Massaman Curry Lab
What moves together in 95 massaman curry recipes
When a recipe uses more of one ingredient, does it tend to use more or less of another? For every pair of the 8 ingredient ratios (grams per can of coconut milk), this page ranks all 95 recipes on each and measures how closely the two rankings agree: Spearman's rho, from −1 (the more of one, always the less of the other) through 0 (unrelated) to 1 (always rise together).
28 pairs in all: 13 rise together at rho 0.20 or more, none move in opposite directions at −0.20 or less, and 6 are close to unrelated (between −0.10 and 0.10). This describes recipes as they are published; it does not say what changing one amount would do to a recipe.
Every pair at once
Read a cell by its row and column. Caramel cells rise together, slate cells move in opposite directions, and the deeper the color the stronger the relationship. Pale cells with grey numbers are not distinguishable from no relationship with 95 recipes (p of 0.05 or more).
All 28 pairs as a table
| Pair | Rho | Strength | p |
|---|---|---|---|
| Fish or soy sauce and sugar | 0.56 | moderate | under 0.001 |
| Potatoes and onions | 0.43 | moderate | under 0.001 |
| Tamarind or lime and sugar | 0.41 | moderate | under 0.001 |
| Curry paste and meat, fish or tofu | 0.37 | weak | under 0.001 |
| Peanuts and tamarind or lime | 0.34 | weak | under 0.001 |
| Fish or soy sauce and tamarind or lime | 0.31 | weak | 0.002 |
| Curry paste and potatoes | 0.29 | weak | 0.004 |
| Meat, fish or tofu and potatoes | 0.28 | weak | 0.005 |
| Meat, fish or tofu and sugar | 0.27 | weak | 0.008 |
| Meat, fish or tofu and fish or soy sauce | 0.25 | weak | 0.013 |
| Peanuts and sugar | 0.23 | weak | 0.025 |
| Meat, fish or tofu and onions | 0.23 | weak | 0.028 |
| Curry paste and fish or soy sauce | 0.22 | weak | 0.033 |
| Peanuts and fish or soy sauce | 0.20 | very weak | 0.056 |
| Curry paste and sugar | 0.20 | very weak | 0.057 |
| Meat, fish or tofu and peanuts | 0.19 | very weak | 0.066 |
| Potatoes and sugar | 0.17 | very weak | 0.10 |
| Curry paste and onions | 0.15 | very weak | 0.13 |
| Curry paste and tamarind or lime | 0.14 | very weak | 0.18 |
| Potatoes and fish or soy sauce | 0.12 | very weak | 0.24 |
| Potatoes and tamarind or lime | 0.12 | very weak | 0.25 |
| Onions and peanuts | 0.12 | very weak | 0.26 |
| Curry paste and peanuts | 0.10 | negligible | 0.36 |
| Meat, fish or tofu and tamarind or lime | 0.09 | negligible | 0.38 |
| Onions and tamarind or lime | 0.04 | negligible | 0.68 |
| Potatoes and peanuts | 0.02 | negligible | 0.83 |
| Onions and sugar | −0.01 | negligible | 0.89 |
| Onions and fish or soy sauce | 0.01 | negligible | 0.86 |
The three clearest relationships
The strongest pairs that reach rho 0.20 either way with a p-value under 0.01. Each dot is a recipe; the line joins the median of the second ingredient within each quarter of recipes ranked by the first.
Recipes with more fish or soy sauce tend to use more sugar
rho 0.56 across 95 recipes · moderate · p under 0.001
The third of recipes with the most fish or soy sauce use a median of 13 g sugar per can of coconut milk; the 33 with no fish or soy sauce, 0 g.
Recipes with more potatoes tend to use more onions
rho 0.43 across 95 recipes · moderate · p under 0.001
The third of recipes with the most potatoes use a median of 110 g onions per can of coconut milk; the third with the fewest potatoes, 0 g.
Recipes with more tamarind or lime tend to use more sugar
rho 0.41 across 95 recipes · moderate · p under 0.001
The third of recipes with the most tamarind or lime use a median of 12 g sugar per can of coconut milk; the 38 with no tamarind or lime, 0 g.
What moves on its own
Every ingredient moves with at least one other: none has its strongest correlation under 0.15.
The pairs closest to unrelated: onions and fish or soy sauce (0.01), onions and sugar (−0.01), potatoes and peanuts (0.02) and onions and tamarind or lime (0.04).
How the numbers are made
Spearman's rho is the ordinary correlation of the two rankings, so it measures whether one amount tends to rise with the other, in a straight line or not, and a single extreme recipe moves it little. A recipe with none of an ingredient ranks lowest on it. An ingredient fewer than 12 recipes use is left out of the matrix.
The p-value is the chance of a rho at least this far from zero if the two amounts were unrelated. With 28 pairs tested, about 1.4 would pass 0.05 by chance alone even if no ingredient were related to another, which is why the relationships explained above must pass 0.01 and reach 0.20.
Every number is divided by the same count of cans, so a recipe with few cans for its other ingredients reads high on every one of them at once. That nudges every correlation toward positive: part of a positive rho can be the coconut milk, not the two ingredients. A negative one holds despite it.
A correlation describes the recipes as published. “Recipes with more curry paste tend to use…” is a statement about what cooks write down together, not a rule that adding one forces the other. Plot any two ingredients to see the recipes behind a number.