Characteristics of employees of small, manufacturing businesses by occupation: Informing evidence-based intervention planning

Abstract

Objectives

We examined characteristics of employees in six occupational categories in small, manufacturing businesses (20–150 employees).

Methods

We analyzed survey data from 47 businesses (n=2577 employees; 86% response rate) and examined relationships between job type and socio-demographic, health, and organizational support characteristics. Analyses were adjusted for age and gender, and company as a random effect.

Results

Smoking rates were highest for production workers (33%), production managers (27%), and support staff (28%) and lowest for managers (11%) (p<.001). Job stress was higher for production workers and support staff than managers (p<0.0001). Managers perceived social capital (p<0.001), safety climate (p<0.0001) and support for smoking cessation (p<0.001) higher than production managers, production workers, and support staff.

Conclusion

Differences in characteristics by occupation call for integrated interventions that target working class employees, leverage the influence of production managers, and enhance organizational support.

INTRODUCTION

Occupational gradients in smoking behavior, occupational exposures, and chronic disease morbidity and mortality have been extensively documented. Consistently, higher rates have been found among working class occupations than in managerial and administrative positions. ^1–8 High rates of smoking in blue-collar workers have been reported over the past five decades. ^9–14 In the United States the smoking rates among production workers and production managers is 28% while among managers the rate is 16%.¹⁵ Smoking rates of administrative support staff differ by gender; the rate among males is 23% and among females 19.5%. Among all US adult workers the smoking rate is 20%.^15,16

Exposure to both personal and occupational risk factors places the health of production workers in double jeopardy both from personal behaviors such as smoking and exposure to occupational hazards encountered on the job. ^1,2,17 Integrated workplace health protection and promotion programs provide an important approach to addressing and ameliorating occupational gradients in health. ^6–8 Businesses that seek to influence employees’ personal health behaviors such as smoking, exercise and nutrition may be more successful if they also minimize workplaces hazards that cause job-related injuries or illnesses. ¹⁸ Thus a comprehensive approach to employee health is needed. ¹⁹ Few companies of any size offer integrated programs that address both personal behavioral and occupational risk factors. ^6,20–23

Small businesses with 20–149 workers employ 20% (22,866,725) of the country’s private sector workers²⁴ and are likely to play an increasingly important role in workplace health because of the trend for new companies to start smaller and stay smaller over time. ²⁵ Eleven percent (2,401,908) of the workforce in businesses with 20–149 employees is engaged in manufacturing occupations. ²⁴ Manufacturing workers are more likely to smoke and be exposed to occupational risk factors^3,26 that pose independent health risks, and in combination can lead to elevated risk. ^20,23

Small businesses differ from large companies in ways that might have implications for the health of employees. ²⁷ Small businesses have a more personalized work culture^28–30 and less hierarchical structure. ³¹ While they are more likely to employ low-wage workers who are at higher risk for chronic disease, ¹⁴ they are less likely to offer either health protection or promotion programs^{6,21,28,31,32} and have lower rates of employer-provided health insurance. ³³ We describe characteristics of workers in small, manufacturing businesses that participated in the baseline survey conducted as part of Wellness Works, a group randomized trial designed to test an integrated workplace health protection and promotion intervention. Specifically, we compare employee socio-demographic and job characteristics; individual health variables; and employee perceptions of organizational support (POS) (including measures of employer health orientation, social capital, safety climate, and social norms for smoking cessation); for six job categories: managers, production managers, production workers, support staff, research and development/engineering, (R&D/engineering) and sales.

The purposes of this paper are to: (1) describe socio-demographic and job characteristics, health behaviors and POS of a cohort of employees in understudied small, manufacturing businesses by job type; (2) report characteristics of production managers, an occupational category of focus in this study; and (3) discuss strategies for applying these findings to worksite intervention planning.

METHODS

Research Design

Wellness Works was a group-randomized trial in which small manufacturing workplaces were randomized either to an immediate or delayed intervention condition following the completion of a baseline survey of all employees. This paper describes results from the baseline survey.

