Why teleworking works for today's workforce

Teleworking

Telecommuting's emergence as a reciprocal strategy for retaining valuable workers while reducing the costs of operating an enterprise is yielding an entirely set of anticipated benefits for businesses, and creating entirely unforeseen effects on employees, their social networks, and relationships with peers and superiors. Baby Boomers, the foundation of the nations' largest demographic trend in its history, are now beginning to retire, leaving many highly skilled jobs open. The rate of baby Boomer retirement has in some industries been faster than the graduation rate of professionals to take their place. This dynamic has led to a shortage of skilled workers in many industries. The challenge of retaining these workers has led to a generation of employees who have been showered with perks, programs, and incentives to stay in jobs not easily filled. Telecommuting in certain industries is seen as the ultimate perk; the chance to work from home while having a job that also has value within a longer-term career. Telecommuting in many industries including Information Technologies (it) professions is used as a symbol of giving employees trust and freedom for their skills and efforts; employers look at it as a strategy of reciprocal trust and freedom for exceptional performance the willingness to go the extra mile when needed. In a sense offering telecommuting to those employees that have the most in-demand jobs may also paradoxically create isolation, a sense of lacking support from key resources in corporate, and the development of entirely new sources of stress.

Introduction

The intent of this paper is to evaluate a series of seven hypotheses that specifically focus on the implications of teleworkers' Internet use levels; the established role of the Internet has a contributory, not cannibalistic media platform relative to television and newspaper; the assumption that Internet usage is greater for men vs. women; an evaluation of the demographics of Internet usage by gender, age, income, and position in their organizations; and the assumption that a teleworkers' Internet usage is proportional to their employment status. In addition, the hypothesis that amount of a teleworkers' Internet usage is directly influenced by the density of their job tasks will also be evaluated. Lastly, the hypothesis that the richness of an employee's home computational infrastructure has a significant influence on the amount of time spends on the Internet. This last point is specifically addressing the use of Virtual Private Networks (VPNs) by organizations to create a secured connection to their teleworkers over broadband communications lines.

In addition to all the above hypotheses, the dynamics of how teleworkers trade off the initial appeal of working from home needs to be balanced with sufficient interaction and socialization with others in the work groups to fulfill affiliation, achievement, and social needs. This specific area of the paper concentrates on the work/life balance achieved by teleworkers, including the measurement of isolation behaviors (DV) as a result of Internet usage in the home. The contributory factors of employment status and job density will specifically be tested through statistical analysis based on the results of the primary research completed.

From the studies completed of telecommuters' demographics, a polarity is beginning to emerge of remote workers who are members of this trend. Of the 20.7 million employees worked at home at least one day of the cited year according the U.S. Bureau of Labor Statistics (2005), the challenge of demographic research is to discover through research the demographic segments of the most frequent telecommuters. While the definition of a telecommuter demographic model or taxonomy has not been specifically defined in previous research, there are dozens of studies that attempt to define telecommuters demographically. Bailey and Kurland (2002), and other researchers have been quick to define the traditional demographic segmentation criteria to telecommuters while Ford and Butts (1991) have proposed from their research that the polarity of telecommuting demographics are comprised of professional jobs requiring highly unique skills and insights on the high end, and routine, high quantity tasks that require little training or expertise on the low-end. It is the intent of this paper to generate new insights into the dominance of genders as it relates to telecommuting, with the hypothesis that men spend more time on the Internet while fulfilling their job responsibilities than women. In addition, this paper will concentrate on the 31-40 age groups' demographic, specifically evaluating their use of the Internet in the context of their telecommuting jobs. What is emerging from the literature review is that the higher-end professional positions are dominated by male telecommuters, and the lower-end, clerical jobs are dominated by women telecommuters. This paper will look to further validate these previous research findings discussed in this literature review.

Ford and Butts (1991) has defined these two segments of telecommuters as the higher-end professional telecommuting jobs, and the lower paying clerical jobs as follows. The researchers first define higher paying telecommuting jobs by their nature require intensive thought, efforts marked by a high level of initiative, and collaboration with others in their workgroups.

