|
|
|
|
A team of national assessment experts, under the leadership of Peter Ewell of the National Center for Higher Education Management Systems, designed the National Survey for Student Engagement. Indiana University's Center for Survey Research administers the survey which was first administered nationally in the spring of 2000. A sample of first-year and fourth-year students at participating colleges and universities were sent web and paper surveys, depending upon choices made by college and university administrators. U.Va. choses web-only participation, which allows for a larger number of students to be surveyed. For more on the possibility of response bias from web surveys, see below. For more from NSSE on the survey's conceptual framework and design process, click here. Some studies have shown that web surveys can produce response bias effects. However, such effects do not seem likely to apply to U.Va. in recent years. The University has had a strong commitment to its computer networking and information technology programs. Most students have their own computers according to data collected in a four-year panel study conducted by IAS (the Longitudinal Study of Undergraduate Education), 70% of first-year students and 80% of fourth-year students reported having their own computers with them at school. Networked computer labs with internet access are available throughout the grounds for those without computers. Perhaps the best evidence that U.Va. undergraduates universally have web access comes from IAS's surveys over the past three years. Very few U.Va. undergraduates respond to surveys on paper, when given a choice. For on-grounds student surveys conducted by IAS over the past three years, only 14 paper surveys were requested and returned. Of the more than 2,000 surveys completed, 99 percent of students living on grounds choose to complete their survey via the web. In terms of response effects on NSSE outside of the University, survey administrators conducted thorough response mode testing and found consistent effects showing that web respondents score higher on most of the survey items. However, this effect was small for the vast majority of the variables with the notable exceptions concerning technology and computer-related items. A more in-depth discussion of response effects in the NSSE can be found at the NSSE's website. More than 660,000 first- and fourth-year students from 529 participating colleges and universities were invited to participate in NSSE 2005. . Approximately 245,000 students responded nationally. The overall response rate was 39%. UVa invited 3622 students to participate and 2079 responded. The adjusted response rate (excludes bad addresses and ineligibles) for UVa was 58%. UVa. employed oversampling procedures for students on need-based financial aid, as well as African-American students. After a random sample of students was chosen, an additional 524 students on need-based financial aid were added to the sample, including all students designated as "high need" (200% of poverty level or below) and a sample of additional need-based financial aid recipients. After the oversample on financial aid was chosen, all 1st- and 4th-year African-Americans not already in the sample were added (this amounted to an aditional 298 students). These procedures insure a sufficient number of cases to allow the University to analyze the results by students who qualified for need-based financial aid, as well as the results for African-American students. NSSE did not include any of the oversampled students in their national statistics, or the University's statistics as reported in the means, frequencies and benchmarks. A total of 1583 cases (853 1st-year and 730 4th-year) students are reported in the NSSE means, frequencies and benchmark reports. The 496 students who responded to the survey and were part of the need-based financial aid and African-American oversamples were excluded from these results. Separate statistics and reports will be issued by IAS for the oversampled group in fall 2005. As is often the case, women were overrepresented in both the national and UVa. samples. Part-time students were underrepresented in the national sample. To reduce this response bias a weighting procedure, by gender and enrollment status, was employed. For a detailed description of the weighting procedures NSSE employed, click here to view the NSSE 2005 Annual Report. Over 206,000 first- and fourth-year students at 366 participating colleges and universities were sent surveys in the spring of 2002. About 81,000 students responded, roughly an equal mix of first- and fourth-year undergraduates. The response rate nationally was 41%; the response rate for U.Va. was 54%. A total of 947 students from U.Va. returned surveys. U.Va. employed oversampling procedures for African, Hispanic, and Asian-American students. This procedure insures a sufficient number of cases to allow the University to analyze the results by race. NSSE does not include the oversampled students in their national statistics. The University's statistics do include the oversampled minority students--in particular, Asian and Hispanic-Americans. In percentage terms, Asians were overrepresented in the sample by about 55 percent; Hispanics were overrepresented by 233 percent. In total, the number of oversampled students was 123, or about 13 percent of U.Va.'s sample. NSSE did not weight the data which is reported in the tables in the data section of this site to account for this overrepresentation. Technically, the data should be weighted to reduce the impact of the oversampling. However, to see if the results were skewed by the oversample, IAS weighted the data by race and compared the results. Weighting made little difference. As is often the case, women were overrepresented in both the national and U.Va. samples. Part-time students were underrepresented in the national sample. To reduce this response bias a weighting procedure, by gender and enrollment status, was employed. For a detailed description of the weighting procedures NSSE employed, click here. Over 151,000 first-
and fourth-year students at 276 participating colleges and universities
were sent surveys in the spring of 2000. About 63,000 students responded,
roughly an equal mix of first- and fourth-year undergraduates. The
response rate nationally was 42%; the response rate for U.Va. was
46%. A total of 318 students from U.Va. returned surveys. As is
often the case, women were overrepresented in both the national
and U.Va. samples. Part-time students were underrepresented in the
national sample. To reduce this response bias a weighting procedure,
by gender and enrollment status, was employed. For
a detailed description of the weighting procedures NSSE employed,
click here. |
|
||||||||||||||
