Clients ’ service perceptions of private higher education institutions in South Africa : An importance-performance analysis for strategic managers

Globally, higher education institutions (HEI) are required to be more business-like in their functioning. Students are increasingly being recognised as discerning customers, selecting a provider based on its ability to provide a superior value proposition. Unlike their public counterparts, private higher education institutions (PHEI) in South Africa receive no financial support from the government. To ensure students’ patronage, it is imperative for private providers to determine their expectations and how they rate the PHEIs performance. An importance performance analysis (IP) was done based on data gathered from a quantitative survey on 600 full-time students at three different PHEIs. A comparative IP matrix was also constructed to provide the PHEIs with a visual interpretation of the gathered data. The IP analysis yielded significant negative gaps between performance and importance regarding the attributes. The comparative IP matrix indicates important areas for strategic consideration to the three PHEIs respectively. This should assist management in improving their service offerings. This study should also contribute towards the body of knowledge regarding PHEIs in South Africa.


INTRODUCTION
Globally, higher education (HE) landscapes have been characterised by a move towards the "commodification" of HE.Higher education institutions (HEIs) are required to be more business-like in their functioning and with students, viewed as clients, who are displaying a more consumerist behaviour towards HE, selecting HEIs on the basis of aspects such as value for money, and the future possibilities of securing employment, as opposed to the love of the subject (Maringe, 2006, 466, 467).The marked proliferation of private higher education institutions (PHEI), capitalising on governments' inability to provide for the ever increasing demand for HE, has probably contributed to the above mentioned trend.Whether or not morally justified, most PHEIs are essentially businesses that, in their search for profit, exploit opportunities in the HE industry (Levy, 2008:7).Add to this the fact that, unlike their public counterparts, PHEIs in South Africa receive no state subsidies (De Villiers and Nieuwoudt, 2010:13), and it becomes clear that they have to subscribe to sound business principles in order to ensure their long term survival.
concept philosophy propagates the creation, communication, and the delivering of meaningful superior customer value (Kotler and Keller, 2009: 59-63).This implies that students, as potential customers, select a HEI based on its superior ability to fulfil his or her most important needs (the perceived benefits), compared to the perceived costs of the offering.Based on this premise, in order to design a superior needs satisfaction offering, PHEIs should thus identify what those needs or expectations are.Even if a student select a specific HE, this does not guarantee long term patronage to the point where the student graduates.Students' attrition will inevitably lead to loss of potential long term profit.In line with modern marketing philosophy, the delivery of superior customer value is inextricably linked to customer relations management (Strydom, 2011:287).PHEIs thus need to identify the expectations of students and constantly monitor if the students' expectations are met, as a negative gap between the expectations and the experiences of students at a PHEI may lead to students' attrition or negative word-of-mouth communication.
Numerous studies have investigated aspects in the HE field like students' expectations and experiences, students' satisfaction and service quality (Joseph et al., 2003;National Survey of Student Engagement, 2012;Martilla and James, 1977;Parasuraman et al., 1986;Yooyen et al., 2011).Research on the PHEI landscape in South Africa is limited and the researcher was unable to find any evidence of research regarding the comparison between expectations and experiences of PHEI-students in South Africa.Considering this problem, a research question was formulated: RQ1: What are the differences between the important expectations and perceived experiences of students at different private higher education institutions in South Africa?
The purpose of this paper was to identify the important expectations that students had in deciding on a specific PHEI, and to compare it with the students' perceived experiences regarding these important expectations at the PHEIs that they have chosen and were now studying at.It was envisaged that a comparison of the attributes of their service offerings in terms of their importance and perceived performance and an IP matrix, constructed from this analysis, would provide PHEIs with an indication of important areas for strategic consideration.Data gathered from this study can thus assist PHEIs to enhance their value propositions to their primary customers, namely their students.

LITERATURE REVIEW
The higher education industry and its increasing reliance on a marketing orientation were firstly studied and related to consumer satisfaction.Different approaches to the measurement of consumer satisfaction in HE were further investigated, linking it to the value thereof in management decision making.