Subjects and Data Collection Procedures

Workplace Recruitment

Forty-seven small manufacturing businesses located in Minnesota participated in the study. Staggered recruitment occurred over a period of 2.5 years. Manufacturing businesses with 20 to 150 employees in five Minnesota counties - four located in the Minneapolis/St. Paul metropolitan area and one approximately 60 miles south - were identified from two commercial databases. All businesses were contacted to determine eligibility and interest in participation. In addition to meeting size requirements, eligible manufacturing companies were expected to have latitude to make their own decisions about safety and wellness improvements; anticipate no major changes in the company in the upcoming year; employ at least four smokers; and agree to study procedures. Details of business recruitment are reported elsewhere.³⁴

Survey Procedure

Baseline data were collected in surveys of all employees and were conducted at participating businesses between December, 2010, and November, 2012. In most cases, research staff administered the survey during work time. Also, Spanish and Hmong translations were available.

Measures

Surveys included questions about employee socio-demographic and job characteristics; individual employee characteristics and health behaviors; and employee perceptions of work safety and organizational support.

Demographic and Job Characteristics

Socio-demographic characteristics included age; gender; ethnicity and race; and education level completed. For analyses, race/ethnicity was re-coded into White/non-Hispanic or Other. Education level was collapsed into three categories: High School or less, Vocational training or some college, and College or graduate degree.

Job characteristics included years worked at the company (<1, 1–5, 6–10, >10 years), average numbers of hours worked per week, and job type. We asked respondents to report the number of hours they worked in an average week. To assess job type, respondents were asked to check all of the job types that applied to them from a list of ten. We then grouped the responses to create six categories: 1) management (those who selected management, but not production or production-like jobs); 2) production managers (those who selected both production or production-like jobs and management); 3) production workers (those who selected production or production-like jobs such as quality control, inventory-shipping, maintenance, driving, but did not choose management); 4) support; 5) research and development or engineering sales; and 6) sales.

Personal Health Behaviors, Perceived Stress, and Job Stress

Cigarette smoking status was assessed by standard measures of seven-day point-prevalence of smoking³⁵ and questions typically used to obtain national estimates of the prevalence of smoking. ¹⁵ A respondent was considered a smoker if s/he had smoked 100 cigarettes in his/her lifetime and reported either smoking in the last seven days or being a current daily or some day smoker.

For smokeless tobacco use respondents were asked if they currently used chewing tobacco, snuff, or snus every day, some days, or not at all. This item was taken from the 2010 Behavioral Risk Factor Surveillance Survey (BRFSS). ³⁶ For analyses, those who responded that they used smokeless tobacco daily or some days were considered smokeless users.

Recent alcohol use was assessed by questions taken from the BRFSS: number of days in the past 30 in which they had had at least one drink and the number of drinks containing alcohol they had imbibed on days in the past 30 when they drank alcohol. ³⁶ A total number of alcoholic drinks were calculated by multiplying the number of days in the past 30 they had had at least one drink by reported drinks per day. Risky alcohol use was calculated by dividing the total number of drinks by 30 to generate drinks per day. In keeping with National Institute of Alcohol Abuse and Alcoholism (NIAAA) guidelines, women who averaged more than one drink per day and men who averaged more than two were considered risky alcohol users.

Perceived stress was assessed with the 4-item Perceived Stress scale. ³⁷ Respondents rated the frequency in the past month of experiencing four thoughts and feelings related to the presence or absence of stress. The Cronbach’s alpha for these items was 0.70, indicating acceptable reliability. We calculated a perceived stress total score by summing the four items. High scores reflect high perceived stress.

Job stress was assessed by five items adapted by Sapp et al. ³⁸ from the Karasek Job Content Questionnaire. ³⁹ Two items measured job demands (“My job makes conflicting demands on me”; “I have to work fast to keep up on my job”); one item assessed skill discretion (“I have a lot of say about what happens on my job”); and two items assessed decision authority (“My job requires that I learn new things”; “I have a lot of say about what happens on my job”). As suggested by Sapp et al., we created a weighted measure of job control from the skill discretion and decision authority items and then subtracted the job control measure from the sum of the job demand items. After shifting to make the minimum zero, and scaling to make the maximum 8, the resulting measure of job stress varied from 0 to 8 and was roughly normally distributed.

