Each year the National Survey of Student Engagement (NSSE) collects information from undergraduates at four-year colleges and universities across
the country to assess the extent to which students engage in a variety of
effective educational practices. The NSSE project is grounded in the proposition that student engagement, the frequency with which students participate in activities that represent effective educational practice, is a meaningful proxy for collegiate quality. Launched with a generous grant from The Pew Charitable Trusts, the annual survey is now supported by institutional participation fees. NSSE is cosponsored by The Carnegie Foundation for the Advancement of Teaching and the Pew Forum on Undergraduate Learning.

This overview is divided into several key sections. First, we compare the characteristics of participating institutions and students with institutional and national profiles as well as provide general information on overall response rates. In the second section we present selected findings, including descriptive information about the students who completed the survey and preliminary analyses of patterns of engagement among various groups of students. Finally, we provide suggestions for interpreting the data presented in this report.

Later this fall you will receive national benchmarks of effective educational practice as well as benchmarks for your institution. This information will be based on the aggregated data from 731 different colleges and universities that have participated in NSSE since 2000.

About 348,000 first-year and senior students were included in the NSSE 2003 sample.1 These students were randomly selected from data files provided by 437 participating four-year colleges and universities. A list of these institutions is available in the "Additional Information" tab of the institutional report. NSSE sampling procedures call for sending the survey to an equal number of first-year and senior students with the standard sample size determined by the number of undergraduate students enrolled at the institution. Students at the majority of colleges and universities (73% or 316 schools) had the option of responding either via a traditional paper questionnaire or via the World Wide Web. One-hundred and nineteen (27%) schools opted to be Web-only institutions where students received an introduction letter through the mail and all further contact electronically.

Tables 1 and 2 below show that NSSE 2003 participating institutions and respondents approximate the characteristics of students enrolled at participating schools as well as the national profile of all four-year colleges and universities. The source of the comparative data is the 1999-2000 Integrated Postsecondary Education Data System (IPEDS) database, the most recent complete data file available. However, the IPEDS data are three years old so the comparisons may not accurately reflect certain institutional and student characteristics for the 2002-2003 academic year.

Profile of NSSE 2003 Respondents

Table 2 shows selected characteristics of the students who completed The College Student Report in 2003. The first column represents students who responded to the NSSE survey in 2003. The second column shows the characteristics of students at four-year schools that participated in NSSE 2003, as reflected by 1999-2000 IPEDS data. The third column represents the national profile of students at all four-year colleges and universities from IPEDS data.

Year in School

The sample was equally divided between first-year (50%) and senior (50%) students.

Gender

Women made up two-thirds (66%) of the respondents compared with 55% of the students enrolled at NSSE 2003 schools and 58% nationally (Table 2). The larger proportion of women respondents is consistent with the widely reported survey research findings that women are more likely than men to return questionnaires.

Age

Students 19 years of age or younger compose the largest group (45%), reflecting the fact that half the students selected to receive the survey were in their first year of college. About 37% of respondents were 20-23, 8% were between the ages of 24 and 29, and 10% were 30 years of age or older.

Race and Ethnicity

White, Asian/Pacific Islander, and American Indian/Alaska Native students are slightly over-represented and African American students are slightly underrepresented (Table 2).

Living Arrangements

Forty-five percent of all students lived in campus housing (70% of first-year students, 21% of seniors). The remainder lived within driving distance (42%), within walking distance (12%), or in a fraternity or sorority house (1%).

Fraternity or Sorority

Thirteen percent of men and 11% of women were members of a social fraternity or sorority.

Grades

Just over 41% of all students reported that they have earned mostly A grades. Only 3% of students reported earning mostly C?s or lower.

Parents? Education

Thirty-two percent of all respondents were first-generation college students. Almost two-fifths (39%) had parents who both graduated from college.

Enrollment Status

About 89% of all students were enrolled full-time (Table 2). Approximately 36% of all students attended one or more other institutions in addition to the one at which they were currently enrolled. Of this group of multiple-institution attendees, 15% went to another four-year college, 20% to a community college, 5% to a vocational-technical school, and 4% to some other form of postsecondary education.