The "marketisation" of higher education
It seems that, in many parts of the world, there is a shift towards a market orientation among HEIs.HE is increasingly being viewed as a marketable commodity where the student as primary customer determines the offerings as opposed to HE as a public good (Yooyen et al., 2011:25).Gaziel (2012:290) refers to the shift in HE towards market forces, where governments attempt to relieve HE demand pressures, moving away from "ownership to regulation".In addition to increasing demand pressures, aspects like reduced funding (Altbach et al., 2009:69;Maslen, 2011), the significant growth of PHE (Altbach et al., 2009:69;Levy, 2010:12) and globalisation (Mabizela, 2007:23) have forced HEIs to be more business-like and entrepreneurial in their functioning, relying more on tuition fees and alternative revenue generation to survive.Mabizela (2007: 23) remarks that the role of the government is progressively being replaced by the private sector with the objectives of profitmaximisation, as opposed to social development; this is thus a subscription to "marketisation".Maringe (2006: 477) postulates that the HE environment has become increasingly competitive.He suggests that potential students are becoming discerning choosers in this HE marketplace, and that HEIs should adapt accordingly.According to Virgiyanti et al. (2011:578), globally, HEIs are realising the value of marketing theory and are applying these theories and concepts to gain competitive advantages.Tait andDe Jager (2009:1026) concur with this view by stating that the educational environment "has not only become competitive, but also more comercialised".
Worldwide, enrolment through PHEIs is currently the fastest-growing HE segment (Altbach and Levy, 2005:1).Drawing on Gürüz, Altbach et al. (2009:67-77) point out that 30% of global student enrolment is in PHE.Altbach et al. (2009:67-77), as well as Levy (2008:7), contend that the largest increase in private provision is in "demand absorbing", emanating from excess demand created by the "massification" of HE.They conclude that most forprofit PHEIs seem to operate in this non-elite sector.Gupta (2008:572) notes that this trend is also observed in South Africa, where the number of profit seeking PHEIs in the demand absorbing sector is growing significantly.The increase in non-governmental HE provision may be attributed to the inability of governmental funding to meet the growing demand for HE (Levy, 2009:3).Additionally, Levy (2010:12) refers to the trend of many PHEIs focusing on occupation-oriented fields, and thereby providing programmes that prepare students for a specific job.This is an area, according to him, from which public higher education institutions (PBHEI) have traditionally shied away.In March 2013, there were 87 registered and 31 provisionally registered PHEIs in South Africa (South Africa, Department of Higher Education and Training, 2013:9-72).Even though these PHEIs educate thousands of students, perceptions of PHEIs' inferiority in the South African HE landscape still prevail (Setswe, 2013:101).PHEIs are viewed as profit-seeking enterprises that deliver education of "questionable quality, only in areas of great demand and that they contribute little to research".PHEIs may not include the word "university" in their titles (Bezuidenhout, 2012:151,152).In contrast to PBHEIs, PHEIs do not receive subsidies, tax rebates, or facilities from Government.It thus seems that PHEIs are not truly regarded as full partners in the HE landscape.Without the above-mentioned support from Government and the HE fraternity, private providers are forced to charge more for their programme offerings and mostly cannot afford to award bursaries.Because of these higher fees and lack of bursaries, the demand-absorption capabilities of PHEIs, relieving the pressing need for HE in South Africa, are grossly underutilised (Bezuidenhout, 2012:151,152).With all the above-mentioned additional challenges facing them, the view that HEIs of today should subscribe to a marketing orientation, should thus ring even more true for PHEIs in South Africa.

A marketing orientation for higher education institutions
Modern marketing thought dictates that the customer and the satisfaction of his/ her needs is in essence the reason for enterprises' existence and that all organisational activities should revolve around this philosophy (Kotler and Keller, 2009:59-63;Lamb et al., 2010:5-15).Organisations that adopt this orientation thus aim to create superior needs satisfaction (Venter and Jansen van Rensburg, 2009:9) believing that the creation of superior customer value (perceived benefits minus costs) stands central to the success of any organisation (Safarnia et al., 2011:135).With maximising consumer satisfaction in mind, organisations need ways to measure customer satisfaction levels.
There is a worldwide increase in customer satisfaction studies in HE.Institutions' ability to fulfil students' needs is increasingly being scrutinised and compared.The annual National Student Survey in the United Kingdom (UK), Australia and Ireland allows students, in their respective countries, to complete a survey to evaluate the institution where they study.The results of the surveys,

