Relationship between measures of academic quality and undergraduate student attrition: the case of higher education institutions in the Colombian Caribbean region

This article was downloaded by: [Universidad Del Norte]On: 06 May 2015, At: 15:22Publisher: RoutledgeInforma Ltd Registered in England and Wales Registered Number: 1072954 Registeredoffice: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Click for updates

Higher Education Research &DevelopmentPublication details, including instructions for authors andsubscription information:http://www.tandfonline.com/loi/cher20

Relationship between measures ofacademic quality and undergraduatestudent attrition: the case of highereducation institutions in the ColombianCaribbean regionAnabella Martíneza, Mónica Borjasa, Mariela Herreraa & JorgeValenciaa

a Department of Education, Universidad del Norte, Barranquilla,ColombiaPublished online: 05 May 2015.

To cite this article: Anabella Martínez, Mónica Borjas, Mariela Herrera & Jorge Valencia (2015):Relationship between measures of academic quality and undergraduate student attrition: the caseof higher education institutions in the Colombian Caribbean region, Higher Education Research &Development, DOI: 10.1080/07294360.2015.1024622

To link to this article: http://dx.doi.org/10.1080/07294360.2015.1024622

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the“Content”) contained in the publications on our platform. However, Taylor & Francis,our agents, and our licensors make no representations or warranties whatsoever as tothe accuracy, completeness, or suitability for any purpose of the Content. Any opinionsand views expressed in this publication are the opinions and views of the authors,and are not the views of or endorsed by Taylor & Francis. The accuracy of the Contentshould not be relied upon and should be independently verified with primary sourcesof information. Taylor and Francis shall not be liable for any losses, actions, claims,proceedings, demands, costs, expenses, damages, and other liabilities whatsoever orhowsoever caused arising directly or indirectly in connection with, in relation to or arisingout of the use of the Content.

This article may be used for research, teaching, and private study purposes. Anysubstantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,systematic supply, or distribution in any form to anyone is expressly forbidden. Terms &

Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions

Dow

nloa

ded

by [

Uni

vers

idad

Del

Nor

te]

at 1

5:22

06

May

201

5

Relationship between measures of academic quality andundergraduate student attrition: the case of higher educationinstitutions in the Colombian Caribbean region

Anabella Martínez* , Mónica Borjas , Mariela Herrera† and Jorge Valencia

Department of Education, Universidad del Norte, Barranquilla, Colombia

Undergraduate student attrition is a major concern in higher education. It is usuallyexplained by the impact of student attributes; however, recent developments instudent success literature point to the need of exploring institutional practicesthat may impact a student’s decision to abandon their studies. The current weightof academic quality assurance for Colombian higher education institutions (HEI)and what such measures may mean for how HEI fulfill their missions indicatesthe need to consider a possible relation between such quality measures andundergraduate student attrition. Using official databases from the ColombianMinistry of Education for the year 2009, this study explores through analysis ofvariance the relationship between attrition and three measures of academicquality: accreditation status, professional test scores required to graduate (SaberPro Exam) and the number of research groups at HEI. The scope of the study isthe Colombian Caribbean region and the sample includes 19 HEI. Study resultsdemonstrate that the percentage of accredited undergraduate programs at HEIwas the only measure of quality assurance, out of the three explored, thatshowed a statistically significant relationship with undergraduate student attritionrates.

Keywords: accreditation; attrition; Colombia; higher education; quality assurance

Introduction

Undergraduate student attrition can be considered ameasure of the efficiency and qualityof higher education institutions (HEI) (De Vries, León, Romero, & Hernández, 2011).The expansion of enrollments in higher education has paved the way for the applicationof academic quality assurance measures (Gudmut&Martin, 2008). These measures rep-resent a means for HEI to demonstrate the fulfillment of their mission in teaching,research and service; a key element of this mission is undergraduate student success.Nevertheless, there is scarce research on the relationship between academic qualityassurance measures and undergraduate student attrition. Current research states that stu-dents’ decision to abandon their studies may be influenced by institutional factors, suchas the quality of teaching and the HEI reputation given the weight of such aspects onfuture possibilities to compete in the job market (Fondevila, Del Olmo, & Sierra,2011). This paper presents the results of an exploratory study aimed at describing therelationship between undergraduate student attrition rates of HEI in the Colombian Car-ibbean region and three measures of academic quality: accreditation status, student

© 2015 HERDSA

*Corresponding author. Email: [email protected]†Mariela Herrera (1954–2012).

Higher Education Research & Development, 2015http://dx.doi.org/10.1080/07294360.2015.1024622

Dow

nloa

ded

by [

Uni

vers

idad

Del

Nor

te]

at 1

5:22

06

May

201

5

performance on Saber Pro exams and number of research groups at HEI. The ColombianMinistry of Education (MEN, 2010) defines a student as a dropoutwhen they fail to enrollcontinuously for two ormore semesters. The scope of the study focused on theCaribbeanregion because compared to the rest of the country, this region is lagging in terms ofaccess to, and quality of, undergraduate education. According to the MEN (2010),when academic programs have reached certain standards of quality they are positionedto influence their levels of undergraduate student attrition through the enactment ofproactive efforts. While this may be argued, given the extant research on the role insti-tutions can play in promoting student retention, from a policy and practice perspectivefurther evidence is needed on how the enactment of institutional initiatives related toquality assurance can impact undergraduate student attrition.