This first group of telecommuters is typically in highly unique jobs including computer programmers, engineers, scientists, technical writers, or field sales persons. The researchers define the second demographic segment as those telecommuters who perform routine, highly replicated tasks that require little formal training or expertise. This specific demographic group primarily complete data entry, word processing, telemarketing, and basic accounting tasks that are easily measured and evaluated. The influence of outsourcing and home-sourcing of repetitive processes and tasks has yet to be quantified as part of this segment, yet researchers agree it is a significant influence on this demographic. Christensen (1992) defines the high end of the telecommuting demographic as being comprised of 88% of all telecommuters, with the remaining 12% is administrative, call center, clerical, and support workers who complete tasks both for their own companies and those who have outsourced work to their employers. This 12% figure is expected to increase as a result of greater acceptance of outsourcing of routine, highly proceduralized tasks that can be easily monitored grows. Further demographic research by Olzewski and Mokhtarian (1994) have defined the majority of telecommuters being male (65%) in these high-end telecommuting jobs which is also consistent with the findings from Luukinen (1996), who in a Finnish study found that telecommuters tend to be predominantly be independent professional men and women who are highly educated. On the issue of gender, Christensen (1992) has shown that the 12% of clerical jobs are predominately staffed by women, and is the fastest growing area of telecommuting today.

Telecommuting's initial popularity and enthusiastic support continues to meet with greater skepticism however. Determinants of teleworker employee satisfaction have yet to be conclusively defined, their interrelationships quantified, and the effects of high Internet use required to complete jobs that have a high density of job tasks fully explored. The correlation of the aspects of job density, employment status, work/life balance and overall satisfaction with teleworking needs further research, and will be specifically addressed in the primary research completed in this paper.

Clearly what is needed is a model that can first identify then validate the interrelationships of the variables being measured in this papers' primary research effort. Further, the implications of all measured attributes on teleworker satisfaction and the ability to attain work/life balance, an explicit benefit of teleworking, needs further exploration and validation as well. It is an objective of this paper to illustrate through the use of research and thorough analysis the causality of factors that exemplify telecommuting effectiveness by demographic group and also examine the factors that lead to work/life balance in addition to worker satisfaction while employed in a teleworking arrangement. This paper looks to quantify individual differences between demographics variables (e.g., age, gender, ethnicity, and tenure) and job characteristics (autonomy, flexibility, satisfaction with feedback). In addition to these factors, there is the implicit level of trust that pervades both the informal social networks that are composed of a teleworkers'; co-workers and peers, in addition to the level of trust between a telecommuter and their managers and the broader company formal organization structure. In addition to these two relationships, the telecommuters' level of interaction with or avoidance of family and friends also needs to be evaluated as part of this research effort, specifically focusing on the ability or inability to keep work and life in balance as a result.

In addition to the demographics and job characteristics variables, the role of trust within social networks and within formal organizational structures needs further research throughout the entire field of telecommuting. Herzberg's definition of a two-factor theory in conjunction with researchers Mausner and Snyderman (1959) seeks to differentiate factors that lead to positive vs. negative attitudes about one's job (Pinder, 1998) and sets the foundation for the role of trust between a teleworkers' peer or referent group and also between a teleworker and their managers. The essential aspects of Herzberg's models' ability to accurately predict how specific factors will either enhance or detract from overall satisfaction also is defined and later proven in subsequent research (Herzberg, Mausner, and Snyderman, 1999). The Herzberg two-plane model that includes an upper plane of "motivator" factors, which are those factors that lead to high job satisfaction are compared to the "hygiene" factors that are enablers of stability in any work environment. Motivator factors include achievement, recognition; work itself, responsibility, advancement, and growth, while hygiene factors include company policies, relationships with supervisors, work conditions, salary, relationship with peers, personal life, and relationships with subordinates, status, and security. These factors overall are critical for the development of a motivated and stable workforce. The implications of these factors on the job satisfaction and attitudes of telecommuters form the research foundation of this paper and also look to provide an indication of how teleworkers keep their work/life balance in equilibrium.

Method

The main objectives of this study are:

To provide further understanding of why workers choose to telecommute in the context of their work/life balance objectives.

