A Meta-Analysis of the Effects of Mental Contrasting With Implementation Intentions on Goal Attainment

The study investigates whether mental contrasting combined with implementation intentions (MCII) effectively improves goal attainment across diverse domains and which factors moderate its efficacy. Building on evidence that mental contrasting turns positive fantasies into binding goal commitments and that implementation intentions translate intentions into action via if-then plans, the authors hypothesize a small-to-medium overall effect of MCII on goal pursuit. They aim to extend prior health-focused meta-analyses to a broader range of goals and examine moderators including publication status, age, goal domain, intervention type, measurement type, and expectations of success, given that both component strategies show context-dependent effects.

Prior research shows that fantasizing only about positive outcomes can undermine effort and achievement, while mental contrasting (imagining desired future and obstacles) fosters goal commitment via nonconscious cognitive links, energization, and adaptive responses to feedback. Meta-analytic evidence indicates mental contrasting yields small-to-medium effects on health outcomes (e.g., adjusted Hedges' g ≈ 0.28 to 0.38). Implementation intentions (if-then planning) strengthen cue–behavior links, producing small-to-large effects on achievements, relationships, and health (Cohen's d ≈ 0.14–0.77 across outcomes). Experimental and field studies suggest combining mental contrasting with implementation intentions outperforms either alone in various contexts (academic, health, relationships, personal goals). However, autonomous MCII use can suffer from low-quality obstacle identification or ineffective plans, potentially attenuating effects. A prior meta-analysis limited to health behaviors reported a small-to-medium MCII effect (g ≈ 0.28), motivating a comprehensive meta-analysis across goal domains and potential moderators.

The authors conducted systematic searches across English databases (SCOPUS, PsycINFO, PsycARTICLES, PubMed, Web of Science, ScienceDirect, Springer Link, ProQuest Dissertations & Theses, Google Scholar) and Chinese databases (CNKI, Wanfang, VIP). Search terms included mental contrasting, implementation intention, WOOP, if-then plan, and plan. Inclusion criteria: empirical field interventions (laboratory-only experiments excluded), sufficient data to compute effect sizes, English or Chinese reporting, and designs with an MCII intervention vs. control (either no treatment or matched additional behavior change techniques). Twenty-one articles provided 24 independent studies. Coding captured authors, year, sample size, age group (children, college students, general adults, adult patients), publication status (published vs. theses/dissertations), goal domain (health, academic, relationship, personal), intervention type (experimenter-led face-to-face vs. document-based), dependent measure (self-report vs. objective), and expectations of success. Effect sizes were computed as standardized mean differences (Hedges’ g) between MCII and control. When studies reported multiple outcomes or repeated measures, effect sizes were averaged to maintain independence. Random-effects models were used, anticipating varying true effects across studies. Heterogeneity was assessed via Q and I². Publication bias was examined using funnel plots, Rosenthal’s fail-safe N, Egger’s regression, and trim-and-fill. Analyses were conducted with Comprehensive Meta-Analysis (CMA) 2.0. For moderator analyses, two large studies (Kizilcec and Cohen, 2017; n = 13,909 total) were excluded to prevent dominance in subgroups due to covarying moderators.

- Overall effect: MCII significantly improved goal attainment with a small-to-medium effect size, g = 0.336, 95% CI [0.229, 0.443]. Heterogeneity was moderate: Q(23) = 56.540, I² = 59.321%, p < 0.001. - Publication bias: Funnel plot showed asymmetry; Egger’s test t = 5.46, p < 0.01 suggested bias; fail-safe N = 482 (> critical 130) indicated robustness; trim-and-fill adjusted g = 0.242, 95% CI [0.143, 0.342], suggesting the true effect may be smaller. - Moderator analyses (excluding Kizilcec and Cohen, 2017): - Intervention type: Significant moderator. Experimenter interventions g = 0.465 (95% CI [0.349, 0.580]) vs. document interventions g = 0.277 (95% CI [0.154, 0.399]); Q(1) = 4.797, p = 0.029. - Publication status: No significant difference; published g = 0.404 vs. unpublished g = 0.271; Q(1) = 1.587, p = 0.208. - Sample age: Not significant; general adults g = 0.359; children g = 0.301; adult patients g = 0.521; college students g = 0.447; Q(3) = 2.201, p = 0.532. - Goal domain: Not significant; academic g = 0.255; health g = 0.379; personal g = 0.457; relationship g = 0.609; Q(3) = 3.075, p = 0.380. - Dependent measure: Not significant; objective g = 0.272 vs. self-report g = 0.403; Q(1) = 1.403, p = 0.236. - Expectation level: Not significant; high expectations g = 0.297 vs. low expectations g = 0.481; Q(1) = 1.466, p = 0.226. Limited reporting and possible ceiling effects noted. - Sample: 24 effect sizes from 21 studies; N = 15,907 total in main analysis.

Findings support that MCII is an effective, brief self-regulation strategy for improving goal attainment across diverse populations and domains, with an overall small-to-medium impact. The significant moderation by intervention type indicates face-to-face, experimenter-guided delivery enhances MCII’s effectiveness, likely by fostering interpersonal engagement, ensuring higher-quality identification of wishes, obstacles, and plans, and strengthening commitment. Document-based delivery, while less potent, still yields small-to-medium effects and offers scalable, cost-efficient implementation. Non-significant moderation by age, domain, measurement type, or expectations suggests MCII’s broad applicability, though limited study numbers and covarying moderators warrant cautious interpretation. Publication bias analyses imply some inflation in observed effects; adjusted estimates remain positive and significant, underscoring practical utility. MCII complements its component processes: mental contrasting provides commitment and obstacle identification, while implementation intentions operationalize action, together bridging intention–behavior gaps.

This meta-analysis demonstrates that MCII reliably promotes goal attainment with a small-to-medium overall effect size (g = 0.336). Intervention type moderates efficacy, favoring experimenter-led formats over document-based delivery. The study broadens evidence beyond health behaviors and highlights MCII’s general applicability. Future research should examine intervention duration and follow-up timing, reinforce and increase usage frequency, and evaluate cultural moderators by aligning individualist versus collectivist wishes and obstacles. Optimization strategies include guided, face-to-face facilitation to enhance plan quality, reminders to sustain use, and flexible if-then planning that adapts to changing contexts.

The meta-analysis included a relatively small number of studies, limiting power to detect moderators and generalize results. Moderator analyses excluded two large-scale studies (Kizilcec and Cohen, 2017) to avoid dominance, which may affect reliability of subgroup estimates. Expectations of success were sparsely reported and exhibited ceiling effects, constraining moderation tests. Covariation among moderators across studies may bias single-moderator analyses. Evidence of publication bias suggests the true effect may be smaller than the unadjusted estimate.