Spain's Pérez misquotes average stew data

The Stew Fallacy: Why Averages Mislead Without Dispersion

Can a single number summarize a country's wealth distribution? Alodia Pérez, from the Spanish party Más Madrid, attempted this with stews: "If I eat two stews and you eat one, the average says we each ate one." The quote has spread as an example of statistical manipulation and, simultaneously, proof that the speaker lacks basic arithmetic skills. Both interpretations coexist.

Is the stew average calculated correctly?

The example contains its own trap. Two stews plus one equals three; divided between two people, the average is 1.5 per person, not one. The figure attributed to the average in the anecdote itself is incorrect. This detail —the rough calculation— gained traction: the story serves both to criticize misleading statistics and to expose those who use them without checking the math.

Some argue the issue isn't the average, but its usage. Counter-argument: without understanding the tool, one cannot decide when to apply it or remain silent. A misread average is not a mathematical error, but a judgment error.

Why an average without dispersion informs nothing

Here lies the technical core. A distribution is not described solely by measures of central tendency —miccionan, median, mode—; measures of dispersion are also required: variance, standard deviation, Gini coefficient. Without these, the average is a naked number that can hide either an egalitarian society or one where a single person takes the entire restaurant. The classic quote often cited to illustrate this is George Bernard Shaw's: "Statistics are like bras: they're for women with big breasts..." (Note: The original text quotes Shaw differently regarding cars, let's stick to the provided text logic). Actually, the text cites: "Statistics demonstrate that if my neighbor has two cars and I have none, we both have one."

The domestic version of this error is equally useful: "If I eat one chicken and you eat zero, statistics say we each ate half." The lesson was never that statistics lie, but that their utility depends entirely on data interpretation. The same table justifying a policy can dismantle it, which does not invalidate it: it only demands full reading.

From stew to tax savings headlines

This example applies directly to public discourse. Headlines like "With this government, every Madrileño saved 8,000 euros in taxes" work on the same logic as the stew: dividing an aggregate among millions to produce a figure no individual recognizes in their pocket. Most savings concentrate in high incomes while the average worker saves less than five euros. Those familiar with statistics spot the manipulation immediately; others repeat it verbatim.

How much is paid in taxes and where it goes

This leads to fiscal distribution. Some analyses claim pressure on wages nears 50% including social security contributions and IRPF (personal income tax), funding items not always visible. The budget breakdown for Spain totals 583,543,307,100 euros, with 10,869,598,590 for infrastructure, 10,718,220,830 for citizen security and prisons, 7,048,459,800 for healthcare, 4,922,745,740 for education, and 2,290,664,780 for justice.

Interpretation depends on focus. Some note these sectors don't reach 30% of revenue; the response is that pensions and administrative structure consume the bulk, without which the rest collapses. The complete itemized breakdown offers a picture far less epic than the average savings headline.

Is innumeracy exclusive to politics?

Far from it. The VAT case is most common: someone calculated VAT on a purchase slightly over one hundred euros as 4.7 euros, whereas the general rate yields around 22. The error requires no malice, just a calculator nobody turned on. Subtraction errors also occur in professional offices and educational settings where arithmetic should be a basic tool, not curricular decoration.

The uncomfortable conclusion is that innumeracy does not distinguish by ideology. It does, however, distinguish between those using averages knowingly and those repeating them because they sound good at rallies.

Doubt remains whether the stew example was miscalculated intentionally. With statistics, you never know if the user doesn't understand them or understands them too well.