Undergraduate student attrition is a worldwide phenomenon of concern. In LatinAmerica, Ocampo (2002) reports that most high school graduates do not gain accessto higher education and when they do, a significant number of them drop out duringtheir first year. The report on Higher Education in Latin America and the Caribbean2000–2005 states that the average of young people who have not completed theirdegrees in higher education is 9.1% among men and 8.8% for women, for an overallaverage of 8.5%. In Colombia, it is 6.7% for men and 7.0% for women, for anaverage of 6.9% (Instituto Internacional para la Educación Superior en AméricaLatina y el Caribe [IESALC], 2006). Undergraduate student attrition represents botheconomic and social costs at the individual, institutional and social levels (Castaño,Gallón, Gómez, & Vásquez, 2007; Tinto, 1989). The International Institute for HigherEducation in Latin America and the Caribbean (IESALC, 2006 cited in MEN, 2009a)reports that in 2006 the average rate of student attrition was 55% for each undergraduatestudent cohort. This means that in 2005 undergraduate student attrition cost approxi-mately 11.1 billion US dollars for the 15 countries in Latin America and the Caribbean.

Undergraduate student attrition became a national policy issue in Colombia in 2003(MEN, 2009a). Before 2003, some HEI independently measured their levels of attritionwith the purpose of identifying the causes. These studies resulted in a disparity ofmeasurements, both in terms of time frame and unit of analysis, making it difficultto have an overview of the phenomena on a national, regional and local level, aswell as limited possibilities for conducting follow-up studies. In 2003, the MENadopted a unified methodology to measure undergraduate student attrition andcreated SPADIES (System to Prevent Undergraduate Attrition in Higher Education).Applying the methodology developed by SPADIES, in 2009 Colombia registered anoverall attrition rate per undergraduate student cohort of 44%. There is lower attritionin public HEI (45.3%) when compared to private institutions (52.1%). These ratesinclude attrition in both the technical/technological two-year programs and bachelor/professional five-year programs. The attrition rate of bachelor programs is less(44.2%) when compared to technical (60.6%) and technological programs (52.6%).The situation of undergraduate student attrition in the Caribbean region of Colombia,site of the present study, is not entirely different from the national scenario described.According to a reading of the SPADIES from 2010, the undergraduate attrition rate forthe Caribbean region was 47.8%, which was above the national average (45.30%).

Literature review and research problem

Current studies concur that student attrition is a highly complex and multi-causalphenomenon (Rodríguez & Hernández, 2008). The study of undergraduate student

2 A. Martínez et al.

Dow

nloa

ded

by [

Uni

vers

idad

Del

Nor

te]

at 1

5:22

06

May

201

5

attrition has focused on exploring personal and institutional variables that may influ-ence a student’s decision to abandon their studies. Based on Tinto (1975) much ofthe research is based on an interactionalist model, assuming that the degree of socialand academic integration to the HEI influences students’ decision to abandon school.What influences the level of academic and social integration is explored in studiesthat focus on two groups of variables: (1) specific student attributes and (2) institutionalfactors.

The research focusing on students has explored attributes reflecting demographic,socioeconomic and academic characteristics. Rodríguez and Leyva (2007) state thatin the majority of the cases students at risk of dropping out share academic character-istics such as: enrollment, attendance and grades. These authors further argue that drop-outs who do not have academic problems encounter other difficulties that may influencetheir dropping out of school. For example, poor decisions regarding program of study;major life events (marriage, pregnancy or death) or family issues (moving or loss ofemployment). Rodríguez and Leyva (2007) cluster these issues in categories relatedto personal life stories or family aspects.

On the other hand, the research on institutional factors has focused on exploring theeffect of institutional interventions or ‘policy levers’ (Braxton & McClendon, 2001–2002; Pascarella & Terenzini, 1991) which are specific initiatives implemented withthe purpose of preventing student behaviors that may lead to attrition. In regard to insti-tutional factors, some studies suggest that teaching and student–teacher interactions sig-nificantly influence students’ decisions to abandon higher education (Pérez, 2001).

In Colombia, according to the MEN (2009a), the main cause of undergraduatestudent attrition is academic. This leads us to consider that HEI can play a significantrole in preventing student attrition in the context of assuring academic quality. Weargue that exploring the relation between academic quality measures and student out-comes, such as degree completion, matters because it represents a different approachgiven that such measures focus on strategies institutions can proactively enact toaddress undergraduate student attrition, moving beyond individual student attributesthat are far from institutional control.

Although the phenomenon of undergraduate student attrition has been widelystudied through models that emphasize the impact of variables related to student attri-butes and to specific institutional interventions, the relationship between student attri-tion and measures of academic quality has been scarcely explored. This paper exploresthis relationship in light of three variables: accreditation status, Saber Pro exam resultsand number of research groups and what each of these represents in regard to academicquality: the curriculum, teaching and learning, and research. The following questionguided the study: does the academic quality of HEI represented in terms of accreditationstatus, student performance on the Saber Pro exam and number of research groups ofthe institution have any influence on student attrition?

Conceptual framework

Undergraduate student attrition: concept and models

TheMEN adopted Tinto (1982) and Giovagnoli’s (2002) definition of attrition: the con-dition generated by students who enroll in higher education and do not complete theirstudies. Castaño and colleagues (2007) classify the models that explain undergraduatestudent attrition into four categories. One category considers variables related to

Higher Education Research & Development 3

Dow

nloa

ded

by [

Uni

vers

idad

Del

Nor

te]

at 1

5:22

06

May

201

5

individuals such as age, gender, marital status, family context, health issues and socialintegration, among others (Bean, 1980; Brunsden, Davies, Shevlin, & Bracken, 2000;Nora, Attinasi, & Matonak, 1990; Spady, 1970; Tinto, 1975). Another categoryincludes institutional variables such as lack of financial, academic or psychologicalsupport, limited possibilities for teacher–student interaction outside of the classroom,as well as issues of political and social stability (Adelman, 1999). Socioeconomic vari-ables constitute another category including employment status, student and familyincome, parents’ educational level, among others (Gaviria, 2002; Tinto, 1975). Thelast category involves academic aspects such as achievement, study habits, scores inadmission tests and the quality of the program (Spady, 1970; Tinto, 1975). In sum,there are different models to explain attrition and each focuses on a specific aspectof a multidimensional phenomenon.

