The Core Idea
Borrowed Directly From Newtonian Physics
The gravity model predicts the STRENGTH of interaction between two places using a formula directly borrowed from Newton's law of gravitational attraction: Interaction is proportional to (Population₁ × Population₂) ÷ Distance². Just as larger physical masses exert stronger gravitational pull, larger POPULATIONS generate stronger human interaction (more migration, trade, phone calls, or travel) — and just as gravitational pull weakens with distance, human interaction weakens as distance increases.
This model formalizes something Ravenstein observed empirically in his migration laws (from the earlier Migration Theory lesson): most human interaction — migration, trade, communication — happens preferentially between places that are either large, close together, or ideally both.
💡 Memory Trick
Picture two planets in space. Bigger planets (larger populations) exert a stronger gravitational pull on each other, and that pull weakens rapidly the further apart the planets drift — specifically weakening with the SQUARE of the distance, not just distance itself, meaning doubling the distance cuts the pull to a quarter of its original strength, not merely half. Two massive planets sitting close together experience an enormous mutual pull; two tiny, far-apart planets barely interact at all. Human geography's gravity model borrows this exact mathematical relationship to predict how strongly two CITIES (rather than planets) will interact.
Applying the Formula
Population, Distance, and the Squared Term
1
Population's Role
Larger populations at either location increase predicted interaction — two large cities are predicted to interact more strongly than two small towns, all else being equal, since more people means more potential migrants, traders, and communicators on both ends.
2
Distance's Role — Squared, Not Linear
Interaction decreases as distance increases, but crucially by the SQUARE of the distance, not simply in direct proportion to it — doubling the distance between two places doesn't just halve predicted interaction, it reduces it to roughly a QUARTER of its original strength, since the distance term is squared in the denominator.
3
Real-World Applications
The gravity model is used to predict migration flows between specific cities, international trade volumes between countries, retail shopping patterns (which store customers are likely to travel to), and even telecommunications traffic between regions — anywhere human interaction can be modeled as a function of relative size and distance.
The Model's Real Limitations
What Pure Population and Distance Can't Capture
The gravity model is a genuinely useful FIRST approximation, but it doesn't account for several real-world factors that meaningfully shape actual interaction: cultural and linguistic ties (shared language dramatically increases interaction beyond what population and distance alone would predict), historical colonial relationships (which shape trade and migration patterns independent of pure distance), political relationships and border restrictions (a closed or heavily restricted border can suppress interaction between two large, nearby populations far below what the formula predicts), and transportation infrastructure (a direct flight route can effectively 'shrink' distance in a way ground distance alone doesn't capture).
This is exactly why real-world applications of the gravity model frequently need to be ADJUSTED with additional variables beyond the basic population-and-distance formula — the pure Newtonian version is a starting point for prediction, not a complete, final explanation of actual human interaction patterns.
🖥️ Applied Scenario
A geographer is using the gravity model to predict migration flow between two pairs of cities: Pair A consists of two large cities 200 miles apart, and Pair B consists of two similarly large cities 400 miles apart.
1
You apply the formula to both pairs, noting that since population sizes are similar between the two pairs, the key difference driving the prediction is the DISTANCE term.
2
You calculate that because Pair B's distance (400 miles) is DOUBLE Pair A's distance (200 miles), and the formula divides by distance SQUARED, Pair B's predicted interaction should be roughly ONE QUARTER of Pair A's — not simply half.
3
You predict significantly stronger migration flow (and other interaction, like trade or communication) between the closer Pair A cities compared to the more distant Pair B cities.
4
Conclusion: correctly applying the SQUARED distance term — rather than assuming a simple, direct (linear) relationship between distance and reduced interaction — produces a meaningfully more accurate prediction, exactly illustrating why the gravity model's specific mathematical structure matters, not just its general 'bigger and closer means more interaction' intuition.
📌 Exam Application
Exam questions frequently ask you to apply the gravity model formula given population and distance data for two locations, correctly accounting for the squared distance term. You may also be asked to explain a real-world factor (cultural ties, political relationships, transportation infrastructure) that could cause actual interaction to diverge from the basic formula's prediction.
⚠️ Most Common Gravity Model Mistakes
The most common mistake is treating distance's effect as directly proportional (linear) rather than squared — doubling the distance between two places doesn't simply halve predicted interaction, it reduces it to roughly a QUARTER, and forgetting to square the distance term produces a significantly incorrect prediction. Another frequent error is treating the gravity model's prediction as a complete, final explanation rather than a useful first approximation — real-world factors like shared language, political relationships, and transportation infrastructure can cause actual interaction to diverge substantially from what the basic population-and-distance formula alone predicts.
✓ Quick Self-Test
Given population and distance data for two locations, can you correctly apply the gravity model formula, including the squared distance term? Can you identify a real-world factor that could cause actual interaction between two places to diverge from the model's basic prediction?
Next Lesson
Urban Structure Models
→
← All Human Geography Lessons