Journal of Nursing Education

Major Article 

Early Identification of At-Risk Nursing Students: A Student Support Model

T. Hampton Hopkins, EdD

Abstract

Due to the shortage of nurses in the health care industry, colleges offering associate-degree nursing programs are beginning to pay more attention to attrition and the factors contributing to success. A logistic regression model was used to explain the cognitive and noncognitive variables that contribute to success in a nursing fundamentals course. Although much work is necessary to fully understand first-semester nursing students’ retention and success, an early identification model is explored to better support students as they enter associate-degree nursing programs.

Abstract

Due to the shortage of nurses in the health care industry, colleges offering associate-degree nursing programs are beginning to pay more attention to attrition and the factors contributing to success. A logistic regression model was used to explain the cognitive and noncognitive variables that contribute to success in a nursing fundamentals course. Although much work is necessary to fully understand first-semester nursing students’ retention and success, an early identification model is explored to better support students as they enter associate-degree nursing programs.

Dr. Hopkins is Dean, Student Services and Enrollment Management, Carolinas College of Health Sciences, Charlotte, North Carolina.

Address correspondence to T. Hampton Hopkins, EdD, Dean, Student Services and Enrollment Management, Carolinas College of Health Sciences, 1200 Blythe Boulevard, Charlotte, NC 28203; e-mail: hampton.hopkins@carolinascollege.edu.

Received: April 03, 2005
Accepted: October 20, 2006

Higher education administrators continue to face growing concerns about lower academic achievement and higher attrition rates among associate-degree nursing (ADN) students. In light of the extreme nursing shortages facing the health care industry, focusing on the retention of those students who gain enrollment should become a top priority in nursing education (Sayles, Shelton, & Powell, 2003; Wells, 2003). ADN programs are in a unique position to graduate nursing students in a relatively short time frame, compared with bachelor-degree programs, to meet the growing demand for professional nurses despite the rigors of nursing education (Wells, 2003). Identifying factors that predict student success and applying student support interventions early in the first semester can affect student persistence.

Predicting the success of college students from prematriculation data has long been a subject of debate and discussion. College admission staff have been charged with the increasingly challenging responsibility of identifying students who can be successful (Bauchmoyer, Carr, Clutter, & Hoberty, 2004). Studies have found a strong correlation between high school grade point average (GPA) and cumulative college GPA, a typical measure of student success (Astin, Korn, & Green, 1987; Lotkowski, Robbins, & Noeth, 2004; Schwartz & Washington, 2002; Timpson, Winokur, & Merline, 2002). Other studies have focused on the strength of standardized tests such as the SAT Reasoning Test as a predictor of success in college (Isaacs, 2001; Noble & Sawyer, 2004). Additional studies state that factors such as high school quality, self-efficacy, and other nonacademic variables are equally, if not more, valid predictors of college success (Spitzer, 2000; Tam & Sukhatme, 2003).

As Glynn, Sauer, and Miller (2003) suggested, early prediction of student success (or lack of success) is important in creating effective interventions. Identifying factors that predict success is more likely to be effective if both academic and nonacademic variables are used (Schwartz & Washington, 2002). Academic variables, including SAT or ACT scores (Dowling, 1999–2000; Heriot, 2001; Isaacs, 2001), high school GPA (Noble & Sawyer, 2004), and prior college GPA, are purported to be predictive of college success. Nonacademic variables may include high school quality (Pike & Saupe, 2002; Tam & Sukhatme, 2003); critical thinking skills (Timpson et al., 2002); and demographic variables such as gender, race, or socioeconomic status (Reason, 2003). Sayles et al. (2003), in their correlational study of ADN students, posited questions of the effective use of the Nursing Entrance Test (NET) as a tool for admissions. Although some factors appeared to be predictive of success in the study by Sayles et al. (2003), Simmons, Haupt, and Davis (2004) found even more evidence of the predictive nature of the NET. In fact, both studies agreed that the NET could be an adequate tool for understanding some of the nonacademic variables that affect nursing student success (Sayles et al., 2003; Simmons et al., 2004).

