Competition in the independent school industry is intense because current market forces compel schools to compete with one another (Bisschoff, Du Plessis & Smith 2004:1). The ability of independent schools to survive and grow would be enhanced by knowledge of what motivates parents when selecting independent schools in order to assist these schools in better selecting their marketing efforts (McGovern 2012:16). The marketing of independent schools is evident at some independent schools in South Africa. An example is the Independent Schools’ Marketing Association (ISMA) which is recognised as an affinity group and stakeholder of the Independent Schools Association of Southern Africa (ISASA) with the objective of promoting marketing efficiency amongst the schools served by each of its members (ISASA 2013). Therefore, the marketing of schools has become an indispensable managerial function; without it a school cannot survive in the current competitive environment in which it operates (Oplatka & Hemsley-Brown 2004:377).

School marketing is defined as the means by which a school actively communicates and promotes its purpose, values and products to learners, parents, staff and the wider community (Sferle et al. 2012:765). Literature on school marketing originated in the United States of America and the United Kingdom and major themes covered by research in school marketing include the attitudes of school principals and other stakeholders towards the concept of marketing, the lack of research in school marketing and general literature on how to market a school (Oplatka & Hemsley-Brown 2004:375). According to Soedijati and Pratminingsih (2011:212), the development of a marketing strategy involves the coordination and combination of all elements of the marketing mix. The traditional marketing mix of product, place, price and promotion (also referred to as integrated marketing communication or IMC) needs to be expanded to include three additional Ps, namely people, processes and physical evidence. Within the context of this study, the marketing mix elements can be applied as follows: product could, for example, refer to the curriculum followed by the school, price could include school fees or discounts offered by schools, IMC could include advertising (local newspaper advertisements), sales promotion (bursaries), events and experiences (sport festivals), public relations and publicity (fundraising), direct and interactive marketing (school website), personal selling (exhibitions) and word of mouth (person to person or social media). The additional marketing mix elements can be applied as follows: people could refer to educators, processes could include the enrolment process at the school and physical evidence could include aspects such as an attractive and well-maintained school building.

In response to the independent school marketing mix, independent schools need to understand their consumers and their consumption behaviour (Hoyer & MacInnis 2010:24). The key concept of marketing is customer centrality and to realise this goal, service providers have to understand consumer decision-making: how consumers choose between different offerings and how they experience and evaluate the service offerings (Blythe 2013:3).

The decision-making process of services consists of the pre-purchase stage, the service encounter stage and the post-encounter stage (Wells & Foxall 2012:128). The first stage of decision-making in the pre-purchase stage for prospective parents with a child in Grade 1 is the realisation of parents that the child needs to be enrolled at a school (problem recognition). According to the South African Schools Act (1996), schooling is compulsory for children aged 7 to 15 (or attendance in Grades 1 to 9, whichever comes first). The second stage of the pre-purchase stage involves investigating all possible sources of information regarding the service being offered and includes sources such as school open days and the school website. The third stage in the pre-purchase stage, namely the evaluation of alternatives, involves the act of identifying alternative solutions to a problem and assessing the relative merits and demerits of each solution (include the dimensions, features or benefits that consumers seek in making buying decisions) (Parumasur & Roberts-Lombard 2012:258). The service encounter stage is the next stage in the decision-making process for services. This is the stage at which the parent has chosen an outlet (school) and paid the enrolment fee at a specific independent school. The post-encounter stage is the final stage in the decision-making process for services (Wells & Foxall 2012:128).

Consumers evaluate service performance by comparing what they expect with what they perceive they receive from a particular supplier (Lovelock & Wirtz 2011:74). It also happens that consumers often feel post-purchase dissonance (also referred to as cognitive dissonance), which can be described as the doubt or anxiety experienced after making a difficult and relatively permanent decision (Kotler, Armstrong & Tait 2010:164). In the context of this study, cognitive dissonance takes place when parents doubt they had made the correct choice of an independent primary school for their child. This article focuses on the third stage in the pre-purchase stage, namely the evaluation of alternatives. In the context of this study, the evaluation of alternatives is the choice factors of parents when selecting independent primary schools. Consumers typically form a list or set of alternatives from which a choice is made; decision rules are applied to make a selection from the alternatives (Hoffman & Bateson 2011:90).

