The influence of learning activity on low-skilled workers’ skill improvement in the South Korean manufacturing industry

Abstract

The main purpose of this study was to explore how low-skilled worker’s learning activity influences skill improvement. Using a unique 2007 Human Capital Corporate Panel data-set from the South Korean manufacturing industry, we operationalize skill improvement over time among low-skilled workers. A worker is classified as ‘low skilled’ if he or she has a low education level and poor technical skills. Regression models show that low-skilled workers’ informal learning positively influences their skill improvement. In contrast, we note that supervisors negatively influence skill improvement of low-skilled workers when measuring the change in technical skill proficiency. Quality circle programmes also have a positive influence on skill improvement. In conclusion, skills can be improved through planned interventions that increase collaboration on the job. The results from this study help to highlight the importance of designing learning interventions for low-skilled workers that take account of their underlying education and skills.

Keywords:

• low-skilled worker
• learning
• skill improvement

Introduction

Dramatic changes in scientific production have led to significant changes in technological skill for workers. Many researchers have shown that the spread of computers has increased the need for high-skilled workers by employers and industries (Krueger 1993; Howell and Wieler 1998). Increasing technical requirements for high-skilled workers have many benefits for those who can accommodate changing skill demands, but changing technology raises problems for workers who are not highly skilled. Low-skilled workers are defined as individuals who perform work that is either routine or manual, or are in jobs that require no more than a high school education or no more than one year of work experience (Maxwell 2006). In the manufacturing environment, this definition generally covers entry-level employees with middle or high school education or workers who have remained in front-line jobs and not advanced to more skilled assignments as supervisors or work leaders (Osterman 2001). As Osterman (2001) notes, estimates were that low-skilled manufacturing jobs were about 10% of all jobs in the previous decade in the USA. Globally, low-skilled jobs are almost 45% of all jobs according to the International Labour Organization (ILO 2015).

Researchers have drawn attention to low-skilled workers by noting that earnings growth has been concentrated in high-income or high-skilled workers (Autor and Dorn 2013). Low-skilled workers show moderate wage gains, with middle-skilled workers experiencing the sharpest declines in earnings. Therefore, skill improvement is a precondition for low-skilled workers’ wage improvement. To improve the low-skilled workers’ wages, many researchers have shown that training can help individuals increase skills and earnings (Prince 2008; Washington State Board for Community and Technical College 2005). Low-skilled workers’ learning activities significantly influence their skill improvement (Duleep and Dowhan 2002; Powers and Seltzer 1998).

Along with the ongoing globalization of labour market, low-skilled workers are likely to show lower wage gains because they continue to come into the labour market. This is especially true in the manufacturing industry for South Korea. The declines in the low-skilled workers’ earnings in South Korea lead to increasing focus on how training and development can be used to improve skills for the low-skilled workers as a way to improve long-run earnings gains (Bae et al. 2013). Empirical studies show that learning activities for workers are associated with their improvements in job skills and thus increases their wage gain owing to higher productivity in Korea (Kim and Kim 2008).

In recent times, low-skilled workers have been more likely to have carried out all forms of training in the workplace (Bottrup and Clematide 2005; Kahn and Lim 1998). The South Korean government implemented a job skill development programme in which the government provides support for low-skilled workers’ learning activities and their skill improvement. However, many low-skilled workers have difficulty benefiting from such support because they have to take time away from work (Kim 2007a,b; KRIVET 2013).

Nevertheless, Kim and Kim (2008) showed that many Korean workers have participated in various learning activities officially or unofficially provided in the workplace, but low-skilled workers’ learning activities in Korea manufacturing industries focused on eight learning activities mainly such as informal learning by supervisors, informal learning by co-workers, self-learning through work, formal on-the-job training (OJT) programme, Task Force Team (TFT), Quality Circle (QC), Knowledge Mileage System, and Six-Sigma. The main eight learning activities are incorporated into the survey instrument for the Korean Human Capital Corporate Panel (HCCP) (KRIVET 2013). More generally, while a number of studies (Ahlstrand, Bassi, and McMurrer 2001; Bosworth 2007; Kazis and Liebowitz 2003; Liebowitz and Taylor 2004; Prince and Jenkins 2005; Wachen, Jenkins, and Van Noy 2010; Women Employed Institute 2005; Zeidenberg, Cho, and Jenkins 2010) have described the positive benefits of low-skilled workers’ learning participation, few studies focus on what kinds of learning programmes improve the application of skills on the job. Some researchers have shown that the input of the learning components influences the output in terms of individual or corporate productivity (Frazis and Loewenstein 1999; Loewenstein and Spletzer 1997; Kim 2002). Other researchers have focused on low-skilled students in college (Kazis and Liebowitz 2003; Liebowitz and Taylor 2004; Prince and Jenkins 2005), policy-making for low-skilled workers’ learning (Ahlstrand, Bassi, and McMurrer 2001; Bosworth 2007), and learning activities for low-skilled workers on governmental support (Wachen, Jenkins, and Van Noy 2010; Zeidenberg, Cho, and Jenkins 2010). In short, little empirical work has been done on what kinds of learning activities officially or unofficially provided in the workplace have an impact on improving low-skilled workers’ skills.