In general, Tinto’s model (1975, 1988, 1993) of student integration is the most citedin studies on undergraduate student attrition (Hossler, Ziskin, Moore, & Wakhungu,2008). This model states that attrition is caused by a lack of, or weak, student inte-gration to the social and/or academic environment of the HEI into which they enroll.It recognizes that there may be differences between the students’ interests and expec-tations regarding their experience in higher education, and those of the institutionregarding their academic and social behaviors. Even though the predominant modelto study undergraduate student attrition has been Tinto’s, some researchers havestarted to explore other theoretical possibilities. For example, Bensimon (2007) hassuggested that institutions should assume a proactive stance regarding undergraduatestudent attrition, which implies understanding that attrition is not prompted only bypsychosocial and demographic student characteristics, and emphasizes the need forresearch on what HEI can do in order to prevent undergraduate student attrition. InColombia, academics and educational policy-makers agree that HEI should committo effectively reduce attrition (MEN, 2006). In 2010, 140 presidents of Colombianpublic and private HEI signed a national agreement to reduce attrition. This agreementincludes three strategies to be coordinated by the MEN: to increase financial support forHEI, to support student academic achievement and to strengthen vocational orientation.

Academic quality assurance system in Colombia

The MEN and specifically the National Commission of Accreditation (CNA for itsSpanish acronym) define and measure the quality of a HEI as the distance or differencebetween what a HEI or academic program defines as its goals and its results or accom-plishments. The CNA considers that the academic quality of both programs and HEIshould be assessed mainly by their pertinence to their local and regional contexts.With the purpose of monitoring the quality of HEI in Colombia, the MEN created anational system of quality assurance with two goals: (1) to promote the accountabilityof HEI in response to social and governmental interests by providing reliable infor-mation to students, families, policy-makers and other stakeholders of higher education(2) to foster continuous self-assessment of HEI and their academic programs.

The quality assurance system includes three related subsystems each focusing onone of the following aspects: information, promotion and evaluation. The informationsystem is constituted by five data sources that report on institutional profiles, job statusof graduates and academic quality of HEI. The promotion system focuses on adminis-tering support for HEI in the attainment of infrastructure, research and faculty develop-ment quality standards, among others. The evaluation system aims at verifying the

4 A. Martínez et al.

Dow

nloa

ded

by [

Uni

vers

idad

Del

Nor

te]

at 1

5:22

06

May

201

5

fulfillment of academic, scientific, administrative and infrastructure requirements inacademic programs and HEI. This evaluation takes places at two levels: (1) mandatoryassessment of the minimum conditions required for an academic program to functionand award degrees and (2) voluntary assessment of the conditions leading to the attain-ment of high-quality accreditation for both academic programs and institutions. Theaccreditation process involves both self-assessment (undertaken with input fromfaculty, students, staff and alumni) and external evaluation (undertaken by nationalexperts) of academic and non-academic aspects, and how these together contribute tothe attainment of program and HEI goals. These aspects include mission statements,students, professors, academic processes, student affairs, alumni impact, adminis-tration, financial resources and campus infrastructure. The admission and retentionpolicy, as well as attrition rates discriminated by academic program, are requireddata for both the self-assessment and external evaluation (CNA, 2006). After the exter-nal evaluation, the national experts issue a report that is verified by the MEN to confirmthe attainment of the program or HEI goals. If the MEN considers that the goals are met,it awards accreditation for a specific time period.

In addition to the mandatory assessment of minimum conditions and accreditation,the evaluation subsystem also includes the results of the Saber Pro exam, whichmeasures graduating students’ performance in basic competencies such as writing,quantitative reasoning and reading comprehension. The tests also evaluate specificcompetencies for academic programs as established by Decree 3963 of 2009.Through this exam, the MEN explores the degree of graduating students’ attainmentof program learning outcomes. In Colombia the quality assurance system for highereducation, with all of its different components, is run by the national governmentwho awards accreditation. This is different from other countries such as the USA,England or Australia, where the quality assurance processes for HEI are decentralizedand run mostly by agencies without governmental affiliation.

An important indicator of academic quality is the volume and range of researchactivity of a HEI as a reflection of its capacity to generate new knowledge in thefields represented in its academic structure (Agencia de Calidad, Acreditación y Pro-spectiva de las Universidades de Madrid, 2006; Organización Para La Cooperación yDesarrollo Económicos y Banco Mundial, 2012). Among the entities that feed into thequality assurance system but that are not part of it is the Colombian AdministrativeDepartment of Science, Technology and Innovation (COLCIENCIAS for its namein Spanish). The number of research groups at a HEI recognized by COLCIENCIASand their corresponding products (i.e., research projects, patents and publications)could constitute another measure of quality in HEI. Nevertheless, in Colombiathere is no articulation between COLCIENCIAS and the quality assurance systemdescribed above, which might influence the overall efficiency and efficacy of thehigher education system (Mejía Pardo & Duque Sanchez, 2013). This lack of articu-lation is of concern because it may influence the criteria used to evaluate academicquality of HEI.

