4.1 RESEARCH METHODOLOGY
4.1.8 Research Methodology
This study utilized evaluative survey research in the methodology (Salant & Dillman, 1994). A descriptive 60-item questionnaire was utilized [See, Appendix B].
4.1.8.1 Subjects
This study was principally concerned with the perceptions of admissions personnel regarding access to baccalaureate degree study. Vice-Presidents/Directors’ of Enrollment Management, Dean of Admission, Associate Directors of Admission, and ACT 101 Directors’ were surveyed regarding minority students’ access to four-year institutions of higher education in the state of Pennsylvania. These were subjects of the research. Several ACT 101 Directors were targeted for the study because these personnel possess the institutional authority to recommend admission for minority, first-generation and economically disadvantaged students at participating Pennsylvania four-year institutions. Their decisions to admit are based on the special admissions criteria that are approved by the State of Pennsylvania Legislature under ACT 101. Thus, all of the aforementioned administrative personnel directly or indirectly influence decisions on admission or determine institutional policy on four-year access. Each subject in the study shared these characteristics.
4.1.8.2 Sampling Methods
Several methods were used to access the sample and compile the data for this study: 1. The sample frame of 108 institutions was compiled from information listed on the
Pennsylvania Department of Education (PDE) Website under the caption Educational Names and Addresses (EdNA). This web site also lists each institution’s (e.g., http://www.[XYX].edu.) Web-address.
2. The sampling frame consisted of the 106 four-year public and private baccalaureate degree-granting institutions of higher education and two community colleges in the state of Pennsylvania.
Admissions, Directors of Enrollment Management, and ACT 101 Directors’ in Pennsylvania’s 106 four-year baccalaureates degree-granting institutions of higher education and two community colleges.
4. The list frame for the study was developed from the EdNA on the PDE Website. The names of the sample units were extracted from each institution’s Website
5. A Web-based questionnaire was developed and sent to the target group via email and returned electronically [See, Appendices B, and C].
6. Several institutions had the opportunity to provide one or more responses to the questionnaire based on the targeted subjects of the study.
7. The researcher spent the period of four years [i.e., 2001-05] interviewing and compiling personal business cards from the target population. This was achieved by attending regional College Fairs and various professional higher education Conferences within the state of Pennsylvania. This information was utilized to assist in the development of the List Frame for the survey study.
A convenience sample (i.e., non-probability) survey method was utilized for the survey (Salant & Dillman, 1994). In approximately one-fifth of the 106 Pennsylvania four-year institutions of higher education, the administrative position of Vice-President/Director of Enrollment Management was present. This number was approximately 25 positions. At several institutions in the study, the Vice-President of Enrollment Management also served as the Director of Admissions and was the chief administrator of the Admissions Responsibility Unit. These personnel were the administrative supervisors of Admissions Directors, or served in place of an Admissions Director. Directors of Enrollment Management are often responsible for institutional policy-making on entry decisions.
4.1.8.3 External Validity
The principle of purposive sampling was validated in this study by including a representative sample of admissions personnel from the 108 four-year private and public institutions throughout the state of Pennsylvania (Salant & Dillman, 1994). In this regard, external validity was satisfied in this study based on the geographic representation of the participating institutions [See, Appendix I].
In addition, to secure external validity, it was important in the study to include a representative number of both four-year private and four-year public schools by institutional type (e.g., Research I & II; Private four-year) for the accessible sample frame. This procedure was necessary in order to establish a salient level of external validity in the research results.
4.1.8.4 Internal Validity
Survey-based studies must prevent internal threats in the research design and promote precision by assuring that the instrument is pre-tested and meets the demands for accurately measuring the desired phenomena (Salant & Dillman, 1994; Thomas, 2004). The survey instrument in this study was pre-tested extensively to establish internal consistency and reliability (Isaac & Michael, 1994; Thomas, 2004). The survey instrument was also pre-tested and adjusted for internal stability because it was the principal measurement device in this quantitative-focused study (Thomas, 2004). The survey instrument consisted of a self-developed questionnaire; therefore, the construct validity of the survey items was also an important element for reliability in the research design (Isaac & Michael, 1994).
In this regard, it was important that the survey items in the instrument were developed from the literature review to align with the five research questions, assure content validity, and accurately measure the phenomena under study for the research problem. According to Dillman
(2000), well-designed and pre-tested questions prevent extensive levels of measurement error in the survey.
4.1.8.5 Instrumentation
This study utilized Likert (1932, 1967) scaling methods to measure variables. Means, percentages, and data summations were utilized to calculate and measure the strength of the responses. One of the important elements in developing and pre-testing the Web-survey instrument was to assure that the questionnaire did not “lead the respondent(s)” (Salant & Dillman, 1994) and assure item objectivity (Isaac & Michael, 1994; Eichelberger, 1989). Pre- testing was utilized extensively to fine-tune the questionnaire, correct grammatical issues with the wording of the questions, and to adjust the ‘timing’ of the instrument.
In the questionnaire design, closed-ended, forced choice survey items were utilized (Salant & Dillman, 1994). The Likert-type scaling procedures incorporated a five choice response set in addition to a No Opinion and a Does Not Apply choice set for each question. In the Web-survey design, the five choice response set was intended to measure the degree of disagreement-to-agreement with each statement: For example;
(1) = Low/Negative---to--- (5) = High/Positive – [A] No Opinion – [B] Does Not Apply Disagreement---Agreement 1--Strongly Disagree 2--Disagree 3--Neutral 4--Agree 5--Strongly Agree A—No Opinion B—Does Not Apply
The Web-based questionnaire used in this study was designed in such a way that all of the survey questions had to be completed before the instrument could be submitted
electronically. However, each respondent also had the option to respond to any question/variable through the [A] No Opinion, or [B} DNA variables.
In this research design, the items that were outlined in the following survey question matrices were formatted to address Research Questions 1 through 5. It was important to distinguish when the survey questions were soliciting personal opinion, and when the survey items were requesting the respondents to answer on behalf of their institution. The disaggregated information was then used to segment institutional policy questions from personal perceptions in order to gain clearer insight to the structural and administrative aspects of higher education access.
Concerning the research methods for the study, a Survey Question Matrix [See, Table 4.1] was developed to distinguish the ‘type’ of questions that were presented in the survey. The survey questions consisted of three types; 1) demographic information, 2) questions that required a personal response and, 3) items that asked the respondents to answer on behalf of their institutions regarding access policy.
Table 4.1 Survey Question Matrix
Type of Survey Questions Institutional Information Opinion/Perceptions (#) Institutional/Policy (#) Research Question(s)
Survey Items A and B
1. Access (1-12) 5, 6, 7, 8, 9, 12 1, 2, 3, 4, 10, 11 2. Preparation (13-24) 13, 17, 18, 19, 20, 21, 22, 23, 24 14, 15, 16 3. Admissibility (25-34) 25, 27, 28 26. 29, 30, 31, 32, 33, 34 4. Affordability (35-45) 35, 36, 40, 41, 42, 43,44 37, 38, 39, 45
Type of Survey Questions Institutional Information Opinion/Perceptions (#) Institutional/Policy (#) Research Question(s)
Survey Items A and B
5. Legal Issues (46-55) 47, 49, 50, 51, 52, 53, 54, 55 46, 48 Summary Question 56, Preparation [A]
Affordability [B]
Demographic Data Set C, D, E, & F [Descriptive Information]