How Many Kids Will It Bring?

Lincoln stopped borrowing its student generation rate and measured its own — down to the subdivision, with a margin of error on every estimate.

By Kurt Elder, AICP, GISP, Senior Planner, City of Lincoln Urban Development, and Kevin Nelms, Geospatial Analyst, Lincoln Public Schools

Ask a school board member, a parks planner, or a children's librarian what keeps them up at night, and you will hear the same question in different words: when the next subdivision goes in, how many kids are coming with it, or what will the projected population be in this area?

Four public agencies need that answer. Lincoln Public Schools (LPS) uses it to draw attendance boundaries and plan buildings. Planning practitioners use it to review subdivisions and balance the comprehensive plan estimates with another source. Recreation staff can use it to size neighborhood parks and programs. Libraries can use it to plan branches and youth services. And the nonprofits that serve children and families are asking too (e.g. daycares etc.,).

For years, everyone worked from the same rule of thumb: 0.3 students for every home. It came from the experience of longtime LPS and Planning staff, and it felt about right. But no one could point to where it came from. When we went looking for a source, the closest thing we found was a 2020 analysis from the National Association of Home Builders, “One Public School Child for Every Three Homes,” built on 2018 American Community Survey (ACS) data. It gave state-by-state rates by housing type and by new versus existing construction. Useful, but it could not say whether 0.3 fit Lincoln, let alone a 300-unit apartment project on the south edge of town.

So LPS, City of Lincoln Urban Development, and InformationArt set out to test the number against Lincoln's own records, and to put a range around every estimate. 

We counted every student and every home

This is not a survey. It is a count.

The City's Development Application Tracking System (DATS) divides Lincoln and its growth area into 4,847 small polygons called DATS tracts. The typical one is about 26 acres. Each DATS tract carries its application number, its growth tier from the comprehensive plan, and its housing units by type: existing, approved, and still to be built, sorted into single-family detached (SFD), single-family attached (SFA), duplex, and multifamily.

Every K–12 student enrolled in LPS on October 1, 2025 was matched to a DATS tract by grade level and housing type. In the existing city alone, that is 126,908 housing units and 38,530 students. The rates by housing type below rest on the 125,138 units and 35,527 students in the four main types (67,826 + 9,805 + 6,030 + 41,477 units); mobile homes make up most of the rest. Because we measured the whole population rather than a sample, the citywide averages carry very little sampling error.

From DATS tracts to traffic zones. Each DATS tract sits entirely inside one Traffic Analysis Zone (TAZ), the geography used in transportation modeling; no DATS tract is split between zones. The 4,847 DATS tracts roll up into 509 TAZs, a median of five per zone. The rates are measured at the DATS tract level, and the TAZ gives the tool a larger, familiar geography for mapping and export.

Three outside sources round out the picture. Esri's 2025 Tapestry Segmentation sorts neighborhoods into lifestyle types. Esri small-area population estimates support the population projections. ACS tables from the U.S. Census Bureau supply county fertility (B13002), child migration (C07001), and rental bedroom mix (B25068).

The old number was right. It was also hiding something.

Here is the first surprise: 0.3 holds up. Divide the existing city's 38,530 students by its 126,908 housing units and you get 0.304 students per home. Thirty years of staff intuition landed within two percent of the measured answer.

Here is the second: that average hides a threefold spread. A detached house or a duplex yields close to 0.38 students. An apartment yields about a third of that. Use 0.3 on a subdivision of single-family homes and you will undercount; use it on an apartment complex and you will overcount by more than double.

The precision column tells you how much to trust each rate. The existing-city rates are solid. The new-growth multifamily rate is not yet: it rests on only 21 DATS tracts, so the tool flags it as low confidence and says so on every estimate that uses it.

Two kinds of uncertainty, and why the difference matters

The most useful thing we learned had nothing to do with software.
It was separating two questions that usually get blurred together.

How well do we know the citywide average? This is precision. The rate is students divided by units. Its standard error comes from how much the rate varies across the DATS tracts that contain that housing type, and the coefficient of variation (CoV) expresses that error as a share of the rate:

SGR = K–12 students ÷ housing units        SE = s ÷ √n        CoV = SE ÷ SGR

For single-family detached homes, the spread (s) is 0.2147 across 617 DATS tracts (n). So SE = 0.2147 ÷ √617 ≈ 0.0086, and CoV ≈ 0.0086 ÷ 0.3835 ≈ 2.2%. The tool reports the rate as 0.384 ± 0.009, one standard error. At the 90% level the Census Bureau uses, the margin of error is 1.645 × 0.0086 ≈ ±0.014. The tool rates any CoV under 15% as high confidence.