Primary Major Field

Table 3 shows the percentages of students majoring in different fields broken down by class and gender. More men are majoring in business, engineering, and physical sciences, while more women are pursuing degrees in education, professional schools, and the social sciences.

Response Rates

The average institutional response rate for NSSE 2003 was 43%. 2 The average institutional response rate for paper schools (institutions where students had the option of completing either the paper or the Web version of The College Student Report) was 43%, with a range of 14% to 70% across schools. The average institutional response rate for NSSE 2003 Web-only schools (institutions where students only had the option of completing the survey online) was 44%, with a range of 7% to 78% across schools. About 48% of the NSSE 2003 respondents completed the paper version of The College Student Report and approximately 52% completed it using the Web. Additional information about response rates, including the response rate for your institution, can be found under the Respondent Characteristics tab of the institutional report.

This section is divided into two parts. The first part presents a general view of the nature and frequency of undergraduate student engagement in effective educational practices. The second part briefly summarizes the results from a series of regression analyses examining the levels of engagement of different groups of students, controlling for various student characteristics and institutional factors such as selectivity and sector.

College Activities

Page 1 of The Report includes questions about the nature of the activities in which students engage. A "substantial amount" of engagement is defined to be at least 50% of all students reporting "often" or "very often" (Table 4).

The least frequent activities are those where the percentage of students who responded "never" exceeded 35%, meaning that roughly one third or more of the students had no experiences in these areas during the 2002-2003 academic year (Table 4).

Course Emphasis and Educational Programs

Another way to gain insight into the student experience is to look at the kinds of intellectual and mental activities that institutions emphasize and the types of educational programs in which students take part that complement and enrich their collegiate experience.

• Nearly 80% of seniors said their classes, to a substantial degree, emphasized applying theories or concepts to practical problems (combination of "quite a bit" and "very much" responses).
• More than four-fifths (86%) of seniors said their classes emphasized analyzing ideas or situations.
• Seven of ten seniors completed an internship or other type of field experience.
• About one-quarter of seniors (27%) worked on a research project with a faculty member outside of course or program requirements.
• About 41% of seniors took foreign language coursework.
• One-fifth (18%) of seniors studied abroad.

Community Service and Volunteerism

Two thirds of seniors (66%) did community service or volunteer work during college. Students who belong to Greek organizations were more likely than their non-member peers to perform a service activity. In addition, transfer and older students were less likely to engage in community service than their nontransfer or traditional-age peers. We also found that students who live on or near campus are more engaged in volunteer work than their peers who drive to campus.

Integration of Knowledge and Experience

Deep learning requires the acquisition of knowledge, skills, and competencies across a variety of academic and social activities and integration of these diverse experiences into a meaningful whole. To estimate the degree to which students take part in activities that provide opportunities to integrate their curricular and co-curricular experiences, we created an integration scale composed of six NSSE questions (1d, 1e, 1i, 1p, 1t, 2c). These items represent such activities as incorporating ideas from various sources into a paper, including diverse perspectives in class discussions or writing, and putting together ideas and concepts from different courses. OLS regression models indicate that integration is a very strong predictor of engagement, satisfaction, and self-reported gains, with effect sizes ranging from .22 to .61. For example, the higher the integration score, the more likely a student is to:

• interact with faculty (.58)
• experience diversity (.44)
• report their courses emphasize higher-order thinking (.61)
• engage in active and collaborative learning (.47)
• work harder than they thought they could in response to instructor standards (.45)
• report making substantial gains in a variety of desired outcomes of college (.51)
• be satisfied with the college experience (.31).

Women, seniors, and students attending Baccalaureate-Liberal Arts Colleges tend to engage more frequently in activities that require integration. In contrast, traditional-age students (under 24 years old), student-athletes, and students living on campus are less engaged in integration activities.

We conducted multivariate regression analyses for different groups of students using nine clusters of items from The College Student Report as dependent variables.3 These clusters are:

1. college activities (22 items in question #1);
2. course emphasis on higher-order mental activities (Question #2, items b through e);
3. reading and writing (Question #4);
4. educational programs (Question #7);
5. quality of relationships (Question #8);
6. time-usage (Question #9, items a, b, d);
7. opinions of campus environment (Question #10);
8. educational and personal growth (Question #11);
9. satisfaction with your overall college experiences (Questions #12 through #14).

