Pre-Commitment and Post-Hoc Rationalisation: Why You Must Set Criteria Before You See Options

You’re hiring for a senior role. You’ve reviewed four strong candidates. One of them — the second you interviewed — had a presence that stood out. Articulate, confident, impressive track record. By the time you sit down to make your final assessment, you find yourself emphasising “communication skills” and “executive presence” as your top criteria. These feel like the right priorities for the role. What you don’t notice is that three weeks ago, before you started interviewing, you would have listed “technical depth” and “team-building track record” as the priorities. The criteria didn’t drive the evaluation. The evaluation rewrote the criteria.

This is post-hoc rationalisation at work — and it’s one of the most common, least visible ways that decisions go wrong. You don’t experience it as bias. You experience it as clarity.

The research

J. Edward Russo, Victoria Medvec, and Margaret Meloy published a study in Organizational Behavior and Human Decision Processes in 1996 that mapped this distortion in real time. They tracked how participants evaluated options as information was revealed sequentially. The finding was stark: from the moment a preliminary preference formed — often after just the first or second piece of information — subsequent information was systematically distorted in favour of the leading option. Participants didn’t just weigh information differently. They actually perceived it differently, interpreting ambiguous data as supporting their emerging favourite.

Russo and colleagues called this “predecisional distortion.” It operates before the decision is made, not after — which means it can’t be corrected by “sleeping on it” or “thinking it through more carefully.” The longer the evaluation continues with an active preference, the more distorted the information landscape becomes. By the time you reach a decision, the case for the preferred option feels overwhelming — because you’ve been unconsciously building it with every new piece of data.

Richard Nisbett and Timothy Wilson, in a landmark 1977 paper in Psychological Review, demonstrated that people have remarkably little insight into their own decision-making processes. When asked to explain why they chose one option over another, participants routinely offered confident, plausible explanations that bore no relationship to the factors that actually influenced their choice. They weren’t lying. They were confabulating — constructing rational narratives for decisions driven by processes they couldn’t access.

Ralph Keeney, a decision scientist at Duke and MIT, argued in Value-Focused Thinking (1992) that the standard decision-making approach — generate options, then evaluate — is fundamentally backward. Options should be generated from criteria, not evaluated against them after the fact. When you define what success looks like before you see what’s available, you create a fixed reference point that resists the gravitational pull of whatever option happens to be most vivid, most recent, or most emotionally appealing.

The mechanism

The psychological mechanism is attachment — specifically, what Irving Janis and Leon Mann described in Decision Making (1977) as “bolstering.” Once a tentative preference forms, the decision-maker enters a phase of selective attention: seeking information that supports the preferred option, avoiding or discounting information that challenges it, and raising the importance of criteria on which the preferred option excels.

Bolstering is not a failure of critical thinking. It’s a feature of how the brain manages the cognitive and emotional cost of decision-making. Holding multiple options in genuine tension is uncomfortable and resource-intensive. The brain resolves this tension by quietly promoting one option, then marshalling reasons to justify the promotion. By the time you “decide,” the decision has already been made — the conscious deliberation is largely ceremonial.

Christopher Hsee’s evaluability hypothesis, published in Organizational Behavior and Human Decision Processes in 1996, adds another dimension. Hsee showed that attributes which are easy to evaluate — salary, years of experience, a well-known employer on a CV — dominate when options are evaluated separately. Attributes that are harder to evaluate — creativity, cultural fit, long-term potential — only become influential when options are compared side by side against explicit criteria. Without pre-set criteria, the easy-to-evaluate attributes win by default, regardless of whether they’re the most important.

This means that defining criteria early doesn’t just prevent post-hoc rationalisation — it also changes which features of the options become visible. Hard-to-evaluate but genuinely important attributes only enter the decision if you’ve explicitly named them as criteria before the evaluation begins.

You’re not resisting bias through willpower when you set criteria early. You’re structuring the decision so the bias doesn’t get a foothold.

The practical implications

Three criteria, ranked, is the minimum viable structure. More than five criteria and you lose the ability to make trade-offs — everything feels equally important, which means nothing is prioritised. Fewer than three and you’re likely missing a dimension that will surface as regret later. The ranking is essential: “cost matters more than speed, speed matters more than features” creates a decision rule that can actually resolve ties, which is where most decisions stall.

The criteria must be written down before any options are examined. The written list acts as an anchor, not a memory aid. Unwritten criteria are malleable; they shift to accommodate whatever option is most appealing. Written criteria are fixed — they can be revisited and revised, but the revision is conscious and deliberate, not the invisible drift of post-hoc rationalisation. The physical act of writing creates a commitment that verbal discussion does not.

Review the criteria after the decision to calibrate your process. Once you’ve chosen, compare the criteria you used in practice with the ones you wrote down. If they’ve diverged — if “culture fit” crept in as a top priority when you’d originally ranked it third — that’s useful data. It doesn’t necessarily mean the decision was wrong. It means your original criteria were incomplete, or your evaluation was biased, and knowing which is true improves every subsequent decision you make.

The bigger picture

Most decision processes in organisations are options-first. Someone brings a proposal, a shortlist, a set of candidates — and the evaluation begins. The criteria emerge during the discussion, shaped by whoever speaks first, whichever option is presented most compellingly, and whatever factors happen to be salient that day. This feels natural and efficient. It is also the structure most vulnerable to predecisional distortion.

The alternative — values-first decision-making, as Keeney advocated — requires a few minutes of uncomfortable abstraction before the satisfying concreteness of evaluating real options. You have to sit with the question “what would make this decision successful?” before you know what the options look like. This feels unproductive. It is the opposite.

Every decision you’ve later regretted, where you found yourself saying “I should have prioritised differently” — that’s a post-hoc rationalisation failure. The criteria shifted without your noticing. The option that felt best in the moment wasn’t the one that best served what actually mattered. The fix is not more discipline in evaluation. It’s more discipline in framing — before a single option crosses your desk.