Can Soft Skills Change?
The evidence does not force a choice between fixed personality and limitless self-improvement. Social and emotional characteristics can change, but change is uneven across domains, life stages, interventions, and outcomes. This chapter separates a change in a broad measured trait from a change in a behavior or strategy, then uses the training evidence to identify what a realistic development decision should specify: the target, the setting, the instructor, the practice, and the feedback.
Malleability is measurable, but not unlimited
The World Bank training synthesis finds an average improvement of about 0.46 standard deviations in socio-emotional outcomes across the included interventions. That is evidence that social and emotional capabilities are not completely fixed. It is not a guarantee that any person can quickly install a desired personality, nor does it mean that every intervention changes every domain by the same amount.
The pooled effects differ by the domain being targeted. Interventions show the largest improvements for openness, about 0.65–0.67 standard deviations, and conscientiousness, about 0.40. Effects are moderate for emotional stability, about 0.37, and extraversion, about 0.38. The effect for lower agreeableness is smaller, about 0.16. These are averages from a heterogeneous training literature, not a ranking of which traits are most valuable as people. They suggest that attitudes, flexibility, self-regulation, or goal-directed behavior may respond differently to structured practice than relational or emotional patterns that are more embedded in a person’s social environment.
The distinction between a trait and a target matters. A program may help someone plan tasks, regulate a response under pressure, or practice constructive disagreement without producing a large, stable shift in a broad personality score. Conversely, a score may move without the person transferring the new behavior to work. The most useful development target is often a behavior or strategy connected to a domain and a task: preparing for a difficult conversation, making a plan and reviewing it, asking for feedback, pausing before responding under stress, or experimenting with a more flexible approach to a changing problem. Those targets are observable and revisable. “Become more conscientious” is not.
Life stage changes the average effect without making later change impossible. The synthesis reports larger pooled effects for interventions aimed at children in primary education or earlier, about 0.55 standard deviations, and adolescents in secondary education, about 0.58. Post-secondary effects are somewhat lower but remain meaningful at about 0.41. Early development may offer more time, repetition, and environmental reinforcement. Later learners may have more established routines and competing demands, but they can also connect practice directly to a job transition, leadership role, or other consequential task.
The correct conclusion is neither “personality is destiny” nor “anyone can become anything with the right course.” It is that malleability has a distribution. It depends on the target, the starting point, the duration and intensity of practice, the surrounding environment, and the measure used to judge change. A small average effect can be valuable for a difficult population or a specific behavior, while a larger score change may have little practical value if it does not transfer to the outcome that matters.
The phrase “0.46 standard deviations” also needs interpretation. It is a pooled difference in measured outcomes between intervention and comparison conditions in the included evidence.