The literature raises some concerns regarding the admission and retention of women and minorities. Young and Fisler (2000) stated that the SAT is not an adequate tool as a predictor of college success among women. In addition, evidence suggests that the SAT contributes little to the variance, in addition to high school GPA and high school rank of African American students and other minorities (Schwartz and Washington, 2002; Ungerleider & Maslow, 2001). The authors further suggested that nonacademic variables are better predictors of college success than are SAT scores (Schwartz & Washington, 2002), especially when combined with high school GPA. Similar concerns are found in the admission and retention of nursing students. Wells (2003) and Childs, Jones, Nugent, and Cook (2004) suggested that the underrepresentation of minority students in nursing programs is due in part to the admission processes and retention efforts of schools of nursing.

Some experts have suggested deemphasizing the SAT in the admissions process. Most notably, Richard Atkin-son, President of the University of California system, has been a strong voice in considering student achievement instead of aptitude (Atkinson, 2001–2002). In a supporting study conducted by the University of California, Geiser and Studley (2002) suggested that student success is better predicted using achievement tests such as the SAT II instead of aptitude tests like the SAT I, especially when other academic and nonacademic variables are included. However, these findings are called into question by Dowling (1999–2000) and Heriot (2001), who suggested that the politics of California Proposition 209 are the impetus behind this proposed change. Regardless of the issues, the SAT continues to be a tool used by many colleges for admissions and recruitment efforts (Isaacs, 2001).

A literature review yielded a broad perspective of student retention studies in support of predicting success in college. A common approach to predicting success occurs prior to admission, through the use of predictor variables in the application process. Another approach is to use academic and nonacademic predictor variables to identify those students most at risk for being unsuccessful in college. Schwartz and Washington (2002) found that colleges have the ability to create student support programs for at-risk students when identifying factors affecting success. Therefore, this study sought to identify to what extent academic and nonacademic variables explain first-semester academic success among ADN students, as defined by passing the Nursing Fundamentals course.

Method

Success was operationally defined as achieving an 80 (of 100) or higher as the final grade in the Nursing Fundamentals course, taken during the first semester. Although a final grade of 77 is considered passing, a grade of 80 was used in this study to determine whether a student is facing academic difficulties, which would enable the student to take advantage of support services to improve the chances for success. Identifying the student early in the semester may allow for more time for intervention. The practical use of such a model is especially important in associate degree programs, where time is short, to create interventions in support of student success and retention (Glynn et al., 2003). What follows is a look at the factors that can be used to explain success among nursing students and to begin to better understand those students most at risk for being unsuccessful.

Data Collection

The NET was used to collect several of the variables in this study. The NET, administered prior to enrollment, is an instrument that measures students’ academic ability, stress levels, learning style, and other nonacademic variables and generates individual student reports on seven subscales and a composite scale, which are nationally normed annually. The NET has an internal reliability coefficient of 0.92, as reported in studies conducted by Educational Resources, Inc. (1998). Validity data and reliability data for the NET subscales was not available. The remaining variables were identified through the educational records maintained by the college.

Academic and Nonacademic Variables

To decide which academic and nonacademic variables to include in the study, a number of variables could be used. Ideally in a study such as this, the more variables included will lead to more accurate results and will account for more of the variance, notwithstanding the issue of diminishing returns (Schwartz & Washington, 2002). In reality, the most convenient data are those variables already maintained in the college’s records. In the interest of time and resources, this study used a sample of convenience with variables previously collected during the matriculation process.

The academic variables in this study included students’ SAT and ACT scores. The ACT scores were converted to a standard scale corresponding with the SAT for use in analysis. In addition, cumulative high school GPA and cumulative college GPA were used as predictors of success. The GPAs were collected during the admissions process and were available from the registrar’s office. Cumulative college GPA was included because many of the students in this study had some prior college before enrolling in the nursing program (Wells, 2003). The final two academic variables used in this study were math and reading composite scores identified from the NET. On the basis of NET results, each student receives a math and reading composite score identifying the level of proficiency from pre-junior high school to post-high school. Because SAT scores and GPAs are commonly used as predictors of academic success, they were included in this study. As for the NET scores, prior research indicates the potential of these variables as predictors of nursing student success (Sayles et al., 2003; Simmons et al., 2004).