To define the demographic segments that are emerging that have the potential to create a new definition of telecommuters and the reasons behind their choice of this specific type of work arrangement.

To provide an understanding of what aspects of telecommuting jobs contribute or detract from job satisfaction. In the past, a widespread of ad hoc scales measuring job satisfaction has been used and the psychometric properties of these ad hoc scales are largely unsubstantiated (Kinicki, Schriesheim, McKee-Ryan, & Carson, 2002).

To validate that the Internet has become equally balanced as a media source relative to television and newspaper in the context of a telecommuters' use to stay informed.

This research was designed to measure the level of satisfaction for a telecommuting population and to determine the relationship between hours spent telecommuting and a variety of job characteristics defined in the Job Diagnostic Survey. The sample consisted of telecommuters from a software company who ranged in types of occupations. There was an attempt to collect data from people that ranged from very little hours spent telecommuting to "all" of their work time telecommuting the population was asked to respond to the Job Diagnostic Survey (JDS) online over four-week period of time. The participants were asked to provide identifying information in the demographic section that would allow the researcher to match demographic characteristics to their responses across dimensions of the survey.

Participants were asked to voluntarily participate in the study. All data were collected online. The survey was the JDS as authored by Hackman and Oldham, and the five job characteristics were the primary measures for hypotheses testing (autonomy, feedback, task identity, task significance, & skill variety). The survey included 83 items and 14 demographic questions. Once the data were collected, they were cleaned and analyzed using the statistical analyses described later in the data analysis section. This section will outline the sample, instruments, research design, and data analyses methods.

Sample

The sample consisted of 153 participants. A participant was defined as an individual who was presently in a telecommuting working relationship as a full time employee with their employer, and had been with the organization for at least 12 months. Of the 153 participants, the number of hours spent telecommuting in a given week was not controlled.

The data collection was conducted online. The experimenter put the JDS online, drafted an invitation, as all participation was voluntary, and sent by the VP of HR on behalf of the experimenter. In addition to the data collected from the organization regarding the JDS, participants were asked to provide self-ratings of the importance they placed on telecommuting, their general satisfaction and a variety of other questions that were developed to provide demographic level detail. These items were captured in the demographics section of the survey.

One industry was represented: the industry is best described as software development and sales in the insurance sector. Company a will be the nominal reference throughout this document. Company a was a fast-growing organization that had been in operation for 55 years and employed 500 people. This company provided software that assisted insurance carriers in the process of validating and detecting the accuracy of insurance claims for the auto and medical sectors within the larger insurance sector.

The sample ranged occupationally from software engineers to customer service representatives whom are more "business to business" in that they are solving software and technical issues for the insurance companies rather than an individual consumer per se. There were technical writers, customer service professionals and software engineers in the sample.

The criteria were set by which the participants would be selected. First, all participants had to have been employed by the organization for at least 1 year prior to the study (this information was obtained in the demographics section by asking for "company tenure"). Second, all participants had to have worked in a telecommuting work arrangement for at least 6 months (this information was collected by the VP of HR). Job responsibilities were not a criteria for entry into the survey, rather we asked several questions in the demographic section that were meant to ascertain the tenure, level, and relative job responsibilities at a high level.

Demographic Variables

The demographic variables that will be measured include: (a) tenure, (b) job category, (d) level in company, (e) why the participant chooses to telecommute (f) gender, (g) hours worked per week, (j) hours spent telecommuting per week, (k) number of children, (l) marital status, (m) self satisfaction, (n) presence or absence of a home office, (o) distance from the workplace, (p) whether the participant has a spouse who works from home, and (q) level of education. These 14 questions were added at the end of the survey. Past literature and interesting findings from previous literature were used in selecting many of the demographic variables. For instance, Yap and Tng's (1990) research findings provide rationale for collecting demographic information specific to this study. They found that having an available work-space at home and being married rather than single were demographic characteristics related to positive attitudes towards telecommuting.

