Equality of opportunity is invoked across the political spectrum, which suggests the phrase carries more than one meaning. Measuring it exposes exactly where those meanings diverge.
Two standards travel under one phrase
The narrower standard asks whether formal barriers have been removed: whether rules exclude anyone by category and whether selection considers only relevant qualities.
The broader standard asks whether people start from comparable positions, on the reasoning that a fair race still produces predictable results if the starting lines differ.
Both are defensible, and they are not compatible. A system can fully satisfy the first while failing the second, which is why arguments using the same phrase frequently talk past each other.
Separating circumstance from effort is contested
The broader standard requires distinguishing outcomes caused by circumstances a person did not choose from those caused by their own decisions. The boundary is genuinely difficult to draw.
Parental income and place of birth fall clearly on the circumstance side. Ambition and persistence appear to be choices, yet both are shaped heavily by environment during childhood.
Where the line is drawn determines how much inequality counts as unfair. Analysts who place effort broadly find far less unfairness than those who treat most traits as inherited.
Mobility statistics measure something specific
The most common empirical approach compares children's adult position with their parents'. High mobility is treated as evidence that origins matter less.
Relative mobility asks how people are ranked against their peers, so improvement for everyone leaves it unchanged. Absolute mobility asks whether children are better off than their parents were.
The two often move in opposite directions. A country can show broad income growth across generations while rankings stay almost as fixed as before.
Timing changes the answer
Measuring outcomes at different ages produces different pictures, since earnings trajectories diverge over a career and comparisons made early understate lifetime differences.
Gaps also open and close at different educational stages. Assessment at school entry, at the end of compulsory schooling and after tertiary study each identifies a distinct problem.
Because interventions are usually evaluated over short windows, a programme that shifts an early measure may show nothing on a later one, and both findings can be accurate.
The measure chosen drives the conclusion
Once several defensible measures exist, selecting among them becomes a substantive decision rather than a technical one, and the selection is rarely stated explicitly.
Careful analysis reports several measures together and explains what each does and does not capture. That practice makes disagreement about values visible instead of disguising it as arithmetic.
Definitions embedded in law and official statistics also change over time and differ across countries, so cross-national comparisons require checking that the underlying constructs match.