A survey of faculty users was chosen as the method to study current use of network communications. This method is appropriate because the nature of this study relies on candid responses from a substantial number of faculty subjects of diverse statuses. This can be achieved due to the relatively small demands of time and effort a short, pre-tested, self-administering survey has on the part of faculty who tend to have busy schedules. In this study, the need for basic, self-descriptive information from faculty about their use of network communications, perceptions of risk, and statuses is satisfied by the survey method and does not require other methods such as a controlled experimental setting or the intimacy of face-to-face field work (Anderson and Meyer, 1988).
The setting selected for the study is a large, eastern university with a national reputation as a research institution. Its large size, number of schools and students, and diversity of faculty with their academic statuses satisfies the need for data in this study and is typical of universities across the country. This setting is also typical of universities where research conducted by college faculty is associated with their organizational role. This academic setting also contains a variety of departments and disciplines and exhibits a diversity of organizational and non-organizational statuses of its faculty.
The choice of faculty from within one large university setting rather than several settings was based on the desire to control for many differences between university settings which may make the testing of the hypotheses more difficult (e.g., the stage of implementation, diversity of faculty, and access to the Internet). Added to this are the researcher's knowledge and experience with the setting and the innovation as it is implemented, and the reduction of problems associated with subject access and rapport.
Subjects will be selected from a sampling frame that is entirely composed of full-time faculty who are either tenured or non-tenured (but tenure-track). Only full-time faculty are subject to the varied statuses and perquisites that will be studied. In this setting, all full-time faculty are automatically issued accounts that allow computer and network access. This satisfies the criteria for the environmental factor of ubiquity and access (Fullan, 1991) and the likelihood of response when using e-mail as a method of collecting raw survey data (Kiesler and Sproull, 1986). As of the academic year concurrent with the study, this list is composed of 1,447 faculty that meet these criteria.
Faculty from four schools from this university will be used to develop a stratified sampling frame (n=948). These schools include a college of arts and sciences (n=483), a school of education (n=188), a business school (n=202), and a school of mathematics and computer science (n=75). These particular schools were chosen because they collectively represent the diverse statuses of academic specialty and computer-based disciplines that are being studied.
A stratified, disproportionate random sample will be used to obtain an equal number of faculty from each school and thus preserve a relatively equal representation of the statuses being studied. This type of random sampling has been shown to be advantageous in allowing a smaller, more manageable sample size as long as the categories are stratified by few variables and the classifications closely follow those variables being studied Burroughs, 1971; Anderson and Meyer, 1988). These criteria are satisfied in this study by stratifying the sample based on two status comparisons which are collectively represented in the academic specialties of these schools. The sample size will be based on the school with the smallest number of faculty (n=75). Drawing a disproportionately random sample of 75 faculty from within each of the four schools provides a total sample of 300 faculty.
A randomly chosen sample size of 300 should produce a desired accuracy in the range of +/- .05. This sample size is needed to assure with 95% confidence that the sample proportion will be within +/- .05 of the population proportion (Reaves, 1992).
The survey will be conducted using both a printed and an e-mail version of a self-administered questionnaire. These as well as a cover letter, an e-mail instruction sheet, and follow-up letters are found in appendices at the end of this proposal.
The survey instrument employs a self-administered questionnaire constructed for this study. It is primarily composed of close-ended items that have been pre-tested for clarity and comparison.
Both print and e-mail versions of the questionnaire contain the same content, but only vary in the instructions for indicating subject choice to response category. The print version instructs the subject to circle the number of the chosen response category and the e-mail version instructs the subject to type an X next to the chosen response category.
Besides the advantages in bi-modal data collection that offers a choice of print or e-mail for convenience and self-preference (Kiesler and Sproull, 1986), this study is interested in obtaining descriptive data about how faculty members in this setting respond to the use of network communications for online surveys. A review of studies using online (e-mail) surveys indicates that they are becoming more feasible as greater numbers of faculty are beginning to use network communications. Besides the speed and cost-savings of this data collection method, there is a tendency for e-mail to elicit more candid self-disclosure on the part of online survey subjects (Sproull and Kiesler, 1991).
A field study of seven faculty from the proposed population was conducted in the spring of 1995 to assess the suitability of this setting and to test the first drafts of the questionnaire and gather ideas for modifications to it. This field study employed several methods of data collection: observations of faculty subjects in their use of network communications; a draft questionnaire based on these observations and prior reading of the research literature; and follow-up phone, e-mail, and face-to-face interviews with participating faculty. Another preliminary questionnaire was subsequently developed and administered via e-mail among five of the subjects who were self-acknowledged network communications users.
Findings showed differential responses both among respondents and across scholarly activities. The e-mail questionnaire confirmed the perceived presence of access to technology, clarity of goals, and other factors considered by researchers (Fullan, 1991) to support the implementation of an innovation. Research on planned change in education indicates that support for using innovations is an important component in managing change (Seashore and Miles, 1990). Another finding confirmed results of a similar study (D'Souza, 1992) in which faculty using e-mail for scholarly communications indicated the greatest change in the number and frequency of contacts outside the university.
