Organizational resilience and enrollment trends of independent,for-profit higher education institutions

Abstract

BACKGROUND: From 2010 to 2012, the for-profit sector of higher education in the United States (otherwise known as career colleges) existed in a turbulent environment, characterized by regulatory, media, and public scrutiny. While virtually all career colleges experienced enrollment declines during this period, by 2012 some colleges were starting to see this trend stabilize or reverse, whereas others did not.

OBJECTIVE: The purpose of this study was to determine if the differences in career colleges’ enrollment trends could be attributed to organizational resilience.

METHODS: A quantitative correlation study using a multiple regression analysis was conducted to determine the nature of the relationship between organizational resilience and the enrollment fluctuations of 59 career colleges located throughout the United States.

RESULTS: The correlation between organizational resilience levels and enrollment fluctuations was fair to moderate and significant, r = 0.40, p < 0.05. A multiple-regression analysis revealed that the model significantly explained the impact of the six organizational resilience factors on enrollment fluctuations, F = 4.15, p < 0.01. The R² for the model was 0.32, and the adjusted R² was 0.25. In terms of individual organizational resilience factors, two tested either significantly or moderately significantly: avoidance-skepticism and critical understanding or sensemaking.

CONCLUSIONS: Recommendations for college leaders include monitoring the level of avoidance to ensure a healthy balance of skepticism regarding new situations and incorporating strategies to help organizational members increase their levels of critical understanding or sensemaking.

Keywords

Private colleges adjustment organizational change workplace resilience

1 Introduction

In the decade of 2000 to 2010, for-profit colleges saw an almost meteoric increase in enrollments. Many industries undergo turbulence, especially after periods of fast growth, like for-profit higher education had experienced [1]. In 2010, this sector started to experience such turbulence as the institutions came under regulatory and media scrutiny, were the subject of government accountability audits that questioned their outcomes [2], and were forced to deal with major changes in student loan eligibility rules that caused many of them to overhaul their operations [3]. The challenges created from this turbulent environment had a profound impact on enrollment levels for many of these colleges.

By the end of 2011, just one year after the turbulence began, the for-profit chains such as University of Phoenix, Corinthian Colleges, Devry, and ITT, reported enrollment declines as high as 30% [4]. The trend of enrollment declines continued for some but stabilized and even reversed for others over the following year [3]. It appeared that some colleges had been able to weather the storm of negative regulatory and public scrutiny. A review of enrollment fluctuations showed a wide disparity for both the chains [3] and the independent for-profit colleges [5].

Despite the turbulent environment of for-profits, some of the colleges appeared to be bouncing back, whereas others did not. The independent for-profit colleges were chosen as the target population for this study because their enrollment fluctuation trends differed from the sharp enrollment declines experienced by the larger chain schools [3]. Moreover, they have often been omitted from academic research. The central question of this research, therefore, was what differentiated the enrollment fluctuations of independent for-profits from the chain schools? The research was conducted to determine if a differentiating factor was organizational resilience.

The study of organizational resilience was shown to be pertinent, especially when comparing institutions in the same sector that have experienced the same turbulence or crisis but have had different results postcrisis. Gittel, Cameron, Lim, and Rivas [6] compared airlines’ performance before and after September 11, 2001, and found links to organizational resilience. Chewning, Lai, and Doerfel [7] compared New Orleans businesses that were able to bounce back after Hurricane Katrina to those organizations that perished. More recently, the 2008 economic crisis had been a common stream of research to compare the organizational resilience of financial institutions before and after 2008 [8 –10].

The core line of investigation for this study was based upon similar comparisons of preturbulence and postturbulence, the resiliency to bounce back, and bounce forward. The enrollment levels of independent, for-profit colleges were compared prior to the turbulent regulatory environment of the 2010-2011 school year to their enrollment levels in 2012. The research questions were based upon the hypothesis that those colleges that bounced back from the turbulence had higher levels of organizational resilience than those that did not. The hypotheses dealt first with organizational resilience as a whole and second with the individual factors that shape it. While there are several influences that may impactenrollment fluctuations [4 , 11], organizational resilience was chosen as the focus for this study because the factors that shape it are largely within the control of the organizational leader [12] and, as stated earlier, resilience has been found to be relevant to organizations in other sectors that have been able to bounce back after turbulence [6 –10].

Organizational resilience has been studied in several industries: health care [13 –15], banking [8, 9], engineering [16], hospitality [17], community services [12 , 18–20], airlines [6], real estate [21], pharmaceuticals [22, 23], and petrochemicals [24]. However, only a few studies of organizational resilience have focused on institutions of higher learning [25 –27]. It is notable that many of the organizations studied for their resilience are for-profit institutions [6 , 28–33]. The colleges in this study were both for-profit organizations and institutions of higher learning. Their for-profit status provided a transition between research already conducted on organizational resilience and institutions of higher learning. This study, therefore, fills a research gap for both higher education and for-profit institutions in terms of their workplace resilience.

For-profit higher education institutions are those colleges and universities that operate as for-profit corporations and identify profit as one of their operational goals [5]. Deming, Goldin, and Katz [5] provided an important distinction between the colleges in this sector by defining independent schools as those that operate five or fewer campuses, usually in one state, and chains or chain schools (such as Devry and University of Phoenix), which represent the multistate, multicampus operations that are well known to the average consumer. These institutions, particularly the smaller independent ones, have been commonly called career colleges [5]. Career colleges were the focus of this workplace resilience study as they reflected similar patterns of bouncing back after turbulence as airlines [6] and the banking industry[8, 9].