Perceived Organizational Support (POS)

POS measures were calculated for each employee. Three items assessing employees’ perceptions of their employer’s health orientation were selected from a scale developed by Ribisl and Reischl: ⁴⁰ (1) “This company values healthy workers”; (2) “This company is generally concerned about my health and well-being”; (3) “It is easy to see that management has a commitment to improving employee health.” For each item, the respondent selected a number between 1 (“Strongly Agree”) and 5 (“Strongly Disagree”) that indicated the extent to which s/he agreed with the statement. The Cronbach’s alpha for the items was 0.89, indicating excellent agreement. We calculated an employer’s health orientation total score by summing the three items. Because the total score was extremely skewed, we collapsed this measure into three categories (low, medium, high) that included roughly equal numbers of respondents.

Three items assessing workplace social capital were adapted from those used by Sapp and colleagues: ³⁸ (1) “The people I work with are willing to help each other”: 2) “I trust the people I work with”: and 3) “The managers of this company look out for the people who work here.” For each item, the respondent selected a number between 1 (“Strongly Agree”) and 5 (“Strongly Disagree”) that indicated the extent to which s/he agreed with the statement. The Cronbach’s alpha for the items was 0.82, indicating very good agreement. We calculated a social capital total score by summing the three items. Because the total score was extremely skewed, we collapsed this measure into three categories (low, medium, high) that included roughly equal numbers of respondents.

We assessed safety climate with 18 Likert-style questions designed to measure employee perceptions of nine dimensions related to workplace safety: management, priority of safety, communication, safety rules, employee involvement, personal appreciation of risk, supportive environment, work environment and training. These questions were drawn from validated items developed by Cox and co-investigators and tested in a variety of workplaces. ^41–44 Cronbach’s alpha for the 18-item scale was 0.89, indicating excellent agreement. We calculated a safety climate total score by reversing items as necessary and then summing the 18 items.

Two items measured respondents’ perception of coworkers’ attitudes toward smoking. The questions asked how many coworkers: (1) “Think that smoking is a bad habit”; (2) “Would support people at this workplace who are trying to quit smoking” (Few or none, Some, Most or all). These items were selected from scales developed by Ribisl and Reitchl. ⁴⁰

Analyses

Multiple imputation was performed using SAS Proc MI to “fill out” the dataset for some important variables with small amounts of missing data: safety climate items, race, education level and years working at the business. The variables used for adjusting covariates in the analysis, gender and age, were available for all participants, so no imputation was required.

Missing data for the 18 safety climate items were imputed because this was a central variable with little missing data. When summing to the overall safety score, 224 (8.6%) of respondents had a missing value. Characteristics thought to provide correlated information were included in the imputations; these were race/ethnicity, job status, gender, age, education level, years working at the business, and company in which employed.

The imputation model was deliberately made very full as the main purpose was to reduce bias; hence a full batch of potential correlates was included. ⁴⁵ For our purposes of merely completing the data set to allow a balanced calculation of the scales, we chose randomly one of the five imputed values from the default five imputations for any missing item for an individual.

Analyses were conducted using SAS Version 9.2. All analyses included workplace as a random effect, thereby adjusting for the extent to which individuals in workplaces resembled each other on the variable in question (intraclass correlation). Analyses of the associations between job type and the personal health and organizational variables were adjusted for age and gender of respondents. Analyses of the relationship of job type with dichotomous dependent variables used a generalized linear model procedure (i.e., PROC GENMOD), using the logit link and binomial error distribution. Analyses of the relationship of job type with continuous dependent variables used a mixed linear model procedure (i.e., PROC MIXED) designed to handle Gaussian error terms.

RESULTS

Eighty-six percent (2652/3072) of the employees completed the survey. Return rates for individual businesses ranged between 50% and 100%.

Socio-demographic and Job Characteristics

Socio-demographic characteristics of participants varied significantly by job type on all dimensions reported here. Mean age ranged from 40 to 49 years; managers were significantly older than production managers who, in turn, were significantly older than production workers. Employees were predominantly male (73%) and white (81%), with the exception of support staff who were predominantly female (67%) (Table 1). R&D/Engineers and managers were significantly more likely to have had college or graduate degrees than those in other occupations (p<0.0001); production workers were least likely to have college or graduate degrees. Managers and production managers were significantly more likely than those in other job categories to have worked at the business for ten or more years (p<0.001).

Table 1.

