‘ Festivalscape ’ factors influencing visitors ’ loyalty to an agri-festival in South Africa Orientation

Practical implications and contribution: This research gives valuable insights into the ‘festivalscape’ factors that influence loyalty to an agri-festival in South Africa. Loyalty amongst visitors can be created through the management of a variety of factors. The findings from this research can be used by similar agri-festivals in the country to create loyalty amongst visitors as well as to give exposure to the agri-sector.


Introduction
The agri-tourism sector includes farming activities that can be linked to the tourism industry as an entrepreneurial advantage.McGehee and Kim (2004) state that the notion of agri-tourism is inclusive of accommodation, educational activities, recreation and festivals.Research by Viljoen and Tlabela (2007) shows that there is a new trend amongst farmers to embrace new opportunities and often farmers will embrace tourism as an income generator.Recent years have seen an exponential growth in interest in the field of agri-tourism, with factors such as poor agriculture, commodity prices, increased production costs, globalisation and industrialisation causing many farmers to find new means of remaining profitable (McGehee & Kim 2004;Myer & De Crom 2013).
In South Africa, there has been a sharp increase in agri-tourism since 1994 and the agri-tourism sector has seen steady growth for the past 20 years with sustained future growth being predicted by researchers (Van Niekerk 2013).Provinces with large agricultural sectors, such as the Eastern Cape, Western Cape and the Free State, host festivals with agricultural products (Visser 2007).Joshni (2012) states that the NAMPO Harvest Day is an example of an agri-festival and, specifically, the festival falls into the agri-entertainment category as festivals form part of this subcategory.Currently, the NAMPO Harvest Day is the largest festival of its kind in the southern hemisphere (Visser 2007).This festival serves as a showcase for GrainSA (GrainSA provides support to grain producers in South Africa for long-term sustainability) for agricultural products and is used as a central meeting place for buyers as well as sellers (GrainSA 2013).The festival has over 650 exhibitors and is known for the introduction of new sustainable farming methods and for showcasing the latest technology in the farming industry (GrainSA 2013).The attendance figures at the NAMPO Harvest Day for the past four years are displayed in Table 1.
The number of visitors in the past four years shows a slight decrease in 2012 and an increase in 2013.However, there was a slight decrease in visitor numbers in 2015.Janeke (2011) published in the Vaal Weekly that the reason for the 73 552 visitors in 2011 can be ascribed to the voting day during the NAMPO Harvest Day.It is clear that the festival attracts a substantial number of visitors each year.In order to maintain these visitor numbers, the factors that influence visitors' loyalty towards the festival need to be assessed, especially considering that visitor numbers vary each year.For the purpose of this study, loyalty can be defined as making repeat visits to the same festival and making positive recommendations to friends and family afterwards (Skogland & Sigauw 2004).Determining the factors that influence this loyalty will enable organisers of the festival to determine not only the degree to which the festival attracts visitors, but also how it maintains visitors (Kruger & Saayman 2012).This is especially vital since there has been a sharp increase in competition amongst agri-festivals (Getz 2002).Festivals such as the South African Cheese Agri-Expo Festival, The Royal Show, the Agriexpo, South African International Documentary Festival, Santam Agricultural Farmers Country Festival, the Knysna Oyster Festival, Ficksburg Cherry Festival, Prince Albert Olive Festival and the Robertson Wacky Wine Weekend all compete for similar visitors.However, the majority of these festivals mainly focus on attracting agri-tourists whilst NAMPO Harvest Day primarily focuses on farmers.The purpose of the festival may therefore significantly influence visitors' loyalty.
The proliferation of festivals has increased the level of competition between different cities wishing to organise festivals, including agri-festivals, and attract potentially interested visitors (Grappi & Montanari 2011).It seems increasingly important, therefore, to understand which features of a festival enhance its attractiveness and increase visitors' retention.Without continued or increased participation, many festivals have a difficult time justifying the necessary financial support from the community where they take place as well as from sponsors (Liang, Illum & Cole 2008).Therefore, in order to exploit a festival's potential benefits, festival planners and managers have to manage all the activities involved in the creation and development of a festival efficiently.Festivals in particular have to adopt strategies that are the most effective in increasing customer retention.Repeat attendees represent a key asset, since they are likely to speak positively about the festival than occasional visitors, pay less attention to offers by competitors, and are more tolerant of low levels of satisfaction (Hume 2008).It is therefore important to examine attendees' levels of satisfaction and loyalty as well as the essential attributes associated with loyalty (Liang et al. 2008).
Despite the growing amount of research on festivals, limited research has to date examined the factors and attributes that affect the level of satisfaction and loyalty towards agrifestivals.This study fills the gap in the current literature and determines the combination of factors or 'festivalscape' aspects that influence visitors' loyalty to the NAMPO Harvest Day.Determining these factors can assist the festival to be promoted, organised and managed so that visitors find value in the festival experience (Baker & Crompton 2000;Lee, Petrick & Crompton 2007;Lee et al. 2008;Yoon & Uysal 2005).As the NAMPO Harvest Day is very successful in attracting repeat patronage, knowledge of these factors can especially assist other agri-festivals to expand visitor numbers and increase loyalty.