How much will one project vary? This is variability, and it is far larger. DATS tracts genuinely differ, so the tool's rate cards also show a planning range of plus or minus one standard deviation. For 100 new detached homes, that is 38 students as the best estimate, with a range of roughly 17 to 60. Staff use the best estimate for budgeting and the ends of the range to stress-test school capacity. For a specific project, the Estimator reports a wider, 90% prediction interval: the range within which 9 of 10 comparable Lincoln developments fall. For the sample project in Figure 3, that is 0 to 136 students around a best estimate of 44.

A precise average and a wide project range are both true. Saying which is which is what builds trust with a school board, a developer, or a neighborhood meeting.

Not all apartments are alike

The strongest finding concerns apartments, and it changes how a planner should read a site plan. Student yield from apartments peaks in middle-income neighborhoods and drops at both ends of the income range.

The $30k–$50k band yields 0.1455 students per unit. That is 1.9 times the $100k+ band (0.0777) and 2.3 times the under-$30k band (0.0638), which is dominated by student and transient housing. In practice, 300 apartments in a $30k–$50k neighborhood suggest about 44 students; the same 300 units in a $100k+ neighborhood suggest about 23. That is a 21-student difference on one project, roughly a classroom. The tool therefore asks for a rent level on every multifamily estimate and maps it to these bands.

Inside the tool

The whole thing is a single web page file. It opens in any browser, needs no license, and is written in plain English for people who do not spend their days in spreadsheets. It has six tabs.

Estimate a Development. Staff enter unit counts by housing type, the growth tier or TAZ, and a rent level for apartments. The tool returns students by elementary, middle, and high school, with a best estimate, a planning range, and a confidence level. A printable impact report records every adjustment in a Methodology and Adjustment Log, so a reviewer can trace the number back to the measured rate.

What's in the model. Every estimate is built from five pieces, shown in Figure 2. (1) The base SGR, Lincoln's measured 2025 rate (LPS enrollment ÷ DATS units), is always the anchor and is never replaced. Optional multipliers then adjust it: (2) Tapestry, from ×0.75 to ×1.35, for whether the site's neighborhood type skews younger or older than its tier's average; (3) lifecycle, from ×0.05 to ×1.00, for where the project sits on the roughly 20-year ramp from a near-empty new subdivision to a mature one; and (4) fertility, which compounds Lancaster County's decline of about 2.4% a year in the general fertility rate, partly offset by families moving in, over the chosen build-out horizon. (5) Succession applies only to a year-40 projection; it swaps in the Tapestry multiplier of the segment the neighborhood is expected to become as it matures. Because every adjustment multiplies the base rate, any estimate traces back to the measured number.

Rates and Reference. The rate cards, the two kinds of confidence, and a 60-segment Tapestry reference grid, so users can see the evidence behind each number rather than take our word for it.

Full Build-Out Picture, Data Explorers, and Mapping. The build-out tab applies today's measured rates to every remaining approved and planned unit in DATS. Across all 4,847 DATS tracts, the 36,754 approved, platted, or planned units in the original DATS extract would raise K–12 enrollment from 41,763 to about 48,100 students: 6,363 more, or 15.2% (6,363 ÷ 41,763). Most of that capacity, 23,495 units or 64%, sits in Tier I Priority A, where new apartments currently yield 0.074 students per unit, about half the existing-city multifamily rate. The map draws TAZ results over an OpenStreetMap basemap, can overlay LPS attendance zones and board districts, and lets a user sketch any area to total students by school level in five-year steps. Every table exports as a spreadsheet (CSV) file for use in a geographic information system (GIS).

Let people ask their own questions

A printed report answers the questions its author thought to ask. An interactive tool lets each reader bring their own. Two features do most of that work.

Pick your own area and your own school level. A user can click zones or draw any shape on the map and read the result by school level. In Figure 8, a hand-drawn area near Highway 2 and Saltillo Road selects 11 TAZs with 2,000 housing units and 1,452 households. It holds 756 students today (343 elementary, 178 middle, 235 high), projected to reach 1,324 by 2050. A principal, a parks planner, or a librarian can each draw the area that matters to them, rather than accept the one the analyst chose.

See the assumption, not just the answer. The projection rises in equal steps of about 114 students every five years (568 ÷ 5 = 113.6). That straight line is a stated assumption, not a forecast of how construction will unfold, and showing it openly invites sharper questions. Will this area fill in five years or twenty? Is the nearby elementary school tight in 2030, long before 2050? Will about half a student per household hold (756 ÷ 1,452 = 0.52 today; 1,324 ÷ 2,571 = 0.51 in 2050), or will new subdivisions start younger?