2 Background

2.1 Organizational resilience and its origins

Weick [34] first referred to the need for organizations to be resilient when he compared the destructive fire at Mann Gulch to crises that face organizations, such as bankruptcy or even organizational collapse. Acar and Winfrey [35] synthesized theliterature about organizations that were able to sustain performance and overcome periods of turbulence. These concepts provided a basis for Mallak’s [14] measurement tool that was further tested by Somers in 2009 in the field of emergency services in public works departments [12].

A precursor to organizational resilience is the concept of organizational adaptation. Thomas, Francil, John, and Davies [36, p. 240] asked the following question: “What distinguishes organizations that flourish from those that flounder within the same competitive environment?”. Thomas et al. distributed questionnaires, written with a case scenario methodology, to hospital chief executive officers. Their performance was measured based on occupancy rates, profit per discharge, and the ratio of admissions to beds. The authors found that organizational adaptation (i.e., the ability to change plans based on a wide array of information sources) was positively correlated with performance.

The concepts of adaptation [36] and learning from change are components of theories of complexity management, which established the roots of organizational resilience theory. In his complexity management studies of firms in Brazil and China, Sull [37] asked a similar question that had been posed in the Thomas et al. [36] study more than a decade earlier: What differentiates strategies of organizations that succeed against those that fail within the same environment? Sull matched successful firms with unsuccessful firms within the same industry in order to compare their organizational strategies. Sull concluded that unsuccessful firms practiced linear strategic planning, characterized by planning and implementation, with little modification or contingency planning. The successful firms approached strategic planning in more of a loop. Sensemaking of failures was more common, and learning and adjusting from these failures aided in the creation of new strategies [37]. Overall, the strategies of the successful firms were more complex thanlinear.

Limnios, Mazzarol, Ghadouani, and Schilizzi [38] traced the origins of organizational resilience back even further than complexity theorists, noting that resilience was first mentioned in ecological systems theory in the early 1970s and shifted gradually into socioecological systems before being considered part of complexity theory. The authors also noted evidence of engineering influences in the concepts of robustness and stability and pointed to the fact that organizational resilience is often misinterpreted as the most desirable state.

2.2 Measuring organizational resilience

Mallak developed scales in 1998 to measure organizational resilience for his study of health-care organizations using a target sample of nursing executives in Michigan. Mallak based the initial survey questions from the characteristics describing organizational resilience that had been outlined in research by Weick [34] and Acar and Winfrey [35], specifically bricolage (i.e., ability to work under pressure, ability to access resources, and ability to react appropriately), attitude of wisdom (i.e., healthy mixture of skepticism and curiosity), and virtual role system (i.e., how well members understand each other’s roles and the amount of cross-training betweenroles).

After multiple rounds of reliability, validity, and sample adequacy testing with the groups of nursing executives, six factors were retained that were deemed to be the necessary traits of highly resilient organizations.

Goal-directed solution seeking. Adapting through goal-directed solution seeking was defined as “bringing together the need for goals and a vision to guide creative processes in seeking solutions to problems” [39, p. 586].

Avoidance-skepticism. Mallak’s [39] early definition of avoidance outlined the contradictions between backing off from chaotic situations and the earlier concept of bricolage, entailing confrontation of situations and solving them with the tools at hand. Through different iterations of this theory, however, avoidance became more in line with “approaching new situations with skepticism” [14, p. 152]. This factor was labeled avoidance-skepticism to reflect the wisdom of picking and choosing which situations toconform.

Sensemaking and critical understanding. Sensemaking was defined as using multiple sources of information to scan, interpret, and make sense of the environment and the risks associated with it [36]. Additionally, a component of Mallak’s [14] definition of organizational resilience that traced itself back to the original references of the concept [34] was sensemaking.

Role dependence. Role dependence was identified as a key to functioning resiliently as a team. Organizations high in role dependence have individuals who can easily fill in for a missing coworker [14].

Source resilience. Source resilience was described as a characteristic of resilient individuals and organizations that gather information from varied sources [39]. Mallak [14] compared source resilience to a journalist who uses multiple sources of information to arrive at an unbiased conclusion.

Resource access. Resource access was defined as the ability to acquire and distribute the necessary resources to solve a problem [14]. Resilient organizations, according to Mallak [14], have a resource allocation strategy that gives them access to resources before, during, and after crises. The individuals in that organization would access resources to solve a problem, with or without authorization to do so [14].

These six factors became the basis for the Organizational Resilience Potential scale, which was a tool used in the subsequent research of organizational resilience by Somers [12]. In the study conducted by Somers, municipal public works departments were measured for what he defined as “latent resilience” or “resilience that is not presently evident or realized” [12, p. 13]. Somers hypothesized that organizations that score high on the Organizational Resilience Potential scale will have a “greater propensity to display adaptive behaviors necessary for dealing with emergencies” [12, p. 13]. He questioned the extent to which managers can positively impact organizational resilience. Somers found that managers impact organizational resilience primarily through their information-seeking behaviors and their planning processes. Like Somers’ research, the ways that managers can impact or control aspects of organizational resilience was a focus of this study

3 Method

The purpose of this quantitative correlational study was to examine the relationship between the predictor variable of organizational resilience and the outcome variable of enrollment fluctuations in independent, for-profit higher education institutions. Additionally, the intent of this study was to examine the differences among the six predictor variables that shape organizational resilience [14] and the extent of their relationship with the outcome variable of enrollment fluctuations.

3.1 Research questions

The quantitative correlational design was grounded in the study’s research questions:

To what degree does a relationship exist between organizational resilience and three-year enrollment fluctuations of independent, for-profit higher education institutions?

Which predictor variables that shape organizational resilience (i.e., goal-directed solution seeking, avoidance-skepticism, critical understanding or sensemaking, role dependence, source resilience, and access to resources) have a greater impact on the outcome variable (i.e., three-year enrollment fluctuations of independent, for-profit higher education institutions)?

