Why Corporate Training Fails to Deliver Results

Distinguishing the effects of training through concrete examples

Training in general has been shown to produce certain effects. At the same time, however, the scope of what counts as an “effect” has also been shown to be narrower than many people assume. The clearest way to understand this boundary is to contrast examples that are meaningful for the intended objective with examples that are not.

The following are examples in which training has been confirmed to produce reproducible improvements in actual work environments.

• Laparoscopic surgery simulation
• Increased call volume through call-center job training
• Prevention of accidents and occupational illness through workplace safety training
• Actual aircraft operation through flight-simulator training
• Reduced scrap rates through skills training in manufacturing
• Classification of lesions based on X-ray and CT images
• Cardiopulmonary resuscitation practice

The following are examples for which an effect was once proposed but was later rejected through replication or other verification.

• Improved intelligence through working-memory training¹
• Improved health and job performance through workplace wellness training²
• Achievement of work-related goals through coaching³
• Improvements in productivity, quality, and subordinate absenteeism through management training⁴
• Deterrent effects of sexual-harassment prevention training⁵
• Improved mathematical and general problem-solving ability through chess⁶
• Improved intelligence and academic performance through learning a musical instrument⁶
• Improved cognitive ability through brain-training and cognitive games⁷
• Improved cognitive ability through action video games⁸

Examples near the boundary are also useful for understanding what training can and cannot accomplish. Ethics training produces small to moderate improvements in knowledge and in ethical judgments about cases, but it has not been shown to reduce actual misconduct. Similarly, driving simulators improve vehicle-control skills, but their effects on road driving and accident prevention remain unclear.

Across training research, the hypothesis that task training in one domain improves task performance in another domain is known as “far transfer” and has become a central concern in the field. The examples listed above are findings from research on far transfer.

By contrast, near transfer—improvement on tasks that closely resemble the training task—occurs in many forms of training. The delivery format is not a major constraint: e-learning also works when the practical demands remain within the range of near transfer.

OJT (On-the-Job Training) must be divided into separate cases.
It is effective when limited to its original purpose of skills training, but it is not effective when expansively interpreted to mean all forms of work instruction that do not take the form of formal training.

The range in which “far transfer” can be regarded as attainable has narrowed

The intellectual foundations that enable people to perform work tasks have traditionally been studied under three conceptual categories.

Research first advanced in the area of knowledge and skills. As individual studies accumulated, the assumption that far transfer occurs was rejected. The resulting conclusion—that acquired knowledge and skills can be expressed only in limited ways when the task is posed in the same form or situation as in training—accords with ordinary intuition. Because this point may still be misunderstood, it can be restated more directly: it has already been established that “acquiring a vast amount of knowledge does not raise one’s ability in another domain.”

Cognitive ability was developed around expectations such as, “Repeatedly practicing difficult problems will strengthen the capacity for thinking itself.” Its definition therefore overlaps with far transfer from the outset. Its older roots extend back to the late nineteenth century, and research on methods for strengthening cognitive ability continued for many decades. Ultimately, meta-analyses systematically rejected far transfer from methods including working-memory training and brain-training programs in the list above. By around the 2019 paper, a consensus had formed that general cognitive ability cannot be strengthened through training⁹.

Research aimed at improving the general thinking ability of ordinary adults has largely disappeared, and the central topics of far-transfer research have contracted toward clinical applications and rehabilitation for older adults.

Implications for business practice

The preceding section summarizes the mainstream body of research examining whether training in general transfers to a broad range of everyday tasks. Expectations were certainly high, but the conclusion is that the anticipated effects of training did not remain.

A position that respects academic evidence should not casually assume that cognitive capacities such as “the ability to think” can be acquired through some form of training. The same applies to corporate training, and the opening list should be used to distinguish the range in which effectiveness can reasonably be expected.

Separate from research on cognitive ability itself, there was once a period when cognitive ability was treated as the core method of employee selection in business applications. During that period, paper-and-pencil selection methods became widespread in the form of academic and aptitude tests. This practice relied on research claiming that general cognitive ability was the strongest predictor of job performance.

A 2015 meta-analysis showed that a large part of this explanatory power was merely the result of statistical correction. General cognitive ability is therefore no longer regarded as an indicator that actively predicts job performance.

The original expectations for development and training were excessive

To repeat the point, research on corporate training continues to search for valid far-transfer effects such as those in the opening list¹⁰. Naturally, evaluations of effectiveness are also conducted with actual application in business practice in mind.

The authors nevertheless state that transfer is not “free”; they do not regard it as something that occurs unconditionally through training alone. Even research supportive of training no longer assumes that every participant will benefit. It has confirmed that training is ineffective for people who do not meet the following conditions. Because such participants introduce noise into effectiveness measurement, the position taken is that people who do not fit the program’s prerequisites should be excluded from it.

Cognitive ability is not itself a target of training. The most confusing point is that some training services offer programs that appear to strengthen cognitive ability under claims such as “developing the ability to think.” The continuing research framework in this field does not support this category of program.

Transfer is also understood to emerge only when training is followed by an active set of workplace supports such as the following.

As a result of integrating the accumulated evidence in this way, frameworks for corporate training have become heavily conditional.

Skills training readily satisfies workplace conditions, so its effects can be reproduced when measured in actual work. In more general forms of corporate training, however, it is difficult to isolate what produced the output.

A straightforward reading of the accumulated evidence therefore leads to the natural conclusion that effective training is largely limited to practical skills training, while other forms have no effect.
The original expectations placed on corporate training were too high. From the perspective of later generations, this period will probably be understood as a temporary episode of confusion followed by clarification.

Understanding the capacity to learn depends on Openness

The source of far transfer in work is individual personality. One component of the capacity to learn is the Big Five trait of Conscientiousness.

The other component, general cognitive ability, began from the position that it could be strengthened in the same way as knowledge and skills, but later abandoned that premise. This brings us closer to the interpretation that the concept is converging on the Big Five trait of Openness.

Openness is the most difficult of the Big Five traits to characterize and appears to capture differences in the breadth and ease of associative thought. Research is advancing on its correspondence with brain function, particularly the possibility that it arises from differences in the network structure of the default mode network, which functions independently of immediate responses to the external environment.

Overemphasis on development creates conditions for overstretch

Management research has identified a paradox: “training investment at the organizational level of analysis is viewed so positively, whereas individual training initiatives are viewed so skeptically”¹⁰. This reflects two misconceptions: that companies define the range of capabilities that can be developed through training too broadly, and that merely providing training is sufficient.

The greatest problem with this misunderstanding is that it does not guarantee feasibility in frontline execution. Many businesspeople currently assume that completing training will enable them to perform their work successfully, but research has shown that training does not itself provide the capacity to execute the work.

The preceding discussion also makes the appropriate direction clear: improving work feasibility should begin with assigning people to roles that fit them. Even academic and aptitude tests, which have now lost much of their explanatory power, were originally designed for selection and placement rather than development.

When organizations overemphasize development and underuse reassignment, they create more workplaces that demand excessive stretching of employees’ capabilities. Consistent with this structure, mental-health disorders have increased at a rate disproportionate to changes in the size of the labor force.

The basic tool for improving work feasibility from this point forward is Big Five personality analysis. By introducing Decider, companies can easily analyze employees’ personalities, including Openness, the central topic of this article. Reassignment was previously difficult to implement because there was no objective method for describing each person. Just as GPS has become indispensable for navigation, it is now reasonable to say that without a Big Five analysis tool, an organization cannot even reach the starting point for considering appropriate placement.