It is necessary to specify what kinds of learning activities in the workplace are effective at improving low-skilled workers’ skills. This research focuses on low-skilled workers employed in the manufacturing industry and excludes workers in the finance or service industries. Focusing on workers in manufacturing allows us to better measure the extent of skill improvement, where learning activities are focused on technical training. That is, this study focuses on the manufacturing industry because of the emphasis on production in manufacturing as opposed to finance or service industries. This focus on making goods in manufacturing distinguishes the sample of workers we study from those in service industries. The purpose of this study is to investigate the following: How learning activities provided for low-skilled workers are associated with improvements in job skills.

Literature review

Low-skilled worker defined

As we note in the Introduction, the standard definition of the low-skilled worker is an individual who performs work that is either routine or manual, or in a job that requires no more than a high school education or no more than one year of work experience (Maxwell 2006). This standard is operationalized in different ways by government. The International Labour Organization (ILO) provides an annual estimate of the number and percentage of low-skilled workers. The ILO definition is based on an analysis and labelling of ‘low-skilled’ jobs as occupations that are non-routine manual, non-routine cognitive and routine occupations. These are opposed to jobs that are high skilled. Assignment is based on a methodology from Acemoglu and Autor (2011) and Jaimovich and Siu (2012). Using these techniques, researchers have been able to track the occupations that are low skilled, determining that up to 45% of all jobs in the world are low skilled (ILO 2015). This strategy allows us to document the continued existence of low-skilled work, and the correlation between high skill and high wage (and general economic growth). However, the empirical strategy used to classify based on national level data on employment is not easily used in companies or by human resource development (HRD) scholars. National level data provide a good framework for understanding the scope of the problem, such as the fraction of all workers that are low skilled. However, individual nations or companies need data on individuals from panel surveys such as the HCCP to understand the employment characteristics of low-skilled workers, as opposed to low-skilled occupations or jobs.

According to Falkinger and Grossmann (1999), using an economic perspective, the concept of low-skilled workers is defined by a comparison between high- and low-skilled workers. In other words, low-skilled workers are conceptualized as those who hold low-level skills compared with high-skilled workers in the labour market. This perspective only stresses the relative knowledge and skill level, but pays little attention to any other components except skill level. In recent years, the discussion of low-skilled workers has focused on ‘those who are in a vulnerable situation in relation to the competence demands of modern society and economy’ (Illeris 2006, 16). However, by taking education into account, we can be precise about what makes workers low skilled.

From a traditional viewpoint, low-skilled workers are early school leavers who have not completed any formal education or school qualification (Illeris 2006). Low-skilled workers can include those without formal training or education beyond high school (Hale 2004) as well as ‘no more than a high school education and no more than one year of work experience’ (Maxwell 2006, 1). The fact that low-skilled workers participated in a formal education system for less time dictates that they do not have the basic knowledge and skills of society, and continue their low-skilled status (Osterman 2007). Low-skilled workers without high school diplomas or credentials are likely to face more challenges in the workforce because of a mismatch between their skills and those required in open position descriptions (Capps, Fortuny, and Fix 2007). The skills gap often is reflected in workers’ inability to get jobs (Torraco 2007).

Low-skilled workers can remain low skilled because what they know about work does not match what is needed (Illeris 2006). They need to respond to changes in skills requirements for specific jobs or they face lay-offs (Falkinger and Grossmann 2001). They also have difficulties accepting new and innovative technologies compared with high-skilled workers (Booth and Snower 1996).

Consequently, the definition of low-skilled workers needs to include those who hold little knowledge and skills required in the workplace. Simultaneously, considering the economic perspective on the basis of low-skilled workers’ relative level of skills, low-skilled workers should be defined as early school leavers or high school graduates with a low level of workplace skills.

Learning activities

Learning participation and skill improvement in human capital

Human capital is a key concept of individual, organizational, and national productivity. While human capital emerged in Europe after the Second World War as a way to explain growth, it expanded in the 1960s to non-Western nations as a concept leading to greater use of tools such as education and training to improve national level human capital (Arndt 1988). Therefore, human capital is one of the production elements required to add value to national economic activities (Schultz 1961). Researchers have discussed human capital as education, skills, abilities, and knowledge internalized within the human mind (Garavan et al. 2001; Youndt, Subramaniam, and Snell 2004). Consequently, human capital is a stock of measurable and changeable production elements internalized to create added value because it is continuously updated.

An example helps us understand the depth of low skills as a reflection of both low education and low skills. Studies of the shipbuilding industry show the changing role of ship inspectors, who often did very basic tasks that may be defined as low skill including detecting open-to-surface discontinuities of weld zone in a ship by striking with hammer. In the process of detecting the discontinuities of ships, Korean shipbuilding companies in the 1970s introduced new procedures such as radiographic nondestructive testing (Lee 2010). As a result of these shifts in specific skills, Korean shipbuilding firms eliminated jobs mostly among low-skilled workers, as firms substituted technology for low-skilled workers. However, many new jobs were added and these paid much more on average.