Literature review
The literature review is divided into three sections.The first section focuses on loyalty and the theories related to the concept.This is followed by a discussion on the 'festivalscape' factors that may influence loyalty in a festival context, as identified by previous research, as well as the importance of this research.

Loyalty and related theories
Loyalty is a multidimensional concept and has been addressed in numerous different ways in marketing (Oliver 1999;Olsen 2002;Zeithaml, Berry & Parasuraman 1996) and tourism literature (Chi & Qu 2008;Lee et al. 2007;Pritchard & Howard 1997).Loyalty has been a crucial objective of service providers since the high retention of customers or a low defection rate determines long-term profit levels (Zeithaml et al. 1996).Following Oliver (1999), a re-patronising intention can be defined as a pledge to repurchase a potential product (or service) in the future, despite situational influences aimed at causing switching behaviours.Lovelock (2010:151) defines loyalty as customers' 'willingness to continue patronizing a business over a long-term, purchasing and using its goods and services on a repeated and preferably exclusive basis, and voluntarily recommending the firms' products to friends and associates'.Loyalty therefore represents 'irrational behaviour as a result of a deeply held commitment to re-patronise a preferred product/service consistently' (Oliver 1997:392).It is a feeling of allegiance causing customers to prefer a particular brand almost to the extent that the competition gets eliminated (Schoemaker & Bowen 2003:48;Skogland & Sigauw 2004:224).Loyalty or behavioural intention has been measured by (1) positive word-of-mouth, (2) recommendations to others, (3) repurchase intention and (4) high tolerance for price premium (Baker & Crompton 2000;Chi & Qu 2008;Cronin & Taylor 1992;Lee 2009;Zeithaml et al. 1996).
A traditional three-dimensional loyalty framework is dominant in the literature, featuring a behavioural, an attitudinal and a composite dimension; behavioural intention and attitudinal loyalty have been used interchangeably in the literature of marketing and tourism.Behavioural loyalty is synonymous with repeat purchase behaviour, underlying how people make repeat purchases rather than why they buy.Behavioural measures have more commonly been used because of easier implementation from readily available data on customers' repeat purchase history compared with other loyalty measures (Oppermann 2000).Attitudinal loyalty has been proposed as a complement to the use of behavioural indicators of loyalty because of the shortcomings of using behavioural indicators alone (Dick & Basu 1994).Attitudinal indicators provide insights on why people re-patronise a product or service.They focus on understanding consumers' preferences, liking and positive attitudes that are relatively stable over time (Lee, Kyle & Scott 2012).Attitudinal loyalty has also been used synonymously with psychological commitment (Iwasaki & Havitz 1998;Jacoby & Chestnut 1978;Park 1996).Oliver (1997;1999) proposes that the attitudinal dimensions of cognitive and conative loyalty develop in sequence.Consumers first become cognitively loyal based on a belief that one brand is preferable because its attributes are superior to those of the alternatives (conative loyalty).Next consumers form an emotional attachment to the brand through cumulative satisfaction after having used it (attitudinal loyalty).Composite loyalty implies that neither the behavioural nor the attitudinal loyalty approach alone describes loyalty entirely (Day 1976;Lutz & Winn 1974).
Instead it suggests that loyalty should be simultaneously considered from a behavioural and an attitudinal perspective; in other words, a true loyal customer must both purchase the brand and have a positive attitude towards the brand at the same time (Backman & Crompton 1991;Bowen & Chen 2001;Dick & Basu 1994;Petrick 2004).
Odin, Odin and Valette-Florence (2001) propose two approaches, namely the determinist and operational or stochastic approach.The aim of the determinist approach is that loyalty is treated more as an attitude, and the psychological commitment of the purchase is investigated, whereas the stochastic approach explains that loyalty is behaviour, which just means that when a consumer buys the same brand repeatedly then it is said that the customer is loyal.Modern researchers made use of loyalty theories as transactional satisfaction, trust and value, which can also become the determinants of loyalty (Agustin & Singh 2005).Ribbink et al. (2004) as well as Agustin and Singh (2005) explain that transactional satisfaction is an evaluation of how a customer perceives a product or service whilst trust is a belief that there is a trust relationship between the client and the service provider.Value is related to a material thing considered to be a fair exchange in return for a product a customer has purchased and can be defined as the right price for the right quality such that satisfaction is experienced.