Measuring customers' satisfaction in HE
There is no generally accepted approach to the measurement of customer satisfaction (Al-Alak and Alnaser, 2012:157).Numerous researches in the HE field have followed different measurement approaches that are prevalent in general customer satisfaction studies.Yoohen et al. (2011, 26) mention that service encounter evaluations in HE lead to varying degrees of "satisfaction" or "dissatisfaction" outcomes.The disconfirmation of expectations paradigm (DP) proposes that meeting or exceeding a customer's expectations will result in satisfaction, but perceived organisational performance that falls short of the customer's expectations will lead to the disconfirmation of those expectations and, ultimately, dissatisfaction (Oliver, 1980: 460-469;Lamb et al., 2010:5).Patterson and Johnson (1993:91) refer to the extensive use of this "post-purchase, evaluative judgement concerning a specific transaction" model in HE.
They also cite numerous studies that have indicated that the disconfirmation construct is a robust indicator of satisfaction/ dissatisfaction.Using the DP as conceptual framework, Parasuraman et al. (1985) have developed a similar conceptual model that examines dimensions of service quality, namely SERVQUAL (SQ), as opposed to the DP approach, which was, according to them, more suited to product related measurement (Boshoff and Du Plessis, 2009:37).The SQ approach advocate that perceived service quality is determined by a customer's comparison of his/her expectations and the organisation's perceived performance, resulting in performanceexpectation gaps (Patterson and Johnson, 1993:92).
Whereas the DP paradigm pertains to specific consumption experiences, SQ refers more to an overall attitude towards service quality.Unlike the SQ approach, the DP identifies disconfirmation as intervening factor that has a distinct effect on a customer's satisfaction (Patterson and Johnson, 1993:93).Citing Mukherjee and Nath, Yoohen et al. (2011, 28) state that the SQ model is the most widely followed measurement approach in HE.Joseph et al. (2003:22) suggest that the importanceperformance (IP) paradigm provides an alternative to SQ, contending that, similarly to DP, IP allows for a detection of specific variance in different service offerings, as opposed to a more general attitudinal approach of SQ.Where DP and SQ is concerned with comparing expectations with perceived performance, IP, a methodology based on the original IP theory of Martilla and James (1977), investigates the importance that customers attach to different variables and then compare it with the customers' perceived experience of each specific variable, resulting in a P-I score (performance minus importance).These P-I scores can then also be plotted on an IP grid or matrix that indicates important areas for HEIs to consider in improving their service offerings (Boshoff anddu Plessis, 2009: 47, Kitcharoen, 2004:21;Olimpia, 2012:475).The IP grid is a marketing instrument suited to decision making concerning customer satisfaction improvement (Olimpia, 2012:474).
All three above approaches have unlocked theoretical, conceptual and operational critique (Arokiasamy, 2012: 57;Boshoff and du Plessis, 2009:48;Patterson and Johnson, 1993:92;Olimpia, 2012:476).Despite its alleged shortfalls, these measurement approaches do, at the very least, provide an indication to HEIs of their primary clients' expectations and how they evaluate their service offerings.According to Khodayari and Khodayari (2011: 42), most studies on HE identify service quality as antecedent to customer satisfaction.Most customer satisfaction/ service quality studies in HE also identify significant gaps between expectations and perceived performance.(Joseph et al., 2003:14;Khodayari and Khodayari, 2011: 43;Yooyen et al., 2011:33;Yorke and Vaughan, 2012: 18;).Service quality, specifically in service organisations like HEIs, refers to a customer's evaluative judgement of his/her expectations and perceived experiences, resulting in gaps between these constructs (Boshoff and Du Plessis, 2009:37).Based on this viewpoint, a lack of service quality or elements thereof, as identified at most HEIs, should have a significant influence on student satisfaction and thus on their resulting post enrolment communication and on their decision regarding the continuation of their studies.Currently, there is no study that has investigated the expectations and perceived performance evaluations of PHEI-students in South Africa.The focus of this paper is not to define the dimensions of service quality or critically evaluate the different service quality measurement approaches, nor is it to provide a pin-point accurate measurement of customer satisfaction or dissatisfaction.The focus of this paper is rather to alert PHEIs in South Africa as to the importance that students assign to different attributes in their choice of a PHEI compared with their perception of the PHEIs performance regarding each attribute.

Key research objectives
1. To conduct an analysis of the importance-performance evaluations of various service offering attributes among respondents at different private higher education institutions in South Africa 2. To construct an importance-performance matrix of various service-offering attributes among respondents at different private higher education institutions in South Africa.