From an economic perspective of human capital, an individual’s learning participation is an important way to measure the extent to which the individual develops his or her knowledge and skills (Kim et al. 2014). On the basis of learning knowledge and skills required to improve the application of skills to their job, specifically, the worker may experience an increase in earnings because the worker learns how to do his or her job better. Many researchers have shown that workers’ high level of learning participation increases their ability to do their job better because the workers increase their knowledge and skills to more efficiently or effectively (Vinokur et al. 2000).

Components of learning activities

According to the European Commission (2006, 9), learning activities are defined as ‘any activities of an individual organized with the intention to improve his/her knowledge, skills, and competence’. Learning activities are distinguished from non-learning activities by intention and organization (Gnahs et al. 2002). The activities are purposeful learning and should be organized by the learners so that they are involved in the transfer of information (Merriam, Cafarella, and Baumgartner 2007).

The learning activities of adults are described as either informal or formal institutionalized learning (Livingstone 2001; Merriam, Cafarella, and Baumgartner 2007). Informal learning encompasses daily life activities closely linked to work, family, community, and any other life-related activities (Fenwick, Nesbit, and Spencer 2006; Spencer 2006). For instance, workers are able to improve their knowledge and skills by working with their supervisors or co-workers. Formal institutionalized learning is structured, systematic, typically takes place in educational institutions, and usually provided with a curriculum (Fenwick, Nesbit, and Spencer 2006). For instance, workers can improve their knowledge, skills, and competency using a curriculum in educational settings provided by their employer or organization. In addition to a regular curriculum, the employer or organization can provide the workers with any structured programme closely linked to their learning (KRIVET 2013).

By focusing on detailed learning activities, we see that low-skilled workers’ informal learning activities are closely associated with skill improvement (Kahn and Lim 1998). According to Livingstone (2001, 2), informal learning is ‘undertaken on one’s own, either individually or collectively, without either externally imposed criteria or the presence of an institutionally authorized instructor’. Informal learning can be categorized as self-directed learning, incidental learning, and socialization based on intentionality and consciousness (Schugurensky 2000). By focusing on learning activity in the workplace, low-skilled workers will individually or collectively implement informal learning. However, status within the firm can potentially influence participation in informal learning. In comparison to formal learning, informal learning experiences resulted in the acquisition of more tacit-related knowledge and skills such as language acquisition, cultural norms, and manners (Occupational Information Network 2013). Specifically, KRIVET (2013) categorized informal learning in the workplace into informal learning by supervisors or co-workers and separated out learning that individuals engage in on their own.

Informal learning by supervisors means the process through which low-skilled workers acquire more tacit-related knowledge and skills by communicating with their supervisors in their organizations. Informal learning by co-workers focuses on the process of communicating with low-skilled workers’ co-workers to acquire more tacit-related knowledge and skills. Self-learning through work means low-skilled workers are acquiring tacit-related knowledge and skills through their work in the workplace. In other words, low-skilled workers can acquire more tacit-related knowledge and skills through informal learning activities, which includes either communicating with their supervisors and co-workers or learning through their work.

Along with the informal learning on low-skilled workers’ skill improvement, the components of institutionalized learning in the workplace are closely associated with skill improvement (Bottrup and Clematide 2005). Berg and Chyung (2008) described the similarity between formal and institutionalized learning when they are sponsored by a firm. Low-skilled workers acquire more explicit-related knowledge and skills by participating in institutionalized learning activities (Merriam, Cafarella, and Baumgartner 2007). Low-skilled workers conceptualize their own practice by creating a distance between their daily work situation and their own practice in new methods as well (Bottrup and Clematide 2005). Therefore, low-skilled workers can acquire more explicit-related knowledge and skills to improve their knowledge and skills.

According to KRIVET (2013), institutionalized learning can be categorized as formal OJT programmes, TFT, QC, Knowledge Mileage System, and Six-Sigma. A formal OJT programme is a form of individualized learning activity conducted at a low-skilled workers’ job (Cho 2005). Low-skilled workers learn their job-related knowledge and skills through officially designated employees in the organization. Low-skilled workers participate more often in OJT programmes (Rothwell and Kazanas 1990). Second, TFT is a temporary activity in which individuals selected in the organization solve an important project (KRIVET 2013). Low-skilled workers can be members of a TFT and can learn the knowledge and skills required in the project. Third, QC is a small group of workers in an organization who carry out similar jobs, and work to identify and solve quality-related problems (Majumdar and Manohar 2011). Low-skilled workers as members of QC can learn the knowledge and skills of how to conduct quality control. Fourth, a Knowledge Mileage System is a system in which low-skilled workers receive compensation and evaluation when accumulating new knowledge and skills (KRIVET 2013). Finally, Six-Sigma is an approach that recognizes and fixes shortcomings, mistakes, and failures in the work process (Snee 2004).