In general, the results reported below are similar to those reported in previous years.

Year in School

Compared to first-year students, seniors were more engaged in effective educational practices. That is, they were more engaged in college activities, did more reading and writing, reported greater course-emphasis on higherorder mental activities, and spent more time on educationally productive activities. Therefore, it?s no surprise seniors report greater gains on all educational and personal growth items. First-year students held higher opinions of their campus, perceived their relationships with peers, faculty, and administrators to be more positive and the campus environment to be more supportive, and were more satisfied with their overall college experience.

Gender

Women engaged more frequently in good educational practices than did their male counterparts.

Race and Ethnicity

African American and Hispanic students generally were a little more engaged in college activities, more frequently took advantage of enriching educational programs, reported greater course-emphases on higher-order mental activities, and had higher self-reported gains in educational and personal growth than their peers. Asian students also reported increased educational and personal growth and greater participation in educational programs. Compared with other groups, White and Hispanic students had the most favorable opinions about campus climate and the quality of relations among people on campus.

Age

Younger, traditional-aged students (18-24 years) reported participating more frequently in enriching educational programs, spent more time in educationally productive activities, and perceived their campus environment to be more supportive. However, older students did not differ much from their younger counterparts in educational and personal growth. Older students reported more positive relationships with other students, faculty, and administrative personnel, and were more satisfied with their overall college experience.

Transfer Students

Overall, transfer students were less engaged in effective educational activities than their non-transfer peers. Transfer students tended to be older and had more external responsibilities such as working for pay off-campus and caring for dependents. Transfer students spent more time preparing for class and believed their coursework provided more emphasis on cultivating higher-order thinking abilities than did their peers, yet they interact with faculty members and engage enriching educational programs at levels lower than their counterparts. Transfer students were also less satisfied with their quality of relationships with peers, faculty, and administrators and perceived the campus environment to be less supportive.

Fraternity and Sorority Members

Taking into account selected student and institutional characteristics, members of Greek-letter social organizations were more engaged than non-members in all areas of good educational practice. In terms of reading, writing, and the nature of exams, Greek students were more similar to their non-Greek peers than in other areas.

Student-Athletes

Student-athletes, compared to their peers who did not participate in intercollegiate athletics, were more engaged in a variety of educationally effective activities. In general, athletes were similar to their non-athlete peers in participating in enriching educational programs and taking classes that emphasized higher-order thinking skills. However, student-athletes had more positive perceptions of the campus environment and reported more positive relationships with other constituencies on campus.

Parents' Education

Students whose parents hold college degrees were more engaged than first-generation college students in enriching educational programs, reading and writing, and a variety of college activities and spent more time on educationally productive activities. However, students with college-educated parents did not differ from their first-generation counterparts in terms of their opinion of the campus, quality of relationships, as well as the overall satisfaction with the college experience. Differences in engagement between first-generation students and their counterparts were even greater when a student?s parent held a graduate degree.

NSSE continues to pilot survey items for future administration. This year a set of questions about information technology was attached to the end of the on-line survey. Thus, only students responding to the on-line version of The College Student Report were asked these questions.

Student responses to the information technology questions indicate that many students use information technology regularly for personal and academic uses as well as to communicate with students and instructors. More than half of all students reported that their instructors frequently (often or very often) use information technology in the classroom. Students also used information technology in the following ways:

• Men were more likely to use information technology when working with other students on academic work.
• Part-time students were less likely to communicate electronically with other students or their instructors. However, they were more likely to use information technology to obtain resources from libraries at other institutions.
• Over one-third (34%) of education majors reported that their peers frequently copied and pasted from the WWW without attribution as did about one-quarter (24%) of arts and humanities, engineering, physical science, and social science majors.

Before sharing your NSSE results institution-wide, become familiar with the nature of the data and "story line" of your school?s performance. Here are some things to consider.

