PBA Home > Institutional Research & Analysis > Surveys > NSSE 2006 > Method

National Survey of Student Engagement 2006

Population and Sample

The CU-Boulder population for the 2006 NSSE included:

• Freshmen - all first-time freshmen who started at CU-Boulder in summer or fall 2005 and who were still enrolled in spring 2006 (n = 4,895)
• Seniors - all seniors enrolled spring 2006 (n = 4,861)

Students on study abroad and students who requested a privacy block on their records were removed from the population file, which was sent to NSSE. From this file, NSSE staff selected a random sample of 2,000 freshmen and 2,000 seniors to participate in the survey. An additional 600 seniors in the 24 largest majors (e.g., economics, psychology) and small colleges (e.g., journalism) were oversampled to allow us to characterize these majors/colleges separately. For analyses that include seniors, the data have been weighted to compensate for oversampling and for different response rates over academic majors.

How Students Were Contacted

Students were contacted on several occasions during the 2006 NSSE administration. The following table shows the date and form of contact (e.g., letter, e-mail) for each occasion:

Response Rates by Class Level

The following table lists CU-Boulder response rates by class level. The overall response rate for CU-Boulder was 39%, which was somewhat higher than that of the 2002 NSSE administration (34%).

¹ These sample numbers are smaller than the samples of 2,000 freshmen and 2,600 seniors drawn by NSSE in fall 2005 due to student attrition from CU-Boulder by the time of survey administration in spring 2006.

Response Rates by Class Level and Demographic Subgroup

The following table contains CU-Boulder response rates by class level and demographic subgroup:

It can be seen that, regardless of class level, CU-Boulder females responded to the survey at a higher rate than did males (39% vs. 31% for freshmen; 49% vs. 37% for seniors, respectively). White students responded at slightly higher rates than did racial/ethnic minority students (35% vs. 34% for freshmen; 43% vs. 39% for seniors, respectively).

Response Rates by College

Response rates over colleges/divisions ranged from 22% (Architecture & Planning) to 46% (Engineering) for freshmen, and from 33% (Arts & Sciences: Other, and Architecture & Planning) to 52% (Journalism) for seniors.

Respondents Versus Non-Respondents

The table below shows percentages of the sample, non-respondents, and respondents by various demographic characteristics (i.e., gender, race/ethnicity, college/division, and GPA category). The table also shows response rates in the last two columns.

Respondents and non-respondents are similar with respect to race/ethnicity and college/division, irrespective of whether they are freshmen or seniors. For example, 15% of freshmen non-respondents are racial/ethnic minorities, versus 14% of respondents, and 16% of both freshmen non-respondents and respondents are enrolled in Arts and Sciences Natural Sciences. However, there are differences between non-respondents and respondents with respect to gender and cumulative GPA, a pattern which has been observed in other surveys on this campus. For example, 40% of senior non-respondents are female, compared with 52% of senior respondents, and 44% of senior non-respondents have a GPA above the median (3.14), whereas 58% of senior respondents fall into this category.

The fact that females and students with higher GPAs are somewhat more likely to respond may have some effect on the survey results. We previously investigated the likelihood of such an effect in the 2000 NSSE administration. At that time, analyses of relationships among GPA and and 2000 NSSE item responses demonstrated these relationships to be weak. We have observed similar, weak relationships for the 2006 NSSE data, with a  median (over items) GPA and item response correlation coefficient of .03. The largest GPA/item correlations occurred between GPA and frequency of asking questions in class or contributing to class discussions, frequency of tutoring other students, and ratings of the quality of relationships with faculty members (r = .17 for each). Weak relationships similarly exist among gender and NSSE item responses (median r over items = .03). The largest gender/item correlations occurred for performing community service or volunteer work (.19), frequency of including diverse perspectives in class discussions and assignments (.16), and participating in study abroad (.15). Because GPA/item and gender/item relationships are weak, no statistical adjustments have been made to attempt to correct for overrepresentation of females and high-GPA students among the respondents.

Scales

The scales used in the 2006 analyses are similar to those used in 2002. However, because some items used in the 2002 NSSE administration differ from those in the 2006 administration, the 2002 and 2006 scales differ somewhat. For example, a question about discussing grades or assignments with an instructor is included in the faculty engagement scale for 2006, but was not included in 2002. To view a table containing item numbers and names for each item within each of the scales, see the 2006 NSSE Items table. Note that the "study abroad" scale differs from the other scales, in that it consists of only one item.

CU-Boulder NSSE 2006 Main Page

Last revision 06/02/08

l:\ir\survey\NSSE\06\Writeups\NSSE06method4.doc