While these information subsystems provide key information related to the qualityof HEI, we argue that for the purposes of demonstrating the effects of quality standardson institutional goals, such as undergraduate student retention, there needs to be a stron-ger articulation among them. Accreditation and Saber Pro Exams are indicators of theevaluation subsystem but they are currently not being considered in conjunction whenassessing the quality of academic programs or HEI. Furthermore, other indicators suchas research activity need to be included in these subsystems so that the information it

Higher Education Research & Development 5

Dow

nloa

ded

by [

Uni

vers

idad

Del

Nor

te]

at 1

5:22

06

May

201

5

generates can inform institutional decision-making. The study presented in this paper isan initial attempt in this regard.

Methodology

Data sources and study variables

This study was undertaken with data of HEI of the Caribbean region of Colombia.According to the National System of Information for Higher Education (SNIES forits initials in Spanish), downloaded in August 2011, the Colombian Caribbeanregion, made up of seven provinces (Atlántico, Bolívar, Cesar, Córdoba, La Guajira,Magdalena and Sucre) reported 19 HEI and 183 undergraduate programs that at thatpoint enrolled students and awarded degrees.

The study results stem from a multivariate analysis undertaken with a data set fromfour governmental sources for the 2009 academic year. These were as follows: (1)selected MEN databases (SNIES and SPADIES); (2) The Colombian Institute for theEvaluation of Education (ICFES for its initials in Spanish), specifically its databaseon the results of the Saber Pro national exam; (3) the CNA, which is responsible forthe accreditation of academic programs as well as HEI and (4) COLCIENCIAS, respon-sible for the information regarding research groups and their activities. The year 2009was chosen given the availability of the data for all study variables. Table 1 presentsstudy variables and indicators.

Study variables and their corresponding indicators were determined according tothe conceptual framework. The total attrition rate by periods, as a dependent variable,is an indicator of the internal efficiency of the quality of higher education given itsassociation with retention (MEN, 2009a). Regarding the independent variables, thefirst two, accreditation and Saber Pro Test Scores, correspond to indicators of theColombian Quality Assurance System for Higher Education. The percentage of accre-dited programs was considered because accreditation is an expression of academicquality given the different criteria of both academic and non-academic nature thatare taken into account throughout the accreditation process. The indicator correspond-ing to the percentage of programs with scores in the Saber Pro test above the nationalaverage may point out issues associated with the quality of the teaching and learning,important elements of academic integration that may ultimately affect student retention

Table 1. Selection and operationalization of variables.

Category Source Variable IndicatorVariabletype

Attrition SPADIES Attrition Total attrition rate by period(TAP)

Dependent

CNA/SNIES Certification ofquality

Percentage of accreditedprograms (PAP)

Independent

ICFES Saber Pro TestScores

Percentage of programs withscores in the Saber Pro testabove the national average(PPSP)

Independent

COLCIENCIAS Research Number of research groups(NRG)

Independent

Source: Authors.

6 A. Martínez et al.

Dow

nloa

ded

by [

Uni

vers

idad

Del

Nor

te]

at 1

5:22

06

May

201

5

(Tinto, 1975). The number of research groups, although not included in the ColombianQuality Assurance System, was considered as an indicator of academic quality becauseof its impact on both the program and the institution’s reputation (Arber, Barrere, &Anlló, 2008 cited by UNESCO, 2010; Tapper & Filippakou, 2009). Currently, thereis pressure on Colombian HEI to improve their rankings where research productivityis weighted significantly higher when compared to other institutional characteristicssuch as the quality of teaching. In this context, we decided to include research as anindicator of academic quality with the purpose of exploring the relationship betweenteaching and research, and its influence on undergraduate student success.

The analysis of these variables seeks to verify the following hypothesis: the academicquality of the HEI has a positive and statistically significant effect on undergraduatestudent attrition (H1). This study assumed that an increment in an institution’s indicatorsof academic quality, expressed through: (1) accreditation status (2) Saber Pro test resultsand (3) the capacity to generate research may be associated to decreasing levels of attri-tion rate by academic period. We expect that institutions with higher levels in these indi-cators of academic quality will have lower levels of undergraduate student attrition, andgiven the study’s methods, will result in negative coefficients.

Sample

The study sample is constituted by HEI of the Colombian Caribbean region offeringundergraduate programs during 2009. This implies a total of 19 HEI, with seven ofthem (37%) being public and 12 (63%) private. Table 2 summarizes HEI fulfillingthis criterion.

Data analysis

The inferential analysis took place in two phases. The first phase consisted of a bivariateanalysis that separately examined each of the independent variables, as well as theirrelation to the dependent one, with the goal of examining the level of correlationbetween them. The purpose of this phase was to not underestimate variables withlow individual influence, but with significant possible contributions when measuredin conjunction with the others. The second phase consisted in the definition of aregression model. The technique used was multiple regression of the form:

Y = b0+ b1X1+ b2X2+ bk Xk + u. (1)

This technique allowed for hypothesis verification, as well as the quantification of thecontribution of the independent variables, in conjunction with the dependent variable.The study sought to prove whether the conditions of academic quality met by the HEIcould affect the variability of undergraduate student attrition. Taking into account theneed to guarantee that the dependent variable (TAP) followed a normal distribution, itwas necessary to transform such a variable. For this, the study followed Yeo and Jonson(2000, cited by Martínez, Castañeda, & Domínguez, 2006) who point out that in caseswhere the data do not follow normal distribution, a transformation of the model isadvised in order to obtain a distribution close to normal. In this regard, Martínez andcolleagues (2006) point out that the Box–Cox transformation is a commonly used tech-nique. Box and Cox’s proposal is based on the use of methods of maximum

Higher Education Research & Development 7

Dow

nloa

ded

by [

Uni

vers

idad

Del

Nor

te]

at 1

5:22

06

May

201

5

verisimilitude to determine the most appropriate power to meet the assumption ofnormality in a group of data. Taking into consideration what these authors suggest totransform the model proposed in Equation (1), the transformation’s parameter of theindependent variable would be given by the following expression:

l = 1+ (TAP−0.95 − 1)(−0.95∗0.148814−1.95) . (2)

The difference between the distribution of the original variable and the transformedvariable applying the expression (2) is represented in Figure 1.

