The Most Popular Marketing Psychology Theory Has a Bias-Corrected Effect Size of Zero
A Theory Everyone Uses, Almost Nobody Questions
Construal Level Theory (CLT) is everywhere in marketing advice. The core idea: when something feels psychologically distant — far in the future, geographically remote, socially unfamiliar, or hypothetically unlikely — your brain processes it abstractly. When it's close, you think concretely.
Far away = "why" (purpose, values, vision). Close up = "how" (steps, logistics, features).
The implication for marketing seems clean: use abstract benefit-driven messaging for cold audiences (psychologically distant), concrete feature-driven messaging for warm audiences (psychologically close). Match your message to their distance.
Thousands of marketers have built their copy frameworks on this. The problem is what happens when scientists try to replicate the findings.
---
The Original Evidence Looked Strong
Soderberg et al. (2015) published a meta-analysis in Psychological Bulletin covering 106 papers and 267 experiments. They found:
- Psychological distance on abstraction: d = 0.40 (95% CI: 0.32-0.48)
- Downstream consequences: d = 0.26 (95% CI: 0.16-0.36)
- All four distance dimensions showed similar effects (temporal, spatial, social, hypothetical)
That d = 0.40 is a respectable medium effect. Enough to feel confident building practical applications on it.
Then Maier et al. (2022) ran the same data through a publication bias correction model.
---
The Publication Bias Bombshell
Maier, Bartos, Oh, Wagenmakers, Shanks, and Harris (2022) reanalyzed the Soderberg dataset using Robust Bayesian Meta-Analysis (RoBMA), which accounts for publication bias across 12 different model specifications.
The results:
- Very strong evidence for publication bias: BF > 10^100
- Moderate evidence for the absence of the effect: BF = 7.75 (data 7.75x more likely under null hypothesis)
- Bias-corrected effect size: d = -0.002 (95% CI: -0.113 to 0.056)
Read that again. The uncorrected d = 0.40 becomes d = -0.002 after correcting for publication bias. The effect essentially disappears to zero.
Schimmack (2022) ran an independent z-curve analysis on 3,970 test statistics from 200 CLT papers. He found an expected discovery rate of just 14% against an observed rate of 74% — a 500% inflation. This expected discovery rate is lower than discredited candidate gene studies. Up to 31% of published significant results may be false positives.
---
What Actually Failed to Replicate
This isn't just statistical critique. Specific replication attempts have failed:
Wakslak et al. (2006) — the likelihood-construal link. Original studies: N = 20 and N = 34. Replication attempts (Calderon et al., 2020): N = 115 and N = 120. Result: failed to find the effect. Bayesian analyses supported absence.
Benschop et al. (2021) — the standard construal level prime. Used the commonly employed manipulation in a lab experiment. The prime failed to produce significant differences on the Behavior Identification Form. No downstream effects found.
Yang et al. (2021) — spatial distance effects on mental construal. Failed to replicate using cued distance manipulation.
The definitive test is still in progress: the CLIMR project (Construal Level International Multilab Replication) has 78 labs across 27 countries, targeting ~7,500 participants with 97.2% power to detect d = 0.20. Results are not yet published as of early 2026.
---
What This Does NOT Mean
It does not mean the intuition behind CLT is worthless. The idea that people think differently about close versus distant events has face validity. You probably do think about tomorrow's meeting differently than a conference six months from now.
What it means is that the measured effect sizes are likely much smaller than reported, the standard experimental manipulations may not reliably shift how people process information, and building precise marketing frameworks on CLT-specific findings is building on sand.
The theory may be directionally useful. It is not a reliable engineering tool.
---
What Actually HAS Replicated: Implementation Intentions
Here is the good news. One of the most practically useful ideas connected to CLT — implementation intentions — has robust evidence that has survived scrutiny.
Gollwitzer and Sheeran (2006) meta-analyzed 94 independent tests with 8,000+ participants and found d = 0.65 (medium-to-large) for if-then plans on goal achievement.