Research Instruments used for measuring Job Characteristics and Satisfaction

The Job Diagnostic Survey (JDS), developed in 1975 by Hackman and Oldham, is based on the Job Characteristics Model, a theory of how job design affects work motivation. The JDS is a diagnostic tool designed to measure the characteristics of jobs in organizations and the reaction of people to their jobs. Specifically, the JDS measures: a) objective job dimensions (including skill variety, task identity, task significance, autonomy, and feedback); b) supplementary job characteristic scales (including feedback from others and dealing with others); c) individual psychological states resulting from these dimensions (including meaningfulness of work, responsibility for work outcomes, and knowledge of results); and d) affective reactions of employees to the job and work setting (general satisfaction, internal work motivation, and satisfaction with opportunity for development and growth; Hackman & Oldham, 1975).

All five of these job characteristics contribute to "experiences," those experiences are referred to by Hackman and Oldham as "psychological states." These experiences of work, or psychological experiences, then contribute to affective outcomes like internal work motivation, quality work performance, satisfaction with the work, and/or absenteeism and turnover.

The JDS taps into overall level of motivation and satisfaction through 83 items broken down into seven sections. Sections one through six employ 7-point response scales. While section seven used 5-point response scales. The scales for each section intend to measure a specific job characteristics which include: (a) skill variety, (b) task identity, - task significance, (e) feedback from the job itself, (f) feedback from others, (g) dealing with others, and (h) autonomy.

Hackman & Oldham's (1980) use a specific scoring scheme to score their items. For each measured scale of the JDS, the items yielded a sum score for the scale. All reversed scored items were also taken into consideration by the scoring scheme.

Table 1 presents the internal consistency reliability of each of the objective and supplementary job characteristic scales and critical psychological states measured in the JDS (Hackman & Oldham, 1980). The results section of this study will report the specific internal consistency for this population. The median correlations are also included in this table, referred to as the median off-diagonal correlation. The median off-diagonal correlation is the "median correlation of the items scored on a given scale with all of the items scored on different scales of the same type of variable" (p. 164). For example, the median off-diagonal correlation for task identity (.12) is the median correlation of all items measuring task identity with all items that measure the other six job dimensions as well. One indication that there is discriminant validity of the items is the median off-diagonal correlations are smaller than the internal consistency reliabilities.

The JDS was primarily developed to use in work redesign, specifically the diagnosis of job prior to their redesign and assessing the effects of redesigning jobs on the people who do them (Hackman & Oldham, 1975). Although, Hackman and Oldham (1975) found that the diagnostic use of the JDS did not stop there and found that the JDS could be used to tap into overall level of motivation and satisfaction of employees.

The JDS is designed to measure the theoretical link of employee perceptions of skill variety, task identity, task significance, autonomy, and feedback to job satisfaction. "If the Job Characteristics Model is correct and these characteristics are related to job satisfaction, then analyzing jobs in terms of these characteristics may influence employees' job satisfaction" (Renz, 1995, p. 14).

Based on the developing literature of the JDS and its application to measuring satisfaction, the researcher used this instrument to expand the general understanding on work satisfaction of telecommuters and the Job Diagnostic Survey. In doing so, the researcher investigated telecommuters working arrangement, the time spent telecommuting, and perceptions of satisfaction with feedback and satisfaction with autonomy, as well as skill variety, task identity and task significance.

Furthermore, autonomy and feedback were identified in this study as important variables in understanding the effects that time spent telecommuting may have on perceptions of satisfaction. Autonomy is operationally defined by the JDS as "the degree to which the job provides substantial freedom, independence and discretion of the employee in scheduling the work and in determining the procedures to carrying it out" (Hackman and Oldham, 1976). Because this definition includes the concept of flexibility, flexibility will not be measured separately in this study.

The JDS is comprised of items that measure the core job dimensions in Hackman & Oldham's (1976) Job Characteristic Model. Specifically these characteristics include: skill variety, task identity, task significance, autonomy, and feedback, as well as a general satisfaction sub-section will be measured. The JDS is used for two purposes: (a) it is designed to measure the theoretical link of employee perceptions of skill variety, task identity, task significance, autonomy, and feedback to job satisfaction and (b) Recent literature is starting to explore the JDS with other variables such as work satisfaction and performance (Fried, 1991). The JDS was chosen for this study because the JDS was found to tap into overall level of motivation and satisfaction of employees (Hackman and Oldham, 1975). Furthermore, Fried (1991) has used the JDS to explore other variables such as work satisfaction and performance.