Secondary objectives
To determine whether there are significant differences in the importance that students from different ethnic backgrounds have assigned to the top three overall most important service offering attributes identified in this study.The above secondary objective alludes to the fact that the sample of this study exhibited quite a diverse racial make-up.

Hypotheses
To address the above secondary objective, the following hypotheses were developed to provide a specific testable expectation of empirical reality: 1. H 1(0) : There are no significant differences in levels of importance that students from different ethnic backgrounds have assigned to security and safety conditions on campus.2. H 1(a) : There are significant differences in levels of importance that students from different ethnic backgrounds have assigned to security and safety conditions on campus.3. H 2(0) : There are no significant differences in levels of importance that students from different ethnic backgrounds have assigned to employment prospects.4. H 2(a) : There are significant differences in levels of importance that students from different ethnic backgrounds have assigned to employment prospects.5. H 3(0) : There are no significant differences in levels of importance that students from different ethnic backgrounds have assigned to well-equipped computer facilities.6. H 3(a) : There are significant differences in levels of importance that students from different ethnic backgrounds have assigned to well-equipped computer facilities..

RESEARCH METHODOLOGY
The social research conducted in this cross-sectional study fell within a quantitative paradigm and was descriptive in nature.In addition to a literature review, the study included an empirical investigation by means of a self-report survey design.The quantitative self-report survey design enabled the researcher to include a large number of respondents in the study.The importance-performance (IP) measurement approach as was followed in the survey (Joseph et al., 2003;Kitcharoen, 2004;Martilla and James, 1977;Olimpia, 2012).Martilla and James (1977:79) propose that this "low-cost, easily understood" technique provide organisations with an indication of where to focusing terms of their resources and marketing mix.The above authors state that value of the techniques lies in determining the relative importance and performance of attributes to prioritise marketing actions.This approach has enabled the researcher to identify the underlying relative importance of various service offering attributes to respondents in their choice of PHEI and their evaluation of the PHEIs offering in terms of these attributes.The gaps between the dimensions of importance and corresponding performance were identified and incorporated into an IP-grid or matrix.An IP-grid can provide a way to graphically present data-interpretation for organisations (Martilla and James, 1977:79).Citing Slack and Barsky, Kitcharoen (2004:21) postulates that attributes of a service offering that are deemed (by customers) to be of high importance, will play a critical role customers' overall satisfaction.By identifying these important elements or attributes and their corresponding perceived performance, PHEIs should receive an indication of the most important performance areas that need improvement.

The sample framework
The target population consisted of full-time students registered at three PHEIs.A purposive sample was employed; consequently, 600 full-time students (200 per PHEI) were targeted across all the academic departments of each PHEI.458 usable questionnaires were received back.55.3% of the respondents were female versus 44.7% male.This is in line with an overall female-to-male student ratio of 54.5 to 45.5% found in a study of 82 PHEIs that yielded a head count of 65 755 students registered at private providers in 2010 (Tladi, 2012:12).

The research instrument
The primary data collection method involved a quantitative survey with a paper-based, self-administered questionnaire as the research instrument.Interval scale questions were posed to respondents to determine the importance of 45 attributes to them in selecting a PHEI by rating it on a five-point scale from "not important at all" to "very important".These choice-related attributes were selected from various existing choice-studies in HE worldwide.A similar scale was used to request the respondents' perceived performance evaluation (from "not good at all" to "excellent") of the 45 attributes regarding the PHEI where they studied.To facilitate the content validity of the Bezuidenhout and Jager 59 data collection instrument, the researcher consulted established questionnaires and followed the guidelines of good questionnaire design and asking questions.The questionnaire was presented to a panel of experts for their scrutiny.It was also subjected to a pre-test study among students attending a PHEI.The internal reliability of the instrument was measured by using Cronbach's Alpha technique to indicate the internal consistency of the instrument.An overall correlation coefficient of 0.96 that was obtained provides a strong indication that respondents would have scored the same at different points in time (Field, 2005, 666).