In 2024, Sheeran, Listrom, and Gollwitzer updated the analysis with 642 independent tests. The effect held. Key moderators that increase effectiveness: contingent if-then format (vs. general plans), high motivation to pursue the goal, and plans rehearsed at least once.
Implementation intentions are the ultimate concrete intervention. They take an abstract goal ("I want to build systems for my business") and translate it into a specific situational response ("If it's Tuesday at 9am, then I will record one FAQ video").
This matters because it gives you a practical framework that doesn't depend on CLT's questionable effect sizes:
- First, establish why (connect to purpose). This draws on CLT's desirability framing — even if the distance-construal link is weaker than thought, people still respond to meaning and purpose.
- Then, specify how (provide if-then plans). This draws on implementation intentions — which have robust, replicated evidence across 642 tests.
The "why then how" sequence works not because CLT proves it works, but because it combines a reasonable theoretical framework with a well-replicated action mechanism.
---
Concrete Language and Truth Perception
One CLT-adjacent finding worth preserving: Hansen and Wanke (2010) showed across four experiments that concrete statements are judged as more true than abstract statements containing the same information. The effect was amplified when people were in a concrete mindset.
This aligns with broader fluency research: concrete, vivid language is easier to process, and processing ease increases perceived truth (connecting to the illusory truth effect from our previous post).
Practical implication that doesn't depend on CLT replicating: When you need people to believe what you're saying — product descriptions, testimonials, case studies — use concrete, specific language. "Revenue increased 23% in 6 weeks" beats "significant improvement in business outcomes."
---
The Honest Framework for Burned-Out Solopreneurs
Here is where the science and the human reality meet:
When you're running on empty, you don't have the cognitive bandwidth for abstract strategic visions. You need someone to tell you exactly what to do next. That's not weakness — it's predictable from how proximity affects processing. Your "inability to think strategically" when burned out isn't a character flaw. Your brain is in close-proximity mode, focusing on immediate concrete demands.
What helps:
- When you have energy: Connect with abstract purpose. Why are you building this? What does the end state look like? This isn't CLT magic — it's basic goal commitment.
- When you're depleted: Don't force strategic thinking. Switch to concrete if-then plans. "If I sit down at my desk, then I will record one 3-minute FAQ video." Implementation intentions work even when motivation is low — that's the whole point of automating behavior through if-then links.
- Don't mix levels within a single message. Whether this is CLT or just good communication, don't pitch an abstract vision in the middle of step-by-step instructions. Pick one level and stay there.
---
What This Means for Your Marketing
Be honest about uncertain science. CLT is a case study in why you should never build your marketing strategy on a single theory. The original evidence looked solid. The replication evidence says the effects may be near zero. Showing this kind of intellectual honesty builds more trust than pretending everything is settled.
Use the "why then how" sequence. Not because CLT proves it works, but because it combines reasonable framing with well-replicated action mechanisms (implementation intentions, d = 0.65).
Lead with concrete language when credibility matters. Concrete beats abstract for perceived truth. This finding has practical value regardless of CLT's replication status.
Match your communication to your audience's current state. This is good practice whether CLT's mechanism is real or not. Someone in crisis needs concrete next steps, not a vision board.
Watch the CLIMR results. When 78 labs across 27 countries report their findings, we'll have the best evidence yet on whether CLT's core predictions hold. Until then, treat the theory as a useful thinking tool, not a proven law.
---
Research basis: Soderberg et al., 2015 (Psychological Bulletin, 106 papers, d = 0.40 uncorrected); Maier et al., 2022 (RoBMA reanalysis, d = -0.002 bias-corrected); Schimmack, 2022 (z-curve, 14% expected discovery rate); Gollwitzer & Sheeran, 2006 (d = 0.65, 94 tests); Sheeran, Listrom & Gollwitzer, 2024 (642 tests); Hansen & Wanke, 2010 (concrete language and truth perception); CLIMR project (78 labs, 27 countries, results pending).