'Festivalscape' factors influencing loyalty
Building on 'servicescape' (Bitner 1992;Booms & Bitner 1981), Lee et al. (2008:57) conceptualise 'festivalscape' as the 'general atmosphere experienced by festival patrons' in order to look into festival quality.Similarly to 'servicescape' dimensions, the environmental dimensions of a 'festivalscape' can be described using three categories: ambient conditions (temperature, air quality, noise, music, odours, etc.), space and facilities (layout, equipment, furnishings, etc.) and signs, symbols and artifacts (signage, etc.) (Bitner 1992;Lee et al. 2008).All these environmental aspects of a festival work together to shape the general festival ambience and may therefore affect, for example, the way visitors perceive a festival (Darden & Babin 1994;Lee et al. 2008), their satisfaction and their loyalty (Lee et al. 2008).During a specific festival, the performance quality by the employees, service providers or the activities and programmes is directly under the control of festival organisers; experience quality is not since it is affected by other factors, such as visitors' moods whilst attending the festival (Liang et al. 2008).
Various studies have identified the 'festivalscape' aspects that influence visitors' satisfaction and loyalty at festivals.Saleh and Ryan (1993) discovered that festival programme content was the most crucial factor in attracting visitors to a festival whilst Crompton and Love (1995) and Baker and Crompton (2000) captured four dimensions of festival quality: generic features (festival characteristics), specific entertainment features, information sources (e.g.printed programmes and information booths) and comfort amenities for festival visitors.Information sources and comfort amenities were found to be hygiene factors, or a basic set of conditions, and generic features and entertainment features strongly predicted behavioural intentions.Crompton and Love found that the ambience of the environment, the source of information on the site, comfortable amenities, parking and interaction with vendors were the most important factors.Cole and Illum (2006) and Cole and Chancellor (2009) conclude that quality of amenities, quality of programmes and quality of entertainment influence, directly and indirectly, festival attendees' level of satisfaction, experience and revisit intentions.Entertainment broadly encompasses (live) music and amusement activities like games, depending on the type of festival, and was considered the most important attribute in order to enhance visitors' experiences and satisfaction.Lee et al. (2008) identified seven dimensions of 'festivalscape' that may influence visitor satisfaction: programme content, staff demeanour, facility availability and quality, food perceptions, souvenir availability and quality, convenience and information availability; they further found that controllable characteristics, such as food quality and planned programme content, served as antecedents of festival satisfaction.Liang et al. (2008) examined the behavioural intentions of festival visitors and identified that enjoyment, socialisation and history appreciation all influence the intention to attend festivals.Özdemir and Çulha (2009) found that attributes that lead to higher satisfaction and loyalty were festival programme and the quality of facilities, including sufficient facilities at the festival site, clean festival area, well-organised festival programmes, adequate resting areas, adequate size of the festival area and pleasant atmosphere of the festival site.Yoon, Lee and Lee (2010) identified five festival quality dimensions (informational service, programme, souvenirs, food and convenient facilities) and found that the programmes and convenient facilities alone are antecedents of value for repeat visitors, whereas food, souvenirs, programme and convenient facilities are value antecedents for first-timers.Cole and Chancellor's (2009) conclude that the quality dimensions are programme-related features (i.e.signage, free gifts, children's activities, arts and craft exhibitions and printed programmes and schedules), amenity features (i.e.food and beverages, places to sit, restrooms, overall cleanliness of the festival) and live entertainment.Yoon et al. (2010) assert that merely understanding visitor motives (e.g.escape, socialisation and cultural exploration) is insufficient to guarantee visitor satisfaction and loyalty; the comprehension of motives also needs to be factored into building attractive festival qualities and grasping the flow of festival visitor behaviour (e.g.satisfaction and loyalty).Lee, Lee and Choi (2010) measured festival programme, informational service, festival product, convenient facilities and natural environment and found that the festival programme most strongly predicted value experienced at a festival.Since festival value is a starting point in affecting loyalty, Yoon et al. (2010) adopted five dimensions of festival quality, namely informational service, programme, souvenirs, food and facilities.The festival programme was the dimension most strongly associated with value whilst souvenirs, food and services similarly influenced value, thereby contributing towards festival loyalty.Son and Lee (2011) identified three festival quality factors: general features, comfort amenities and socialisation.General features, including festival quality attributes such as diversity of activities, entertainment sound system, promotion and information, festival atmosphere, entertainment stages, accessibility, safety and security and food and beverage, had the greatest impact on revisit intention.
Wan and Chan (2013) determined the influence of location and accessibility, food, venue facility, environment and ambience, service, festival size, entertainment and timing (organising the festival during good weather) on festival visitor's satisfaction; food, environment, entertainment and service quality had the greatest influence on satisfaction.Grappi and Montanari (2011) focused on programme content, staff behaviour, location and atmosphere, information and facilities, hotel and restaurant offerings and souvenir availability and found that a festival's programme content affects both visitors' emotions and hedonism more strongly than other aspects.Mason and Paggiaro (2012) Lee et al. (2008) also argue that the size of a festival may influence its popularity and festivals that combine food, drink and music are often able to create a playful consumption material.
Collectively, the results of the studies above show that a variety of 'festivalscape' aspects may influence festival visitor loyalty.Programme content was the most influential factor found in the majority of studies.The results further show that the type of festival and the visitors it attracts determines the importance and influence of the 'festivalscape' factors.
When comparing an arts festival with an agri-festival, different loyalty factors will be present.Different tourists also attend various festivals and therefore the demographic characteristics of the visitor and their travel motives will vary.Whilst many studies have focused on the identifying the 'festivalscape' factors at food festivals, to date and to the authors' knowledge, no study has been done at an agrifestival such as NAMPO Harvest Day.