The data collection and ethical considerations
A representative of each of the three institutions of this study was identified and contacted by the researcher, after which each PHEI gave formal permission for the study.The representatives distributed 200 questionnaires among the respective academic departments of each PHEI.Lecturers, who were used as field workers, handed out the questionnaires to students attending classes, which they completed there and then in the lecture rooms.
The respondents were fully informed about the nature of the study being conducted and given a choice of either participation or non-participation.The anonymity of all participants was protected.The respondents were allowed to withdraw from the study at any time.An informed consent form accompanied the questionnaire.This form was signed by all the participants.

The data analysis
The data were analysed with the help of SPSS, version 18.0.The statistical analysis of data included descriptive, as well as inferential statistics.The relative importance and performance rating of the 45 attributes in the questionnaire were determined by calculating the means of the ratings per group of respondents for each of the PHEIs respectively.The sum of the means of all three PHEIs' respondents was also calculated to identify the 10 overall most and 10 overall least important attributes.A performance minus importance (P-I) gap analysis for each PHEI was done based on the 10 overall most important and 10 overall least important attributes.A comparative IP grid between the three PHEIs of this study was constructed to indicate areas for strategic consideration.Univariate tests (ANOVA or analysis of variance) were employed to test the set null hypotheses (see section 3) regarding the effects of categorical (independent) variables on individual dependent interval variables (Field 2005, 288, 571-573, 725).A significance level (α) of 0.05 was chosen as a cut-off point for rejecting or accepting all the null-hypotheses of this study.

RESULTS
200 questionnaires were distributed to three PHEIs respectively.458 usable questionnaires were returned out of the 600 that were distributed, yielding a very high response rate per institution and a high subsequent overall response rate (76%).
Table 1 depicts the overall (combined mean scores of the three PHEIs) 10 most important attributes to students when they decided to study at a PHEI.Table 2 provides a look at the 10 overall least important attributes to PHEIstudents.From this, the following noteworthy observations: 1. Overall, the respondents identified safety and security conditions as being the most important attribute in their choice of PHEI.The second most important overall variable in this study was employment prospects.2. Interestingly, institutional facility factors such as wellequipped computer facilities and spacious, well-equipped classes ranked overall third and fourth, respectively, among the choice variables.
3. The availability of facilities, including those for recreation, sport, dining halls, hostels and private accommodation, were ranked the lowest in importance among the 45 choice variables by the respondents at all three PHEIs.4. Tradition was, by a large margin, the least important choice variable.Tables 3 to 5 provide a P-I gap analysis (mean score of perceived performances minus the mean score of perceived importance) of the 10 overall most important attributes to students for each of the three PHEIs respectively and Tables 6 to 8 a gap analysis of the 10 overall least important attributes to students per PHEI.

Salient observations include:
1.There are significant negative gaps for all 10 most important attributes for all three PHEIs.Negative gaps indicate that students' expectations with regards to a specific service offering attribute have not been met and vice versa.Higher negative gaps indicate higher levels of dissatisfaction (as discussed earlier in the study).2. The gap analysis for PHEI two displays much lower negative gap scores.This may allude to this private provider's superior ability, in relation to the other two PHEIs, in addressing the important expectations of students.The negative gaps though indicate that PHEI two still needs to improve regarding all 10 attributes.3. Reasonable class fees consistently display the biggest negative gap. 4. The P-I gap analyses of the 10 least important attributes yield quite a few positive scores.
Figures 1 and 2 display I-P matrices for the three PHEIs of this study.The importance and perceived performance of attributes (based on mean scores) of  each of the PHEIs are plotted on a matrix.Figure 1 depicts a matrix of the 10 overall most important attributes (attributes 1-10) to students and Figure 2, the 10 overall least important attributes (attributes 36-45).The matrix depicts four quadrants (Martilla and James, 1977: 78): 1. Quadrant A: Concentrate here.This quadrant represents attributes with high importance to students, but where they report low satisfaction with the PHEIs       factor theory of Herzberg, as applied to student satisfaction (DeShields et al., 2005:131,132), PHEIs should take caution not to neglect these potential "dissatisfiers".4. Quadrant D: Possible overkill.In this quadrant, students do not rate the attributes as being very important, yet they are pleased with the PHEIs performance.Overall, there are only three attributes among the PHEIs in this area.
Table 9 depicts a description of the top three overall (combined mean scores of the three PHEIssee Table 1) most important attributes to students when they decided to study at a PHEI, as well as the mean scores of the importance that were assigned to each of these three attributes by three different ethnical groups respectively.From this, the following noteworthy observations were seen: 1. Regarding security/ safety on campus: Black and coloured students have assigned a relatively higher (based on mean score) importance to this attribute.Table 10 indicates a statistical rejection (α = 0.00) of the null hypothesis (first hypothesis, see section 3), thus confirming that there are significant differences between different ethnic groups' performance rating of this attribute.
2. Regarding employment prospects: All ethnic groupings exhibit a high importance rating, with black students displaying the highest mean score.Table 11 indicates a statistical acceptance (α = 0.188) of the null hypothesis (second hypothesis, see section 3), thus confirming that no significant differences exist between different ethnic groups' performance rating of this attribute.
3. Regarding well equipped computer facilities: Black students exhibited a significantly higher assigned importance rating.This notion is confirmed by the statistical rejection (α = 0.00, see Table 12) of the null hypothesis (third hypotheses, see section 3), thus confirming that significant differences do exist between the different ethnic groupings' importance rating of this attribute.