These research questions mandated that only higher education institutions deemed as independent (i.e., five campuses or less) and for-profit were considered for this study. As a result, because institutions had to meet certain criteria to be included in the sample, criterion or purposive sampling was used [40].

3.2 Participants

The target population for this study included independent, for-profit higher education institutions. The target sample, or research participants in this study, were leaders (e.g., C-level executives, presidents, vice-presidents, and directors) from 475 independent, for-profit higher education institutions across the United States.

Participating colleges were recruited from a list of independent, for-profit higher education institutions obtained from the IPEDS (Integrated Post-secondary Education Data System). This database is maintained by the U.S. Department of Education’s National Center for Education Statistics. It contains the list of all postsecondary institutions in the United States that have a program participation agreement with the U.S. Department of Education. These institutions are also known as Title IV institutions. Beauty schools, barbering schools, and cosmetology schools were not included in the target sample.

The unit of analysis for this study was the institution, more specifically any institution from the IPEDS list with a unique unit identification number (unit ID), which was assigned to postsecondary institutions included in the IPEDS. Any institutions that were part of a group that had more than five unique unit IDs were eliminated from the target sample [5]. Because the institution was the unit of analysis for this study, multiple responses from any one institution were considered as one response by calculating the mean of those responses from that respective institution. Even though individuals completed Mallak’s [41] version of the resilience instrument, keeping the institution as the unit of analysis was an appropriate choice given the purpose and research questions that pertained to organizational resilience and not individual resilience [42, 43].

3.3 Instrument

Mallak’s [41] resilience instrument was sent in a survey web link via e-mail to leaders of independent, for-profit higher education institutions. This instrument contained 32 questions using a 6-point Likert-type scale that ranged from 1 (strongly disagree) to 6 (strongly agree).

The resilience instrument [41] required only slight modification: Items 13 and 20 containing the word hospital were changed to organization. The modified version, used with permission, was reviewed by Dr. Larry Mallak in order to ensure content validity. Construct validity had already been determined through the confirmatory factor analysis previously conducted by Mallak [14]. Because purposive sampling is a nonprobability sampling technique, external validity was not expected in this research design.

Because of the small modifications made to this survey, and because this instrument had never been tested in a for-profit, higher education institution, the test-retest reliability process used by Somers [12] was replicated. Like Somers [12] had done in his study with 11 managers of public works departments, this study used 11 managers from for-profit higher education who were not included in the target sample. They completed the survey and retook it 10 days later. A Pearson’s product-moment correlation between this first and second administration was calculated to determine the reliability of the slightly modified resilience instrument. The correlation between test and retest was r = 0.93. Using Litwin’s [44] criterion that correlations above 0.70 represent acceptable reliability, the instrument was determined to be reliable.

3.4 Data collection

E-mail addresses of leaders of the 475 target institutions in the final purposive sample were obtained from public sources that included institution websites, accreditation agencies, and industry association databases. In many cases, there were multiple e-mail addresses for a single institution because any organizational leaders, including chief executives, vice-presidents, managers, and directors, were eligible to complete the survey. Recruitment e-mails were sent to these leaders of the 475 target institutions explaining the purpose of the research, including a web link to the survey via Survey Monkey. Follow up e-mails were then sent in one-week intervals for four consecutive weeks to achieve a greater response rate. This strategy allowed for an increase from 32 respondents after the first distribution to 65 respondents after the fourth follow-up e-mail.

Each questionnaire required the participant to list the name and address of the institution. This demographic information was then matched to the unit ID from the IPEDS database. Because this information was necessary to obtain enrollment data, used as the outcome variable, any returned survey not including the name and address of the institution was removed prior to data analysis.

3.5 Data analysis

Data analysis was conducted based upon the remaining 59 responses. There were two sources of data collected for this study: resilience data and enrollment data. The resilience data constituted the predictor variable, and the enrollment data represented the outcome variable:

Resilience data. The 32-item resilience instrument [41] served as the source of the resilience data. The total organizational resilience score was the total score of all 32 items. Other than two surveys without institution names, no surveys were returned with missing data. The maximum possible score was 192, which indicates a response of strongly agree for 25 items, and strongly disagree for seven items. The scores for each of the six factors that shape organizational resilience was the mean of the items that mapped to each of these factors. A mean was necessary because the number of items mapping to individual factors was not equal.

Enrollment data. The enrollment data were retrieved for each distinct IPEDS unit ID survey returned. The data were obtained from the IPEDS for 2010 and 2012. The IPEDS 12-month enrollment measure was used instead of the fall enrollment data in order to capture numbers that included students in less conventional and short programs that start throughout the year [5]. The terms “12-month enrollment” and “total enrollment” are used interchangeably throughout this article. The outcome variable for both research questions of this study is three-year enrollment fluctuations. Fluctuations were calculated using the following equation: Enrollment fluctuation ratio = 2012 total enrollment / 2010 total enrollment.

It is important to note the background behind the years chosen for this enrollment fluctuation. In late 2010, for-profit higher education started experiencing a crisis as it was faced with tough regulatory sanctions and negative press [5]. A comparison of enrollment levels in 2012 suggested that some institutions had recovered from this crisis, whereas others had not [3]. Therefore, an enrollment fluctuation ratio using the benchmark years of 2012 versus 2010 was a reasonable measurement of enrollment fluctuation. A ratio less than 1.0 reflected a decline in enrollment two years after the crisis. A ratio of more than 1.0 reflected an increase in enrollment two years after the crisis.

3.6 Correlational design

A correlational procedure was used to determine the relationship outlined in the first research question: To what degree does a relationship exist between organizational resilience and three-year enrollment fluctuations of independent, for-profit higher education institutions? The null hypothesis indicated that a relationship would not exist between organizational resilience and three-year enrollment fluctuations of independent, for-profit higher education institutions. The alternative hypothesis indicated that a relationship would exist between organizational resilience and three-year enrollment fluctuations of independent, for-profit higher education institutions.