Such adjustment made it possible to approximate to a normal distribution of thedata.

Results

The bivariate analysis presented in Figure 2 indicates that as the percentage of accre-dited programs increases, the total attrition rate by period decreases. This indicatesan inverse proportional relationship between the independent variable TAP and thedependent variable PAP. In regard to the second independent variable considered inthe study (percentage of programs with scores in the Saber Pro test above the nationalaverage (PPSP)) results show that the total attrition rate by period increases as theaverage program scores on these exams decreases. In regard to the variable NRG

Table 2. Number of undergraduate programs by HEI in the Colombian Caribbean in 2009.

Higher Education Institutions (HEI)Number of on-campus undergraduate programs

functioning in 2009

Corporación Universitaria de la Costa 17Corporación Universitaria de Sucre 3Corporación Universitaria del Caribe 19Corporación Universitaria Rafael Núñez 13Fundación Universidad del Norte 22Fundación Universitaria Tecnológico

Comfenalco5

Institución Universitaria Bellas Artes yCiencias de Bolívar

4

Universidad Autónoma del Caribe 22Universidad de Cartagena 29Universidad de Córdoba 25Universidad de La Guajira 19Universidad de Sucre 8Universidad del Atlántico 30Universidad del Magdalena 30Universidad del Sinú – Elías Bechara

Zainum21

Universidad Metropolitana 11Universidad Popular del Cesar 22Universidad Simón Bolívar 15Universidad Tecnológica de Bolívar 17

Source: Authors.

8 A. Martínez et al.

Dow

nloa

ded

by [

Uni

vers

idad

Del

Nor

te]

at 1

5:22

06

May

201

5

(number of research groups), the results show that as the number of research groupsincreases, the total attrition rate by period decreases, but this tendency is muchweaker when compared to the correlation between TAP and PAP.

In addition to the information presented in Figure 2, Table 3 includes the correlationcoefficients of the dependent variable TAP in relation to the independent variablesPAP, NRG and PPSP. The data in the table show that the coefficient of lineal corre-lation between the total attrition rate by period (TAP) and the percentage of accreditedprograms (PAP) is −0.44735547, which indicates the existence of a linear, negative andsignificant relation. In regard to the correlation coefficient between total attrition rate byperiod (TAP) and number of research groups (NRG) is −0.2024642 which also demon-strates a negative lineal relation.

In addition, the existence of a positive correlation coefficient between PAP andPPSP may be evidence of a possible serial correlation. This was assessed in theregression model, which constitutes the second phase of the analysis. The regressionmodel was run including all variables with the purpose of not underestimating possiblejoint influence of these on the dependent variable PAP.

The following step was taken to evaluate the regression model by introducing thevariables one at a time, setting aside those with no significant contributions to the varia-bility of the dependent variable, following the model’s functional form:

TAP(l) = b0 − b1NRG− b2PAP− b3PPSP+ m. (3)

Figure 1. Smooth probability density function: original variable and transformed variable.

Figure 2. Dispersion of TAP in relation to PAP, NRG and PPSP.Source: Authors.

Higher Education Research & Development 9

Dow

nloa

ded

by [

Uni

vers

idad

Del

Nor

te]

at 1

5:22

06

May

201

5

The functional form of the model shows that if the indicators of quality are improvedthen the undergraduate student attrition rate for the studied period (2009) shoulddecrease. The adjusted curve by applying the prior procedure was

TAP(l) = 0.890708− 0.304447∗PAP. (4)

R2 = 0.500399Adjusted R2 (adjusted for df) = 0.412234Standard error = 0.0556937Mean absolute error = 0.0373052Durbin–Watson Statistic = 2.09591 (p = 0.5677)

The p-value of the independent variables NRG (0.2773) and PPSP (0.1140) makesevident the lack of a statistically significant relationship between these variables andthe dependent variable. Therefore, these two variables were not taken into accountfor the final model given their p-values. Regarding the percentage of accredited pro-grams (PAP) which was included in the final model, the results of the analysis of var-iance (see Table 4) show the existence of a significant relationship betweenundergraduate student attrition rates and the PAP of each of the HEI included in the

Table 3. Correlation Matrix for TAP, PAP, NRG and PPSP.

Variables Measure

Indicator

TAP PAP NRG PPSP

TAP (total attritionrate by period)

Correlationcoefficient

1 −0.44735547 −0.2024642 −0.23223993

p-Value 0.0547972 0.40582214 0.33869102PAP (percentage of

accreditedprograms)

Correlationcoefficient

−0.44735547 1 0.32178677 0.72566382

p-Value 0.0547972 0.1791184 0.00043693NRG (number of

research groups)Correlation

coefficient−0.2024642 0.32178677 1 0.20764286

p-Value 0.40582214 0.1791184 0.39365813PPSP (percentage

of programs withscores in theSaber Pro testabove thenational average)

Correlationcoefficient

−0.23223993 0.72566382 0.20764286 1

p-Value 0.33869102 0.00043693 0.39365813

Source: Authors.

Table 4. Analysis of variance.

Source Sum of squares df Mean square F ratio Significance (p-value)

Model 0.0528147 3 0.0176049 5.68 0.0070Residual 0.0527304 17 0.00310179Total correlation 0.105545 20

Source: Authors.