Benefits of festival loyalty
It is imperative to know the combination of 'festivalscape' factors that influence loyalty (Bowen & Chen 2001).Loyalty can increase profits, promote the business or festival, create business referrals, increase sales and motivate repeat purchases (Bowen & Chen 2001:213;Edvardsson et al. 2000:919).When visitors experience higher levels of loyalty, it increases their willingness to pay more for a product or service at an event or festival.Loyalty also influences the visitors' behavioural intentions indirectly through satisfaction with a product or service (Cole & Chancellor 2009:323).Chen andTsai (2008:1115) and Edvardsson et al. (2000:920) state that behavioural intentions include evaluations during the stay, such as the experience, perceived quality, perceived value and overall satisfaction, whilst future behaviour intentions include evaluations such as the intention regarding repeat visits, willingness to recommend and positive word-ofmouth referrals.
Numerous advantages of loyalty also exist, such as: costeffective marketing, which can be realised as there is no need to replace a customer when visitors are loyal (Anderson & Srinivason 2003:124;Rust, Lemon & Zeithaml 2004:109), providing more goods and services with better quality and the prices associated with better quality as loyal customers are not price sensitive (Dowling & Uncles 1997:71;Zeithaml 2000:68).Price insensitivity occurs when the client still prefers a brand even though the price fluctuates; creating brand advocacy refers to a situation in which the customers will continue their loyal attitudes despite other products or services on the market.Customers in this phase will also provide positive recommendations towards the product or services rendered (Zeithaml 2000:68) and forecasting becomes easier.Assessing these factors, cost estimates can be predicted, as there will be loyal customers that will always invest in the products or services, which ensures a strong customer base.This will also give greater peace of mind to businesses as businesses will always have an income and also have a sense of competitive advantage (Salanova, Agut & Peiro 2005:1227).

Method of research
A structured questionnaire was used to collect the data.This section describes the methodology applied in the research as well as the results.

The questionnaire
The questionnaire consisted of two sections.Section A captured socio-demographic and behavioural details such as gender, age, home language, occupation, home province and preferred accommodation, as well as spending behaviour.
Spending behaviour included the number of persons paid for, length of stay and the expenditure of visitors on the different aspects of the trip.Section B captured the 'festivalscape' aspects that might contribute to visitor loyalty at the NAMPO Harvest Day.43 items were measured on a five-point Likert scale (1 = strongly disagree and 5 = strongly agree).The loyalty factors were based on the 'festivalscape' factors discussed in the literature review (and subsections), namely programme content, staff, facilities, food, convenience, and information availability, to name but a few.Statements were adapted for the festival and specific items distinct to the festival in terms of its programme and setup, for example agricultural and farming aspects, were included.Since travel motives also assist in identifying loyalty factors, motivations were also included in this section.These motives were based on the work of Crompton and McKay (1997) and included aspects such as to seek out new experiences, to spend time with friends and relatives, for rest and relaxation and to gain new knowledge.

Trustworthiness
The convergent construct validity of the variables in Section B's scale was tested by means of factor analysis to determine the combination of factors with which the variables are most consistent (Zikmund 2010:308).The last criterion for a measurement is reliability.The reliability of Section B's scales is also measured by means of the alpha coefficient that represents the average of all possible split-half reliability for a construct (Zikmund 2010:306).The alpha coefficient indicates the quality of the scale; in other words, a score ranging from 0.60 to 0.70 indicates fair reliability, a score between 0.70 and 0.80 indicates good reliability and a score over 0.80 indicates excellent reliability (Zikmund 2010:306).