Among the other institutionalized learning activities, QC has unique characteristics. QC is defined as a group of workers who implement the same or similar work to hold meetings regularly for the purpose of identifying, analysing, and solving work-related problems (Hutchins 1985). As the core purpose of QC, workers organize small groups and implement their activities to find a solution regarding work-related problems. All activities may usually be led by a supervisor or manager on the basis of the workers’ autonomy (Chung and Kim 2011). Collaboration within a group of workers is one of the essential factors to identify, analyse, and solve work-related problems more successfully as well (Park 2002). Under autonomous circumstances, the workers can understand what problem is in their mind, internalize why the problem should be solved sincerely, and show how the problem should be solved positively (Montana and Charnov 2008). The workers can present better outcomes on the collaboration with other workers in comparison with other solutions mixed by various opinions on the independence among workers as well (Park 2002). Consequently, QC is a learning activity closely linked to workers’ real job settings.

A review of the literature leads to the conclusion that in order to study the efforts to improve skills for low-skilled workers, the models need to consider both learning activities officially or unofficially provided for low-skilled workers in a firm. According to KRIVET (2013), Korean worker’s learning participation can be divided into informal learning by supervisors, informal learning by co-workers, self-learning through work, a formal OJT programme, TFT, QC, Knowledge Mileage System, and Six-Sigma. That is, most Korean companies focus on the learning activities mentioned above for workers in the workplace. On the basis of this issue from the perspective of low-skilled workers’ learning elements, a hypothesis associated with low-skilled workers’ learning activity to improve their skill is as follows:

Hypothesis:

Learning activities provided for low-skilled workers are associated with improvements in job skills (Figure 1).

Figure 1. Research design.

Method

Data

The HCCP is a nationally representative sample of industry, companies, and workers in Korea. The HCCP has biennially obtained data every two years since 2005, and the data from 2007 include low-skilled workers and their learning activities. This research is based on HCCP data collected in 2007, because variables related to low-skilled workers’ learning activities were included in this year. The Korea Research Institute for Vocational Education and Training (KRIVET) started an investigation of the educational experience of Korean adolescents and college students as research subjects in 2005 (KRIVET 2013). The original purpose of the investigation was to understand the HR possessed by Korean companies, the endeavours of companies to develop their HR, the situation of HRD within the companies, and its impact on the companies’ performance (KRIVET 2013). The HCCP sample, that is, employers and employee, was selected by a three-stage stratified sampling of the industry, scale, and company, and most questions were closely related to understand the situation of human resource management (HRM) and HRD in the companies.

The samples of this study are 1319 low-skilled workers who responded to all questions related to dependent and independent variables. The original sample number was 1447, but 128 low-skilled workers did not respond to all questions related to learning programmes provided for the low-skilled workers in a firm, and these were excluded from the study. The HCCP uses core technical ability – namely, the proficiency of a worker to perform his or her job in the workplace as a low-skills deficiency in the core-technical aspects of job skills (KRIVET 2013; Kim and Kim 2008). The low-skilled workers of this study are those who have deficiencies in the core-technical aspects of their job skills.

This research focuses on low-skilled workers employed in the manufacturing industry and excludes workers in the finance or service industries. Focusing on workers in manufacturing allows us to better measure the extent of skill improvement, where learning activities are focused on technical training. In order to control for the impact of job mobility on skill improvement, the samples are limited to those individual workers who have remained in the same job in a specific firm. All low-skilled workers, furthermore, are individuals with the following educational credentials: apprenticeship participants, high school graduates, vocational high school graduates, and other technical high school graduates. We excluded college-educated workers.

Table 1 provides descriptive data on the samples. In Table 1, the ratio between low-skilled male and female workers is approximately 7:3. Most low-skilled workers are in their twenties to forties. This age distribution is closely linked to the fact that most workers are employed in the manufacturing industry and are likely to retire before they turn 60. The periods of previous schooling of low-skilled workers is nearly equally distributed between three high school or technical school groups, with a small fraction only finishing middle school or less. Finally, 216 firms are represented in the data and most employ less than 1000 workers.

Table 1. Demographic distribution of low-skilled workers.

Section	Frequency	Ratio
Gender
Male	969	67.0
Female	478	33.0
Age (years)
Below 19	21	1.5
20–29	456	31.5
30–39	514	35.5
40–49	362	25.0
50–59	92	6.4
Above 60	2	0.1
Periods of previous schooling
Below middle school	133	9.2
Graduation of high school	415	28.7
Graduation of vocational high school	409	28.3
Graduation of technical high school	490	33.9
Firm size
100–299	503	34.8
300–999	675	46.6
1000–1999	106	7.3
Above 2000	163	11.3

Variables

Dependent variable

In the HCCP questionnaire, a dependent variable was originally developed to measure the skill level at the outset of working for the company as well as the current skill level, both of which are categorized into beginner (= 1), apprenticeship (= 2), and skilled worker 1 (= 3), skilled worker 2 (= 4), skilled worker 3 (= 5), skilled worker 4 (= 6), and skilled worker 5 (= 7). According to the HCCP questionnaire, the skill is the core-technical ability that the worker has to perform his or her job in the workplace, and the skill level serves to rank in ascending order how many skill holders have the core-technical skills (KRIVET 2013). For instance, a skilled worker of level 5 has more core-technical skills in his or her job and can perform the job more proficiently than a skilled worker of level 4. While other countries such as the USA categorize skills into basic, complex problem-solving, and resource management skills (Occupational Information Network 2013), the HCCP defines skill level in reference to the research outputs (Kim and Kim 2008; Park and Kim 2006). Therefore, the terminology (e.g., skilled worker 1–5) was defined by the number of core-technical skills that each worker has completed.

