2. REVIEW OF TELECOMMUTING LITERATURE

2.1 Advantages and Disadvantages

If, according to some predictions, 7.5 and 15 million workers will be telecommuting by 2002, the effects of telecommuting are likely to make a significant impact (USDOT 1993). It is important to remember that these impacts, both positive and negative, will vary based on the frequency of telecommuting (from once a month to five days a week) (Mokhtarian 1991). Although many telecommuting plans are driven by the needs of the transportation system, it is the employees and the employers who are directly affected by the change in work location. Thus, the advantages and disadvantages can be categorized according to three groups: the employees, the employers, and the transportation system.

Employees who telecommute are likely to accrue many benefits from this work arrangement (see Table 2-1). Most importantly for some, the time, expense, and stress associated with the daily commute is reduced. Telecommuting allows employees to schedule their own work hours in order to accommodate activities that take place during the normal workday. Also, the work environment is often improved; there are fewer distractions from co-workers, and the employee can modify the alternate workplace to suit his or her needs. Since the supervisor is no longer present while work is done, telecommuting often results in greater job autonomy and may change performance evaluation procedures to focus on work output instead of the appearance of working. Additionally, mobility-impaired and handicapped persons can greatly benefit from telecommuting by allowing them to be gainfully employed when they would otherwise be excluded from consideration (USDOT 1993). Finally, since telecommuting shortens (or eliminates) the distance between home and work, the time formally spent commuting could be spent with family or friends instead. Depending on the extent to which these benefits apply to a particular employee, there may be a corresponding increase in the employee's job satisfaction (SCAG 1985).

Although there are many possible benefits to telecommuting, there are also some possible negative impacts of this work arrangement (see Table 2-1). One disadvantage is the professional and social isolation from the regular workplace that the employee endures while telecommuting. Absence from the main workplace is also perceived to reduce chances for career advancement since the management cannot directly observe the contributions of the employee. Lack of direct observation may also affect performance evaluations by the supervisor. Difficulty in completing tasks may occur if the services needed are not present at the alternate work site. The changes associated with the telecommuting option may affect performance or cause overworking. In addition, telecommuting often removes the commute boundary between work and home which can be an important psychological factor in the motivation to work. The transfer of the workplace to the home may lead to loss of benefits (for example, worker's compensation applying only at the regular workplace) (USDOT 1993). Lastly, conflicts with household members over home activities may result from telecommuting (Bernardino, et al. 1993). Thus, the preference for telecommuting will likely depend on how important the underlying factors of the advantages and disadvantages are to the particular employee.

The employer, the other important player, must also see some advantages to telecommuting before the organization can be persuaded into the adoption of a telecommuting program (see Table 2-2). The most often mentioned benefit for employers is increased quantity and quality of work due to improvements in employee job satisfaction. Important cost savings can accrue from reduced overhead for office space and a reduction in absenteeism and sick leave. Telecommuting is also seen as an important benefit for employees, so they are more likely to accept a position with or remain an employee of the organization. Improved retention can also save the organization money spent on the recruitment, relocation, and training of new employees. In addition, changes in employee scheduling can be handled more effectively since employees can work from home during odd hours and in remote locations. Telecommuting has been used effectively in emergency situations, such as after the Loma Prieta earthquake in the San Francisco Bay area, when the regular workplace is not easily accessible to the employees (USDOT 1993). Finally, implementing a telecommuting program can fulfill some requirements of clean air mandates that require employers to reduce the pollution caused by its employees during their commutes (SCAG 1985).

Employers usually have many concerns about a telecommuting program (see Table 2-2). The management may hold a negative attitude towards offering alternative work arrangements which acts as a barrier to the introduction of telecommuting. The primary objection to telecommuting comes from its requirement of remote supervision; many supervisors feel uncomfortable if they cannot observe the employee at work. The security of proprietary information may be at risk if the employee is working outside the regular workplace. The costs associated with implementing and operating a telecommuting program may be prohibitive since duplicate equipment may be required at the alternate workplace. Additionally, the employer may be concerned about being held liable for injuries that are not work-related if they occur at the home during telecommuting work hours. Finally, employee unions may oppose certain telecommuting plans if they feel the worker is not adequately protected (USDOT 1993).