3.7 Multiple-regression analysis

The second research question was approached using a multiple-regression procedure: Which predictor variables that shape organizational resilience (i.e., goal-directed solution seeking, avoidance-skepticism, critical understanding or sensemaking, role dependence, source resilience, and access to resources) have a greater impact on the outcome variable (i.e., three-year enrollment fluctuations of independent, for-profit higher education institutions)? The null hypothesis indicated that there would be no differences on the impact of each of the predictor variables that shape organizational resilience on three-year enrollment fluctuations. The alternative hypothesis indicated that there would be differences in the impact of each of the predictor variables that shape organizational resilience on three-year enrollment fluctuations.

4 Results

4.1 Eligibility to participate and response rate

After three follow-up e-mail requests, 65 distinct responses were returned from the original 475 institutions invited to participate, for a response rate of 13.7%. The term distinct responses indicated the responses from individual colleges. In total, 68 responses were returned, but some were from the same college. Although this response rate seemed low, it was expected; the percentage of return fell within the range of current response rates for online surveys of 10% to 20% [45]. Additionally, the 65 responses exceeded the 55 minimal number of responses to obtain a statistical power of 0.80 at an alpha level of 0.05 and an effect size of 0.15, according to the a priori G*Power analysis [46]. Six of the 65 responses were eliminated prior to data analysis because they did not meet the criteria for inclusion.

4.2 Participant demographics

Organizational leaders from 59 for-profit, higher education institutions completed the scale on behalf of their institution and met the criteria to be included in the study. Organizational leaders, as opposed to their subordinates, were selected to complete the resilience instrument for two reasons. First, by limiting survey distribution to C-level executives, presidents, vice-presidents, and directors, the researcher made an attempt to confirm that the person completing the survey had an organization-wide perspective. Second, the concept of managerial control over resilience factors came into play. Because the six organizational resilience factors measured in this survey have been determined as within the control of an organizational leader [12], it is appropriate that survey completion be limited to leaders of the colleges.

The 59 participating institutions represented career colleges from 22 different states. Table 1 shows that respondents were not limited to any particular region and included the Pacific Northwest, New England, southern states, the Midwest, and Hawaii. The 2012 enrollment population of these institutions varied widely from very small colleges with less than 100 students to large colleges with more than 10,000 students. Less than half of the participating institutions (24) were one location or campus, with the remaining 35 participants being part of organizations with between two and five campus locations. All 59 participants represented for-profit institutions.

4.3 Descriptive results of the organizational resilience data

Table 2 shows the distribution of total organizational resilience for the 59 colleges. Organizational resilience scores have been reported in Table 2 as both total scores (i.e., total of all 32 items) and mean scores (i.e., mean of all 32 items). Although two colleges scored lower than 100 of a maximum 192, the majority had moderate levels of organizational resilience; almost 85% of colleges had total organizational resilience scores between 121 and 150. The frequency percentage column reflects a central tendency to the data.

4.4 Descriptive results of the enrollment data

Based upon the 12-month enrollment figures of the 59 survey respondents for 2010 and 2012, the enrollment fluctuation ratios were calculated (i.e., 2012 enrollment / 2010 enrollment). Table 3 shows a summary and frequency distribution of these ratios. The mean enrollment fluctuation ratio for this group of 59 colleges was 1.15. The median enrollment fluctuation ratio was 1.02, indicating that about half of the participating colleges had enrollment declines and the other half had enrollment increases. The wide variance in ratios highlighted in Table 3 indicated that even the independent for-profit institutions varied in how much their enrollments fluctuated; the standard deviation of 0.42 around a mean of 1.15 is a reflection of this wide range in enrollment fluctuation (see Table 3).

4.5 Results of research questions

The first research question posed was: To what degree does a relationship exist between organizational resilience and three-year enrollment fluctuations of independent, for-profit higher education institutions? A Pearson’s product-moment correlation coefficient (r) was calculated to assess the strength and direction of the relationship between these two variables. The predictor variable of organizational resilience was correlated with the enrollment fluctuation ratio (i.e., the outcome variable). The correlation was fairly positive, r = 0.40. It was also significant at p < 0.05, resulting in a rejection of the null hypothesis. The coefficient of determination (r²) was calculated at 0.16, indicating that 16% of the variability in enrollment fluctuations could be explained by variability in organizational resilience. The results indicated a relationship between organizational resilience and three-year enrollment fluctuations of for-profit, higher education institutions in this study.

The Pearson product-moment correlation would have only proven to be the appropriate statistical test if a linear relationship existed between the two variables in this study. If the relationship had been found to been to be curvilinear, the Pearson correlation would have underestimated the strength of the relationship present in the data [40]. In order to ensure that the relationship between organizational resilience and the enrollment fluctuation ratio was linear, the scatterplot for this correlation was reviewed. This examination provided assurance of linearity, and there was no evidence of a curvilinearrelationship.

The second research question was: Which predictor variables that shape organizational resilience (i.e., goal-directed solution seeking, avoidance-skepticism, critical understanding or sensemaking, role dependence, source resilience, and access to resources) have a greater impact on the outcome variable (i.e., three-year enrollment fluctuations of independent, for-profit higher education institutions)? A multiple regression was used to determine how the predictor variables collectively related to the outcome variable, and which had the most impact. A traditional significance level of p < 0.05 was used in this analysis to indicate statistical significance, with a level of p < 0.10 indicating marginal significance [40].

