The Selection Mirage

Differential Effective Thresholds, Selective Survivor Excellence, and the Mis-Inference of Fairness

Jonathan Latim · Concept Paper · September 2026

A system need not exclude a population outright in order to impose a substantially greater effective burden upon it. Two populations may pursue the same outcome under formally identical rules while facing systematically different requirements for entering, remaining, recovering from error, or advancing. Selection then changes the composition of the people who remain visible, and those survivors can be used as evidence about the fairness of the process that produced them.

That pattern is the Selection Mirage.

The Missing Comparison

1 Same starting point Same spread of ability before selection

Same starting point Two identical rows of twenty people, A and B, spread by one relevant ability. In this model the groups start with the same spread of ability. ABlessrelevant abilitymore

2 Different thresholds B faces a higher bar; outlines mark who drops out

Different thresholds A faces a lower bar and B a higher one. Below each bar, people drop out, shown as empty outlines. 4 of A's twenty are filtered out; 12 of B's are, disproportionately its ordinary members. ABA's barB's bar

3 What the observer compares Survivors only

What the observer compares Only the survivors enter the comparison: 16 from A and 8 from B. B's survivors are fewer, and B's average on the relevant ability is higher than A's. The filtered-out people are absent from this comparison. ABA's averageB's average

B's higher survivor average can make B's burden look no greater than A's. In this model both groups started with the same spread of ability, but B faced the higher bar. The comparison leaves out everyone the bars removed: 12 from B, 4 from A.

A simplified model, illustrative, not data: both groups start with the same spread of one relevant ability, and B faces a higher threshold. In real cases, survivor quality alone does not establish a higher bar: different starting points, preparation, or self-selection can produce it too.

The Ordinary Test

Among people who were comparable before the relevant selection process materially altered them, does group membership systematically change the performance required, the mistakes tolerated, or the opportunities for recovery needed to obtain and maintain the same outcome?

In shorthand: Among comparable people, does group membership raise the bar or narrow the margin for ordinary error?

What This Does Not Establish

  1. Exceptional survivors do not, by themselves, prove that a differential filter existed. Different source distributions, different preparation, and independent self-selection are alternative explanations that must be considered.
  2. A disparity does not, by itself, prove discrimination or a Selection Mirage. The relevant evidence concerns the selection process, not the outcome alone.
  3. The framework specifies a question. It does not presume the answer. A Selection Mirage must be established empirically, not assumed from survivor quality or disparity.

The Concept Paper

The full concept paper, including the formal model, prior-art analysis, empirical predictions, failure cases, and boundary conditions.

The concept paper will be linked here when available on SSRN.

How to Cite

Latim, J. (2026). The Selection Mirage: Differential Effective Thresholds, Selective Survivor Excellence, and the Mis-Inference of Fairness. Concept paper. Available at https://latim.org/selection-mirage/.

ORCID: 0009-0002-1051-6994
SSRN Author ID: 13078568