The nonacademic variables in this study included demographic data obtained from the registrar’s office. Gender, race, and age were identified as the demographic variables most important in this study on the basis of the literature review and in consideration of the need for higher representation of minorities in nursing (Childs et al., 2004; Wells, 2003). The remaining nonacademic variables resulted from students’ completion of the NET and included three critical thinking scores, five stress level scores, two learning style scores, and one testing score. Each of these variables was rated on a scale of 0 to 1, producing a percentage score on the variable for each student. Again, these variables are included because the constructs they represent have been identified as potential predictors of success (Childs et al., 2004; Sayles et al., 2003; Simmons et al., 2004; Timpson et al., 2002).

Sample

After Institutional Review Board approval was obtained, the data were collected from the school of nursing tracking database and other existing educational records maintained by the college. Data were included for 383 ADN students at a small, private college of health sciences in the southeastern United States. This model focused on first-semester success, exclusively as measured by students’ final grade in the Nursing Fundamentals course. It is thought that most students who are successful in the first semester of college will ultimately graduate (Reason, 2003). The intent of the study was to identify those factors that significantly explain the first-semester success of nursing students. Most of the students in this study were female (89.6%) and White (62.1%). The mean age was 28.2 years (SD = 7.68), ranging from 18 to 59 years. The mean SAT score was 1014.2 (SD = 124.56), the mean high school GPA was 3.06 (SD = 0.55), and the mean college entrance GPA was 2.89 (SD = 0.53). The mean final grade in the Nursing Fundamentals course was 84.7 (SD = 4.35). Descriptive statistics for all variables are presented in Table 1.

Descriptive Statistics (N = 383)

Table 1: Descriptive Statistics (N = 383)

Procedures

The current study examined the likelihood that students would achieve success in the Nursing Fundamentals course on the basis of 19 predictor variables. All analysis was conducted using SPSS software version 12.0. Using descriptive statistics, it was determined the distribution of data was generally within normal ranges (skewness and kurtosis). This was expected, given that the population of the college is mostly female and White, with a mean age of 31 years. All students (N = 383) were included in the sample, and mean scores were substituted for missing data according to established protocols for multivariate statistics (Grimm & Yarnold, 1995).

Results

Using a simple correlation, it was determined that several of the variables were significantly correlated. SAT score, high school GPA, and NET score were all generally correlated and were included in an exploratory factor analysis. A principal component analysis was used to identify the uncorrelated predictor variables to reduce the potential for multicollinearity (Glynn et al., 2003). A skree test identified six factors with eigenvalues greater than one. As a result of the final factor loading only on the NET testing variable, it was decided to limit the number of factors. A second principal component analysis was run, limiting the factors to five, and each was retained for further analysis. An orthogonal rotation (varimax) was used, and the resulting model accounted for 61.79% of total variance. The five factors were identified using a loading factor of 0.40 to determine relationships. The five factors were:

  • Reasoning.
  • Learning style.
  • Analytic.
  • Anxiety.
  • Commitment.

The factor pattern matrix and communalities are presented in Table 2, with each of the above factors identified to better represent the constructs being measured.

Factor Pattern Matrix and Communalities

Table 2: Factor Pattern Matrix and Communalities

A direct logistic regression was used to arrive at the probability of successfully completing the Nursing Fundamentals course. Of the new nursing students who matriculated between 2001 and 2004 and were included in the analysis, 68 (17.75%) were unsuccessful (final grade below 80) in the Nursing Fundamentals course, and 315 (82.25%) were successful (final grade of 80 or above). The variables were entered in a specific order based on perceived importance in contributing to success in the course and included the five factors identified in the factor analysis. The regression test of the full model with all predictors was statistically significant, indicating that the variables reliably predicted those who would be successful, χ2(9, N = 383) = 33.10, p < 0.01. The variance accounted for is low (Cox and Snell R2 = 0.08, Nagelkerke R2 = 0.14). Predicted success was good, with 99% of the successful students correctly identified and an overall success rate of 82.5%. However, the model did not do a good job of predicting those students who would be unsuccessful (5.9%). The regression coefficients, Wald statistics, odds ratios, and significance are presented in Table 3. The Wald criteria indicate that only the Reasoning Factor and Analytic Factor reliably explain student success. In other words, for each point the Reasoning Factor score increases for each student, the student is 1.58-fold more likely to be successful in the course. Likewise, as the Analytic Factor increases, a student is 1.80-fold more likely to be successful.