Check The Representativeness of Your Respondents

An essential early step is comparing your student respondents? demographic characteristics, summarized in the Frequency Distribution section, with your institutional data files for first-year and senior students. Women and some historically underrepresented groups are somewhat over-represented among NSSE 2003 respondents. Check to see if this is also true in your case and whether your respondents differ in any other ways from the profiles of your first-year and senior students. The determination of student year in school ("first-year" or "senior") is based on the information from the electronic file that your school provided to us last fall. The Frequency Distribution section contains students? responses to this question on The Report, which in a few cases may differ from the institution?s classification.

Another way to gauge representativeness is through sampling error, an estimate of the margin by which the "true" score for your institution on a given item could differ from the reported score for one or more reasons, such as differences in one or more important characteristics between the sample and the populations. For example, if 60% reply "very often" to a particular item and the sampling error is ± 5% there is a 95% chance that the population value is between 55% and 65%. Keep in mind that sampling error is based on the population of interest. If you want to estimate the sampling error for first-year male students, it must be calculated using the numbers of all first-year male students and the first-year male respondents (as contrasted with all undergraduates or all male and female first-year students). Increasing the number of respondents relative to the total population reduces sampling error. For this reason some schools are increasing their sample size using NSSE oversampling.

Look for Patterns in Item Differences

In addition to focusing on items with medium to large effect sizes, look for patterns in your students? responses. For example, are your students consistently above or below the mean of your comparison group in certain areas of engagement? Are the differences explainable, perhaps a function of your school?s mission, the nature of the undergraduate program, or certain students? characteristics?

Also, don?t rely exclusively on statistical significance tests to identify areas that warrant attention. A consistent pattern of scoring above the mean, even though all items may not reach statistical significance, may indicate your institution is doing the right things in terms of good educational practice. At the same time, some institutions have very high expectations for student engagement and may fall short of their own aspirations even though comparisons with other institutions are favorable.

The Results Are Unweighted

The data in the Means Summary Report comparisons are not weighted. That is, no adjustments were made to correct for potential bias in students? responses to approximate the populations of first-year and senior students at your school and other colleges and universities in your comparison groups. Later this fall, when we prepare the five national benchmarks of student engagement, we?ll use appropriate weighting techniques, similar to those employed in previous years, to make the appropriate adjustments. That said, the unweighted and weighted results for most NSSE items tend to be very similar at the institution, comparison group, and national levels. Some possible exceptions may be the reading, writing, and time on task questions (e.g., study hours, caring for dependents) at schools that have substantial proportions of part-time students, as they take fewer classes per term and cannot be expected to read and write as much as full-time students. Keep this in mind when interpreting the results.

Look Carefully At Items With Large Effect Sizes

In the Means Summary Report an asterisk (*) marks those items where your students? responses differ at a statistically significant level from students at schools in your respective comparison group(s) or at all NSSE 2003 institutions. The more asterisks reported for a particular item indicate a smaller probability that the difference noted is due to chance (p < .01 for consortia comparisons, p < .001 for Carnegie and national comparisons). Even so, the actual magnitude of some item score differences may seem trivial, even though they are highly reliable and statistically significant. For this reason, we also report the effect size associated with those item comparisons that are statistically significant. The effect size represents the magnitude of the discrepancy in the student or institutional behavior represented by the item. When the effect size is large, or a pattern of moderate effect sizes exists, it?s likely that the quality of the student experience is appreciably different and, therefore, may be of practical as well as statistical significance in the respective area of student engagement.

Finding large effect sizes is not that common in most areas of non-experimental educational and social science research including the NSSE project. If your results include some medium or large effects, something may be going on that warrants immediate attention, especially if other empirical or anecdotal information corroborate the NSSE data. Here are some general guidelines for determining the relative importance of a Cohen?s d effect size:

If Your School Is In A Consortium

If your school belongs to a consortium that used additional questions, the responses to these additional questions are included in the Means Summary Report and Frequency Distribution sections. These data are also in the institutional data file. Answers to such questions as "What is your reason for working off campus?" and "Who is your academic advisor?" have categorical response options that are meaningless when displayed in the Means Summary Report format. For this reason the response cells for such questions are empty. When presenting the results to categorical questions to colleagues and others, please use the information in the Frequency Distributions.