The degree of workers’ skill improvement was defined by the difference between the core technical skill level at the beginning of working for this company and the current core technical skill level. For instance, the degree of a worker’s skill improvement is two (= 4 minus 2), if their core technical skill level at the beginning of working for the company is ‘apprenticeship (= 2)’ and their current core technical skill level is ‘skilled worker 2 (= 4)’. A low-skilled worker is defined as an individual classified as either a beginner or an apprentice and with schooling levels below high school when the worker started working for the company. Therefore, low-skilled workers’ skill improvement (SKILL) is based on the difference between the skill level when starting work for this company and the current skill level of the worker, and is limited to those workers whose periods of previous schooling were below high school and had a skill level of either beginner or apprentice (Figure 2).

Figure 2. Hierarchical nature of classification system from a skilled worker 1 to 5.

Independent variable

Independent variables were largely divided into two sections: firm scale and learning participation. In Table 2, the firm scale means how many workers employed in the respondent’s firm, and the learning participation factor includes eight variables (SUPER, COWORK, SELFLEARN, OJT, TFT, QC, KNOW, and SIXSIGMA). The learning participation elements of this study were based on self-reported data and the respondents’ perception about whether the respondents of this study have participated in one of the eight learning activities mentioned above, and are thus focused on only participation or non-participation in the learning activities. For instance, if a worker has experience participating in ‘self-learning through work’ more than once, the worker’s SELFLEARN value is 1; however, if the worker has no such experience, the value is zero. FIRM is an ordinal variable representing how many workers are employed in the respondent’s firm (100–299 = 1, 300–999 = 2, 1000–1999 = 3, and above 2000 = 4). For instance, if the number of workers employed in the respondent’s firm is more than 100 but below 299; the value of this variable is 1. If the number is more than 1000 but below 1999, the value of this variable is 3.

Table 2. Definition of variables.

Variable	Definition
Dependent
SKILL	Degree of workers’ skill improvement (continuous)
Independent
FIRM	How many workers employed in the respondent’s firm?
	(100–299 = 1, 300–999 = 2, 1000–1999 = 3, above 2000 = 4)
Learning
SUPER	Informal learning by supervisor (participation = 1, otherwise = 0)
COWORK	Informal learning by co-workers (participation = 1, otherwise = 0)
SELFLEARN	Self-learning through work (participation = 1, otherwise = 0)
OJT	Formal OJT programme (participation = 1, otherwise = 0)
TFT	Task Force Team (participation = 1, otherwise = 0)
QC	Quality Circle (participation = 1, otherwise = 0)
KNOW	Knowledge Mileage System (participation = 1, otherwise = 0)
SIXSIGMA	Six-Sigma (participation = 1, otherwise = 0)

Validity and reliability

As can be seen in Table 3, each result of the factor and reliability analysis of the independent variables is presented with the Component, Eigenvalue, and Cronbach’s alpha. A factor analysis is performed using a principal component analysis with a varimax rotation, and the standard of factor loading is based on .5.

Table 3. Validity and reliability for variables.

Variable	Total	Component
		1	2
SUPER		.078	.820
COWORK		.099	.840
SELFLEARN		.126	.751
OJT		.648	.255
TFT		.702	.075
QC		.582	.212
KNOW		.631	−.082
SIXSIGMA		.649	.064
Eigenvalue		2.658	1.512
R^2	52.118	33.221	18.897
Cronbach’s alpha	.706	.751	.637

Using a standard of factor loading above .5, the analysis showed that low-skilled workers’ learning activities can be divided into two factors, informal learning (SUPER, COWORK, and SELFLEARN) and formal learning (OJT, TFT, QC, KNOW, and SIXSIGMA). The coefficient of Cronbach’s alpha was acceptable (Cronbach’s alpha = .706). Therefore, the analysis showed that each component can be categorized into informal learning activities and institutionalized learning within the framework of the organization.

Data analysis

The data analysis used three procedures. First, a zero-order correlation analysis was applied to examine the relationship between independent and dependent variables. Second, a comparison was made among low-skilled workers’ learning activities (e.g., no learning at all, informal supervisor learning, and other forms of learning) on their skill improvement to describe the effect of learning as a whole. Finally, a multiple linear regression analysis (MLRA) with ordinary least squares (OLS) was applied to explore how low-skilled workers’ learning activities influenced their skill improvement, considering the scale of the firm where the worker is employed.

Results

Zero-order correlation

Zero-order correlations among the variables are presented in Table 4. Informal learning by supervisors was significantly related to low-skilled workers’ skill improvement (r = .056, p < .05), and informal learning by co-workers was significantly related to low-skilled workers’ skill improvement (r = .229, p < .01). Self-learning through work was significantly related to low-skilled workers’ skill improvement (r = .159, p < .01), as was a formal OJT programme (r = .084, p < .01). Finally, TFT, QC, Knowledge Mileage System, and Six-Sigma were significantly related to low-skilled workers’ skill improvement (r = .089, r = .145, r = .073, r = .119; p < .01, respectively). Zero-order correlation measures the bivariate relationship between the independent and dependent variables (Nathans, Oswald, and Nimon 2012).