Once the questionnaire had taken shape, two more field tests were conducted using faculty subjects of different statuses from the proposed setting and one faculty subject from another large university. The first group of three suggested changes which are noted below. These changes were then made and the second group of three examined the updated questionnaire and called for no further changes. A cover letter to subjects asked them to provide comments that address any problems that they felt might be present, including the following areas: clarity, appearance, ease of use, item order, and wording. Changes made as a result of this field testing include adding a brief justification of the study in the cover letter, and slight changes in the wording of several questionnaire items (as compared by the samples listed in the Appendix of this proposal).Also, the definition of "scholarly work or communications" was added below the Section 1 heading as follows: "The term, scholarly work or communications, can include contact with colleagues or other scholars, gathering support for scholarly work, conducting research, and publishing or disseminating knowledge."
Clarity in the e-mail instruction sheet was improved through the use of bold styling, instructions to use the Reply command in the e-mail program before inserting responses in the questionnaire, and also using the Save command to make a copy of the questionnaire.
This section of the proposal describes how the major variables are operationalized in the survey questionnaire. The questionnaire is divided into three sections. Each section tests one of the major variables of the causal model used in this study.
The first section of the questionnaire measures the use of network communications, the dependent variable in this study. The level, type, frequency, and changes in the use of network communications are measured as follows: overall use of networks versus traditional media for scholarly communications, the degree of use at work and/or at home, the frequency of use by application types of network communications, and how long it has been used on a regular basis. Lastly, this section measures how much current communications involves network communications compared to traditional forms in each of four areas of scholarly activities: contacting colleagues and other scholars, gathering support for scholarly work, conducting research, and disseminating or publishing results of scholarly work.
The second section of the questionnaire measures perceptions of perquisites as the study's set of intervening variables.
These include achieving tenure, achieving promotions, receiving pay raises, gaining professional recognition, and pursuing scholarly interests. Corresponding perceptions of perks in each of the four main areas of scholarly activities are also measured in each of the four questions in this section. These closed-ended questions use a five-point scale (from very disadvantageous to very advantageous) with the additional response category of "Not Applicable."
The third section measures the various statuses associated with the study's independent variables and with other organizational and non-organizational variables that have been shown in similar studies to be related to implementation of educational innovations. Organizational variables include rank, tenure, years of teaching experience, field of study (discipline), administrative and/or committee experience, and scholarly publication experience. Non-organizational variables include age and gender. These and other questions in this section also provide demographics and background information, such as scholarly publications output in traditional and online venues. Some of these questions determine the statuses used in status comparisons such as tenure versus non-tenure, hard versus soft science, and computer versus non-computer related specialty.
Data collection procedures will conform to those found by Dillman (1978) to be most effective in obtaining responses. These procedures include the use of a cover letter and follow-up letters that are carefully worded and timed in a manner that supports the completion and return of the questionnaire. Based on previous surveys of computer network users (Kiesler and Sproull, 1986), this survey employs a bi-modal approach to data collection: offering a print version of the questionnaire, but allowing subjects the option for completing and returning an e-mail version of the questionnaire (which is identical in content and highly similar in format).
The printed version of the questionnaire, a cover letter, and a separate page of instructions for subjects using the e-mail version will be delivered via interdepartmental mail to each of the full-time faculty subjects. In the cover letter, the subject is informed that an e-mail version of the questionnaire had been sent within a few days after the mailing of the print questionnaire. The subject can elect to use either the print or the e-mail version. One week after the initial mailing, a follow-up postcard will be mailed to all subjects as a reminder to respond for anyone who has not yet responded. Two weeks after mailing the follow-up postcard, a second follow-up in the form of a cover letter and questionnaire will be mailed to all subjects to remind them to respond or to use the enclosed questionnaire in case they have misplaced the original. This second cover letter also informs subjects that an e-mail version of the questionnaire has been sent. Finally, a third follow-up letter will be mailed during the seventh week after the initial mailing as a reminder to complete and return the questionnaire and also to ask why the subject has not responded.
To conform to the Human Subjects Committee guidelines for "Exempt From Review" status (and thus, not requiring the subjects' signature on consent forms), the following procedures will be followed in this study. The researcher will be the only person administering the survey and analyzing the data and no other person will be allowed to view the returned questionnaires or the survey mailing list. Additionally, subjects will be informed in the cover letter to the survey questionnaire that their responses will be only represented by an identification number and that their name will not (nor should not) be present on the processed questionnaire. They will also be informed that the data from their responses will be used only in the aggregate.
To provide confidentiality, the cover letter will include a unique ID number which will be used with whatever version of the questionnaire (print or e-mail) the a subject chooses. On the separate e-mail instruction page, the subject will be informed that the completed e-mail questionnaire can be returned in either of two ways that will separate personal information from the questionnaire. The subject can send the completed questionnaire via e-mail as an "attached" file or send the completed e-mail questionnaire back as an e-mail message with the letters fs in the Subject line of the mail header. Both of these methods ensure that the questionnaire will be stored in a separate file area upon receipt where any name or e-mail address will be stripped and deleted before processing the data.
The ID code on each returned questionnaire will be removed once the respondent's name is checked off. When the deadline for all returned questionnaires is reached and no further processing is performed, the master mailing list will also be destroyed. This procedure permanently erases all connection between subjects and their responses. The researcher will obtain the approval of the Committee on Human Subjects prior to conducting the proposed survey.
Selection of the University Setting
Selection of Faculty Subjects within the University Setting
Instrumentation
Survey instrument
Preliminary Field Study
Operationalization of Variables
Dependent Variable
Intervening Variables
Independent Variables
Data Collection Procedures
Protection of Human Subjects' Rights
Data Processing