Sampling method and survey
The survey was done by means of a self-administered questionnaire; probability sampling method and a stratified sample was used.The questionnaires were distributed at the festival in Bothaville from 13 to 16 May 2014.The survey was conducted on the festival grounds, especially at various relaxation points on the festival grounds to limit bias.Only adults and every second visitor were requested to complete questionnaires.Trained fieldworkers informed respondents of the purpose of the research and the voluntary nature of the study.The questionnaires were progressively handed out towards the end of the festival to give a more accurate account of the different types of visitors at the festival as well as their spending.The number of questionnaires distributed over the duration of the four days was 90 questionnaires in day 1, 120 questionnaires on day 2, 130 questionnaires on day 3 and 160 questionnaires on day 4. A total of 422 completed questionnaires were returned.The attendance figure for the NAMPO Harvest Day in 2013 was 72 376 visitors (GrainSA 2015).In a population of 70 000 (N), 382 respondents (n) would be seen as representative and the number of completed questionnaires (n = 422) was thus more than adequate (Krejcie & Morgan 1970).

Profile of respondents
The majority of the respondents were male (65%) with an average age of 39 years and were mostly Afrikaans speaking.They were predominantly farmers (31%) with a gross income of between R221 001 and R305 000 and were mostly from the Free State.The respondents had an average of one person per group that spent an average of two days at the agri-festival.The majority of the respondents were only visitors (49%) who had initiated the visit themselves (43%) and preferred to attend the agri-festival for the stalls (60%).
The respondents heard about the agri-festival on the radio (42%) and their primary farming interest is mostly mixed farming (35%).The respondents attending the Harvest Day feel strong about loyalty as the majority indicated that they attended the NAMPO Harvest Day in previous years (68%) with an average of three times prior.The highest spending was on the purchasing of machinery that was an average of R88 885 followed by the purchasing of seeds and crops (R49 087) and purchasing of farm implements (R72 213).

Statistical analysis and results
Using an Oblimin rotation with Kaiser normalisation, two principal axis factor analyses were performed on the 40 items that can contribute to loyalty and on the three items related specifically to loyalty, as indicated by Lee et al. (2008:58), to explain the variance-covariance structure of the set of variables through a few linear combinations of these variables.The Kaiser-Meyer-Olkin measure of sampling adequacy was used to determine whether the covariance matrix is suitable for factor analysis.Kaiser's criteria for the extraction of all the factors with eigenvalues larger than 1 were used.All items with a factor loading above 0.3 were considered to be contributing to a factor, whereas those with loadings lower than 0.3 were not correlating significantly to this factor (Pallant 2005:116;Steyn 2000:2).In addition, any item that cross-loaded on two factors, with factor loadings greater than 0.3, was categorised in the factor where interpretability was best.A reliability coefficient (Cronbach's alpha) was computed to estimate the internal consistency of each factor.The average inter-item correlations were also computed as another measure of reliability, which should lie between 0.15 and 0.55 (Clark & Watson 1995:310).
A structural equation model (SEM) was used to indicate the relationship between the identified factors and visitor loyalty at the NAMPO Harvest Day.A SEM is used to determine whether an individual model is valid and not to prove or find a model that is suitable for a certain situation (Lee et al. 2008:59).A SEM analysis estimates effects between different variables and serves as a flexible approach to modelling different data sets using a wide variety of estimation methods and in the process accommodates means, pattern of means, latent interaction and clustered data (Byrne 2011:14).The results of the statistical analysis are subsequently discussed.

Factor analysis results: Loyalty
Since the aim of the research is to determine the most influential factors that contribute towards visitors' loyalty to the NAMPO Harvest Day, the items relating directly to loyalty (positive word-of-mouth, recommendations to others and re-patronisation intentions) were extracted from the other items and dealt with as a single factor.These three factors correspond with the core aspects of loyalty as identified in previous research (Baker & Crompton 2000;Chi & Qu 2008;Cronin & Taylor Kotler et al. 2006;Lee 2009;Lee et al. 2008;Zeithaml et al. 1996).The pattern matrix of the principal axis factor analysis using Oblimin rotation with Kaiser normalisation grouped the three items under the one factor that was labelled Loyalty (Table 2).The one factor accounted for 77% of the total variance.The factor had a relatively high-reliability coefficient of 0.85.The average inter-item correlation coefficients with a value of 0.66 also imply internal consistency.Moreover, all items loaded on the factor with a loading greater than 0.3 and relatively high factor loadings indicate a reasonably high correlation between the delineated factors and their individual items (see Table 2).The Kaiser-Meyer-Olkin measure of sampling adequacy of 0.915 showed that patterns of correlation are relatively compact and yield distinct and reliable factors (Field 2005:640).Bartlett's test of sphericity also reached statistical significance ( p < 0.001), supporting the factorability of the correlation matrix (Pallant 2007:197).The factor score was calculated as the average of all items contributing to the particular factor so that they could be interpreted on the original five-point Likert scale of measurement.The factor received a mean value of 4.22 with a reliability coefficient of 0.85 and an average inter-item correlation of 0.66.Based on the mean value, respondents strongly agreed with this factor.