Data and analysis are waypoints, not destinations. They point back to the families already living in these neighborhoods, the schools, parks, and libraries they use, and the public conversation about what comes next.

A tool people can question moves that conversation from “what does the model say?” to “what do we know about this place, and what should we do?”

From students to people

The same data project estimates total population. The tool starts from Esri's 2026 estimate for the DATS service area, which covers Lincoln, its extraterritorial jurisdiction (ETJ), and surrounding Lancaster County communities such as Waverly, Hickman, Bennet, and Firth: 339,319 residents in 137,808 households. It then adds the population implied by the remaining pipeline, using a Lincoln-calibrated ratio of about 8.15 residents per K–12 student (a trimmed median across DATS tracts). The result is 394,696 residents at build-out, 55,377 more than today, or 16.3%.

Households grow from 137,808 to 159,808, an increase of 22,000, or 16.0%. Average household size barely moves: 339,319 ÷ 137,808 ≈ 2.46 persons today, against 394,696 ÷ 159,808 ≈ 2.47 at build-out.

The build-out population was derived from student records, independently of the Planning Department's methods, yet it lands within 1.3% to 3.7% of the comprehensive plan's 400,000 to 410,000 target. (400,000 − 394,696 = 5,304, and 5,304 ÷ 400,000 = 1.3%; 410,000 − 394,696 = 15,304, and 15,304 ÷ 410,000 = 3.7%.) Two methods that start from different data and arrive at nearly the same place are a useful check on each other. The tool exports five-year projections from 2030 to 2050 for every area, alongside students.

The next step is to break those projections out by age group. With children under five, school-age youth, working adults, and seniors projected by small area, the City and its partners can run location-allocation analysis in GIS, which chooses the set of sites that best serves projected demand within a travel time or distance. That is how you decide where the next school, park, library branch, or child care center should go before the households that need it arrive.

Two sets of totals? Here's why.

Readers who open the tool will find two sets of baseline numbers, and both are right. The Full Build-Out Picture tab (Figure 6) counts all 4,847 DATS tracts: 41,763 students today and about 48,100 at build-out. The Data Explorer tabs (Figure 9) count only the 3,497 DATS tracts that link to the interactive TAZ map: 38,625 students today and 44,653 at build-out. The 3,138-student gap (41,763 − 38,625) sits in DATS tracts that are not in the map's TAZ layer. Remaining units differ for a separate reason: the Build-Out tab's 36,754 comes from the original DATS extract, while the Data Explorers compute from a newer one that shows about 36,950. The 196-unit difference is new applications added to the pipeline between the two pulls. Neither view is an error. The Build-Out tab gives the system-wide totals used in this article; the Data Explorers trade some completeness for the ability to map, filter, and export by area. Each Data Explorer tile now shows the full-system figure beside its own, so a user can see the difference and why it exists.

What this means for your community

First, the limits. The rates are a 2025 snapshot. Build-out figures apply them flat to remaining units; they do not model when units are absorbed or how a subdivision ages from young families to empty nesters. The Estimator's optional adjustments address those effects for individual projects. All results are planning-grade, not a guarantee for any parcel.

Now the lessons, and they travel well:

1. Use your own records. School enrollment joined to local permit or application data beats a borrowed national rate every time, and it can be refreshed every October.

2. Report uncertainty in two parts. A precise average and a wide project range are both true. Saying which is which builds trust with boards and developers.

3. Don't treat all apartments alike. In Lincoln, multifamily yield varies by more than a factor of two with neighborhood income (0.1455 ÷ 0.0638 = 2.3).

None of this required a consultant's budget or proprietary software. It required three things most Nebraska communities already have or can get: a school district willing to share an enrollment file with home addresses under a data-sharing agreement, a permit or application record that tags housing units by type, and a few weeks of GIS and analysis time. A city of 15,000 with one school district has fewer DATS tracts to build, not a harder problem.

If you want to learn more, feel free to reach out to the authors. Contact Kurt Elder at [email protected] and Kevin Nelms at [email protected].

The project is a cooperative effort of Lincoln Public Schools, City of Lincoln Urban Development, and InformationArt.

Data sources: LPS administrative enrollment, October 1, 2025; City of Lincoln DATS, extract of February 25, 2026; Esri 2025 Tapestry Segmentation and small-area estimates; U.S. Census Bureau ACS tables B13002, C07001, and B25068; National Association of Home Builders, “One Public School Child for Every Three Homes,” May 2020.