Table 4. Zero-order correlation matrix for variables.

	SKILL	SUPER	COWORK	SELFLEARN	OJT	TFT	QC	KNOW	SIXSIGMA
SKILL	1
SUPER	.056*	1
COWORK	.229**	.579**	1
SELFLEARN	.159**	.431**	.491**	1
OJT	.084**	.229**	.214**	.215**	1
TFT	.089**	.117**	.144**	.156**	.398**	1
QC	.145**	.175**	.185**	.185**	.397**	.243**	1
KNOW	.073**	.042*	.064*	.076**	.197**	.289**	.174**	1
SIXSIGMA	.119**	.129**	.147**	.128**	.246**	.296**	.263**	.304**	1

Note: *p < .05, **p < .01.

Comparison among low-skilled workers’ learning activities on their skill improvement

Table 5 includes the difference between the skill level when a low-skilled worker started work for the company and the current skill level of the worker, with the number of observations that received no learning at all, informal supervisor learning only, and other forms of learning. The number of those who received ‘no learning at all’ declined over 2%, indicating that low-skilled workers’ skill improvements may increase to a certain extent without going through any learning programme (N_{B-1, A>0} = 117). This means that other factors, with the exception of learning activity officially and unofficially provided for the low-skilled workers in a company, can be associated with the workers’ skill improvement. Next, the number who received informal supervisor learning (N_B-2 = 19) indicated that informal supervisor learning needs support to improve low-skilled workers’ skills. Therefore, the two findings show that informal supervisor learning might not be associated with low-skilled workers’ skill improvement.

Table 5. The comparison among low-skilled workers’ learning activities on their skill improvement.

A	B-1	B-2	B-3
−1	1	0	2
0	63	3	150
1	22	0	98
2	27	2	256
3	31	3	245
4	22	4	216
5	13	4	118
6	2	3	35
Total	181	19	1120

Note: A, the difference between the skill level when a low-skilled worker started work for his or her company and current skill level of the worker; B-1, the number of observations that received no learning at all; B-2, the number of observations that received informal supervisor learning only; B-3, the number of of observations that received the other forms of learning.

The effect of learning activities on low-skilled workers’ skill improvement

The next regression analysis showed how all independent variables are associated with low-skilled workers’ skill improvement. Table 6 includes coefficients, t-statistics, beta weights, and tolerance and variance inflation factor (VIF). The results of this MLRA with OLS showed that all independent variables explained 8.0% of the dependent variable (R² = .087, adjusted R² = .080).

Table 6. Influence of learning programme on low-skilled worker’s skill improvement.

	B	t	β	Tolerance	VIF
Constant	1.755	13.159
FIRM	.113*	2.291	.063	.911	1.098
SUPER	−.491**	−4.150	−.139	.625	1.599
COWORK	.880**	7.138	.246	.586	1.705
SELFLEARN	.281*	2.342	.073	.713	1.402
OJT	−.102	−.795	−.025	.705	1.418
TFT	−.093	.459	.014	.764	1.309
QC	.337**	3.082	.091	.797	1.254
KNOW	.112	.438	.013	.844	1.185
SIXSIGMA	.260	1.914	.056	.819	1.221

Note: Dependent variable, SKILL; R² = .087, adjusted R² = .080; Durbin–Watson statistic = 1.568; *p < .05, **p < .01.

Second, the coefficient FIRM was consistently positive and highly significant (p < .05). This result shows that low-skilled workers’ skill improvement is positively associated with how many workers are employed in the low-skilled workers’ firm, in comparison to other learning activities provided for the workers in the firm.

Third, the coefficients COWORK, SELFLEARN, and QC were consistently positive and highly significant (COWORK and QC: p < .01; SELFLEARN: p < .05). Compared with COWORK, SELFLEARN, and QC, the coefficient SUPER was consistently negative and highly significant (p < .01). This analysis shows that a low-skilled worker’s participation in informal learning by supervisors is negatively associated with skill improvement, but informal learning by co-workers, self-learning through work, and QC positively influences skill improvement. OJT, TFT, KNOW, and SIXSIGMA showed little significance as well.

In terms of beta weights, this analysis showed that the influence of COWORK on the dependent variable is highest in comparison to other independent variables, when all other independent variables are constant (β = .246). SUPER has the next rank-ordered significant effect on the dependent variable, whereas the smallest effect is of KNOW (SUPER: β = .139; KNOW: β = .013). The results show that the association with informal learning by co-workers or a supervisor on low-skilled workers’ skill improvement is higher than formalized learning experiences such as QC.

Overall, these findings show that low-skilled workers’ skill improvement is positively influenced by how many workers are employed in the low-skilled workers’ firm. It shows that low-skilled workers’ informal learning components positively influence their skill improvement, but learning activities for a low-skilled worker on a certain hierarchy in an organization can negatively or slightly influence the skill improvement. Finally, it shows that institutionalized learning for a low-skilled worker positively influences the skill improvement when the learning activity is based on autonomy for a quality improvement of their work.