DISCUSSION AND RECOMMENDATIONS
From an investigation of the relevant literature (see section one and two) certain issues are evident: 1.With the recent drastic changes in the HE landscape and dwindling subsidies, HEIs need to be more businesslike in order to survive.This means that HEIs should subscribe to the modern marketing concept, advocating the imperativeness of consumer satisfaction.The marked  proliferation of consumer satisfaction studies in HE bears testament to the notion that students are discerning consumers who will select those institutions that would potentially satisfy their needs in the best way.2. There is no one generally accepted method that researchers use to measure customer satisfaction in HE (see section 2.3).In contrast to a more general attitudinal approach to customer satisfaction of the SERVQUAL (SQ) technique, this study followed an importanceperformance (IP) approach.This approach based on the original IP theory of Martilla and James (1977), enables researchers to detect specific variance in regarding different service offering variables.The perceived (by students) importance subtracted from the perceived performance of a specific service related attribute yields a score that indicates students' possible satisfaction/ dissatisfaction regarding this attribute.High negative scores indicates high dissatisfaction and vice versa.These scores are plotted on an IP matrix to specifically indicate attributes that managers need focus on to enhance customer satisfaction.Both the SQ and PI have its supporters and detractors in the literature (see section two).The researcher though supports the notion of Martilla and James (1977:79) that the "low-cost, easily understood" IP technique provides organisations with an indication of where to focus in terms of their resources and marketing mix.
3. There are currently only a few studies concerning PHEIs in South Africa, especially in the fields of student choice and satisfaction.There is no continuous research to document the historic and current development of private higher education provision in South Africa.There are numerous PHEIs in South Africa, serving thousands of South African and international students.Many of the facets of these institutions still need investigation.
From the empirical study, the following is evident regarding the 45 choice attributes: 1. HEIs should consider safety and security (see Table 1 that depicts the ten overall most important choice attributes), as well as employment prospects, as very important choice attributes.The respondents in this study rated these attributes as the most and second most important, respectively.This is in contrast with many international studies, where studies did not even include security and safety conditions in their surveys.The only other study where respondents also recognised the very high importance of campus safety and security was a local study of 1500 respondents at six PBHEIs (Wiese, 2008, 233).This possible pattern in recent studies may allude to the influence of the current crime situation in South Africa on the psyche of HE students (Bezuidenhout, 2012:163).The current high unemployment rate in South Africa might have also influenced the respondents in their high importance rating of employment prospects.The high importance of employment prospects is comparable to the Wiese-study mentioned above, as well as the comprehensive Cooperative Institutional Research Programme (CIRP) longitudinal study that has already investigated more than 15 million American students from about 1900 HEIs since 1966 (HERI, 2012).
The respondents of this study identified facility variables, including spacious, well-equipped computer facilities and spacious, well-equipped classes, as being very important.
From the IP gap analysis of the collected data, the following is evident: 1.There are significant negative gaps for all 10 most important attributes for all three PHEIs (see Tables 3, 4  and 5).This is consistent with numerous studies, referred to earlier in this study, that conclude that perceived performance is markedly lower than the importance of attributes in HE.This should have a significant negative influence on student satisfaction and thus on their resulting post enrolment communication and on their decision regarding the continuation of their studies.The ten overall most attributes especially yielded very big negative P-I gaps with PHEI 1 and 3; this should be of great concern to them (see Tables 3, 4 and 5).All three PHEIs should specifically consider strategies around their price of their offerings, as this attribute consistently displayed the biggest negative P-I gap.This may be as a result of PHEIs in South Africa not receiving any state subsidies like their public counterparts (see section two).
2. The positive P-I gap analyses of the 10 least important attributes indicates where perceived performance is thus higher than the importance of attributes.This does not necessarily mean that perceived performance is very high, but rather that the importance attached to these attributes is so low.
The IP matrix that was constructed from the collected data indicates the following: 3. The PHEIs should take cognisance of the numerous attributes in quadrant A in the comparative IP matrix (see Figures 1 and 2) that require their urgent attention.It is imperative that the PHEIs seek to improve their performance with regards to these attributes as identified in their respective analyses (see section 5).The respondents at PHEI two appears to be, in general, more content than their counterparts at the other PHEIs.This could be seen in the significant lower number of attributes in quadrant A (concentrate here) and the subsequent higher number of attributes in quadrant B (keep up the good work).
From an empirical investigation of differences between ethnic backgrounds with regards to attribute importance, the following is evident: 4. Black and coloured students have placed a significantly higher premium on safety and security than their white counterparts (see section 5 and Table 10).No similar study is available to compare this with.5. Students from all three ethnic groupings have assigned a similar high importance on employment prospects.This is in line with comprehensive international studies (CIRP, as mentioned earlier) and a local study (Wiese, as mentioned earlier in this study).6. Black students have rated the availability of wellequipped computer facilities as significantly more important than the other two ethnic groups.The researcher can only speculate that this may allude to the possibility that more black students do not have computers.