Employee socio-demographic characteristics by job type (n= 2565)^*

Variable	n	% or Mean (SD) of Employees	Job Type						z or chi-square	p
			Managers	Production Managers	Production Workers	Support Staff	R&D/Engineering Staff	Sales Staff
Number (%) of participants in each job type	2565		231 (9%)	369 (14%)	1545 (60%)	220 (9%)	101 (4%)	99 (4%)
Age (mean)	2496	43.5 (12.16)	49.3a	45.2 b	42.7 c	45.2 b	40.2 d	45.3 b	16.95	<0.0001
Male Sex	2565	73%	69% cd	85% a	77% bc	33% e	83% ab	63% d	29.25	<0.0001
White Race	2565	81%	97% a	83% c	76% d	91% ab	92% ab	95% a	20.67	<0.001
Education
High school degree or less	821	32%	5%	27%	42%	18%	5%	11%
Vocational or some college	1102	43%	27%	49%	44%	48%	26%	41%
College or graduate degree	641	25%	68% a	24% d	14% e	34% c	69% a	48% b	34.83	<0.0001
Tenure at business
< 1 year	344	13%	7%	7%	15%	14%	20%	19%
1–5 years	839	33%	11%	26%	33%	41%	45%	29%
6–10 years	438	17%	15%	15%	18%	21%	10%	22%
>10 years	938	37%	52% a	52% a	34% b	25% b	26% b	29% b	24.49	<0.001
Works more than 40 hrs per week	2539	43%	77% a	67% b	31% d	36% d	62% b	50% c	32.28	<0.0001

^*Estimates and analyses are not adjusted for gender and age. Job type information was not available for 85 employees.

Note: Values within a row that share the same letter are not statistically significantly different (p>.05).

Personal Health Behaviors and Experience of Stress

Production workers (32%), production managers (26%), support staff (28%) and sales employees (20%) reported the highest rates of smoking while managers and R&D/engineers reported low rates (11% and 14%, respectively) (p<0.001) (Table 2). There was no significant difference between job types in smokeless tobacco use and the differences in risky alcohol use, although statistically significant (p<0.05), were not striking. Levels of perceived stress were highest for production workers, production managers and support staff (p<0.0001) and production workers and support staff reported experiencing higher rates of job stress than other occupational categories (p<0.0001).

Table 2.

Employee personal health behaviors and experience of stress by job type (n= 2565)

Variable	n	% or Mean (SD) of Employees (unadjusted)*	Job Type**						z or chi-square	p
			Managers	Production Managers	Production Workers	Support Staff	R&D/Engineering Staff	Sales Staff
Personal Health Behaviors
Smoker	2496	29%	11%a	26% bc	32% c	28% bc	14% ab	20% b	25.20	<0.001
Smokeless tobacco user	2442	8%	6%	8%	8%	7%	6%	13%	8.20	ns
Risky alcohol user	2316	13%	18%a	16% a	13% ab	11% b	11% ab	15% ab	12.42	<0.05
Stress
Perceived stress (mean)	2464	9.68 (3.01)	8.54a	9.72b	9.90 b	9.57 b	9.01 a	9.24 a	9.50	<0.0001
Job stress (mean)	2410	3.30 (1.91)	2.26a	2.52ab	3.72c	3.76c	2.78b	3.01b	46.41	<0.0001

^*Estimates for all employees are not adjusted.

^**All estimates and analyses by job type are adjusted for gender and age. Job type information was not available for 85 employees.

Note: Values within a row that share the same letter are not statistically significantly different (p>.05).

Perceived Organizational Support

Managers (52%) and production managers (43%) were more likely than production workers (38%) and R&D/Engineers (33%) to rate their company highly on concern about employee health (p<0.01) (Table 3). Managers (58%) were most likely and production employees least likely (34%) to report high levels of workplace social capital (p<0.001). Safety climate scores were highest among managers and lowest among production workers, with production managers falling in between these two job categories (p<0.0001). Production workers’ perceptions of coworkers’ attitudes toward smoking indicated a less supportive environment for quitting than did the perceptions of all other job categories (p<0.001). More managers than other categories of workers reported that most or all employees think smoking is a bad habit (p<0.001).

Table 3.