Procedure

To initiate contact and explain the study, a "request for participation" the researcher and the VP of HR met to discuss the nature and utility of the study. Once permission was granted by the organization (VP of HR and the CEO) to survey employees, four steps were taken. First, the researcher and the organization determined a time frame for data collection. Second, the organization provided the researcher with the eligible employees and their email addresses. Third, a method and criteria of communicating with the participants was agreed upon. Fourth, the sales employees were sent a survey invitation by email. The survey email invitation briefly explained the purpose of the study and the company's involvement and contained a hyperlink that took the participants directly to the survey. In the email invitation, participants were given the researcher's contact information to use if they had questions about the study.

The data for the organization were recorded automatically by SurveyMonkey™ and the company never possessed the data. The survey was developed and administered using SurveyMonkey™ . The order of the items from the five scales was exactly as the JDS appears in the appendix of the "Work Redesign" book published by Hackman and Oldham (p. 303-306).

Before participating in the survey, the participants read the consent agreement and the subject bill of rights. Next, the participants were asked to choose to participate if they agreed to the stated conditions. At the end of the survey participants were taken to a "thank you" page, where a letter of thanks and a Web address were given. The letter indicated that a debriefing of a summary of the findings at an organizational level would be given to the VP of HR by the researcher but that none of the findings would be shared with anyone at an individual level of analyses.

The researcher allowed approximately 4 weeks for responses. At the beginning of Week 3, the researcher or the organization sent a reminder email invitation to all sales representative employees, who were again invited to participate in the survey on a voluntary basis.

The duration of survey completion was approximately 15 to 20 minutes, with a total of 83 items and 14 demographic questions.

Research Design

The research design was selected to examine the correlations between predictor variables and a criterion variable (satisfaction). The present study is a cross functional study in that the random assignment of participants to levels of the independent variable (hours spent telecommuting) was not possible. Individuals N. Of hours spent telecommuting occurred naturally and the researcher had no ability to control or influence the independent variable. The design was a within-subjects design, in that all employees participated in the same measures and there was no differing manipulation of the predictor variables across participants. The JDS survey yields interval level data in the use of a Likert scale. Hours spent telecommuting was treated as continuous variable that was later grouped into low, average, or high.

Variables and Controls

This section describes the variables and controls pertinent to the present study. Predictor variables and criterion variables, and potential confounds are identified.

Telecommuting telecommuter is defined as anyone who has been using their home as an alternative work site for some or all of the work-week for at least six consecutive months. Additionally, all participants must have worked for their employer for at least one year, whether in a telecommuting or an in-house capacity. This one-year tenure requirement is intended to increase the probability that employees are reliably able to assess their job satisfaction.

The independent variable, time spent telecommuting, is treated as a continuous variable. Telecommuting was specifically measured as a self reported number of hours on spends "working out of the office" during a typical week. Although all participants will be full time employees, I will not assume that all participants work forty hours a week. A study by Cree and Sorenson (2000b) found that in one large, multi-national company, telecommuters worked on average more than 56 hours a week while their office worker counterparts worked an average of 47 hours a week. This finding demonstrates the importance of considering hours worked beyond the typical 40-hour work-week. In this study the researcher does as "the total number of hours spent working" to potentially contrast with the self reported telecommuting number.

Predictor Variables

There were six predictor variables in this study: (a) task identity (b) task significance - skill variety (d) autonomy (e) Job feedback (f) hours spent telecommuting. (a) - (e), were continuous and defined by the sum score of the items as defined by the JDS scoring scheme (Hackman and Oldham, Work Redesign p. 310). Variables (a) - (e) were developed by summing the items from distinct sections on the survey as prescribed by Hackman and Oldham in their "Work Re-Design" book and was consistent with the scoring scheme provided by the Roy W. Walters and Associates consulting firm (Whitney Industrial Park, Whitney Road, Mahwah, NJ 07430). The "sub-scales" consisting of three items were used to construct the variables.