LIMITATIONS
This study has investigated the "institutional" factors that influence service quality and student satisfaction.It did not venture into the "individual" factors like student motivation; preparedness; personal background, etc. (Round, 2005: ii).It thus addressed the "what" rather than the "why" of student behaviour.The study was limited to three PHEIs in South Africa.The results can thus not be extrapolated to all PHEI-students in South Africa.IP's value lies in identifying the relative importance of attributes of service offerings and then ascertaining the organisation's performance, as perceived by its customers, especially regarding the most important attributes, resulting in a practical, easy to understand, indication of important areas for strategic consideration.This method thus provides indications, rather than pin-point accurate measurements.It is doubtful that any social research method in this field will in any case yield exact measurements.

Conclusion
The study provided a strong indication that the perceived performance of PHEIs are significantly lower than the corresponding importance that respondents attach to attributes, resulting thus, similarly to numerous studies, in negative P-I gap scores.The study identified a rank-order list of important choice attributes to respondents at three PHEIs and indicated how respondents rate these attributes in terms of the PHEIs performance.A comparative IP matrix was constructed to provide the PHEIs with a graphical means to interpret the gathered data.The IP grid indicated areas for strategic consideration to the three PHEIs, respectively.This will assist the PHEIs to customise their service offerings to the needs of their primary customers (students).This study may contribute towards the body of knowledge regarding student choice, service quality and student satisfaction, especially regarding PHEIs in South Africa.

Figure 1 .
Figure 1.I-P matrix for three PHEIs concerning the 10 overall most important attributes.

Figure 2 .
Figure 2. I-P matrix for three PHEIs concerning the 10 overall least important attributes.

Table 1 .
Overall 10 most important attributes that have influenced PHEI choice.

Table 2 .
Overall 10 least important attributes that have influenced PHEI choice.

Table 3 .
P-I gap analysis of the overall 10 most important attributes for PHEI 1.

Table 4 .
P-I gap analysis of the overall 10 most important attributes for PHEI 2.

Table 5 .
P-I gap analysis of the overall 10 most important attributes for PHEI 3.

Table 6 .
P-I gap analysis of the overall 10 least important attributes for PHEI 1.

Table 7 .
P-I gap analysis of the overall 10 least important attributes for PHEI 2.

Table 8 .
P-I gap analysis of the overall 10 least important attributes for PHEI 3.
2. Quadrant B: Keep up the good work.Students are generally pleased with the provider's performance.The PHEIs nonetheless need to take cognisance of the fact

Table 9 .
The ethnic-specific importance of the top three overall attributes.

Table 10 .
The ethnic-specific importance of security/ safety on campus.

Table 11 .
The ethnic-specific importance of employment prospects.

Table 12 .
The ethnic-specific importance of spacious, well equipped facilities.