Examining and testing the roles of human resource development (HRD) in the public sector: evidence from US federal agencies

Abstract

Human resources (HR) are a priority in organizations, and managing the HR force effectively is the best way for them to survive and flourish. As a critical and integral component for achieving public agencies’ missions and objectives, a well-developed and effective public human resource development (HRD) program needs to be in place. The following research questions guided this study: (1) What are the role of HRD systems (i.e. off-the-job training and on-the-job training), and HRD effectiveness in public organizations? (2) Do HRD systems and HRD effectiveness have a positive impact on employee attitudes and behaviors? (3) What are the mediating roles of mission congruence on trust in leaders and satisfaction with the organization? This study uses data from the Merit Systems Protection Board to identify the nature of HRD and the antecedents of HRD effectiveness in the public sector, and to determine the resulting organizational effectiveness of HRD. This research finds that some HRD effectiveness and mission congruence play major roles in meaningfully and significantly boosting employee trust in leaders and organizational satisfaction. Finally, a discussion on the main findings, research, and practical implications for public management theory and practice is provided.

Keywords:

• Human resource development (HRD)
• on-the-job training (OJT)
• off-the-job training (Off-JT)
• HRD effectiveness
• mission congruence
• social learning theory (SLT)
• social exchange theory (SET)
• mediating effect

Notes

1. The details on these four processes are as follows:

(1) Attention. To increase learning effectiveness, people must first absorb the crucial components of the model’s behavior. Attention comprises interactions among people and is affected by other factors, such as ‘the model’s status or virtue, person’s sensory capacities, and characteristic of the model’s activities’ (Gibson, 2004).

(2) Retention. To benefit from observing information, people must have long-term retention of the model’s behaviors and activities. If people want to represent a model’s behavior in the long term, they have to memorize the model’s activities as a symbolic form, either imaginably or verbally (Bandura, 1971). Imaginative symbols embody pictures or images of past experiences, while verbal symbols represent intricate behaviors as words.

(3) Behavior production. A behavior process is concerned with obvious actions guided by symbolic representations, which lead investigative learning to performance. To accomplish behavior production, people have to integrate responses provided by attention and retention processes. Then they can reveal their obtained component skills behaviorally. In addition, if a person has these skills, he or she can easily produce new types of behaviors. However, if response elements suffer from deficiencies, behavior production will be impaired. Thus, individuals steadily adapt their behavior based on self-observation and self-correction until an adequate integration with the model is achieved.

(4) Reinforcement and motivation. The reinforcement process not only controls the overt action of corresponding behaviors, but also can affect individuals’ level of attention to learning by managing what they participate in and how they practice and actively display what they have seen. Then the motivation process is influenced by observational learning, in which people are more capable of assuming the modeled activities if those behaviors are considered good performance. In addition, Bandura (1971) argued that when positive incentives are provided, people more easily translate learning into behavior.

2. As we used data from a single source, we tried to examine for common method bias (CMB), and we show this to be of minor or insignificant concern only. To test the common method variance, we employed an analysis exploring the common latent factors (CLF), as posited in Harman’s single factor test. The test included all items from all the constructs in the study put into a factor analysis to determine whether the majority of the variance could be accounted for by one general factor. From the CLF test, we found that the common method variance factor was .24 below the suggested threshold of .50. Hence, we are convinced that CMB is not a major methodological issue in this study (Fornell & Larcker, 1981).