(hours spent telecommuting) was a continuous variable that was self reported, it was later categorized into high, mid, and low. Each job characteristics was also examined as mediator and moderator variables to examine how they may (or may not have) changed the relationship between hours spent telecommuting and general job satisfaction.

The dependent variable of job satisfaction was a continuous variable. This variable was measured on an interval scale from 1 = strongly disagree to 7 = strongly agree. There was one sub-scale that measured general satisfaction with 5 items coming from two sections of the JDS.

Personal characteristics were analyzed on an exploratory basis. The personal characteristics captured in the demographics included Tenure (1 = 0-1 yrs, 2 = 1-3 yrs, 3 = 3-5 yrs, 4 = 5-7 yrs, 5 = over 7 yrs); Job Category (Programmer, Customer Service rep, Clerical, Technical Writer, other); Level (1 = non-management, 2 = management); Why Telecommute (1 = Desire for autonomy, 2 = Family demands, 3 = Ability to be more creative, 4 = Away from management/Culture, 5 = Health concerns, 6 = Other); Self Sat = Overall I am satisfied with my working arrangement (1 = dissatisfied, 2= moderately dissatisfied, 3 = neutral, 4 = moderately satisfied, 5 = satisfied); Number of children in the household (N); Married not married (O = no, 1 = yes); Time telecommuting (N); Total hours worked per week (N); Level of Education (1 = HS, 2 = AA, 3 = BA, 4 = Master, 5 = more than a masters); Gender (1= female, 2 = male); Do you have a home office (1 = yes, 2 = no); Distance for the primary office (1 = less than 10 miles, 2 = 10-20 miles, 3 = 30-40 miles, 4 = 40-50 miles, 5 = over 50 miles); and Spouse at home during the day (1 = yes, 2 = no).

The population consists of employees from one company in a variety of position types. To prevent possible confounding variables, several steps were taken. Since random selection was not possible in the present study, the first step was to determine the criteria for eligibility of participants in the study. For an participant to be eligible for participation the following requirements were set: (a) participants must be full-time employees of the organization, (b) participants at all performance levels in positions must have been provided by the organization, - the participants must have been employed by the organization for at least 12 months prior to study, (d) participation was voluntary and was positioned as optional by the organization and the researcher. These controls compensated for threats to validity that are present when conducting a study (Keppel, 1991).

This study was administered within the customer service organization, which houses a variety of position types (see demographic section above for position types by name). The organizational sponsor, the Vice President of Human Resources, was asked to provide job descriptions for all participating.

Results

The primary objective of this study was to examine the relationships among hours spent telecommuting, and job characteristics, such as task identity, task significance, skill variety, autonomy, perceptions regarding feedback, and job satisfaction. This chapter reports the results of those analyses. It is divided into six major sections: Demographics, Power Analyses, Descriptive Statistics (missing data, outliers, and skewness), Scale Analysis, Research Hypotheses Testing, and Exploratory Analyses.

Demographics total of 153 employees participated in this study. All employees responded to demographic questions. Of the 175 surveys distributed 153 were returned. Response rate of 86% is above the norm for typical satisfaction survey results.

Information about tenure was collected in the demographics section. From the total sample of employees, 56.9% reported tenure of over 7 years, 2.0% reported a tenure of less than 1 year, 14.4% reported a tenure of 1 to 3 years, 15% reported a tenure of 3 to 5 years, and 11.8% reported a tenure of 5 to 7 years.

Participants were asked to report their "job category." The categories were; programmer, customer service representative, clerical, technical writer, or, other. 50% of the respondents were programmers, 2% were clerical, 10% were customer service representatives, 2% were technical writers, and 36% were other. Within "other" the respondents could type in there job type and of the 55 respondents who choose this category all typed in a response and no one dominant job category emerged.

Participants self reported education level. 63% had at least a B.A. degree11% had a Masters degree or higher. All other had either an a.A. Or H.S. degree.

