
build and compare exponential models of Missouri’s population from census data, use logarithms to find doubling times and projection dates, and judge the models against current estimates.
Instruction
The U.S. Census Bureau has counted Missouri’s population every ten years. The official counts:
| Year | Population | Growth factor from previous census |
|---|---|---|
| 1950 | 3,954,653 | — |
| 1960 | 4,319,813 | 1.092 |
| 1970 | 4,676,501 | 1.083 |
| 1980 | 4,916,686 | 1.051 |
| 1990 | 5,117,073 | 1.041 |
| 2000 | 5,595,211 | 1.093 |
| 2010 | 5,988,927 | 1.070 |
| 2020 | 6,154,913 | 1.028 |
Is it exponential? An exponential model needs a roughly constant growth factor. The decade factors fall from 1.092 to 1.041, jump back to 1.093 in the 1990s, and drop to 1.028 in the 2010s. Converting each to an annual rate (A2.IF.A.2: factor1/10) gives about 0.89% a year in the 1950s, 0.90% in the 1990s and only about 0.27% a year in the 2010s. A single exponential model cannot capture that.
Two models. Model A uses the endpoints: (6,154,913/3,954,653)1/70 ≈ 1.00634, so PA(t) = 3,954,653(1.00634)t, with t in years after 1950. It matches 1950 and 2020 exactly but predicts about 5.42 million for 2000, below the actual 5.60 million: the curve is too low in the middle because growth was not steady. Model B uses only the last census decade, 2010-2020: annual factor (6,154,913/5,988,927)1/10 ≈ 1.00274, or about 0.27% a year, so PB(t) = 5,988,927(1.00274)t with t in years after 2010. Because Missouri grows slowly, keep five decimal places in these factors; rounding 1.00274 to 1.003 shortens the doubling time below by about twenty years.
Logarithms answer the planning questions (A2.SSE.A.2). Doubling time for Model A: ln 2/ln 1.00634 ≈ 109.7 years. For Model B: about 253.3 years. When does each model reach 7 million? Model A: solve 3,954,653(1.00634)t = 7,000,000, so t = ln(7,000,000/3,954,653)/ln 1.00634 ≈ 90.4, around 2040. Model B, counting from 2020: t = ln(7,000,000/6,154,913)/ln 1.00274 ≈ 47.0, around 2067. Same data, different base years, and projections twenty-seven years apart.
Reality check. The U.S. Census Bureau publishes a population estimate for every state for July 1 of each year between censuses. Its Vintage 2025 estimate puts Missouri at 6,270,541 on July 1, 2025, a growth rate of about 0.36% a year since the 2020 census. Model A predicts about 6.36 million for that date (too high) and Model B about 6.24 million (too low): the two models bracket the estimate, and neither rate has held. That is the most important finding a report can make: exponential models describe constant-rate growth, and Missouri’s growth rate has not been constant. Compare the latest Census Bureau estimates with your models’ predictions. Your report should argue for which model, if any, a state planner should use for a ten-year forecast, and what it would need to be trusted.
📖 Read for this lesson
The reading opens over the lesson; close it to come back to exactly where you were. The checks below draw on it.
Formative check
Work through these before moving on. They are not graded — they tell you whether the standard below has landed yet.
Missouri grew by a factor of 1.0277 from 2010 to 2020. What annual growth rate is that?
What is the doubling time for growth of 0.27% per year?
Simulation & tools
Fitting an exponential to the census
In a blank Desmos graph, add a table with x1 = years after 1950 (0, 10, …, 70) and y1 = population in millions from the lesson table. Type y1 ~ a bx1 to fit an exponential by regression and record a, b and the implied annual growth rate. Then enter Model A, y = 3.955(1.00634)x, and Model B, y = 5.989(1.00274)x − 60. Record each model’s value at x = 50 and x = 80. Question: which of the three models is closest to the 2000 count, which gives the largest 2030 projection, and what does the pattern of misses (above or below the data) tell you about the shape of Missouri’s growth? No device: compute the same values with a calculator and plot the census points by hand.
Open Desmos ↗ · Desmos · free to use; Desmos terms
Stage 1: choose two base years for your models and explain why each is a reasonable choice. Predict, before calculating, which will fit 2020 better and which will project a larger population for 2035.
How confident are you that you can build, compare and critique exponential models of real data using logarithms?
Unit 1 checkpoint
Do this before you start the task. It covers the whole unit and tells you what to review first.