10 A. Martínez et al.

Dow

nloa

ded

by [

Uni

vers

idad

Del

Nor

te]

at 1

5:22

06

May

201

5

sample with a confidence interval of 95.0%. In addition, the results indicate the theor-etical coincidence of the coefficients since it is expected that the variable TAP(λ)

decreases with higher levels in the indicators of institutional academic quality. Inthis sense, the parameter shows that the increase in the PAP in a HEI correlates witha reduction of the attrition rate by academic period. The R2 shows that the modelexplains a 50.0399% of the variability in TAP(λ). The adjusted statistic R2 (which ismore convenient to compare models with different numbers of independent variables)is 41.2234%.

Taking into consideration the definition of the study’s variables, the originalhypothesis can be expressed in the following manner: H1: The PAP of the HEI has apositive and statistically significant effect over the HEI attrition’s rate by academicperiod. To verify this hypothesis, the null version: H0 β1≤ 0 was posed against itsalternative H1 β1≥ 0 (when considering the calculated p-value of the regression com-pared with a selected). Given that the p-value (0.0070) is lower than 0.05 there isenough evidence to reject the null hypothesis. Therefore, the alternative hypothesisresults are more plausible: the PAP may have a positive and statistically significantinfluence over the HEI undergraduate student attrition rate by academic period in theColombian Caribbean region.

Discussion

The study’s results show that there is a positive and statistically significant relationshipbetween accreditation and undergraduate student attrition. Nevertheless, accreditationstatus in itself is not a guarantee of lower undergraduate student attrition. The valueof R2 seems to agree with the literature claiming that attrition is a complex phenomenoncaused by multiple factors such as individual circumstances, which were not taken intoaccount in this study. Nevertheless, study results in this regard agree with claims thatinstitutions could make some efforts to lower their attrition rate (IESALC, 2006;Ocampo, 2002). Based on study results, we argue that the study of undergraduatestudent attrition through specific and isolated institutional interventions is limited inregard to the role that an institution can play because it does not include institutionwide efforts aimed at assuring the quality of academic programs specifically in termsof their pertinence and relevance. This is precisely what the accreditation process inColombia addresses: the degree to which academic programs and HEI meet the stan-dards and goals they have set out for themselves in regard to teaching, research andservice.

Studies have demonstrated that some institutional factors associated with the aca-demic quality of higher education may impact undergraduate students’ permanencein an institution (Pineda, Pedraza, & López, 2011; Yorke & Longden, 2004). Theseinclude innovation in professors’ pedagogical practices, the organization and adminis-tration of the academic program, student–teacher ratio, academic support available tothe students and extracurricular activities, among other aspects. These factors coincidewith the criteria enforced by the CNA to award accreditation, which include both a self-study and external evaluation of mission statement, academic processes and studentaffairs, impact of the program on alumni, administration, financial resources andcampus infrastructure, among others. The admission and retention policy, as well asattrition rates disaggregated by academic program are data required for both the self-assessment and external evaluation (CNA, 2006). Undergraduate student attrition,and specifically what is being done at the institutional and program levels to address

Higher Education Research & Development 11

Dow

nloa

ded

by [

Uni

vers

idad

Del

Nor

te]

at 1

5:22

06

May

201

5

it, constitutes one of the indicators that are considered to bestow accreditation inColombian higher education.

Study results indicate that accreditation may go beyond the inventory of adminis-trative and operational processes required for the functioning of academia to possiblyinfluence key indicators in higher education, in this case attrition. In ColombianHEI, accreditation can be the guarantor for engaging in continued institutional improve-ment given the temporary nature of the accreditation status and the process involved inrenewing it. While obtaining accreditation implies the attainment of a specific status,the process involved requires the analysis of how academic and non-academic variablescome together and impact the quality of a program. In this regard, the level of under-graduate student attrition represents an indicator of institutional effort to provide qualityeducation to all of its students.

Regarding Saber Pro scores, study results show that there is not a significantrelationship between them and undergraduate student attrition. The results questionif this measure of academic quality is appropriate to relate to undergraduate studentattrition due to its nature and the timing of when students take it, meaning that it ishighly likely that the students who take the Saber Pro exam are those who werenot at risk of abandoning their studies in the first place. Finally, regarding thenumber of research groups, results also showed no statistically significant relation-ship between this variable and the rates of undergraduate student attrition. Tinto(1975, 1988, 1993) considers the attainment of students’ integration as a fundamen-tal aspect to prevent attrition, and the student–teacher relationship an important partof student academic integration. Ballantine (1989) and Moreland-Young (1983)show that in HEI with a strong emphasis on research, teaching tends to be neg-lected. This study’s results show that this is not necessarily the case in the Carib-bean region of Colombia; that is, a greater number of research groups does notcorrelate with increased student attrition. Nevertheless, Arber and colleagues(2008) state that the level of its research productivity positions Colombia in asecond-tier group when compared to other countries in Latin America such asArgentina and Mexico. In addition, the indicator itself – number of researchgroups – may not reflect substantive characteristics related to the quality ofresearch such as time dedication, and impact of knowledge, among other aspects.

Regarding limitations of the study, as with any quantitative study that uses second-ary sources of information, the quality of the data cannot be guaranteed. In addition,statistical models may be limited by not including all variables that could be influencingthe phenomenon. Furthermore, the majority of the variables used to operationallydefine academic quality were determined by the quality assurance model adopted bythe MEN and the availability of information in its databases.

Study results point to future research possibilities. Qualitative studies can furtherexplore what happens within each HEI to comprehend the relation between accred-itation and attrition from the perspective of institutional actors. Furthermore, giventhe different institutional characteristics that are considered for the accreditationmodel in Colombia, quantitative studies could specify the influence of each ofthose aspects on undergraduate student attrition. Taking into consideration that theresults correspond to the Colombian Caribbean region in a specific time frame, itwould be pertinent to undertake the same analysis at a national level in subsequentyears in order to identity trends in other regions and in more recent academicperiods.