There were two assumptions being made about the nature of the data being analyzed in the multiple-regression analysis [40]. The first assumption was no intercorrelation among the predictor variables. Multicollinearity has been known to exist if two or more of the predictor variables are too highly correlated with each other, causing inferences about these predictor variables to be untrustworthy [40]. The intention is that multicollinearity will not exist. Using a cutoff of 0.80, which has been defined as an accepted maximum [47], the correlations of all six predictor variables were analyzed. The highest correlation between all predictor variables was 0.75, which did not meet the accepted maximum cutoff for concern of 0.80.

A second assumption was model specification, which is the assumption that no variables have been erroneously included or excluded and that the relationship among the variables represents a linear relationship [40]. Scatterplots of overall organizational resilience and enrollment fluctuations, as well as scatterplots representing bivariate correlations of each of the six resilience factors with enrollment fluctuations, were examined. There was no evidence of possible curvilinear relationships.

Table 4 displays a summary of results of the multiple-regression analysis of the six predictor variables (i.e., organizational resilience factors) acting upon the outcome variable (i.e., enrollment fluctuation ratios). The adjusted R² indicated that 25% of the variance in the enrollment fluctuations could be explained by the six organizational resilience factors and that this value was statistically significant, F = 4.151, p < 0.01. However, three-quarters of the variance was left unexplained by this model. As posited, the focus of the study involved the variables under the control of the college leaders. Therefore, variables that were expected to explain large amounts of variance in enrollment fluctuations, but uncontrollable by the leader, were left out of this analysis, such as program offerings, marketing strategy, and local regulatory influences.

The intent of this analysis was that leader-influenced variables shaping organizational resilience had a positive impact on organizational resilience. Table 4 also displays the Beta (β) values. One of these items, avoidance-skepticism, was significant at the p < 0.05 level. This item was the principal determinant of changes in enrollment fluctuations. Another item, critical understanding, resulted in p < 0.10, and was considered marginally significant. The remaining four predictors–goal-directedsolution seeking, role dependence, source resilience, and resource access–did not test to a level of statistical significance.

To assess the variability in enrollment fluctuations determined by the significant predictor variable of avoidance-skepticism and the marginally significant predictor variable of critical understanding, the squares of semipartial correlations were calculated and appear in Table 4. The only variable with significant predictive value was avoidance-skepticism which, once the variance explained by the remaining variables were controlled, was responsible for almost 8% of the variance in enrollment fluctuations. Critical understanding or sensemaking was marginally significant and was responsible for almost 6% of the variance in enrollment fluctuations.

Results indicated differences in the impact of each of the factors that shape organizational resilience on three-year enrollment fluctuations of independent, for-profit higher education institutions. Specifically, avoidance-skepticism and critical understanding or sensemaking had more impact on enrollment fluctuations than the other four factors that shape organizational resilience, since they were significant at p < 0.05 and marginally significant at p < 0.10 respectively. As a result, the null hypothesis was rejected.

5 Discussion

5.1 Organizational resilience levels

There was a wide range in the total organizational resilience levels of the 59 respondents. With a maximum possible score of 192, including the reverse score of some items, the lowest score was 81 and the highest score was 162 (M = 134.4, SD = 15.28). More than three-quarters of the institutions had organizational resilience levels higher than 121, which indicates at least a moderate level of resilience in most participating institutions.

5.2 Retrieving enrollment data

Enrollment data were retrieved from the IPEDS. This system is a database maintained by the National Center for Education Statistics, which is an arm of the U.S. Department of Education. Even though the researcher is aware that the enrollment information was self-reported by the institutions, the IPEDS has built-in checks and balances to ensure the accuracy of the data, including oversight by a technical review panel that regularly audits the submissions [48].

These unit IDs for each participating institution were used to retrieve the 12-month enrollment figure for 2010 and 2012, which became the numerator and denominator for the enrollment fluctuation ratio. The researcher had several options from which to choose in the retrieval of enrollment numbers, including fall undergraduate enrollment and full-time equivalent enrollment. The 12-month enrollment figure was chosen as the metric for this calculation because enrollment over a 12-month period more accurately reflects the enrollment patterns of career colleges [5].

5.3 Enrollment fluctuations

Fain [3] documented the wide variances in enrollment fluctuations between the large-chain for-profit colleges, implying that some were able to bounce back sooner than others. Deming et al. [5] suggested that, because the independent for-profit colleges (i.e., those with five campuses or less) did not experience the explosive growth of the chain schools, they likely would have less enrollment fluctuation. Even though there is a wide variance in enrollment fluctuations among the 59 participants of this study, a more accurate picture is painted in Table 3: Over 80% of institutions had ratios less than 1.5, with only 11 institutions with ratios of more than 1.5, and only three that doubled in size from 2010 to 2012, with a ratio of 2.0 or more. In light of the heavy frequency of colleges with ratios between 0.51 and 1.5 (i.e., 47 of the 59 participating colleges), the findings of this study suggest Deming et al. [5] may have been accurate in their conjecture.

5.4 Validity and reliability

The resilience instrument [41] was tested for construct validity in the assessment of internal consistency of items. Construct validity is made more certain when there is evidence of a relationship among test items [49]. This relationship can be tested with Cronbach’s alpha, especially when testing an instrument that uses a Likert-type scale [40]. Using all 59 responses, an overall Cronbach’s alpha was calculated for the 32 items on the instrument, yielding an alpha coefficient of 0.89. This value indicates a high level of internal consistency among the items because it is higher than the 0.70 level that Huck [40] indicated as a generally acceptable alpha coefficient. It matches the Cronbach’s alpha scores ranging from 0.70 to 0.89 that Mallak [14] calculated during instrument development. It is also slightly higher than the Cronbach’s alpha of 0.72 that Somers [12] calculated in his modification of Mallak’s instrument. Based upon these findings, the resilience instrument [41] was found to produce scores that represented a valid measure for organizational resilience in this study.