Employee perception of organizational support by job type (n= 2565)

Variable	n	% or Mean (SD) of Employees (unadjusted)*	Job Type**						z or chi- square	p
			Managers	Production Managers	Production Workers	Support Staff	R&D/Engineering Staff	Sales Staff
Employer’s Health Orientation
Low (3–10)	986	30%	20%	26%	30%	29%	36%	26%
Medium (11–12)	758	31%	28%	31%	31%	31%	32%	31%
High (13–15)	735	40%	52%a	43%a	38%b	40%ab	33%b	43%ab	15.32	<0.01
Social capital
Low (3–10)	744	30%	15%	30%	33%	28%	23%	29%
Medium (11–12)	799	32%	27%	33%	33%	33%	31%	33%
High (13–15)	931	38%	58% a	37% b	34% c	39% b	46% b	38% b	25.59	<0.001
Safety climate score (mean)	2559	69.0 (11.7)	74.4a	70.0b	67.4 c	68.2 bc	69.6bc	69.4 bc	16.61	<0.0001
Perception of Smoking Environment: How many employees report most or all of their coworkers:
- are supportive of others’ quit smoking efforts	2476	41%	62%a	44%b	34%c	46%b	55%ab	53%ab	24.31	<0.001
- think smoking is a bad habit	2475	49%	67%a	51%b	45%b	49%b	60%ab	50%b	21.63	<0.001

^*Estimates for all employees are not adjusted.

^**All estimates and analyses by job type are adjusted for gender and age. Job type information was not available for 85 employees.

Note: Values within a row that share the same letter are not statistically significantly different (p>.05).

DISCUSSION

We found that several socio-demographic characteristics, health behaviors, and perceptions of organizational support differed significantly by job type among a cohort of employees in under-studied small, manufacturing businesses. Occupational gradients in variables such as smoking behavior, job stress, and POS have been reported from surveys of the general population^12–14 and large worksites. ^46,47 To our knowledge, however, this is the first study to examine this array of characteristics by job type in a cohort of employees of small, manufacturing businesses. In addition, we focus on an occupational category, production managers, that has been reported in the literature, ^19,48,49 but not extensively described. The National Institute of Occupational Safety and Health (NIOSH) has identified these middle managers as being in a position to influence both employee health behaviors and the health and safety environment. ¹⁹ We suggest that specificity by occupation may provide a more nuanced approach to customized intervention planning called for by several investigators. ^11,20,48

Production workers and support staff

In the area of personal characteristics and health behaviors, production workers reported the highest rate of smoking (32%), although rates reported by production managers (26%) and support staff (28%) were also considerably higher than that of managers (11%). Also, production workers reported the lowest levels of education (p<0.0001); higher smoking rates among those with both lower educational status and in similar blue-collar occupations have been reported by others. ^12–14,50

Both perceived stress and job stress were high in production workers and support staff. Job stress is of concern because it has been associated with higher risk for chronic disease, ^7,51 and smoking behavior. ^52,53 Perceived stress has been identified as a common barrier to quitting smoking. ⁵⁴ Support staff reported health behaviors and perceptions of the work environment similar to those reported by production workers in spite of their educational differences. This might be explained in part by the levels of perceived stress and job stress and low POS experienced by both groups. In addition to high rates of perceived stress and job stress, significantly fewer of these workers (p<0.001) reported high levels of workplace social capital. High levels of social capital might decrease the impact of adverse work characteristics on psychological distress⁵⁵ and have positive effects on health outcomes.⁵⁶

Compared with other workers, production workers and support staff reported lower support among co-workers for non-smoking. Others have found that social norms supporting non-smoking are less prevalent among blue-collar workers. ⁴⁷

Based on the evidence from the literature together with findings from Wellness Works, effectiveness of smoking cassation interventions for these high prevalence employees might be enhanced if they address both perceived and job stress and if the worksite adopts interventions to strengthen organizational support for health and social capital. ⁵⁴

Production managers

Production managers report characteristics similar to managers in some areas, to production workers in others and fall midway between the two in other domains.