Factor analysis results: 'Festivalscape' factors
The second exploratory factor analysis was done on the remaining items or 'festivalscape' aspects that can contribute towards visitors' loyalty.The pattern matrix of the principal axis factor analysis using Oblimin rotation with Kaiser normalisation identified nine factors that were labelled according to similar characteristics (Table 2).The nine factors accounted for 73% of the total variance.All factors had relatively high-reliability coefficients ranging from 0.83 to 0.93.The average inter-item correlation coefficients with values of 0.477 and 0.847 also imply internal consistency for all factors.Moreover, all items loaded on a factor with a loading greater than 0.3 and relatively high factor loadings indicate a reasonably high correlation between the delineated factors and their individual items.The Kaiser-Meyer-Olkin measure of sampling adequacy of 0.915 also showed that patterns of correlation are relatively compact and yield distinct and reliable factors (Field 2005:640).Bartlett's test of sphericity reached statistical significance ( p < 0.001), supporting the factorability of the correlation matrix (Pallant 2007:197).
Factor scores were calculated as the average of all items contributing to a particular factor so that they could be interpreted on the original five-point Likert scale of measurement.Agricultural exposure and education received the highest mean value of 4.

Results of the structural equation model
The correlation analysis indicated that all nine factors have a positive influence not only on loyalty but also on each other, indicating a strong relationship between the factors.A detailed discussion of the correlation results can be obtained from the authors on request.The SEM analysis is used as an approach to confirm the analysis and also to ensure that there is a covariance between the variables (Byrne 2011:7).The purpose of using SEM analysis within this study is to determine the relationship between the 'festivalscape' factors as identified in the factor analysis and loyalty at the NAMPO Harvest Day.For the purposes of this study the maximum likelihood estimation (ML) procedure was used within the Analysis of Moment Structures program (AMOS).AMOS uses a graphical interface to construct the hypothesised paths.The ML method finds a set of free parameters that maximises the likelihood of the data given the specified model (Hair et al. 2010:663) and delivers estimates that are the most precise of the estimates available with minimum variance (Savalei & Bentler 2006:341;Wang & Wang 2012:15).
The validity of the measurement model depends on the goodness-of-fit results (Briggs, Coleman & Morrison 2012:377;Malhotra 2013:717).The goodness-of-fit measures indicate how well the specified model reproduces the observed covariance matrix amongst the observed variables (Hair et al. 2010:664).One of the measures is the chi-square test, which should be non-significant (i.e.p > 0.05 ;Muijs 2004:377).Since this test will detect even slight deviations from the data with large samples, other fit indices are also necessary to consult measures that are not sensitive to sample size (Briggs et al. 2012:377).A comparative fit index (CFI) closer to 1, root mean square residual (RMSEA) closer to 0 (preferably ≤ 0.08) and a goodness of fit (GFI) above 0.90 reveals a good fit for a model (Malhotra 2013:718-719;Wang & Wang 2012:18).
The fit indices in Table 4 provide the evidence of a good fit, since the RMSEA is not below the expected margin of 0.08, but is still in an acceptable range.Values for the CFI should be between 0.0 and 1.0 and values closer to 1.0 indicate a good fit.The CFI was 0.819, which means that it is acceptable.In addition, the relative or normed chi-square (x 2 /df) is considered acceptable, as the value was 3.987 and an acceptable ratio for the chi-square divided by its degrees of freedom is between 2.0 and 5.0.
The SEM that measured the relationship between all nine 'festivalscape' factors and loyalty.The inspection of the standardised coefficients for regression paths was the method applied for examining the relationship between the identified latent variables.Table 5 shows the maximum  5).