Logistic Regression Coefficients, Standard Errors, Wald Statistics, and Odds Ratios

Table 3: Logistic Regression Coefficients, Standard Errors, Wald Statistics, and Odds Ratios

Discussion

The purpose of this study was to develop a model of student support that explained success in the Nursing Fundamentals course. The predictive validity of the study was 82.5%, as measured by the percentage of students correctly identified as successful or not successful in the course. This model provides for the early identification of those ADN students who can be successful in the Nursing Fundamentals course, thus allowing for a better understanding of those who may be at risk. Caution is suggested when using this as the only model because many factors may influence course grades and success (Jones, Manuel, & Borges, 2003).

Atkinson (2001–2002) suggested the whole student is considered when admissions decisions are being made; this holds true when predicting success, especially because the variance accounted for in this model is low. However, if there is a chance, no matter how small, to influence one student to be successful in the Nursing Fundamentals course, it should be explored. After the at-risk student is identified, a mechanism for referring this student to existing support systems already on campus should be implemented to help the student succeed. It is recognized that some support systems may not be in place at all colleges and that faculty may have to create such systems for their students. Support systems may include peer tutoring, course content review sessions, personal and academic counseling, study skills workshops, and other student support programs. Tinto (1993) said that student retention is influenced by a number of internal and external opportunities, academic and social factors, and other variables that come together to form the student experience. He further stated that it is important to create “intrusive interventions” (p. 182) for those students at risk of failure. In other words, it is the responsibility of the college to identify, monitor, and aggressively intervene with those students who may be unsuccessful (Glynn et al., 2003; Tinto, 1993). By using this model as an early identification of success, those students who underachieve on the math and reading components of the NET may be referred for developmental education. Likewise, those with low GPAs and SAT scores may also be identified for tutoring and other student support services.

Limitations

Some limitations of this study include the low variance accounted for in predicting student success. There appear to be other variables that have not been identified that influence the success of students in the Nursing Fundamentals course. Multiple retention studies have identified numerous variables that may affect success: work commitment, parents’ education, study habits, motivation, attachment to college, socioeconomic status, and others (Cokley, Bernard, Cunningham, & Motoike, 2001; Glynn et al., 2003; Reason, 2003; Schwartz & Washington, 2002). Finally, the NET is understudied in the literature as a tool for predicting success in nursing education, and its validity may be questionable (Sayles et al., 2003). In addition, at this particular college, the NET is not required as a component of admissions; therefore, social desirability responding may be of concern if students are not cautious in how they respond.

Implications

One contribution of this study is that it adds another perspective to student success, especially for ADN students. It further contributes to the knowledge of the use of the NET as a tool for explaining success; however, further studies using the NET are recommended. Opportunities for future study are limitless. It is suggested that future studies look at program retention (instead of success in one course) over the 2-year term of the student. It would further benefit the nursing education field to continue the study of these students to determine success in passing the NCLEX-RN® and including relevant variables during the college experience (i.e., college GPA, math and science GPA). Finally, as a tool for determining student success, ongoing studies of these variables and other nonacademic variables will continue to add to the knowledge about retention.