62% of respondents were "non-management," 22% were "lead or professional," and 16% were management. Regarding gender: 87% were male and 13% were female.

93% of the sample reported that they typically work more than 40 hours in a week and 7% stated they work between 30-40 hours per week. See the below table for the reported hours spent telecommuting:

Table 1: Hours Spent Telecommuting of hours

Percent

Less than 5 hours

5-10 hours

10-15 hours

15-20 hours

20-25 hours

25-30 hours

30-35 hours

35-40 hours

More than 40 hours

Participants were asked to self-report their marital status and the sample consisted of 36% married, 52% single and 12% were un-married but lived with a significant other. 26% of the sample had no children, and the breakdown of children for the sample goes as follows: 18% had 1 child living with them, 23% had two children living with them, 24% had three children living with them, and 10% had four or more children living with them. Of those 77% of the population had a wife or significant other in the home (did not live alone).

The sample were asked; "overall, how satisfied with your working arrangement are you"? 7% were dissatisfied, 4% were moderately dissatisfied, 12% were neutral, 26% were moderately satisfied, and 51% were satisfied.

When asked; "why telecommute"? The sample offered the following frequency of responses:

Table 2: Why telecommute?

Choice of responses

Desire for autonomy

Family demands

Ability to be more creative

Away from management/Culture

Health concerns

When asked the importance of telecommuting the sample responded in the following way: 4% responded "not at all important," 13% stated it was a "nice to have," 20% stated; "it was a nice benefit for working here," 28% claimed they depend on the working arrangement, and 35% claimed it was a "must have."

Power power of.80 is reasonable for behavioral sciences and decreases the chance for a Type II error (Cohen, J., 1992). Keppel (1991) suggested that a significance level of.05 helps to reduce the chance of Type I error. This study used multiple regression and correlation statistics; therefore, a significance level of.05 was used. For this research, a power analysis demonstrated that a significance level of.05, power of.95, and effect size (r) of.50 required a sample size of 176 participants. Participation from 153 employees was obtained. Data analysis pertaining to the relationship between time spent telecommuting, the job characteristics, autonomy, feedback, and satisfaction included all 153 participants from one company. The sample size for data analysis came in under the minimum goal of 176 by 23 participants.

Descriptive Statistics

First, the data were examined for accuracy of data entry, missing values, and outliers. No inaccurate data entries were found. The original sample included 156 response sets; 3 response sets were discarded. To prevent potential confounding variables, participants were required to have been employed by the organization for at least 1 year. Based on this criterion, 3 participants were discarded for having worked with the organization less than 1 year. All participants agreed to the conditions of the survey via informed consent.

Second, the data were examined for outliers. Since the data were never hand transferred in any way manually, and the respondents "clicked" on a forced choice there were no outliers.

Correlation Matrix (Median Correlations)

SkillVarietyAdd TaskIdentityAdd TaskSignificanceAdd AutonomyAdd JobItselfAdd Correlation SkillVarietyAdd 1.000.296.507.314.461 TaskIdentityAdd.296 1.000.226.323.238 TaskSignificanceAdd.507.226 1.000.421.420 AutonomyAdd.314.323.421 1.000.374 JobItselfAdd.461.238.420.374 1.000

Of the five job characteristics measured in the JDS the first three (skill variety, task identity, and task significance) are seen as contributing to the experienced meaningfulness of work (Turner and Lawrence, 1965). Autonomy has been traditionally viewed as contributing to the experienced responsibility for outcomes of the work and Feedback from the job itself contributes to knowledge of the actual results of the work activities.

These experiences of work, or psychological experiences, then contribute to affective outcomes like internal work motivation, quality work performance, satisfaction with the work, and/or absenteeism and turnover.

From the pattern of responses in the survey, and the scoring key recommended for the JDS in Hackman and Oldham's 1976 book entitled "Work Redesign," sub-scales were created for skill variety, task identity, and task significance, autonomy and feedback or job itself, and general satisfaction. The justification for formulating these subscales occurred from the scale analyses.

To determine if the four JDS context satisfaction sub-scales should be combined to form a single overall, or, general satisfaction scale, the relationships among the short scales were examined. The correlation results indicated that the JDS satisfaction subscales did significantly correlate with each other at the level of.05 in two-tail test.