Research into choice factors of parents selecting independent schools is limited and includes international studies by the Independent Schools Queensland Survey (ISQ 2011), the Independent Schools Council of Australia (ISCA 2008), Robinson (2008:17), Denessen, Driessena and Sleegers (2005:354), Foskett and Hemsley-Brown (2001:83) and Gorard (1999:31). The ISQ (2011) identified amongst others discipline, quality educators, the reputation of the school and academic performance as important choice factors. According to the ISCA, choice factors include the school facilities, educators, a nurturing and caring environment, small class size and discipline. Robinson identified academic standards, discipline and small classes and Denessen et al. identified academic performance, school climate and individual attention to the child as the most important choice factors cited by parents. Choice factors such as tradition, subject choices, facilities as well as the school principal are amongst the top 20 factors identified by parents (Foskett & Hemsley-Brown 2001:83). According to Gorard, the common reasons for choice of parents are academic reasons, location of the school, the management style of the school, discipline as well as security factors. As far as could be ascertained, no research exists with specific reference to choice factors that parents considered in the selection of independent primary schools in the South African context. Limited research done in the South African context on school marketing was the study by McAlister (2005:2) where the researcher aimed to understand management's attitudes towards marketing and the study by Malherbe (2004:82) where the aim was to establish whether independent schools in Gauteng have embraced a culture of marketing. The aim of a study by Vigar-Ellis (2013:10) was to understand the choice criteria of parents when enrolling their child at a secondary boarding school in South Africa.

Following the literature findings, the empirical investigation was exploratory and quantitative in nature. The population comprised all Grade 1 parents from the 556 independent primary schools in Gauteng.

A structured questionnaire based on literature with a non-comparative five-point Likert scale was used to gather the data. With the Likert scale, respondents indicate their attitudes by checking how strongly they agree or disagree with carefully constructed statements, ranging from very positive to very negative attitudes (Zikmund & Babin 2007:214). The scales used in this study ranged between ‘not at all important’, and ‘extremely important’, ‘strongly disagree’ and ‘strongly agree’ and ‘not useful’ and ‘extremely useful’.

Respondents were provided with a self-administered questionnaire or an Internet survey, both of which included self-developed items. Suitable scale items were considered reliable and validity was established. Items and constructs in the questionnaire were developed from the literature review and the exploratory research approach.

A non-probability sampling approach was followed and the judgment sampling technique applied. Judgment sampling choosing the sample based on a person's judgment about some appropriate characteristics required of the sample (Zikmund & Babin 2010:424). The data were gathered over a 12-week period (September, October and November) in 2012. During this time, either the father, mother or guardian of a child in Grade 1 who matched the sampling frame was asked to participate in the survey. A total of 669 questionnaires were completed and could be used in the analysis. The questionnaire was pretested in a pilot study involving 10 respondents from different independent primary schools in Gauteng who matched the sampling frame. The internal consistency reliability test was used to measure reliability. Validity was ensured by having the content of the questionnaire align with the research aim and objectives.

Data analysis included a combination of descriptive statistics to determine the relative importance of different choice factors. The statistical test used for the hypothesis was multivariate analysis of variance (MANOVA). The Pillai's trace was the test statistic used to assess the overall significance of the MANOVA (Pallant 2007:286). Because the MANOVA shows only an overall significant difference and does not indicate where a significant trace result is found, it was followed by a univariate analysis with the application of the Scheffe post-tests. These tests were performed to reveal more specific differences between groups on each of the identified choice factors. The significance level for this study was set at a 95% confidence level, thus α = 0.05. The data obtained from the questionnaires was coded, captured and edited. The Statistical Package for Social Sciences (SPSS version 18) was used to analyse the results. The data obtained from the field were coded and captured in the statistical software package. Cronbach's alpha coefficients were calculated to determine whether the results for each scale used in the questionnaire (i.e. the choice construct and its different factors) were reliable.

Cronbach's alpha was used as a measure of the internal consistency reliability for the measurement of the importance of the attributes. Cronbach's alpha values above 0.7 are deemed to have a good strength of association and are acceptable (Hair et al. 2007). According to Hair et al. (2010:708), when calculating Cronbach's alpha, results just exceeding 0.60 will reflect lower levels of acceptability. The reliability statistics for the questionnaire are presented in Table 1.

It is evident from Table 1 that the Cronbach's alpha coefficients for all eight constructs investigated are above the lower limit of acceptability, 0.60. This confirms that the collected data that were used were internally consistent (reliable).

Validity can be attained by aligning the content of the questionnaire with the research objectives (Van Tonder & Ehlers 2011:164). Consequently, the following activities were performed to ensure validity and reliability: a self-administrated questionnaire was designed, which was directly aligned with the research aim and objectives. The necessary adjustments were made to the questionnaire, based on feedback obtained from the pilot study. A sufficiently large sample size was used to increase the accuracy of the results.

The expertise of STATKON (the Statistical Consultation Services) of the University of Johannesburg was employed to analyse the data gathered and to ensure the correct measurements of the results.

Factor analysis was used to assess the structural validity of the choice factors construct and exploratory factor analysis was used in an attempt to gain insight into the structural validity. The 29 items of the choice factor scale were according to Pallant (2007:181) subjected to principal component analysis (PCA); prior to performing PCA, the suitability of data for factor analysis was assessed using the Kaiser-Meyer-Olkin (KMO) test for sampling adequacy and Bartlett's test of sphericity results. This is presented in Table 2.

It is evident from Table 2 that the 29 statements are suitable for factor analysis: KMO of 0.856, which is greater than 0.6 and Bartlett's test of sphericity is 0.000, which is smaller than 0.05 (Pallant 2010:187).