References

  • Astin, AW, Korn, W & Green, K1987. Retaining and satisfying students. Educational Record, 68, 36–42.
  • Atkinson, RC2001–2002, Winter. Achievement versus aptitude in college admissions. Issues in Science and Technology, 31–36.
  • Bauchmoyer, SM, Carr, MP, Clutter, JE & Hoberty, PD2004. Predicting academic and national board dental hygiene examination performance based on academic factors [Electronic version]. The Journal of Dental Hygiene, 781, 39–45.
  • Childs, G, Jones, R, Nugent, KE & Cook, P2004. Retention of African-American students in baccalaureate nursing programs: Are we doing enough?Journal of Professional Nursing, 202, 129–133. doi:10.1016/j.profnurs.2004.03.002 [CrossRef]
  • Cokley, KO, Bernard, N, Cunningham, D & Motoike, J2001. A psychometric investigation of the academic motivation scale using a United States sample. Measurement and Evaluation in Counseling and Development, 34, 109–119.
  • Dowling, WC1999–2000, Winter. Enemies of promise: Why America needs the SAT [Electronic version]. Academic Questions, 6–24. doi:10.1007/s12129-999-1036-0 [CrossRef]
  • Educational Resources, Inc. 1998. The Nurse Entrance Test: Technical and developmental report. Shawnee, KS: Author.
  • Geiser, S & Studley, R2002. UC and the SAT: Predictive validity and differential impact of the SAT I and SAT II at the University of California [Electronic version]. Educational Assessment, 81, 1–26. doi:10.1207/S15326977EA0801_01 [CrossRef]
  • Glynn, JG, Sauer, PL & Miller, TE2003. Signaling student retention with prematriculation data. NASPA Journal, 411, 41–67. Retrieved June 24, 2004, from http://www.naspa.org
  • Grimm, LG & Yarnold, PR1995. Reading and understanding multivariate statistics. Washington, DC: American Psychological Association.
  • Heriot, G2001, Fall. The politics of admissions in California [Electronic version]. Academic Questions, 29–36. doi:10.1007/s12129-001-1033-4 [CrossRef]
  • Isaacs, T2001. Entry to university in the United States: The role of SATs and advanced placement in a competitive sector. Assessment in Education, 83, 391–405. doi:10.1177/107319110100800404 [CrossRef]
  • Jones, BJ, Manuel, RS & Borges, NJ2003, November. Beyond counting heads: The use of multiple prediction formulas to aid admissions decisions. Paper presented at the American Association of College Registrars and Admissions Officers Thirteenth Annual Strategic Enrollment Management Conference. , Boston, MA. .
  • Lotkowski, VA, Robbins, SR & Noeth, RJ2004. The role of academic and non-academic factors in improving college retention. Retrieved September 17, 2005, from http://www.act.org/path/policy/pdf/college_retention.pdf
  • Noble, JP & Sawyer, RL2004. Is high school GPA better than admission test scores for predicting academic success in college?College and University Journal, 794, 17–22.
  • Pike, GR & Saupe, JL2002. Does high school matter? An analysis of three methods of predicting first-year grades. Research in Higher Education, 432, 187–207. doi:10.1023/A:1014419724092 [CrossRef]
  • Reason, RD2003. Student variables that predict retention: Recent research and new developments. NASPA Journal, 404, 172–191. Retrieved June 24, 2004, from http://www.naspa.org
  • Sayles, S, Shelton, D & Powell, H2003, November/December. Predictors of success in nursing education [Electronic version]. The ABNF Journal, 116–120.
  • Schwartz, RA & Washington, CM2002. Predicting academic performance and retention among African American freshman men. NASPA Journal, 394, 354–370. Retrieved June 24, 2004, from http://www.naspa.org
  • Simmons, LE, Haupt, G & Davis, L2004. The usefulness of the Nurse Entrance Test (NET) for prediction of successful completion in a nursing program. Shawnee, KS: Educational Resources, Inc. Retrieved November 4, 2004, from http://www.eriworld.com/secure/net/netstudyreport.pdf
  • Spitzer, TM2000. Predictors of college success: A comparison of traditional and nontraditional age students. NASPA Journal, 381, 82–98. Retrieved June 24, 2004, available from http://www.naspa.org
  • Tam, MS & Sukhatme, U2003. The importance of high school quality in university admissions decisions. College and University Journal, 783, 3–8.
  • Timpson, WM, Winokur, M & Merline, AM2002. Skills for success in college: What students need from high school. International Electronic Journal for Leadership in Learning, 64, 1–10. Retrieved June 21, 2004, from http://www.ucalgary.ca/~iejll
  • Tinto, V1993. Leaving college: Rethinking the causes and cures of student attrition (2nd ed.). Chicago: University of Chicago Press.
  • Ungerleider, D & Maslow, P2001. Association of Educational Therapists: Position paper on the SAT [Electronic version]. Journal of Learning Disabilities, 344, 311–314. doi:10.1177/002221940103400405 [CrossRef]
  • Wells, MI2003. An epidemiologic approach to addressing student attrition in nursing programs. Journal of Professional Nursing, 193, 230–236. doi:10.1016/S8755-7223(03)00069-3 [CrossRef]
  • Young, JW & Fisler, JL2000. Sex differences on the SAT: An analysis of demographic and educational variables [Electronic version]. Research in Higher Education, 413, 401–416. doi:10.1023/A:1007099012828 [CrossRef]

