Home Home | About Us | Sitemap | Contact  
  • Info For
  • Professionals
  • Students
  • Educators
  • Media
  • Search
    Powered By Google

So What is a Good Transfer of Training Estimate? A Reply to Fitzpatrick

Alan M. Saks
York University

 In the October issue of TIP, Robert Fitzpatrick called attention to the fact that many studies on transfer of training cite an unscientific and dated estimate of training transfer. The estimate in question comes from an article published almost 20 years ago about the importance of management support for the transfer of training (Georgenson, 1982).

As noted by Fitzpatrick (2001), Georgenson (1982) was simply speculating when he said: How many times have you heard training directors sayI would estimate that only 10% of content which is presented in the classroom is reflected in behavioural change on the job (p.75). It is doubtful that Georgenson ever expected researchers to take this estimate seriously, let alone refer to it in the introduction section of their scientific papers. Moreover, although Georgensons estimate had nothing to do with the dollar cost or investments in training, most citations report that 10% of training investments result in transfer of training.

According to Fitzpatrick (2001), some might argue that reporting the 10% estimate is just introductory fluff and doesnt really matter while others will argue that it does matter. The fact of the matter is, however, that the 10% estimate was not based on scientific data and is dated. To continue to report it is not only bad science, but it is also misleading and most likely incorrect.

What then should a good transfer of training estimate look like? To begin with, it should incorporate the two main conditions of transfer of training: the generalization and maintenance of newly acquired knowledge and skills on the job (Baldwin & Ford, 1988). Thus, when one provides an estimate of the transfer of training, it should be specific as to the length of time following training. Second, given the increasing concern over the added value of training programs and return on training investments (Salas & Cannon-Bowers, 2001), a transfer estimate should also consider the percentage of training investments that result in transfer. And third, because training outcomes pertain to trainees (behavior) and organizations (results), estimates of the transfer of training should consider behavior and results criteria.

In order to obtain an estimate of the transfer of training along these lines, Monica Belcourt and I conducted a paper-and-pencil survey of 150 members of a training and development society. The participants reported working an average of 10.5 years in training and development and 10 years in their current organization. Their organizations employed an average of 5001,000 employees and comprise over a dozen sectors including manufacturing, service, and government.

We asked respondents several questions about the transfer of training in their organizations. To measure transfer generalization and maintenance, we asked them to indicate (on a scale from 0% to 100%) what percentage of employees in their organization effectively apply and make use of what they learn in training programs on the job immediately after, 6 months after, and 1 year after attending a training program. On average, respondents indicated that 62%, 44%, and 34% of employees apply what they learn in training on the job immediately after, 6 months after, and 1 year after attending training, respectively.

To obtain an estimate of the percentage of training investments that result in a change in employees and the organization, we asked what percentage of the organizations investments in training result in a positive change or improvement in employees (behavior and performance) and the organization (performance and effectiveness). Respondents indicated that an average of 51% of training investments result in a positive change in employees and an average of 47% result in a positive change in the organization.

While these estimates are based on the judgments of trainers and might be inflated due to a self-serving bias, they are nonetheless derived from recently collected survey data. They suggest that the transfer problem is not nearly as bad as the 10% estimate suggests. However, they do indicate that some 40% of trainees do not transfer immediately after training and this rises to close to 70% after 1 year. In addition, only about 50% of training investments result in an improvement in employees and the organization.

So perhaps it is time to close the book on the 10% estimate. Not to worry though, as I suspect that the estimates reported herein should provide researchers sufficient grist (or fluff) for the introductory section of their papers!


   Baldwin, T. T., & Ford, J. K. (1988). Transfer of training: A review and directions for future research. Personnel Psychology, 41, 63105.
   Fitzpatrick, R. (2001). The strange case of the transfer of training estimate. The Industrial-Organizational Psychologist, 39(2), 1819.
   Georgenson, D. L. (1982). The problem of transfer calls for partnership. Training and Development Journal, 36(10), 7578.
   Salas, E., & Cannon-Bowers, J. A. (2001). The science of training: A decade of progress. Annual Review of Psychology, 52, 471499.

January 2002 Table of Contents | TIP Home | SIOP Home