The effect of telecommuting on the transportation system is potentially significant in terms of travel reduction (see Table 2-3). Specifically, the elimination (or shortening) of the commute trip will reduce the amount of peak-period travel. Additionally, empirical studies have shown an overall reduction in the number of trips on telecommuting days and a shift in trip-making to off-peak periods (USDOT 1993). Since the employee stays close to home, any trips made will more likely be short and, consequently, more likely to use non-polluting travel modes (such as walking or bicycling). Unfortunately, the establishment of a telecommuting arrangement can disrupt rideshare groups containing other employees who do not share the same telecommuting schedule (SCAG 1985). Although the reduction in peak-period travel may result in short-term congestion mitigation, the latent demand for travel on congested roadways will replace the reduced travel over the long run. To the extent that travel is reduced (whether relatively or absolutely), comparable fuel savings, emissions reductions, and improved traffic safety should result.

2.2 Forms of Telecommuting

Telecommuting from a center can mitigate many of the disadvantages of telecommuting from home. The trip to the center, although shorter than the commute to the regular workplace, still provides an important barrier between work and home. Noise and interruptions from household members may limit the ability to telecommute; however, a center would be free of these distractions. In addition, the center offers the opportunity for social and professional interaction. For the employer, the center has a more professional image. Telecommuting costs may be reduced at the center through the sharing of equipment and easier definition and control of employer liability. Importantly, the supervisor may feel more confident that the employee is actually working. Although telecommuting from a center does include more travel per trip than telecommuting from home, working from the telecenter may still provide a greater transportation benefit if it is used more frequently (Bagley, et al. 1994). Even though home-based telecommuting has been the most popular to date, it is unclear from empirical studies which form of telecommuting will become the most popular.

2.3 Telecommuting Decision Theory

As mentioned above, drives and constraints are important factors in the decision to telecommute. Drives motivate the individual to seek a change in behavior and are rooted in the individual's lifestyle and stage in the life cycle. Mokhtarian and Salomon (1994a) have identified five drives: work, family, independence (or leisure), ideology, and travel. The work-related drive includes the desires to work more, reduce work-related stress, and gain independence from supervisor control. The family and leisure drives are similar: the first reflects the aspiration to spend more time with family members, while the latter shows a desire to spend more time on oneself. Some people may be driven to supporting certain causes, such as environmental awareness, which constitutes the ideology drive. Finally, the travel drive incorporates the perception of having a difficult commute and difficulties associated with travel limitations. Again, these drives affect not only the preference to telecommute but also the attitudes and perceptions of the individual.

Constraints, on the other hand, are distinct from drives in that they inhibit the preference or choice of a particular alternative rather than provide the impetus to consider the alternative. In other words, the absence of constraints does not necessarily result in telecommuting since the individual must also have a reason driving him/her to consider telecommuting as a work option. Constraints can also act in the opposite sense as facilitators, where the presence of the factor encourages, or facilitates, the decision for an alternative. In addition, some constraints are strict yes-or-no factors while others vary continuously. Finally, the constraints on an alternative can be external, caused by the work environment, or internal, brought about by psychological considerations (Mokhtarian and Salomon 1994a).

Mokhtarian and Salomon (1994a) have identified three external (awareness, organization, and job) and one internal (psycho-social) types of constraints that may affect the decision to telecommute. The awareness-related constraints are the lack of knowledge of telecommuting as a possible alternative and misunderstanding about the telecommuting concept. The organization may also constrain the decision to telecommute through the lack of employer support and supervisor disapproval. The job itself may restrict the decision to telecommute if it is not suitable, requires high technology, or is too expensive to implement. Finally, internal constraints may prevent the individual from considering telecommuting as a viable alternative. These constraints include the need for personal interaction, conflicts with household members, lack of self-discipline, fear of demanding special treatment from the employer, and the perception of the commute as beneficial separator between work and home life.

2.4 Preference Modeling Efforts

Although a few studies have investigated the employer's decision to offer telecommuting (Bernardino and Ben-Akiva 1994, Yen and Mahmassani 1994), research to date has focused on the employee's adoption of telecommuting. Early studies concentrated on identifying the advantages and disadvantages of telecommuting in the areas of work, household, and travel. In one such study, DeSanctis (1984) surveyed employees and managers at a computer services firm concerning their attitudes about telecommuting. The study identifies the benefits and problems of telecommuting from both viewpoints and suggests that manager unwillingness is a major constraint to the adoption process. In another study, Hartman, et al. (1991) investigated telecommuting satisfaction and productivity at eleven organizations. Telecommuting satisfaction was found to be positively correlated with supervisor support and negatively correlated with family disruptions.