Discussion

The growth in high-skilled jobs and decline in low-skilled jobs create inequalities between high- and low-skilled workers. Informal learning has been recognized as one of the ways to improve the skills of workers (Crossan, Lane, and White 1999; Kahn and Lim 1998). HRD professionals use different formal and informal interventions to improve learning outcomes in the workplace (Crossan, Lane, and White 1999). This study investigated the contribution of both informal and formal learning participation to improve the skills of workers over time. The study has importance in that it focuses specifically on those individual workers who are low skilled and have limited formal education credentials. This study specifically analysed whether the firm size and various types of learning activities in the workplace are significantly associated with their skill improvement.

As the first stage, this study carried out a zero-order correlation between low-skilled workers’ skill improvement and their learning activities with informal as well as institutionalized learning in the workplace. The zero-order correlation showed that both informal and institutionalized learning activity components in the workplace are positively correlated with the low-skilled workers’ skill improvement. Therefore, previous research (Dale and Krueger 1999; Kahn and Lim 1998; Leslie and Brinkman 1998) of how the two factors, that is, low-skilled workers’ skill improvement and their learning activities, were associated with each other was supported.

Second, this study showed that low-skilled workers’ skill improvement may be increased without any learning programme provided to them. This is supported because low-skilled workers’ improvement depends on the role of the supervisor, and in the specific instance, supervisors’ support was not critical to changes in skill improvement. This means that other factors and peer learning, with the exception of learning activity officially and unofficially provided for the low-skilled workers in the company, can be associated with the workers’ skill improvement.

Third, the hypothesis that learning activities provided for low-skilled workers are associated with improvements in job skills is partly supported. Considering the size of the firm where workers are employed, informal learning activity and QC institutionalized in the workplace were significantly associated with their skill improvement. The hypothesis that low-skilled workers’ learning activities will have different associations with skill improvement in larger companies was supported by the prior literature (Occupational Information Network 2013; Kahn and Lim 1998).

Informal learning activities for low-skilled workers were more effective than institutionalized learning in the workplace. To clarify this analysis, it is necessary to understand that this result is closely linked to just-in-time learning, a method of learning in which all workers frequently improve their job-related knowledge and skills. For the purpose of effectively implementing the just-in-time learning, workers’ learning activity can be focused on immediate and daily informal learning. In comparison, formal or institutionalized learning takes more planning in the workplace (Brandenburg and Ellinger 2003). Our study finds a stronger association between informal learning activity than formalized learning and skill improvement in the workplace.

In terms of informal learning activity for low-skilled workers, low-skilled workers’ informal learning activity by co-workers is most significantly associated with their skill improvement among other informal learning components in this study, while low-skilled workers’ informal learning activities by supervisors are negatively associated or might be unassociated with their skill improvement. To clarify, it is necessary to understand the relationship between low-skilled workers and the others around them in the organizational culture. Informal learning is closely linked with relationships and peer networks, and requires support and a sense of community to be successful (Berg and Chyung 2008; Boud and Middleton 2003; Doornbos, Simons, and Denessen 2008; Lohman 2006; Quin 2009). Because of the hierarchical nature of firms, we note that it is difficult for low-skilled workers to inform their supervisors of training needs and potentially receive counselling from supervisors (Wayne and Ferris 1990). Co-workers, as opposed to supervisors, can be an important resource for informal learning to receive counselling on work and learning (Boud and Middleton 2003; Doornbos, Simons, and Denessen 2008).

On the other hand, the results of this study showed that participation in QC induced low-skilled workers to improve their skills, but OJT, TFT, Knowledge Mileage Systems, and Six-Sigma do not contribute to their skill improvement. The QC includes a systematic measurement for quality control and how to integrate support for a systematic quality measurement on the basis of autonomy among participants (Breyfogle and Meadows 2000; Nonthaleerak and Hendry 2006). The concept of QC gives workers a role in improving organizational outcomes through learning and is not a function limited to high-skilled workers (Nonthaleerak and Hendry 2006; Wiklund and Wiklund 2002). That is, the QC focus is on providing low-skilled workers with learning opportunities and systemizing a system to support this opportunity on their autonomy, while OJT is an unplanned training programme, TFT and Knowledge Mileage System are viewed as opportunities that can encourage unplanned learning, and Six-Sigma may be far from the participants’ autonomy. It is difficult to improve low-skilled workers’ knowledge and skills in the long term, since OJT, TFT, Knowledge Mileage System, and Six-Sigma are operated in firms where low-skilled workers have a high level of instability in jobs and roles (De Grip and Van Loo 2002; De Grip and Wolbers 2006).

Conclusion: implications and limitations

This study describes the impact of low-skilled workers’ learning activity on skill improvement. Consequently, this study showed that low-skilled workers’ skill improvement was partially associated with informal and institutionalized learning activities for low-skilled workers within the framework of the organization. The study focuses on the relationship between the improvement of core-technical ability through informal and institutionalized learning activities. Reviewing the results of this study in detail, some informal learning activities such as informal learning by co-workers and self-learning through work for low-skilled workers were positively associated with their skill improvements. Low-skilled workers’ skills could be improved through planned interventions on the participants’ autonomy such as QC provided by the organization as well. The core of these results shows that low-skilled workers’ learning is embedded in the social networks of the firm, and low-skilled workers’ skill improvement will depend on the larger relationships among workers.

