs were compared and ranked, and on neighborhood indicator projects done by a few other organizations across the country, most notably the Urban Institute in Washington D.C. But we took this previous work several steps further by broadening the range of topics and data in our report.

We also didn’t prepare this neighborhood project just by ourselves. We consulted with neighborhood leaders across the city, planners and department heads at City Hall, as well as researchers at the University of Missouri-Kansas City and a statistics expert at Rockhurst College.

They all had a hand in shaping our work.

.

How neighborhoods were defined:

Our first step was defining neighborhoods.

The city’s planning department has identified 240 distinct neighborhoods and countless more homes associations. Most are too small to collect large amounts of data on. So The Star, with the advice of neighborhood leaders and UMKC, grouped adjacent neighborhoods and homes associations into 42 clusters. For instance, adjacent neighborhoods at one end of midtown, such as Coleman Highlands, Volker and Valentine, were grouped together in a cluster called Midtown West.

Then we apportioned those 42 clusters into six “sections” of the city. Those sections are identifiable parts of town, such as the Southeast Side or the Southland, based on similar character and geography.

Our intent is to profile the top-performing neighborhood cluster in each of those six sections of the city.

We think this is a fair way to compare neighborhoods. That’s because Kansas City is blessed with an abundance of neighborhood choices, from new up north to older out south. We sought to avoid comparing old to new, but instead highlight the most livable areas in different sections of town — until the last day of the series. Next Sunday we will rank all the neighborhood clusters citywide to give a broader perspective.

Most neighborhood clusters ended up with estimated 2005 populations of between 8,000 to 15,000 residents.

But some clusters were a little smaller or larger. Briarcliff, for example, was the smallest cluster with an estimated 3,400 residents. Having a small number of residents was an advantage in some ways, but also was a disadvantage because some stats are cumulative totals, where bigger clusters tended to do better.

.

How neighborhoods were compared:

Next, we looked for ways to quantify what residents valued in their neighborhoods. After discussions with neighborhood leaders, academics and others, we came up with 34 quality-of-life measures sorted into nine general categories. This is how neighborhood clusters were measured against each other:

.

Crime:

■ Violent crimes (murder, rape, robbery, etc.) per 1,000 residents.

■ Property crimes (burglary, vandalism, etc.) per 1,000 residents.

■ Change in violent crimes from 2001 to 2005.

■ Change in property crimes from 2001 to 2005.

In all cases, the data came from the Kansas City Police Department and was compiled by UMKC’s Center for Economic Information.

.

Education:

■ Public school elementary math scores from the Missouri Assessment Program (MAP) test. Schools counted for a neighborhood cluster if they were located there or had a significant attendance area in that cluster. Attendance areas for Kansas City School District magnet and charter schools extended two miles. Scores were compiled from the Missouri Department of Elementary and Secondary Education’s database.

■ Public school elementary reading/communications scores. We used the same grouping of schools and data sources as above.

■ Private and public elementary school teacher-pupil ratios. Catholic schools counted for neighborhood clusters within 2 to 6 miles, depending on whether a school was considered a neighborhood school or region-serving. The ratios were taken from the National Center for Education Statistics, which includes the Private School Universe Survey.

■ Private and public high school teacher-pupil ratios. High schools, magnet and charter schools in the Kansas City School District counted for neighborhood clusters within four miles, and private and parochial schools counted for neighborhood clusters within six miles. The data came from the National Center and from the Missouri education department database.

■ Private and public high school ACT scores. We applied the same groupings of schools as with high school teacher-pupil ratios. Scores for private schools came from the National Center for Education Statistics, while scores for public schools from the Missouri education department. However, all Catholic high schools had the same average ACT score because that’s the only way the Catholic Diocese of Kansas City-St. Joseph released the information.

■ Proximity to private elementary and high schools. Schools were included in a neighborhood cluster if they were within four miles of a neighborhood cluster (or 6 miles in the case of region-serving high schools such as Pembroke and Rockhurst.) Schools included were based on a listing by the National Center for Education Statistics, plus web sites for local private school groups and the Catholic Diocese of Kansas City-St. Joseph. Schools received double points if they included both an elementary and high school, but only half points if they served only one sex.

.

Housing:

■ Change in home resale values this decade, from 2001 through 2005. The data came from the region’s Multiple Listing Service and was compiled in neighborhoods by UMKC’s Center for Economic Information.

■ Permits for new and rehabbed single-family homes, totaled for 2004 and 2005, per 1,000 residents. This data was compiled for neighborhoods by the city’s Planning and Development Department.

■ Housing balance. This was computed by how close a neighborhood cluster’s range of low-, middle- and upper-priced homes, and its number of apartments, compared to the city’s averages. Differences between a cluster’s and the city’s percentages in these four measures were added up for one total, with lower being better. The data came from the 2000 U.S. census.

■ Structure condition rating. This was computed as the percentage of homes determined to be substandard or worse, scoring less than 3.5 (on a 1 to 5 scale, with 5 being best) in neighborhood surveys done by UMKC in 2000 and 2001. A lower number is better.

.

Community:

■ Number of historic properties and public art objects, based on properties and neighborhood districts listed in the National and Kansas City registers of historic places, and based on the city parks department’s listing of public fountains, monuments and sculptures.

■ Litter index scores, averaged from 2003 through 2005. This was taken from Bridging the Gap’s Keep Kansas City Beautiful annual surveys in neighborhoods.

