{"id":167017,"date":"2026-01-26T09:43:52","date_gmt":"2026-01-26T17:43:52","guid":{"rendered":"https:\/\/lukeford.net\/blog\/?p=167017"},"modified":"2026-01-26T09:51:30","modified_gmt":"2026-01-26T17:51:30","slug":"the-replication-crisis-2","status":"publish","type":"post","link":"https:\/\/lukeford.net\/blog\/?p=167017","title":{"rendered":"The Replication Crisis"},"content":{"rendered":"

ChatGPT says: Alliance Theory<\/a> would interpret the replication crisis less as a sudden discovery that many social-science findings are \u201cfalse,\u201d and more as a structural shift in the alliance incentives that govern what kinds of claims get produced, rewarded, and protected.<\/p>\n

Start with what the old alliance rewarded.<\/p>\n

For decades, the dominant academic coalition in the social sciences was built around:<\/p>\n

Prestige journals
\nGrant agencies
\nElite departments
\nPeer networks
\nMedia amplification<\/p>\n

Status inside this alliance came from novelty, theoretical cleverness, and policy relevance. A result that was:<\/p>\n

Counterintuitive
\nMorally resonant
\nPolitically useful
\nMedia-friendly<\/p>\n

conferred far more alliance value than a result that was merely robust.<\/p>\n

Replication, by contrast, produces almost no alliance prestige. It:<\/p>\n

Does not create new theory.
\nDoes not advance careers quickly.
\nDoes not generate headlines.
\nDoes not signal creativity or moral leadership.<\/p>\n

So the incentive structure selected for interesting, story-driven, ideologically congenial findings, not for slow, confirmatory, boring accuracy.<\/p>\n

Alliance Theory says this is exactly what happens when a community\u2019s status economy is decoupled from its truth-testing function. The group still talks in the language of epistemic rigor, but what actually determines success is:<\/p>\n

Who your allies are.
\nWhich narratives your work supports.
\nHow well your results fit the coalition\u2019s moral and political self-image.
\nWhether your findings can be used as ammunition in broader cultural conflicts.<\/p>\n

Why replication suddenly became a \u201ccrisis.\u201d<\/p>\n

A new sub-alliance formed.<\/p>\n

Methodologists
\nStatisticians
\nOpen-science advocates
\nYounger scholars locked out of elite status
\nResearchers outside the old prestige networks<\/p>\n

They had a different rival map. Their enemy was not \u201creactionary politics\u201d or \u201cbackward social attitudes,\u201d but:<\/p>\n

Questionable research practices
\np-hacking
\nPublication bias
\nCareerist story-telling
\nIdeological filtering<\/p>\n

Once this group gained tools, pre-registration, large-scale replication projects, open data, they could publicly demonstrate that many high-status findings collapsed under scrutiny.<\/p>\n

Alliance Theory predicts the reaction.<\/p>\n

The old elite alliance responded with:<\/p>\n

Minimization.
\nMoral reframing.
\nAttacks on replicators\u2019 motives.
\nClaims that \u201ccontext matters\u201d or \u201ctruth is complex.\u201d<\/p>\n

Not because everyone was dishonest, but because acknowledging widespread failure would delegitimize the very status hierarchy that had elevated them.<\/p>\n

So what looks like an epistemic dispute is also an alliance struggle.<\/p>\n

One coalition optimizes for:<\/p>\n

Narrative power
\nMoral signaling
\nPolicy influence
\nMedia resonance
\nInternal status reproduction<\/p>\n

The other optimizes for:<\/p>\n

Procedural rigor
\nTransparency
\nError detection
\nMethodological humility
\nCross-camp credibility<\/p>\n

These are different loyalty tests.<\/p>\n

Why false results flourished in the first place.<\/p>\n

Alliance Theory\u2019s blunt answer is:<\/p>\n

Because they were useful.<\/p>\n

They supported:<\/p>\n

Progressive moral narratives.
\nInstitutional agendas.
\nGrant priorities.
\nJournal brands.
\nPublic intellectual reputations.<\/p>\n

In a coalition that defines itself as advancing social justice or enlightenment, results that flatter that self-conception are unconsciously protected. Skeptical results threaten not just theories but alliances, careers, and moral identity.<\/p>\n

So the system drifted toward:<\/p>\n

Overclaiming.
\nUnderpowered studies.
\nFlexible analysis.
\nStory-first, method-second publication.<\/p>\n

Not through conspiracy, but through selection.<\/p>\n

What the \u201ccrisis\u201d really marks.<\/p>\n

It marks a partial realignment of the alliance structure governing knowledge.<\/p>\n

The old prestige-and-narrative coalition is being challenged by a procedural-and-credibility coalition.<\/p>\n

The fight is not just about statistics.
\nIt is about who gets to define what counts as \u201cgood science,\u201d who gets status, and which moral and political narratives get epistemic protection.<\/p>\n

Alliance Theory would predict three long-term outcomes.<\/p>\n

Partial reform.
\nSome replication norms become institutionalized, but elite narrative production continues with softer constraints.<\/p>\n

Field bifurcation.
\nA high-prestige, theory-and-politics oriented social science alongside a lower-prestige but more method-rigorous replication culture.<\/p>\n

New equilibrium.
\nIf external trust and funding become tied to reproducibility, the alliance incentives may finally realign toward truth-tracking again.<\/p>\n

In short, the replication crisis is not just a methodological failure.
\nIt is a visible crack in an alliance system that had come to reward being interesting, useful, and morally aligned more than being reliably correct.<\/p>\n","protected":false},"excerpt":{"rendered":"

ChatGPT says: Alliance Theory would interpret the replication crisis less as a sudden discovery that many social-science findings are \u201cfalse,\u201d and more as a structural shift in the alliance incentives that govern what kinds of claims get produced, rewarded, and … Continue reading →<\/span><\/a><\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_monsterinsights_skip_tracking":false,"footnotes":""},"categories":[16281],"tags":[],"class_list":["post-167017","post","type-post","status-publish","format-standard","hentry","category-college"],"_links":{"self":[{"href":"https:\/\/lukeford.net\/blog\/index.php?rest_route=\/wp\/v2\/posts\/167017","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/lukeford.net\/blog\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/lukeford.net\/blog\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/lukeford.net\/blog\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/lukeford.net\/blog\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=167017"}],"version-history":[{"count":2,"href":"https:\/\/lukeford.net\/blog\/index.php?rest_route=\/wp\/v2\/posts\/167017\/revisions"}],"predecessor-version":[{"id":167028,"href":"https:\/\/lukeford.net\/blog\/index.php?rest_route=\/wp\/v2\/posts\/167017\/revisions\/167028"}],"wp:attachment":[{"href":"https:\/\/lukeford.net\/blog\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=167017"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/lukeford.net\/blog\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=167017"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/lukeford.net\/blog\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=167017"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}