12 A. Martínez et al.

Dow

nloa

ded

by [

Uni

vers

idad

Del

Nor

te]

at 1

5:22

06

May

201

5

Conclusions

HEI can proactively address undergraduate student attrition through institutional effortsrelated to academic quality assurance. Study results show that among the measures ofacademic quality in the Colombian higher education system considered both individu-ally and collectively only accreditation is related to undergraduate student attrition forHEI in the Caribbean region. This could be the case because accreditation may guaran-tee to students, prior to their admission, that their expectations will be met given theattainment of quality standards on the part of the program and HEI. Regarding theother two variables, study results do not support a strong correlation with undergraduatestudent attrition. Nevertheless, there is a minor association between higher scores inSaber Pro exams and lower levels of undergraduate student attrition. Results also indi-cate a positive correlation between accreditation and Saber Pro exam results, meaningthat accreditation is associated with a higher percentage of students obtaining betterscores than the national mean. This may be the case given that the variables that inter-vene in the attainment of high scores on this exam may be similar to those consideredfor accreditation. As of 2013, Saber Pro exam results are weighed into accreditation.Finally, regarding the third variable considered, the number of research groups doesnot seem to be correlated with attrition or accreditation. Nevertheless, we argue thatresearch itself should be situated in the quality assurance system enacted by theMEN, given what research may represent for the attainment of academic quality.

Given the voluntary nature of accreditation in Colombian HEI and low number ofacademic programs in the Colombian Caribbean region that have attained accreditation,from a policy perspective it would be pertinent to consider whether accreditation shouldbecome mandatory for academic programs and become a guarantor for engaging incontinued institutional improvement. Institutional efforts to obtain accreditation maypay off in performance measures such as undergraduate student retention leading tothe conclusion that in regard to reducing attrition, HEI can assume a proactive role.

ORCID

Anabella Martínez http://orcid.org/0000-0002-8812-7270Mónica Borjas http://orcid.org/0000-0001-7823-477XJorge Valencia http://orcid.org/0000-0003-1402-2369

ReferencesAdelman, C. (1999). Answer in the toll box: Academic intensity, attendance patterns and bache-

lor’s degree attainment. Washington, DC: U.S Department of Education.Agencia de Calidad, Acreditación y Prospectiva de las Universidades de Madrid. (2006).

Estudio internacional sobre criterios e indicadores de calidad de las universidades.Madrid: Author. Retrieved March 8, 2014, from http://goo.gl/O3dywX

Arber, G., Barrere, R., & Anlló, G. (2008). Measuring R&D in developing countries:Measurement priorities and methodological issues in Latin America. UNESCO (2010).Measuring R&D: Challenges faced by developing countries. Montreal: UNESCO.Retrieved June 12, 2014, from http://www.uis.unesco.org/Library/Documents/tech%205-eng.pdf

Ballantine, J. (1989). University teaching around the world. Teaching Sociology, 17(3), 291–296.Bean, J. (1980). Dropouts and turnover: The synthesis and test of a casual model of student attri-

tion. Research in Higher Education, 12(2), 155–187.Bensimon, E. (2007). The underestimated significance of practitioner knowledge in the scholar-

ship on student success. The Review of Higher Education, 30(4), 441–446.

Higher Education Research & Development 13

Dow

nloa

ded

by [

Uni

vers

idad

Del

Nor

te]

at 1

5:22

06

May

201

5

Braxton, J.M., & McClendon, S.A. (2001–2002). The fostering of social integration throughinstitutional practice. Journal of College Student Retention: Research, Theory andPractice, 3(1), 57–71.

Brunsden, V., Davies, M., Shevlin, M., & Bracken, M. (2000). Why do HE students drop out? Atest of Tinto’s model. Journal of Further and Higher Education, 24(3), 301–310.

Castaño, E., Gallón, S., Gómez, K., & Vásquez, J. (2007). Análisis de los factores asociados a ladeserción y graduación estudiantil en la Universidad de Antioquia. Lecturas de Economía,65, 9–36.

Consejo Nacional de Acreditación. (2006). Lineamientos para la acreditación institucional.Serie Documentos CNA #2. Bogotá: Corcas Editores.

De Vries, W., León, P., Romero, J., & Hernández, I. (2011). ¿Desertores o decepcionados?Distintas causas para abandonar los estudios universitarios. Revista de la EducaciónSuperior, 11(4), 29–49.

Fondevila, J., Del Olmo, J., & Sierra, J. (2011). Universidades privadas, ¿un subsistema difer-ente?: cómo desarrollan los estudiantes sus preferencias hacia la universidad. Revista deComunicación Vivat Academia, (114), Madrid. Universidad Complutense. Retrieved June26, 2014, from http://pendientedemigracion.ucm.es/info/vivataca/numeros/n114/PDFs/UNIVERSIDADES.pdf

Gaviria, A. (2002). Los que suben y los que bajan. Educación y movilidad social en Colombia.Bogotá: Fedesarrollo, Alfa omega.

Giovagnoli, P. (2002). Determinantes de la deserción y graduación universitaria: unaaplicación utilizando modelos de duración. Documento de Trabajo 37. La Plata:Universidad Nacional de la Plata. Retrieved March 28, 2013, from http://www.alfaguia.org/alfaguia/files/1318958623Determinantes%20de%20la%20desercion%20y%20graduacion%20universitaria%20una%20aplicacion%20utilizando%20modelos%20de%20duracion.pdf

Gudmut, H., & Martin, M. (Eds.). (2008). Accreditation and the global higher educationmarket. Paris: UNESCO.