Descriptive Statistics (N = 383)

VariableMeanSDMinimumMaximumSkewnessKurtosis
SAT score1014.22124.56670.001430.000.730.66
High school GPA3.060.551.174.75−0.130.64
College GPA2.890.530.834.00−0.590.98
Math composite score0.700.150.311.00−0.18−0.53
Reading composite score0.650.120.180.91−0.580.57
Critical thinking score
  Inferential reading0.670.170.201.00−0.45−0.32
  Main idea of passage0.770.160.141.00−0.820.96
  Predicting outcomes0.550.150.090.91−0.190.06
Testing skills score0.390.050.200.600.062.60
Stress level score
  Family0.280.110.000.700.462.08
  Social0.440.140.000.80−0.251.16
  Money and time0.530.140.001.000.271.64
  Academic0.470.130.000.90−0.572.29
  Work0.240.090.000.701.065.41
Learning style score
  Auditory0.570.160.001.00−0.622.25
  Visual0.620.150.121.00−0.061.68
Age28.157.6818.0059.001.302.02
Nursing Fundamentals course final grade84.664.3563.8097.20−0.762.32

Factor Pattern Matrix and Communalities

VariableReasoningLearning StyleAnalyticAnxietyCommitmentCommunalities
SAT score0.510.030.60−0.010.080.63
High school GPA0.02−0.090.770.030.090.61
Math composite score0.290.100.63−0.06−0.040.50
Reading composite score0.97−0.050.140.020.050.96
Critical thinking score
  Inferential reading0.830.010.080.000.090.71
  Main idea of passage0.57−0.040.15−0.02−0.120.37
  Predicting outcomes0.73−0.110.130.050.030.57
Testing skills score0.220.05−0.090.18−0.010.10
Stress level score
  Family−0.050.150.140.80−0.230.74
  Social−0.030.190.140.330.790.79
  Money and time−0.070.110.000.32−0.820.79
  Academic−0.170.490.280.070.240.41
  Work0.10−0.15−0.110.790.220.72
Learning style score
  Auditory−0.010.79−0.020.090.090.65
  Visual−0.01−0.830.110.110.200.74

Logistic Regression Coefficients, Standard Errors, Wald Statistics, and Odds Ratios

PredictorβSEWalddfORp
Reasoning0.460.1410.5611.580.00
Learning style0.120.140.8011.130.37
Analytic0.590.1614.1911.800.00
Anxiety0.010.130.0111.010.94
Commitment−0.080.140.2910.930.59
College GPA0.490.273.3911.640.07
Age0.020.020.7411.020.39
Race−0.330.380.7210.720.40
Gender0.640.451.9811.890.16
Constant0.201.030.0411.220.85
Authors

Dr. Hopkins is Dean, Student Services and Enrollment Management, Carolinas College of Health Sciences, Charlotte, North Carolina.

Address correspondence to T. Hampton Hopkins, EdD, Dean, Student Services and Enrollment Management, Carolinas College of Health Sciences, 1200 Blythe Boulevard, Charlotte, NC 28203; e-mail: .hampton.hopkins@carolinascollege.edu

10.3928/01484834-20080601-05

Sign up to receive

Journal E-contents