Which shows the monthly factor for 1.05t (t in years)?
log2 32 = and log 0.001 = .
log7 50 = ?
log(M + N) = log M + log N.
With log 2 ≈ 0.3010, log 8 ≈ .
In proving logb(MN) = logb M + logb N, which exponent law is used?
The solution of 5 · 2t = 80 is t = .
Solve 3 · 100.2t = 600.
The graph of y = log2 x has a vertical asymptote at x = 0.
The inverse of f(x) = 10x is f−1(x) = .
Doubling time at 7% a year is closest to
Submit: Missouri Population Modeling Report (4-6 pages) + 5-minute presentation
Task: Using the 1950-2020 census counts, compute decade and annual growth factors and argue whether Missouri’s growth has been exponential. Build at least two exponential models from different base years, graph them against the data with labeled axes (Desmos or a spreadsheet), and tabulate their errors. Use logarithms to find each model’s doubling time and the year it predicts 7 million. Compare the models with the latest U.S. Census Bureau population estimates and make a recommendation for a ten-year planning forecast.
Submission workflow
- Stage 1 — Plan. Growth-factor table, your two base-year choices with reasons, and a prediction of which model will fit recent years better.
- Stage 2 — Peer-reviewed draft. A partner recomputes one doubling time and one projection date and checks your graph labels; submit a revision note.
- Stage 3 — Final report. 4-6 pages with graphs, error table, logarithmic solutions shown step by step, the Census Bureau estimate comparison and a recommendation.
- Stage 4 — Five-minute presentation. Present your recommendation and answer two questions. A recorded presentation is an equal-credit alternative.
How it is scored
| Criterion | What proficient looks like |
|---|---|
| Logarithmic solutions (A2.SSE.A.2) | Doubling times and projection dates are solved with logarithms correctly and shown step by step. |
| Rewriting (A2.IF.A.2) | Decade factors are converted to annual rates correctly and interpreted. |
| Model building (A2.BF.A.1 / A2.FM.A.1) | Models are built correctly from stated base years and data. |
| Graphs (A2.BF.A.3, A2.FM.A.1) | Models and data are graphed together on labeled axes with appropriate scales. |
| Rate of change (A2.IF.A.1) | Average rates of change are computed and used to argue about the shape of growth. |
| Argument (MP.3, MP.4) | The recommendation follows from the evidence and names the models’ limits. |
Missouri’s Population: The Census Counts, and the Estimates That Arrived Afterward
U.S. Census Bureau, 2020 Census apportionment data file (resident population of Missouri, 1950–2020); U.S. Census Bureau, Population Division, Vintage 2025 state population estimates and components of change (NST-EST2025-POP and NST-EST2025-ALLDATA, released January 2026) and Vintage 2024 county estimates for Missouri (CO-EST2024-POP-29, March 2025); retrieved November 5, 2026 · Public domain (U.S. federal government data)
Why this reading. Your task ends by comparing your models with “the latest Census Bureau estimates.” That comparison is the finding, so it cannot depend on a download working. The Bureau’s July 1, 2025 estimate for Missouri is below, with the release it came from, and so is the arithmetic of the comparison, worked out to the last digit so you can check every step of your own.
Table 1. Decennial census counts for Missouri
| Census date | Population | Factor over the decade | Implied annual rate |
|---|---|---|---|
| April 1, 1950 | 3,954,653 | — | — |
| April 1, 1960 | 4,319,813 | 1.092 | 0.89% |
| April 1, 1970 | 4,676,501 | 1.083 | 0.80% |
| April 1, 1980 | 4,916,686 | 1.051 | 0.50% |
| April 1, 1990 | 5,117,073 | 1.041 | 0.40% |
| April 1, 2000 | 5,595,211 | 1.093 | 0.90% |
| April 1, 2010 | 5,988,927 | 1.070 | 0.68% |
| April 1, 2020 | 6,154,913 | 1.028 | 0.27% |
The annual rates are the decade factors rewritten on a one-year clock: factor1/10 − 1. Read the last column downward and the case against a single exponential model is already made. A constant-rate model requires a constant factor, and Missouri’s slows from 0.89% a year in the 1950s to 0.40% in the 1980s, jumps back to 0.90% in the 1990s, and falls to 0.27% in the 2010s.