5.5 First research question

Although there was a moderate link determined between organizational resilience and enrollment fluctuations in this study, it is important to note the coefficient of determination (r²) was 0.16. This indicated that just 16% of the variability of enrollment fluctuations could be explained by variability in organizational resilience. Priest [11] pointed to many factors that could impact higher education enrollment levels, including the local economy, program offerings, marketing strategies, and the institutions’ recruitment practices. Based on these many possible influences on enrollment levels, the low coefficient of determination was expected.

5.6 Second research question

Before analyzing the individual predictor variables, the effectiveness of the full regression model was evaluated. This was accomplished by calculating the R² and the adjusted R². The R² and the adjusted R² indicated the amount of variance in the outcome variable that was explained by the predictor variables, with the adjusted R² correcting for an overestimation error in the R² based on sample data [40]. The R² was 0.32, the adjusted R² was 0.25, and these values were statistically significant, F = 4.151, p < 0.01. This indicates that 25% of the variance in the enrollment fluctuations could be explained by the six organizational resilience factors.

Notably, three quarters of the variance was left unexplained by this model. This variance again supports Priest [11], who identified factors affecting enrollment levels that are largely outside the control of the leader. It is recommended that further research identify the impact of both controllable and uncontrollable factors on enrollment fluctuations.

As indicated in Table 4, there were two organizational resilience factors with values of significance: Avoidance-skepticism was significant at the p < 0.05 level, and critical understanding or sensemaking was marginally significant at p = 0.07. Avoidance and critical understanding had the highest Beta coefficients at 0.30 and 0.27, respectively. The sr² values indicated that 8% of the variance in enrollment fluctuations could be explained by avoidance-skepticism and 6% could be explained by critical understanding. The other organizational resilience factors of goal-directed solution seeking, role dependence, source resilience, and resource access did not test to a level of statistical significance.

Mallak [14] conducted a confirmatory factor analysis in the development of the original resilience instrument. He found that avoidance accounted for 7.1% of the variance in the instrument, and critical understanding accounted for 6.6%. Although the relative importance of these two factors from this study’s multiple-regression analysis are in line with Mallak’s findings, there is a major difference in the result for goal-directed solution seeking. Goal-directed solution seeking accounted for the most variance (28.5%) in Mallak’s factor analysis. Goal-directed solution seeking did have the third highest Beta coefficient in this study’s multiple-regression analysis; however, unlike Mallak’s finding, it did not test to a level of statistical significance, p = 0.28.

Given that the internal-consistency measures from this study were similar to Mallak’s [14] (i.e., both had Cronbach’s alpha values of 0.89), and that the same instrument was used with only slight modifications, the lack of statistical significance of goal-directed solution seeking in this researcher’s analysis was surprising. This finding is also not supported by other relevant research. Goal-directed solution seeking was found to be related to success of small business owners [33]; many of the participants of this study were themselves owners of their colleges. In another study, 82% of managers rated the ability to adapt creatively in problem solving as an important indicator of performance [50]. It is recommended, therefore, that future studies supplement quantitative research using the resilience instrument [41] with qualitative research focusing on discussions with participants about problem solving, adaptation, and innovation.

There is ample support in the literature, however, that reflects the importance of avoidance-skepticism and critical understanding or sensemaking, particularly with links to performance. For example, the Kagaari [51] study focused on avoidance-skepticism in higher education. The highest performing employees of the universities in this study were compared; it was determined that the high performers regularly questioned the status quo, which is a particular characteristic of skepticism. Another study found that firms that demonstrated a mistrust for oversimplification of a problem and showed less deference to expertise had stronger performance, even while operating within a hostile economic environment [52]. These two studies linked the concept of avoidance-skepticism to performance metrics, much like the current study has demonstrated the impact of avoidance-skepticism on the performance metric of college enrollment levels.

Correspondingly, the multiple-regression analysis in this study also showed that critical understanding or sensemaking was of relative importance in explaining the variance in enrollment fluctuations. The findings from this study are in support of other studies that have demonstrated the link between sensemaking and performance metrics. Sull [37] paired successful firms with unsuccessful firms from the same industries to determine what impacted the differences in performance. He determined that the successful firms strove for a critical understanding, particularly of their past failures. Alam and Nandan [10] compared small accounting firms that were able to bounce back after the 2008 recession with those that did not survive. They found that, during the recession, the surviving firms demonstrated more critical understanding than their unsuccessful counterparts.

5.7 Limitations

Because the population being studied included independent, for-profit higher education institutions, the characteristics of organizations that fit into this category were already identified in previous research [5]. As a result, purposive sampling was used. As with all nonprobability sampling, generalizability was in question. All 59 participants in this research shared the same characteristics, as they all met the inclusion requirements for the purposive sampling: They all were for-profit, higher education institutions located in the United States with five or fewer campuses. Although generalizability may be possible to all other colleges that meet those criteria, it is unlikely that the findings of this study could be applied to any other higher education institutions that did not.

One of the inclusion criteria was that the organization had to have been able to be contacted by e-mail because the survey links were sent in that manner. This criterion required at least one valide-mail address for each organization, and the e-mail address had to be one that went directly to a college leader. This requirement, in itself, limited the sample. Restricting the survey distribution to email also likely contributed to the small response rate.

As noted, only 16% of the variability in enrollment fluctuations could be explained by the variability in organizational resilience. Priest [11] noted several factors that may impact enrollment fluctuations, including the local economy, the relevance of program offerings, and the marketing budget. Other factors, such as the size and age of the institution, may also impact enrollment fluctuations. A limitation of this study is that these factors were not statistically controlled because the data were never retrieved from the participating institutions.

One-time data collection may be seen as a limitation of this study. The intent of this research was to illuminate organizational resilience in career colleges through a pre-turbulence (2010) and post-turbulence (2012) comparison. Analogous research in other industries comparing pre-crisis and post-crisis performance collected data only once [7 –10]. A long-term perspective of organizational resilience in for-profit higher education is therefore suggested for future research.