Take Into Account Possible Mode-of-Administration Effects

Our analyses show that a mode-of-administration effect slightly favors schools where a high percentage of students completed The College Student Report via the Web. However, the differences that favor the Web mode have very small effect sizes. This phenomenon has also been noted by others using the Web for survey research and is discussed in more detail in Appendix A. We still don?t know for sure whether this pattern of responses is a function of the mode of administration itself (e.g., something about responding via the Web induces students to slightly inflate their responses), a function of certain institutional features (e.g., technology investment), or whether students who complete the survey via the Web are different in some ways including engaging more frequently in good educational practices. Evidence of the last of these is that the Web effect is most prominent on the three technology?related items ("used email to communicate with an instructor," "used an electronic medium to discuss or complete an assignment," and self-reported gain in "using computing and information technology"). We are continuing to monitor this issue and will alert you if our analyses lead us to modify our conclusion that the Web mode has little practical impact on student responses to The College Student Report.

Review Responses to Experimental Questions (if applicable)

In an effort to test potential survey items for future administration, a small set of experimental questions related to technology were added to the NSSE online survey. These questions were attached to the end of the survey and only students responding to the online version received these extra questions.

For schools that chose to participate, responses to the experimental questions about technology are included in the institutional data file. However, due to their experimental nature and the fact that only students completing the survey online received the technology items, these questions are not included in the Frequency Distribution and Means Comparison Reports. Rather, frequencies and means by Carnegie type and at the national level are provided in a separate file named "Technology Item Summary by Carnegie and National" to inform institutional comparisons.

When reviewing your institution?s experimental item results, please pay attention to the number of respondents. If the number is small compared with your overall respondent group, interpret your results with extreme caution.

Consortium, Carnegie, and National Comparisons Do Not Include Oversampled Students

NSSE?s minimum sample sizes are determined by undergraduate enrollment (i.e., less than 4,000 students = 450; 4,000 to 15,000 students = 700; greater than 15,000 students = 1,000). It is possible to add students to the minimum sample size by oversampling in one of two ways: (1) all Web-only schools are oversampled using an algorithm based on undergraduate enrollment; and (2) some institutions request oversampling, which requires an additional fee. An increasing number of schools are using the oversampling option to add students to their sample, reduce sampling error, insure an adequate number of respondents to analyze the information by major field, race and ethnicity, or other variables.

NSSE?s policy is to use only respondents from the institution?s standard random sample when developing the national benchmarks of effective educational practice and sector and national norms. This protects against the possibility that colleges and universities with oversamples might unduly influence the results. However, if your school requested a NSSE oversample, the responses of all your students (standard sample and NSSE oversample) are included in your institution's reports and data file.

1 The NSSE 2003 number of respondents reported in the "Overview" does not include the additional students who were oversampled. Oversampling was done at Web-only institutions and at schools that requested more of their students be surveyed than dictated by the NSSE sampling strategy, which is a function of institutional size. All in all, 147,166 students responded to the NSSE 2003 survey.

2 The NSSE 2003 average institutional response rates most likely underestimate the actual adjusted rate. Student postal service and e-mail addresses were based on fall 2002 enrollment information provided by the institutions. An unknown number of students in the sample were no longer eligible to complete the survey because they had dropped out or transferred to another institution. Even though first-class postage was used to guarantee the return of survey packets that could not be delivered, experience suggests that packets were not returned for some students who were no longer in school or residing at their fall 2002 address. In addition, at Web-only schools institution-provided email addresses were used to send students their invitation to participate in NSSE 2003. We have found that many students have multiple e-mail accounts (e.g., Yahoo, AOL, Hotmail). Some institutions have more difficulty tracking these multiple email accounts and some students may not forward their institution assigned e-mail. Therefore, the actual response rate for Web-only institutions, when corrected for the unknown number of students who were no longer in school or did not receive the invitation to participate, may be several percentage points higher than 44%.

3 The regression of each cluster of items on a group characteristic is net of the following student and institutional controls: class, residence, gender, enrollment status, race/ethnicity, age, major, parental education, 2001 Barron?s admissions selectivity, sector, and 2000 Carnegie Classification.