The study by Bernardino, et al. (1993) is an example of a modeling effort using stated preference. The respondents (N = 54) were recruited from USENET newsgroups that discussed telecommunications applications. The researchers used an ordinal probit model to characterize the preference to telecommute under scenarios with different levels of schedule flexibility, equipment and services costs, and salary. The most significant results showed that the preference to telecommute decreased if the employee was required to provide the equipment and the salary decreased, and the preference increased for the presence of children in the household.

A series of studies conducted at the University of Texas, Austin used data from a total of 72 organizations located in 3 Texas cities. In an exploratory analysis, Mahmassani, et al. (1993) performed cross-tabulations between both attitudes about telecommuting and the stated preference to telecommute with sociodemographic variables. Results from the attitude cross-tabulations revealed that females were more likely than males to feel that telecommuting would improve household relations and improve the quantity of work done, higher educated respondents felt the work at the office was more important, and the presence of children also increased the likelihood of respondents feeling that telecommuting would improve household relations. The stated preference to telecommute was stronger for those who were female, had computer skills, had children, had long commute times, and had prior experience with telecommuting.

The second study, by Sullivan, et al. (1993), used a multinomial logit specification to model the stated preference to telecommute full-time, telecommute part-time, possibly telecommute, or not telecommute at all. The results of the model included the significant variables from the exploratory analysis as well as short length of time at employer, less face-to-face communication, end work before 5:30 pm, married, high income, and below age 50. The multinomial logit models for the data from each city were found to be equivalent to the model using the pooled data.

Most recently, Yen and Mahmassani (1994) have used a generalized ordinal probit specification to model the stated preference based on both factor-analyzed attitudinal variables as well as previously studied sociodemographic variables. The researchers found similar sociodemographic variables (children, computers at home, computer skills, face-to- face communication, and commute distance) to be significant. Additionally, attitudes about job suitability, family effects, and social interactions were significant explanatory variables.

Research at the University of California, Davis has investigated telecommuting as a work option in six departments of the City of San Diego (N = 626). Mokhtarian and Salomon (1994c) used factor analysis to create variables for attitudes about the workplace and life in general. Both the factor-analyzed variables and sociodemographic variables were included in a binary logit model of the preference to telecommute. Although the stated preference procedure (specifying different scenarios) was not used, the respondents were assumed to gain understanding about the effects of telecommuting from the survey instrument if they did not already possess that information. Respondents were found to prefer telecommuting from home in order to have disability / parental leave opportunities, reduce stress, receive personal benefits, and reduce commute stress. Longer commute times, job suitability (as measured by the amount of telecommuting allowed by the position), and the lack of household distractions also improved the likelihood of preferring to telecommute. However, if workplace interactions were important and / or the commute was seen as a beneficial separator between home and work, the respondent was less likely to prefer to telecommute. The attitudinal variables were found to be important in the decision to telecommute; six of the nine significant variables were constructed using factor analysis. Further research using the same data set is currently underway to model the preference to telecommute from a center and to model preference for home-based versus center-based telecommuting (Bagley 1995).

Mokhtarian and Salomon (1995) have also modeled the revealed preference, or choice, to telecommute from home using a binary logit specification. Thirteen percent of the respondents were currently telecommuting for at least some portion of the work day. Variables for the number of hours of overtime worked and the factor for commute stress increased the likelihood of the employee to choose telecommuting. Two other factors, misunderstanding of telecommuting and office discipline, decreased the chance of choosing telecommuting. Other significant variables that discouraged the selection of telecommuting included lack of manager support, job unsuitability, and an index for needed equipment and services.

The findings of the above modeling efforts are incorporated throughout the present study. The survey instrument (described in section 3) includes the attributes identified by other studies that distinguish preferrers and non-preferrers of telecommuting. In addition, the survey has questions on both center-based and home-based telecommuting in order to distinguish between the two forms and compare with previous studies of home-based telecommuting. Also, the survey measures the respondent's attitudes about job satisfaction and various work characteristics. These attitudes are factor analyzed (section 4), and then the resulting factors are used as independent variables in the models (section 5) since attitudinal variables were found to be important discriminators in studies by Mokhtarian and Salomon (1994c, 1995) and Yen and Mahmassani (1994). Although the control groups members have chosen their telecommuting status, the center-based telecommuters were surveyed prior to the start of telecommuting, and as a result, the models use preference as the dependent variable like the models in most of the other studies.

Return to the Abstracts of Non-Refereed Publications page.
Return to the Main page.