6 Conclusion

Regarding the links of organizational resilience to enrollment fluctuations, there were two major findings of this study. First, there was a fair to moderate significant relationship established between organizational resilience and enrollment fluctuations, r = 0.40, p < 0.05. A coefficient of determination at 0.16 indicated that 16% of the variability in enrollment fluctuations could be explained by organizational resilience, an expected result given the many possible causes of enrollment fluctuations [11] and the limitations of this study.

Second, two of the six factors shaping organizational resilience tested at a level of significance or marginal significance: avoidance-skepticism and critical understanding or sensemaking. Avoidance (p < 0.05) was found to be responsible for 8% of the variance in enrollment fluctuations. Critical understanding tested at a level of marginal significance, p = 0.07, and was found to be responsible for 6% of the variance in enrollment fluctuations. Although the links between these two constructs and performance have been made in the literature, the fact that goal-directed solution seeking did not test at a level of significance contradicts the findings of Mallak [14] and is worthy of further study.

The findings from this study indicated that avoidance and critical understanding tested significantly in terms of their impact on enrollment fluctuations. There is evidence in the literature to suggest that these two organizational resilience factors may be inextricably connected. Çakar and Ertürk [43] demonstrated a negative correlation between these two constructs and Raney [19] found that training in critical understanding also changed the level of avoidance. Lengnick-Hall and Beck [39] included both concepts in their definition of constructive sensemaking, which was essentially the healthy balance of confidence and skepticism used in organizational decision-making processes. Further research would be warranted to determine the true relationship between these two factors, especially as they relate to organizational resilience.

In addition to enrollment fluctuations, there are several other performance metrics that are of relevance to for-profit higher education; however, they were not within the scope of this study. With continuing changes in federal regulations, metrics such as graduation rates, graduate employment rates, and student loan default rates will play increasingly larger roles in analyses of performance of higher education [53]. College leaders and their regulators would benefit from studies investigating the relationship between organizational resilience and these other performance metrics.

College leaders can arm themselves with the knowledge that there is a relationship between organizational resilience and enrollment fluctuations. This knowledge is not actionable, however, unless the leader knows exactly how to increase the institution’s organizational resilience. Two factors of organizational resilience that were found to have an impact on enrollment fluctuations were avoidance-skepticism and critical understanding or sensemaking. Unlike the broad concept of organizational resilience, these factors can be directly controlled by a leader [12].

Altering the level of avoidance means first being aware of the balance required between high avoidance, implying too much timidity, and low avoidance, which may occur as a result of groupthink. College leaders should be aware of the avoidance levels of each member of their team and also try to gauge how much avoidance is occurring with group decision making. A healthy skepticism toward new and chaotic situations is the more appropriate tactic in higher education [51]. The other factor found to have an impact in enrollment fluctuations, critical understanding or sensemaking, is an element of organizational resilience that can be taught [19 , 37].

Leaders are encouraged to consider ways of improving the levels of resilience in their organizations. The leader may find that improving avoidance and critical understanding within the organization may bolster organizational resilience. This resilience could serve them well if crises – like those experienced in for-profit higher education in 2010 – reoccur.

Conflict of interest

The authors have no conflict of interest to report.

References

Walker

, Holling

, Carpenter

, Kinzig

. Resilience, adaptability and transformability in social-ecological systems. Ecology and Society2004;(2):5–13.

Beaver

. Fraud in for-profit higher education. Society2012;49(3):274–8.

Fain

. Phoenix reloads. Inside Higher Ed. Friday Oct 26 2012. www.insidehighered.com/news/2012/10/26/university-phoenix-down-not-out .

Fain

. More selective for-profits. Inside Higher Ed. Friday Nov 11 2011. www.insidehighered.com/news/2011/11/11/enrollments-tumble-profit-colleges .

Deming

, Goldin

, Katz

. The for-profit postsecondary school sector: Nimble critters or agile predators?Journal of Economic Perspectives2012;26:39–164.

Gittell

, Cameron

, Lim

, Rivas

. Relationships, layoffs and organizational resilience: Airline industry responses to September 11. Journal of Applied Behavioral Science2006;42:300–29.

Chewning

, Lai

, Doerfel

. Organizational resilience and using information and communication technologies to rebuild communication structures. Management Communication Quarterly2013;27(2):237–40.

Fungacova

, Jakubik

. Bank stress tests as an informative device for emerging markets: The case of Russia. Czech Journal of Economics and Finance2013;63:87–105.

Zagelmeyer

, Heckmann

. Flexibility and crisis resistance: Quantitative evidence for German establishments. International Journal of Manpower2013;34:271–87.

10.

Alam

, Nandam

. Organisational change and innovation in small accounting practices: Evidence from the field. Journal of Accounting and Organisational Change2010;6:460–76.

11.

Priest

. The challenge of managing a large university in conditions of uncertainty. London Review of Education2012;10(3):275–88.

12.

Somers

. Measuring resilience potential: An adaptive strategy for organizational crisis planning. Journal of Contingencies and Crisis Management2009;17:12–23.

13.

Horne

, Orr

. Assessing behaviors that create resilient organizations. Employment Relations Today1998;24(4):29–39.

14.

Mallak

. Measuring resilience in health care provider organizations. Health Manpower Management1998;24(4):148–52.

15.

Thomas

, Clark

, Gioia

. Strategic sensemaking and organizationalperformance: Linkages among scanning, interpretation, action and outcomes. Academy of Management Journal1993;36:239–70.

16.

Edson

. A complex adaptive systems view of resilience in a project team. Systems Research and Behavioral Science2012;29(3):499–516.

17.