Additionally the means were compared to the normative data used in the validation study for each scale. Table 3 presents comparisons of the present study data to normative data. All previously validated scales were examined: skill variety, task identity, and task significance, autonomy and feedback or job itself, and general satisfaction.

Table 3: Normative Data Analysis

Job Characteristics

National Norms

Sample

Skill variety

Task identity

Task significance autonomy

Feedback from job

Feedback from agents

Dealing with others

Critical Psychological States

Experienced meaningfulness of work

Experienced responsibility for outcomes

Knowledge of results

Affective Outcomes

General satisfaction

Growth satisfaction

Internal work motivation

Context Satisfactions

Job security

Pay

Co-workers

Supervisors

Individual growth need

From Hackman, Oldham, and Stepina (1979). These norms were based on the responses of 6930 employees who work in 876 different jobs in 56 organizations.

Research Hypothesis Testing

Hypotheses were tested using multiple regression analysis and Pearson product-moment correlations as well as testing form moderation. Before interpreting the results of the correlation matrix, a Bonferroni test was conducted to ensure that inferences of relationships were appropriate. The Bonferroni test was conducted by taking the standard used for the p level.05 divided by the number of correlations found. Based on the results, a new p level was set at.005. First, the data were assessed for multicollinearity. The independent variables were not correlated above r =. (Table), indicating that multicollinearity did not exist (Tabachnick & Fidell, 2001). Multicollinearity was also assessed by examining the Tolerance values (Tolerance = 1-R2) for the independent variables.

Table 5 Correlation Matrix for the Scales

Correlations

Correlation Matrix

SkillVarietyAdd TaskIdentityAdd TaskSignificanceAdd AutonomyAdd JobItselfAdd Correlation SkillVarietyAdd 1.000.296.507.314.461 TaskIdentityAdd.296 1.000.226.323.238 TaskSignificanceAdd.507.226 1.000.421.420 AutonomyAdd.314.323.421 1.000.374 JobItselfAdd.461.238.420.374 1.000

01, two-tailed. **p

The primary objectives of this study were to examine the relationships among hours spent telecommuting, and job characteristics, such as task identity, task significance, skill variety, autonomy, perceptions regarding feedback, and job satisfaction. Specific research questions to be addressed in this study include: (a) Does job satisfaction increase or decrease with the number of hours an employee spends telecommuting? (b) as time spent telecommuting increases, how are perceptions of task identity, task significance, skill variety, autonomy, satisfaction with feedback impacted? - as people spend more time telecommuting, how do feelings of task identity, task significance, skill variety, autonomy, satisfaction with feedback impact the relationship between time spent telecommuting and general job satisfaction?

The following hypotheses were tested in this study. First, based on inconsistencies in the literature (Bailey & Kurland, 2002; Golden, 2005) on whether telecommuting can lead to job satisfaction, number of hours telecommuting and general satisfaction were examined to determine whether a relationship exists.

H1: Number of hours an employee spends telecommuting will correlate with General Satisfaction, as measured by Job Diagnostic Survey, to a statistically significant degree.

Correlations hourstelecommuted

GenSatisfactionAdd hourstelecommuted

Pearson Correlation

Sig. (2-tailed)

GenSatisfactionAdd

Pearson Correlation

Sig. (2-tailed)

Second, research suggests conflicting notions regarding whether employees who telecommute appreciate the increased flexibility that such a working arrangement affords, or whether such arrangements actually have a negative impact on the employee due to decreases in general support and camaraderie from co-workers and supervisors (Chapman et al., 1995; Cooper and Kurland, 2002; Dubrin, 1991; Galinsky et al., 1993; Golden, 2005; Trent et al., 1994). Therefore relationships between Job Diagnostic Survey job characteristics and number of hours telecommuting were examined.

H2a: Number of hours an employee spends telecommuting will correlate to Skill Variety, as measured by Job Diagnostic Survey, to a statistically significant degree.

Correlations hourstelecommuted

SkillVarietyAdd hourstelecommuted