North Carolina State University
2003-2004 Graduating Senior Survey:
Introduction, Methods, and Student Demographic Profile

Introduction

This series of reports presents findings from the 2003-2004 Graduating Senior Survey at NC State. Students from all nine undergraduate academic units who graduated in December 2003 or May 2004 (AY03-04) are included. Survey topics include: student goals and intentions; academic environment and faculty contributions; campus climate; academic and non-academic services for students; NC State's contribution to student knowledge, skills and personal development; student employment and involvement in extracurricular activities; and learning with technology.

This introductory report describes the survey�s methodology and provides a demographic profile of survey respondents in comparison to AY03-04 graduates. Specifically, it compares the gender, race/ethnicity¹, and academic unit of survey respondents with those characteristics of the AY03-04 graduates. A separate report, "2003-2004 Graduating Senior Survey: All Respondents," focuses on each individual survey topic. Tables with gender and racial/ethnic comparisons, as well as comparisons between colleges, are available on the web.

Survey Administration

Students planning to graduate in December 2003 or May 2004 were eligible to complete the Graduating Senior Survey. The web survey was available from the start of the fall semester in mid-August, 2003, until graduation in mid-May, 2004. Individual departments were responsible for informing their students about the survey and for following up with non-respondents. Departments typically connected the survey to the Application for Degree process. Those graduating with more than one major could complete a survey only for their primary major, as defined by university records. As in past years, several colleges and most academic departments took the opportunity to include a departmental "insert" survey along with the AY03-04 GSS to gather information more specific to their individual assessment needs. Students could complete an insert survey for each of their majors. Results from insert surveys are not available on the web, but are provided directly to the respective departments.

Respondents(Table 1)

Of the 3,684 students who graduated in Fall 2003 (N=1,282) or Spring 2004 (N=2,402), 64 percent (N=2,368) completed and returned usable surveys. Overall, spring graduates were more likely than fall graduates to complete the survey (68.9% vs. 55.7%).

The margin of error for the senior survey is low � under 1 percent (+0.7) at a 95 percent confidence interval. That is, if 77.4 percent of the respondents said they would choose NC State again, we can be 95 percent sure that the true figure would be between 78.1 percent (77.4 + 0.7) and 76.7 percent (77.4 - 0.7) if all graduating seniors had responded to the survey³. The margin of error increases as the sample size decreases, so statements for various subgroups, such as the separate figures reported for whites and African Americans, are less precise than statements based on the total sample.

Table 1: Response Rate and Margin of Error by Semester

Gender and Race/Ethnicity (Table 2)

Women are slightly overrepresented among survey respondents. Women make up 43.7 percent of the senior class population, compared to 47.6 percent of the survey respondents. There are no significant racial/ethnic differences between the AY03-04 graduating senior population and survey respondents. White students account for 83.0 percent of the senior population, 8.7 percent are African American, and 8.3 percent are other minorities. Among survey respondents, 83.7 percent are white, 7.6 percent African American, and 8.6 percent other minorities.

Table 2: Demographics of the Graduating Senior Class and Survey Respondents

Academic Unit (Table 3)

Table 3 shows enrollment of the AY03-04 graduating senior population and survey respondents by academic unit. Most colleges are accurately represented in the survey, with similar proportions of survey respondents and graduating class members. However, due to a low response rate, College of Design and College of Education graduates are underrepresented. In contrast, College of Humanities and Social Sciences and College of Textiles graduates are slightly overrepresented.

Table 3: Classification by Academic Unit

Analyses

The data obtained from survey respondents were analyzed using standard statistical methods². These reports attempt to provide a level of detail that makes the data more accessible and interpretable to the user. A primary purpose is to highlight patterns found in responses to related question items or between comparison groups. Such consistencies among items or between groups are usually more important for understanding the data than are the sizes of the differences between individual pairs of ratings or ranks or, to some extent, whether the differences are statistically significant. While some individual small differences might actually be statistically significant, they may not be substantively meaningful. On the other hand, when even relatively small differences yield consistent patterns within a similar series of questions, the results are potentially more telling.

Endnotes:
1. The term "racial/ethnic" is used throughout these reports to recognize the potentially blurred distinction between the individual terms. In application materials students were requested to identify themselves using the following categories: Caucasian, African American or Black (not of Hispanic origin), Native American Indian or Alaskan, Asian or Pacific Islander, or Hispanic (Mexican, Puerto Rican, Cuban, Central or South American, or other Spanish origin or culture, regardless of race). For analysis purposes, these categories were collapsed into "White," "African American," and "other minorities."(back)
2. In analyses not presented in these reports, responses were tested to determine whether there were significant differences between women and men, between white, African American, and other minority students, and between different colleges. Questions requiring categorical responses were analyzed with chi-square tests, and questions with numerically coded responses were analyzed with either T-tests or one-way analysis of variance (ANOVA with Tukey's multiple comparison procedure. Complete results are available from University Planning and Analysis upon request. (back)
3. A 95 percent confidence interval denotes the range of values which contains the true population value in 95 out of 100 possible samples of the graduating senior population. Additionally, margins of error for individual questions are usually smaller than the given margin of error for the sample because the variance of proportions in each survey item is almost always less than the 50/50 figure used in calculating the confidence interval. Thus, the margin of error given is conservative. (back)