The results of this study positively support the theoretical correlation between learning and individual skill improvement in the light of human capital theory in a broad sense (Dale and Krueger 1999; Leslie and Brinkman 1998). Workers’ learning activity is an important element to develop their knowledge and skills from the economic perspective of human capital (Kim et al. 2014). However, when focusing on the detailed elements of learning activity, the results of this study show that the low-skilled workers’ skill improvement was partially associated with informal and institutionalized learning activities (i.e., informal learning by co-workers, self-learning through work, and QC) for low-skilled workers within the framework of the organization (Bottrup and Clematide 2005; Kahn and Lim 1998). This result showed that ‘learning activity with autonomy’, in comparison with other learning activities, may be associated with the low-skilled workers’ skill improvement both positively and significantly.

This study has some practical implications as well. The first implication is that informal learning activity such as learning with a co-worker and self-learning on the job can be powerful tools to better improve low-skilled workers’ skills. Therefore, for low-skilled worker’s skill improvement, it is necessary to establish an environment to enhance self-directed learning on the job. For example, learning contracts can be useful tools to determine a low-skilled worker’s learning needs based on his or her job performance. Through open communication with co-workers, feedback on what should be done provides learning opportunities for low-skilled workers. More powerful contexts can be created for low-skilled workers by an effective feedback system with co-workers in the workplace. However, it was very interesting that informal leaning for a low-skilled worker with a supervisor may negatively influence or be unassociated with their skill improvement. One possible explanation for this is the organizational culture. Employees consider feedback from managers as a negative in a hierarchical structure that is very common in Korean companies. To address this concern, managers need to put forth effort to improve their communication skills.

Second, low-skilled workers should be skilled through planned interventions on autonomy and collaboration in quality control such as QC provided by the organization. While participation in QC induces low-skilled workers to improve their skills, OJT, TFT, Knowledge Mileage Systems, and Six-Sigma were not associated in this specific study with skill improvement. The core of QC is to discover, analyse, and solve work-based problems, and it may work through participants’ autonomy and collaboration (Abo Al-Hol et al. 2006). A supervisor or manager may lead all activities regarding the QC but ensure low-skilled workers’ autonomy, and the workers identify, analyse, and solve work-related problems more successfully on the basis of collaboration with other workers (Chung and Kim 2011; Park 2002). The QC is one of the learning activities closely linked to workers’ real job settings and is able to induce the workers’ voluntary and immersed participation.

In contrast, participation in OJT and TFT does not seem to be associated with low-skilled workers’ skill improvement, because OJT is an unplanned training programme and TFT is viewed as a place that may incur unplanned learning. These show that low-skilled workers have more opportunities to improve their skills in planned intervention than they do in unplanned intervention. Participation in Knowledge Mileage System and Six-Sigma does not seem to play a role in low-skilled workers’ skill improvement, because the two elements may be a planned training programme, but it rarely ensures autonomy and collaboration in the South Korea setting as well. Many workers in South Korea are likely to ‘participate’ in the two training programmes by office rule or top management decree.

This study has certain limitations. First, there is the possibility of omitted variable bias based on the time between the learning activity and the skill improvement. Because low-skilled workers’ learning activities and current skill level are measured at the same time (through the survey instrument a worker answers two separate questions, but on the same survey), it is measured by workers comparing the differences in the skill they displayed at the time of hire and at a time when the survey was given. To solve the issue of a time lag between the learning activity and the skill improvement, further research should focus on how to capture direct measures of worker productivity at different times and incorporate these data into new survey instruments.

Second, selection bias might be a problem. Low-skilled workers are likely to be mandated to participate in certain forms of learning activity because of their poor performance or current skill level. These low-skilled workers may be the least likely to acquire knowledge and skills from forced learning because of their low motivation or attitude. To solve this problem, further research should include low-skilled workers’ psychological conditions, such as motivation or attitude as well as learning elements.

Third, this study did not focus on how to improve the skills of unemployed adults. Many low-skilled workers have lost jobs because of deficiencies of knowledge and skill. The recent economic struggles have left many older workers with long gaps in their employment record. Therefore, it is necessary that future studies focus on low-skilled workers who are unemployed.

Fourth, the main focus of this study was limited to only the core-technical skills of low-skilled workers. Low-skilled workers may have deficiencies in their basic skills such as reading, writing, and math, or their technical job skills. Other skills such as complex problem-solving and resource management skills may improve low-skilled workers’ skills. An increasing number of jobs have focused on further education for developing workers’ basic skills for higher productivity.

Finally, the learning activities of this study were measured separately from the level, frequency, and quality of the participation because the HCCP simply focused on whether the low-skilled workers have participated in the learning activities provided officially or unofficially. Again, this study could be improved by collecting data directly from firms about the participation in different forms of skill improvement.

Disclosure statement

No potential conflict of interest was reported by the authors.