■ Demographic diversity scores. This was computed by how close a neighborhood cluster’s percentages of minorities, young people under 18, older people over 65 and residents in poverty compared to the city’s averages. Differences between a cluster’s and the city’s percentages in these four measures were added up for one total, with lower being better. The data was taken from the 2000 census.

.

Government:

■ Street and infrastructure condition rating. This was computed as a percentage of properties where the streets, sidewalks, curbs and streetlights are determined to be substandard or worse, scoring less than 3.5 (on a 1 to 5 scale, with 5 being best) in neighborhood surveys done by UMKC in 2000 and 2001.

■ Total local property taxes on a $150,000 home. This included school, library and road districts, along with several local and state health levies, plus some special county levies. The levies came from each county’s list of property tax levies by taxing district.

■ Number of water-main breaks in 2005, per 10,000 residents. The data came from the city water department’s list of locations.

■ Major neighborhood-oriented capital improvements spending per-capita, based on neighborhood conservation allocations, which include the Public Improvement Advisory Committee (PIAC). Only projects exceeding $50,000 were considered. They were totaled up for three budget years, fiscal 2003-04 through 2005-06.

.

Services:

■ Retail and service businesses in 2003 per 1,000 residents. This counted groceries, health care, churches and other services (but not restaurants and bars), by ZIP code. The data was taken from census’ ZIP Code Business Patterns. For ZIP codes that covered multiple neighborhoods, businesses were apportioned to a neighborhood based on its proportion of the ZIP code.

■ Net new businesses from 1999 to 2003 per 10,000 residents. This counted the same retail and services businesses as above, and the data came from the same source.

■ Number of licensed child-care centers and registered child-care facilities per 10,000 residents. Facilities counted were those listed in databases kept by Missouri’s Department of Health and Senior Services and the Local Investment Commission. At-home providers counted as one-fifth of a care center, based on a recommendation from The Family Conservancy.

.

Neighbors:

■ Neighborhood activism, as measured by attendance at neighborhood workshops for the city’s FOCUS master plan, from 1998 through 2002. Attendees were listed in neighborhood report prepared by the city’s Planning and Development Department. Neighborhood clusters were compared by attendees per 1,000 population.

■ Neighborhood involvement, as measured by the number of block-party permits obtained from City Hall in 2004 and 2005, per 1,000 residents. Permits are used for closing off a street, and while neighborhoods don’t close a street for every gathering, permits give an indication of how organized a neighborhood is.

■ Neighborhood stability, as measured by percentage of owner-occupied housing in 2004. The data was compiled by UMKC by comparing a county property tax recipient with the address of a home.

■ Voter turnout rates in three recent city elections with national issues — August 2004, November 2004 and April 2005. The turnout rates by precinct came from the Kansas City, Clay County and Platte County election boards.

.

Transportation:

■ Average commute times, taken from the 2000 U.S. census.

■ Tally of Kansas City Area Transportation Authority bus routes in or bordering a neighborhood, per 10,000 population. Routes were based on ATA maps.

■ Traffic accidents per 1,000 residents on major thoroughfares in or bordering a neighborhood in 2005. This was based on Kansas City Police Department’s database of accident locations.

.

Lifestyle:

■ Restaurants and bars within one mile of a neighborhood, tallied in 2003 per 10,000 residents. The data came from census’ ZIP Code Business Patterns.

■ Park acreage within one mile of a neighborhood, tallied per 1,000 residents. Acreage included city, county and even suburban parks (if close enough to the Kansas City border), plus boulevard and parkway medians tallied by the city parks department. Acreage numbers came from the Kansas City Park & Recreation Reference Book, plus county and suburban parks departments.

■ Tally of important recreation and cultural amenities, such as number of city swimming pools, YMCA or community centers, biking or walking trails, public lakes, libraries, etc. This also included a category for special facilities such as skateboard parks, ice rinks, nature centers, etc.

.

How neighborhood scores were determined:

Once we gathered these statistics, they were sent to Rockhurst University’s mathematics department. There, professor and department chair Paula Shorter used a C-score method to standardize each statistical measure into points on a 1-to-100 scale, except for mathematical outliers, when a neighborhood’s performance was way above or below the norm. Scores for neighborhood clusters in every measure were based on how close or far they were to the statistical average, with each measure’s average set at 50.

Additionally, some stats were considered more important than others. Each category of stats was weighted based on a poll done for the Mid-America Regional Council that asked area residents about the importance of various characteristics for their quality of life. The Star only used responses from Jackson, Clay and Platte counties.

For instance, 95 percent of respondents said “safe neighborhoods” were very or somewhat important to their quality of life, and 81 percent said a “safe/healthy environment” was important to them, for an 88 percent average for the crime category. At the opposite end of the spectrum, 37 percent said “good parks and recreation” were important and 28 percent said “strong arts and culture” were important, for a 32.5 percent average for the lifestyle category. Lifestyle became the base weight of 1, and other weights were determined by dividing a category’s average poll results by lifestyle’s results.

Here’s how the weights ended up: crime 2.71, education 2.12, housing 2.08, community 2.02, government 1.80, services 1.71, neighbors 1.43, transportation 1.32 and lifestyle 1.00. This means residents typically thought safety was nearly three times more important to them than parks and culture.

In the end, neighborhood cluster scores for a category were determined by averaging the points for all the stats within that category, then multiplying that average by the category weight. So with crime, for example, points for the four stats were added together, divided by four for an average, then that average was multiplied by 2.71.

This process was repeated for each category to come up with final point totals.

  Comments