NIOSH has specified engagement with mid-level supervisors is one of the Essential Elements of Effective Workplace Programs. ¹⁹ Production managers are well positioned to serve as communication channels between production workers who face safety and health issues on a day-to-day basis and managers who have primary responsibility for workplace safety. We identified one other study that reported socio-demographic characteristics and beliefs of line supervisors. In that study line supervisors were significantly more likely than managers to identify cost, lack of management interest, and production conflicts as barriers to offering wellness programs. ⁴⁸ Perhaps interventions could be designed to enlist production managers as positive influences for health. For example, a three-stage intervention could be planned that first targeted small business management and ownership to enlist their support for an integrated workplace health protection and promotion program. Secondly, with management support, the program could provide production managers with up-to-date, accurate information on tobacco, the relatively high prevalence among blue collar workers, and the potential for health impacts of the synergistic interactions between tobacco smoke and occupational exposures. Resources such as group or telephone support and medications could increase the likelihood of success for their own quit attempts. ⁵⁷ Thirdly, if production managers feel supported by management and have personal experience with smoking cessation they might feel prepared to encourage smoking cessation policies and programs for the workers they supervise.

As noted, production managers’ perceptions of safety climate are different from and intermediate between managers and production employees and support staff. This gap might be narrowed using a two-pronged strategy that first incorporates actual improvement of elements of organizational support such as providing cessation resources for smokers and improving health and safety conditions through job redesign. Second, interventions that enhance communication skills could be provided for managers and production managers to enable them to disseminate information about health and safety improvements effectively; this, in turn, could contribute to strengthening perceived organizational support.

Employees need to have access to channels through which they can state their health and safety concerns. Effective joint worker-management health and safety committees are widely acknowledged as important for health and safety. ⁵⁸ Established Health and Safety committees may serve as channels for interventions to increase employee involvement¹⁹ and to incorporate wellness initiatives aimed at smoking cessation. Integrated workplace health protection and promotion interventions have been shown to improve workers’ perceptions of a company’s safety climate. ^{3,46,47,51,59–61} Also, in a direct comparison of an integrated program with one that addressed wellness only, worker participation rates in activities (21% vs 14%) and contacts with management (25 vs 9 per worksite) were higher in integrated programs. ⁶¹

Managers

Compared with other occupational categories, managers had the highest education levels and worked the longest hours. Also, they had personal health behaviors associated with better health; i.e., lower smoking rates and lower levels of perceived stress and job strain. Managers scored significantly higher than production managers, production workers, and support staff on measures of POS including safety climate, perception of employer interest in employee health, social capital, and support for non-smoking. A case could be made with managers for the value of improving organizational support for health in their businesses because POS has been associated with improvements in employee well-being, increased job satisfaction, perceived competence and reduced stress and burnout. Tools identified for enhancing POS include effective communication, supportive leadership, and health-supporting human resource policies. ^62,63

Strengths and limitations

This study has several strengths. To our knowledge, it is the first to examine and report characteristics of specific occupational categories of workers in under-studied small, manufacturing businesses.

This study includes a large number of worksites (47) and a high response rate to the baseline survey (86%). In addition, the survey sample included a large number of production workers (1545) and an under-reported job category, production managers (369). Together these two categories of workers comprise 74% of the sample. These factors give us confidence in the validity of our results.

These findings, however, must be interpreted in light of study limitations. For reasons related to feasibility, we relied on self-report to determine worker characteristics; therefore, responses were subject to reporting bias common with such measures. The worksites were from a geographically limited area in Minnesota so findings may not generalize to worksites in other areas of the U.S. However the occupational characteristics identified in this study are similar to those found by other investigators in a variety of geographic settings. We used brief measures due to the multi-risk factor focus of the study. The validity and reliability of these measures, however, have been demonstrated in national studies. Worksites were included in the study on the basis of specified eligibility criteria and on their willingness to participate. Thus, the results can be generalized only to similar worksites that may be relatively highly motivated to provide health programs.

Taken together, the differences in characteristics of workers by occupation described here demonstrate that a “one size fits all” workplace health intervention will likely be inadequate in effecting behavioral and organizational change. ^11,19,20,48 This broad array of individual and environmental influences calls for a comprehensive, integrated worksite intervention model^{22,26,64–68} that addresses individual, interpersonal, and organizational influences on health behavior. ^6,23,69 In addition, our findings suggest that integrated workplace health protection and promotion programs in these manufacturing settings may be strengthened by tailoring to evidence-based occupational characteristics with special focus on the role of production managers. On one hand, the nature of the work of production managers requires ensuring safe working conditions, while on the other hand their smoking behaviors suggest they need to be important targets for workplace safety and wellness initiatives. Their participation in and commitment to both smoking policies and programs and worker protection may be crucial to program effectiveness.