Table 2. The estimates that came after the 2020 census
Between censuses, the U.S. Census Bureau estimates each state’s population for July 1 of every year, starting from the 2020 census and adding births, subtracting deaths, and adding net migration from other states and other countries. Each new release (a “vintage”) revises the earlier years. Vintage 2025, released in January 2026, reports:
| Date | Population | Change over the year |
|---|---|---|
| July 1, 2024 (Vintage 2025) | 6,243,544 | +37,588, or +0.61% |
| July 1, 2025 (Vintage 2025) | 6,270,541 | +26,997, or +0.43% |
The year before, Vintage 2024 had put July 1, 2024 at 6,245,466; the newer release revised it down by 1,922 people. Estimates move, and your report should say which vintage it used.
The components for the year to July 1, 2025 are the mechanism a growth factor never shows you. Missouri recorded 68,377 births and 68,200 deaths, a natural increase of only 177 people. Nearly all of the growth came from migration: 12,665 net international migrants (down from 24,076 the year before) and 14,028 net migrants from other states. A state whose births and deaths cancel is not growing “in proportion to its population” at all.
The comparison, worked
Model A fits the endpoints 1950 and 2020: (6,154,913/3,954,653)1/70 = 1.006339, so PA(t) = 3,954,653(1.00634)t, t in years after 1950. Model B fits 2010–2020: (6,154,913/5,988,927)1/10 = 1.002738, so PB(t) = 5,988,927(1.00274)t, t in years after 2010, about 0.27% a year.
| Model A (1950–2020) | Model B (2010–2020) | Actual | |
|---|---|---|---|
| Annual factor | 1.00634 | 1.00274 | — |
| Doubling time, ln 2 / ln b | 109.7 years | 253.3 years | — |
| Year it reaches 7 million | 2040 | 2067 | — |
| Fit at 2000 | 5,424,158 | — | 5,595,211 |
| Prediction for July 1, 2025 | 6,362,539 | 6,243,889 | 6,270,541 |
| Error against the Census estimate | +91,998 (+1.5%) | −26,652 (−0.4%) | — |
The predictions use t = 75.25 for Model A and t = 15.25 for Model B, because July 1, 2025 is a quarter-year past the April census date. Model A is about 92,000 people too high, and Model B is about 27,000 too low: the two models bracket the estimate. Compute the growth factor the state actually recorded from April 2020 to July 2025: (6,270,541/6,154,913)1/5.25 = 1.003551, about 0.36% a year. Model A assumed 0.63%, about 1.8 times that; Model B assumed 0.27%, about three-quarters of it. At the recent rate, Missouri would reach 7 million around 2056, between the two models’ dates of 2040 and 2067.
What a planner would actually need
Your recommendation is the point of the report, so it should say more than “use Model B.” Three things are worth arguing:
- Exponential may be the wrong family here. A constant percentage is a claim about the mechanism — that each year’s growth is proportional to the current population. Missouri’s growth now comes almost entirely from migration, and the international part halved in one year. Migration responds to jobs, housing and federal policy, none of which scale with population.
- A short base period is not automatically better. Model B used only the most recent census decade and still missed, low, because the years since 2020 grew faster than the 2010s did. The 1990s, at 0.90% a year, show how far a single decade can stray from the long-run rate. The problem is not only the window; it is the assumption that any window’s rate stays put.
- Say which direction you are wrong in. Model A errs high because its base period includes the faster growth of the 1950s and 1960s; Model B errs low because the 2010s were Missouri’s slowest census decade on record in this table. The 7-million dates differ by 27 years. A planner sizing schools or roads needs that range, and the reason for it, more than a single date.
One caution on comparing numbers. The census counts are for April 1 and the Bureau’s estimates are for July 1, so the interval used above runs April 2020 to July 2025, which is 5 years and 3 months. Rounding it to 5 makes the annual rate slightly too large: (6,270,541/6,154,913)1/5 = 1.003729, or 0.37%. Recompute your comparison both ways and see whether the conclusion changes — it does not, and showing that is worth a sentence in your report.
Sources: U.S. Census Bureau, 2020 Census apportionment data (historical resident population by state) (www2.census.gov); U.S. Census Bureau, Vintage 2025 state population totals (NST-EST2025-POP) (www2.census.gov); U.S. Census Bureau, Vintage 2025 state components of change (NST-EST2025-ALLDATA) (www2.census.gov); U.S. Census Bureau, Vintage 2024 state population totals (NST-EST2024-POP) (www2.census.gov); U.S. Census Bureau, Vintage 2024 county estimates for Missouri (CO-EST2024-POP-29) (www2.census.gov)