Bharwani

, Mathews

. Risk identification and analysis in the hospitality industry: Practitioners’ perspectives from India. Worldwide Hospitality and Tourism Themes2012;5410–27.

18.

Kuhlicke

. Resilience: A capacity and a myth: Findings from an in-depth case study in disaster management research. Natural Hazards2013;67(1):61–76.

19.

Raney

. Agility in adversity: Integrating mindfulness and principles of adaptive leadership in the administration of a community mental health center. Clinical Social Work Journal2014;42:312–20.

20.

Wicker

, Filo

, Cuskelly

. Organizational resilience of community sport clubs impacted by natural disasters. Journal of Sport Management2013;27:510–11.

21.

Powley

. The process and mechanisms of organizational healing. Journal of Applied Behavioral Science2013;9:42–68.

22.

Koronis

, Ponis

. Introducing corporate reputation continuity to support organizational resilience against crisis. Journal of Applied Research2012;28:283–90.

23.

Olu-Daniels

, Nwibere

. Trust and organizational resilience in the Nigerian oil and gas industry. International Journal of Business and Management2014;9:291–312.

24.

Gavdart

, Garrigou

, Chassaig

. Analysis of organizational conditions for risk management: The case study of a petrochemical site. Work2012;41:2661–7.

25.

Lane

, McCormack

. Resilient leaders: Essential for organizational innovation. International Journal of Organizational Innovation2013;6:7–25.

26.

Matzenberger

. A novel approach to exploring the concept of resilience and principal drivers in a learning environment. Multicultural Education and Technology Journal2013;7:192–206.

27.

Weick

, Sutcliffe

. Managing the unexpected:Assuring high performance in an age of complexity. San Francisco, CA: Jossey-Bass; 2001.

28.

Buys

. Developing corporate strategies to enable resilience in the South African information systems and technology industry. Journal of Applied Business Research2012;28:913–20.

29.

Grobber

, Grubner

, Millig

. Organizational adaptation processes to external complexity. International Journal of Operations and Production Management2006;26:254–81.

30.

Hind

, Front

, Rowley

. The resilience audit and the psychological contract. Journal of Managerial Psychology1996;11:18–9.

31.

Mafabi

, Munene

, Ntayi

. Knowledge management and organizational resilience. Journal of Strategy and Management2012;5:57–80.

32.

Smallbone

, Denkins

, Battisti

, Kitching

. Small business responses to a major economic downturn. Empirical perspectives from New Zealand and the United Kingdom. International Small Business Journal2012;30:754–77.

33.

Owens

, Kirwan

, Lounsbury

, Levy

, Gibson

. Personality correlated of self-employed small business owners’ success. Work2013;45:73–85.

34.

Weick

. The collapse of sensemaking in organizations: The Mann Gulch disaster. Administrative Sciences Quarterly1993;38(4):628–52.

35.

Acar

, Winfrey

. The resilient organization: Sustaining organizational renewal and performance. Journal of Strategic Change1994;3:165–3.

36.

Thomas

, Francil

, John

, Davies

. Identifying the characteristics for achieving sustainable manufacturing companies. Journal of Manufacturing Technology and Management2011;23:426–40.

37.

Sull

. Closing the gap between strategy and execution. MIT Sloan Management Review2007;48(4):30–8.

38.

Limnios

EAM

, Mazzarol

, Ghadouani

, Schilizzi

SGM

. The resilience architecture framework: Four organizational archetypes. European Management Journal2014;32:104–5.

39.

Lengnick-Hall

, Beck

. Adaptive fit versus robust transformation: How organizations respond to environmental change. Journal of Management2005;13:738–57.

40.

Huck

. Reading statistics and research (6th ed.). Boston, MA: Allyn & Bacon; 2012.

41.

Mallak

. Resilience instrument. Kalamazoo: Western Michigan University; 2006.

42.

Hofstede

, Neuijen

, Ohayv

, Sanders

. Measuring organizational cultures: A qualitative and quantitative study across twenty cases. Administrative Science Quarterly1990;35(2):286–302.

43.

Cakar

, Ertürk

. Comparing innovation capability of small and medium-sized enterprises: Examining the effects of organizational culture and empowerment. Journal of Small Business Management2010;48:325–59.

44.

Litwin

. How to assess and interpret survey psychometrics (2nd ed.). Thousand Oaks, CA: Sage; 2003.

45.

Bourque

, Fielder

. How to conduct self-administered and mail surveys (2nd ed.). Thousand Oaks, CA: Sage2003.

46.

Faul

, Erdfelder

, Bucher

, Land

. Statistical power analyses using G*Power 3.1: Tests for correlation and regression analysis. Behavior Research Methods2009;41(4):1149–60.

47.

Cohen

, Cohen

, West

, Aiken

. Applied multiple regression/correlation analysis for the behavioral sciences (3rd ed.). New York, NY: Taylor and Francis; 2003 .

48.

Sykes

. Information required to be disclosed under the Higher Education Act of 1965: Suggestions for dissemination: A supplemental report. Washington, DC: U.S. Department of Education; 2012.

49.

Creswell

. Educational research: Planning, conducting, and evaluating quantitative and qualitative research (4th ed.). Boston, MA: Pearson; 2012.

50.

Koopmans

, Bernaards

, Hildebrandt

, De Vet

HCW

, Van der Beek

. Measuring individual work performance: Identifying and selecting indicators. Work2014;48:229–38.

51.

Kagaari

. Performance management practices and managed performance: The moderating influence of organizational culture and climate in public universities in Uganda. Measuring Business Excellence2011;15(4):36–49.

52.

Bauweraerts

. How do private family firms face the crisis? Empirical evidence from Belgium. International Business Research2013;6(8):91–101.

53.

Cellini

. For-profit higher education: An assessment of costs and benefits. National Tax Journal2012;65:153–79.