Hossler, D., Ziskin, M., Moore, J.V., & Wakhungu, P.K. (2008, November). Institutionalfactors that contribute to student persistence. Views from three campuses. Paper presentedat the Annual Conference of the Association for the Study of Higher Education Jacksonville,Indiana University. Retrieved June 19, 2014, from http://pas.indiana.edu/cb/docs/institutional%20factors%20ashe%2008.pdf

Instituto Internacional para la Educación Superior en América Latina y el Caribe (2006). Informesobre la educación superior en América Latina y el Caribe. 2000–2005. La metamorfosis dela educación superior. Caracas: Editorial Metrópolis. Retrieved May 16, 2014, from http://www.radu.org.ar/Info/2%20IESALC.pdf

Martínez, J., Castañeda, N., & Domínguez, J. (2006).Una propuesta general de transformación anormalidad. Segundo congreso estatal ‘la investigación en el posgrado. ISBN: 970-728-053-0.Retrieved June 26, 2014, from http://www.uaa.mx/investigacion/memoria2006/pdf/ingenierias/Jimenez_Martinez_Jose_de_Jesus.pdf

Mejía Pardo, D.A., & Duque Sánchez, L. (2013). Aproximaciones a la evaluación del Sistemade Aseguramiento de la Calidad para la Educación Superior en Colombia. RetrievedMarch 8, 2014, from http://goo.gl/o4TRX2

Ministerio de Educación Nacional. (2006). Acceder para quedarse: Cobertura con permanenciaeducación superior. Boletín Informativo, 6. Bogotá: Author.

Ministerio de Educación Nacional. (2009a). Deserción estudiantil en la educación superiorcolombiana. Metodología de seguimiento, diagnóstico y elementos para su prevención.Bogotá: Author.

Ministerio de Educación Nacional. (2009b). Decreto 3963 de 2009. Retrieved March 14, 2014,from http://www.mineducacion.gov.co/1621/articles-205955_archivo_pdf_decreto3963.pdf

Ministerio de Educación Nacional. (2010). Educación para la prosperidad. Acuerdo nacionalpara disminuir la deserción en educación superior. Retrieved March 14, 2014, fromhttp://www.colombiaaprende.edu.co/html/directivos/1598/articles-232780_Acuerdo.pdf

Moreland-Young, C. (1983). Teaching analytical and thinking skill in a content course. In P. A.Lacey (Ed.), Revitalizing teaching through faculty development (pp. 41–48). San Francisco,CA: Jossey-Bass.

14 A. Martínez et al.

Dow

nloa

ded

by [

Uni

vers

idad

Del

Nor

te]

at 1

5:22

06

May

201

5

Nora, A., Attinasi, L., & Matonak, A. (1990). Testing qualitative indicators of precollege factorsin Tinto’s model of college withdrawal: A community college student population. Review ofHigher Education, 13(3), 337–356.

Ocampo, J.A. (2002). Panorama social de América Latina 2001–2002. Retrieved May 16,2014, from http://www.eclac.cl/prensa/noticias/comunicados/0/11260/ps2001-2002fin.pdf

Organización Para La Cooperación y Desarrollo Económicos y Banco Mundial (2012).Evaluaciones de políticas nacionales de educación. La educación superior en Colombia.París: Editions OCDE.

Pascarella, E.T., & Terenzini, P.T. (1991). How college affects students: Vol. 1. Findings andinsights from twenty years of research. San Francisco, CA: Jossey-Bass.

Pérez, L. (2001). Los factores socioeconómicos que inciden en el rezago y la deserción escolar.En Chain Revueltas, R. (comp.), Deserción, rezago y eficiencia terminal en las IES (pp. 97–112). México: ANUIES.

Pineda, C., Pedraza, A., & López, C. (2011). Persistencia y graduación. Hacia un modelo deretención estudiantil para instituciones de Educación Superior. Bogotá: MEN,Universidad Javeriana.

Rodríguez, J., & Hernández, J. (2008). La deserción escolar universitaria en México. La experi-encia de la Universidad Autónoma Metropolitana Campus Iztapalapa. ActualidadesInvestigativas en Educación, 8(1), 1–30.

Rodríguez, J., & Leyva, M. (2007). La deserción escolar universitaria. La experiencia de laUNAM. Entre el déficit de la oferta educativa superior y las dificultades de la retenciónescolar. El cotidiano, 22(142), 98–111.

Spady, W. (1970). Dropouts from higher education: An interdisciplinary review and synthesis.Interchange, 1(1), 64–85.

Tapper, T., & Filippakou, O. (2009). The world-class league tables and the sustaining of inter-national reputations in higher education. Journal of Higher Education Policy andManagement, 31(1), 55–66.

Tinto, V. (1975). Dropouts from higher education: A theoretical synthesis of the recent litera-ture. A Review of Educational Research, 45(1), 89–125.

Tinto, V. (1982). Limits of theory and practice in student attrition. Journal of Higher Education,53(6), 687–700.

Tinto, V. (1988). Stages of students departure: Reflections on the longitudinal character ofstudent leaving. Journal of Higher Education, 59(4), 438–455.

Tinto, V. (1989). Definir la deserción: una cuestión de perspectiva. Revista de la EducaciónSuperior, 18(71). Distrito Federal, México, Asociación Nacional de Universidades eInstituciones de Educación Superior (ANUIES).

Tinto, V. (1993). Leaving college: Rethinking the causes and cures of student attrition (2nd ed.).Chicago: University of Chicago Press.

Yorke, M., & Longden, B. (2004). Retention and success in higher education. York: Society forResearch into Higher Education & Open University Press.

Higher Education Research & Development 15

Dow

nloa

ded

by [

Uni

vers

idad

Del

Nor

te]

at 1

5:22

06

May

201

5