Findings and implications
The findings of this study were as follows.Firstly, the particular combination of 'festivalscape' factors found in this research has not been identified in previous research.
Contradicting the finding by most authors discussed in the literature review, programme content was not the most important factor.This finding can be ascribed to the type and nature of the festival (in this case an agri-festival) and the fact that little research in this area has been conducted.This finding therefore confirms the notion advanced by Kruger and Saayman (2012)  It is interesting to note that respondents rated value lower than with the other factors; this could suggest room for improvement in the aspects of this factor that festival organisers need to address.The NAMPO Harvest Day marketers should therefore focus their marketing campaigns on the combination of the above 'festivalscape' factors in order to retain their current market.These findings also provide important information as they contribute to existing literature regarding the aspects that influence visitor loyalty to these type of festivals and can serve as a framework for organisers of smaller, lesser-known agri-events and festivals to learn from the success of this festival and market their events and festivals accordingly.
Secondly, confirming once again the notion that the type and nature of the festival influences visitor loyalty, the correlation analysis indicated that all nine 'festivalscape' factors have a positive impact not only on loyalty but also on each other, indicating the strong relationship between the factors and that the combination of factors needs to be considered to enhance loyalty.The SEM analysis showed a direct positive relationship between loyalty and lifestyle, escape and socialisation as well as between loyalty and agricultural exposure and education.This means that agricultural exposure and education was identified in both the factor analysis and the SEM analysis as more influential, which means that this factor has a significant influence on loyalty to the NAMPO Harvest Day.These findings imply that if the organisers of the festival want to increase loyalty amongst visitors to their festival, their primary focus should be on enhancing these two loyalty factors since they also correlate positively with the other factors.Organisers of the festival should therefore focus on agricultural exposure and education as well as on ways to expand it.The NAMPO Harvest Day is an important agrifestival for farmers in South Africa; therefore, it is important to sustain the festival especially because South Africa is rich in agricultural activities.The educational aspect is one of the key concepts in this agri-festival; therefore, organisers should focus on giving the visitors an educational experience and create awareness regarding the agricultural activities, such as demonstrating how to work with a thresher machine or how to choose the best tractor, at the same time giving the visitors who are not farmers or in the agricultural sector an experience to take home.Agri-festivals should create an interest in the agricultural sector amongst farmers and the general public, for example by providing demonstrations on how to make wine or how to shave a sheep, and in so doing make them part of the experience.Another aspect management needs to consider in order to emphasise the educational aspect whilst at the same time creating an interest in the agricultural sector is by having specialised school packages especially for schools that have an agricultural subject as part of their curriculum.In this way scholars can get the necessary exposure to the industry and foster an understanding of and appreciation for this invaluable sector in the country.Another aspect that agri-festivals should look into is the creation of an opportunity to exchange news about the latest innovations regarding farming and agriculture.Agri-festivals provide farmers the opportunity to exchange information on the latest trends in terms of farming practices and equipment.Agri-festivals such as the NAMPO Harvest Day should therefore create a platform where farmers can interact, exchange ideas and be informed and educated regarding the latest trends in agri-tourism and farming practices.Specialist talks and demonstrations are essential elements to achieving this.
Concerning lifestyle, escape and socialisation, farming is considered a way of life, seeing that many farms have been in families for generations.It therefore makes sense that attending agri-festivals such as NAMPO Harvest Day also forms part of this lifestyle.Agri-festivals should, however, also focus on providing the visitors with entertainment, especially at the rest areas where the visitors can relax, eat and enjoy the entertainment.This will require more management administration, but can provide a competitive advantage for the agri-festival as the majority of the agrifestivals only focus on the agricultural aspects and not necessarily on entertainment.This can also be a means to attract visitors who are not necessarily farmers, thereby increasing the market share of the festival.Agricultural exposure and education can also be coupled with this factor by making auctions and demonstrations fun and interactive and by hosting farm-related competitions.This can also be a way of getting all visitors involved and especially giving non-farmers the opportunity of experiencing farm activities.
However, it is important to note that the remaining seven factors also had a positive influence on loyalty and should therefore not be disregarded.The NAMPO Harvest Day festival already implements various aspects under the factors identified in the factor analysis.Other, smaller festivals should consider the festival's success, learn from it and implement similar measures within their setup.

Conclusion
This research gives valuable insights into the factors that influence loyalty to an agri-festival in South Africa.Loyalty amongst visitors can be created through the management of a variety of factors whilst at the same time taking into consideration the type and needs of the visitors.Since agrifestivals provide a multifaceted product and programme, the different elements need to be managed in cohesion with visitors' needs and preferences for the different types of products on offer.The findings from this research can therefore be used by similar agri-festivals in the country to create loyalty amongst visitors as well as to give exposure to the agri-sector.
From the research it is also evident that agri-festivals such as the NAMPO Harvest Day play a significant role in the agritourism sector and can be used as a means for facilitating tourism to areas such as Bothaville, which usually do not usually receive tourists during the year.Organisers of these types of festivals can consider compiling packages that include accommodation and nearby tourist attractions to encourage visitors to spend more time in the area.These festivals can also consider providing visitors with an authentic 'farm stay' experience and involve local farmers in the area to provide accommodation to visitors for the duration of the festival.This could also be a way of attracting non-farmers to the area.
It is furthermore evident from this research that the factors that contribute to loyalty at an agri-festival differ from, for example, an arts festival.This confirms that the type and nature of the festival greatly influence loyalty and this is a key aspect organisers of agri-festivals need to bear in mind.
Loyalty within the agri-tourism sector will create awareness and keep an interest in the agricultural sector, which is imperative in South Africa.Agri-festivals will also result in the development of rural areas.Various agri-festivals are hosted in South Africa and if managed correctly this can have a huge impact on the tourism industry, especially in South Africa.South Africa can even become the destination for an agri-tourism experience for international tourists.