Of the total 669 respondents from 30 independent schools in Gauteng, of South Africa, who participated in the study, 77% were female, 43% had English as their home language, followed by 35% Afrikaans and 11% Sotho. Almost 40% of the respondents had a postgraduate degree, followed by 15% with an undergraduate degree. The demographic profile of the respondents is presented in Table 3.

Concerning the first research objective, respondents identified 29 variables that they considered when making a decision regarding the independent primary school at which they chose to enrol their child. These variables are categorised into eight factors and are presented in Table 4.

From Table 4, the main finding is that all eight choice factors realised an overall mean score above 3, indicating that all eight factors play an important role in influencing parents’ independent school choice. Five of the eight factors realised an overall mean of above 4 (class size, organisational factors, school characteristics, school leadership and academic factors), indicating that these five factors play a very important role in influencing parents’ selection of an independent school.

Class size (small class size and individual attention to the child) was cited as the most important choice factor by parents when selecting an independent primary school.

The hypotheses focus specifically on reporting on significant differences between demographic groups with regard to the main constructs of the research, namely the factors influencing independent school choice:

A one-way between-groups MANOVA was performed to investigate gender differences of respondents in terms of the importance they attach to choice factors. There was a statistically significant difference between male and female respondents on the combined dependent variables (F [8.658] = 3.9, p = 0.000; Pillai's trace = 0.045; partial eta squared = 0.045). Although a statistically significant result was obtained, the effect size was small (eta squared = 0.045).

The null hypothesis (H₁₍₀₎) was thus rejected, as there is support for H₁.

Female parents (i.e. mothers) agree significantly more than male parents (i.e. fathers) in terms of the importance that they attach to different choice factors in their independent school choice. These differences are present for seven of the eight choice factors, namely academic factors, organisational factors, sport and cultural factors, school characteristics, school leadership, school appearance and class size, and were found to be statistically significantly different by gender:

The above-mentioned hypothesis involves the comparison between three groups, namely Afrikaans, English and other African languages, on a set of choice factors. A one-way between-groups MANOVA was performed to investigate home language differences of respondents regarding the importance they attach to choice factors.

There was a statistically significant difference between home languages of respondents on the combined dependent variables (F [16,1280] = 11.4, p = 0.000; Pillai's trace = 0.25; partial eta squared = 0.12). Although a statistically significant result was obtained, the effect size was medium (eta squared = 0.12).

The null hypothesis (H₂₍₀₎) was thus rejected, as there is support for H_2.

Parents with different home languages differ in the importance they attach to different choice factors in their independent school choice. These differences are present for four of the eight choice factors, namely organisational factors, sport and cultural factors, situational factors and school appearance. Parents with English as their home language rated the importance of organisational factors and sport and cultural factors higher than parents with Afrikaans and other languages as their home language. Parents with other languages as their home language rated the importance of situational factors and school appearance higher than parents with Afrikaans and English as their home language:

The above-mentioned hypothesis involves the comparison between four groups, namely respondents with matric, diploma or higher diploma, undergraduate degree or postgraduate degree, on a set of choice factors. There was a statistically significant difference between the qualification levels of respondents on the combined dependent variables (F [24.1956] = 1.70, p = 0.018; Pillai's trace = 0.06; partial eta squared = 0.02). Although a statistically significant result was obtained, the effect size was small (eta squared = 0.02).

The null hypothesis (H₃₍₀₎) was thus rejected, as there is support for H_3.

Parents with different academic qualification levels differ in the importance they attach to different choice factors in their independent school choice. These differences are present for two of the eight choice factors, namely situational factors and school appearance. Parents with a matric or undergraduate or diploma qualification rated the importance of situational factors and school appearance higher than parents with a postgraduate qualification.

Concerning the second research objective, respondents were asked to rank the top five individual choice factors (from the set of 29 variables). These results are presented in Table 5.

From Table 5 it is clear that the rank total difference between the first rank choice factor (small class size) and the other choice factors (second to fifth) is statistically significant.

The findings of the conducted survey indicate a number of managerial implications for independent primary schools. The primary aim of this article was to recommend marketing guidelines for independent primary schools and provide an understanding of how these choice factors can assist independent schools in South Africa to market themselves better to parents when the latter are selecting an independent school for their children, in order to position independent primary schools more optimally in the market. To accomplish this objective involves the coordination and combination of the marketing mix elements that enable independent schools to meet parents’ needs. These marketing mix elements include product, price, promotion (also referred to as IMC), process, physical evidence and people. For the purposes of this article, only recommendations with regard to product and people are applicable, because only the choice factors applicable to these marketing mix elements and the importance thereof were determined.

The study focused on 30 independent primary schools in Gauteng, South Africa, representing 669 parents of children enrolled in Grade 1. The study therefore cannot be seen as representative of the entire industry in the province or the country. As the study only focused on Gauteng and only independent primary schools, a more inclusive study of all provinces in South Africa (including high schools) is therefore recommended. A comparative study with public schools in South Africa (not included in the study) can also be considered.