Directions for future research
It is recommended that a comparative study between different agri-festivals be done based on visitor profiles, travel motivations and loyalty factors.This research will assist in seeing whether the province where the agri-festival is hosted makes a difference to the type of agri-tourist attracted and the factors contributing towards loyalty at agri-based festivals.Since the factor value yielded such diverse results, research should be conducted on the supply and demand of agri-festivals in South Africa by identifying gaps between what management perceive to be value for money and what the visitors perceive it to be.A mixedmethod approach is recommended here.There is also a vast difference in how management wants to focus on loyalty and how visitors perceive loyalty.This research will assist in addressing these gaps and in providing a value-formoney experience, which is essential for creating loyalty amongst visitors.
Yeoman et al. (2004) and) food, convenience, benefits and communication significantly influence festival visitors' motivation and loyalty behaviour.O'Toole,Harris & McDonnell (2005),Yeoman et al. (2004) and Kitterlin and Yoo (2014))al intention, which includes revisiting and recommendations to others.More recently,Wong, Wu and Cheng (2014)revealed that interaction quality, physical environment quality, outcome quality and programme quality positively affect overall festival quality, whilstKitterlin and Yoo (2014)found that 'festivalscape' aspects such as programme

TABLE 2 :
Factor analysis results of loyalty.

TABLE 3 :
Factor analysis results of the factors that have an influence on loyalty of the visitors to the NAMPO Harvest Day.The NAMPO Harvest Day offers an adequate variety of implements, products and demonstrations.The NAMPO Harvest Day is a well-organised event.The staff at the Harvest Day is willing to assist visitors.The introduction of new products at NAMPO is excellent.There are sufficient facilities on premises.The site is neat and tidy.The atmosphere at the NAMPO Harvest Day is exciting.The site is large enough to accommodate all the visitors.Attending the Harvest Day is value for money.Attendance is an annual commitment.The Harvest Day is the perfect festival to see new products and new agricultural trends.There is adequate and affordable accommodation available.The Harvest Day is the perfect place for purchases of new products, for example seeds and fertilizers.

Factor 5: Lifestyle, escape and socialisation Attending
this type of agri-event is part of my lifestyle.New knowledge about agriculture available at the festival helps me to improve my own farm.The Harvest Day is the perfect opportunity for relaxation.Attending the festival is the perfect opportunity to spend time with family and friends.Attending the Harvest Day is the ideal opportunity to meet new people.

Price and quality of implements, machinery and livestock
The Harvest Day offers affordable prices and quality of implements, machinery and livestock.The location of the Harvest Day is ideal.The Harvest Day offers an excellent variety and availability of implements, machinery and livestock.Good access to the site.

Price and quality of food and beverages
The quality of the food, drinks and refreshments is excellent.The prices of the food, drinks and refreshments are affordable.The variety and availability of food, drinks and refreshments are sufficient.The layout of the premises at the NAMPO Harvest Day is exciting.

TABLE 4 :
Goodness-of-fit indices.The results in Table5indicate that only two 'festivalscape' factors are supported at a 5% level of significance, namely lifestyle, escape and socialisation ( p = 0.001) and agricultural exposure and education ( p = 0.001).Based on the standardised regression, weights of both the factors agricultural exposure and education (0.463) and lifestyle, escape and socialisation (0.284) are positive.Although all the 'festivalscape' factors have an influence on loyalty (based on the correlation analysis) the most significant factors that have an influence on loyalty are agricultural exposure and education and lifestyles, escape and socialisation (Table (Liang et al. 2008;Son & Lee 2011)(Liang et al. 2008;Son & Lee 2011).To the authors' knowledge, to date no research has been done on the specific agricultural aspects of a festival that may influence loyalty, especially in South Africa; the factors agricultural exposure and education, price and quality of implements, machinery and livestock and networking and trade are thus distinct to the NAMPO Harvest Day, over and above the combination of identified factors.

TABLE 5 :
The standardised regression weights, estimates and p-values.
http://www.actacommercii.co.za doi:10.4102/ac.v15i1.307 however provide valuable insights into the factors that have an influence on visitor loyalty to an agri-festival like NAMPO Harvest Day.
This research was conducted at only one agri-festival in South Africa and results can therefore not be generalised.It http://www.actacommercii.co.za doi:10.4102/ac.v15i1.307does