Recreational tourism is the leading global activity representing an average of 77% of global tourism spending between 2017 and 2019 (World Travel and Tourism Council – WTTC 2018 2020). South Africa is synonymous with leisure and nature-based recreational tourism activity (Hammett 2014; South African Tourism 2017; Van Dyk, Tkaczynski and Slabbert 2019). As a result, South Africa is the most attractive tourism destination in Sub-Saharan Africa, accounting for 70% of its tourism Gross Domestic Product in 2019 (World Economic Forum 2019). As a premium recreational tourism destination and one of Africa's largest tourism economies, in 2018, South Africa attracted upwards of 10.5 million inbound tourists and generated an estimated USD4.7 billion in tourism receipts, translating into a cumulative USD32.1 billion in (in)direct tourism revenue (South African Tourism 2019; WTTC 2019).

In an increasingly globalised world, tourists face the challenges and complexities of making their consumptive decisions in an information-rich environment, often resulting in an information overload. However, while this is an interesting phenomenon in various sectors and geographical locations, it is essential to note that one significant impediment to African tourism is a lack of information [also referred to as information asymmetry] which results in tourists' over-reliance on often inaccurate stereotypes and preconceptions to inform their decisions relating to tourism to the continent (Signé 2018). Moreover, the literature (Meeras 2010; Oosterhaven and Van der Knijff 1987) already establishes information asymmetry as an overarching barrier to recreational tourists" decision-making. Hence it is possible to consider that information asymmetry presents African recreational tourism destinations such as South Africa with a unique marketing conundrum. To mitigate the challenges associated with both information overload and, in the case of lesser-known destinations, information asymmetry, tourists are increasingly exploiting Country Images (CIs) and Place Brands (PB) as heuristic cues (cognitive shortcuts) to inform their decision-making (Chaulagain, Wiitala and Fu 2019; Stepchenkova et al. 2018). Hence the present study sought to address the challenge of managing tourist behaviour in an otherwise 'cluttered' and information-rich global tourism market by seeking to better understand the underlying relationship between the CI, PB and travel motives of tourists – more so from an emerging market perspective.

Notwithstanding the increasing number of studies (Chaulagain et al. 2019; Hahm, Tasci, Terry 2018; Nadeau et al. 2008) on CI and its influence on tourist behaviour, what appears to be deficient in the contemporary literature is empirical in-situ evidence related to the effect that CI has as an extrinsic stimulus on the perceptions and travel motives of recreational tourists from an emerging market perspective (Campo Martínez and Alvarez 2010). Furthermore, the influence of intervening perceptual factors such as PB dimensions in the CI, travel motives nexus remains largely unexplored. Therefore, the present study explored the CI – recreational travel motivation nexus mediated by South Africa's PB. The present study makes four significant contributions. First, while CI is an important contemporary subject of academic inquiry within tourism research, discernible evidence exists of a knowledge gap in the literature concerning the relationship between CI and tourist behaviour from an African tourism market perspective. Hence this current study provides empirical evidence that advances the literature and expands the extent of the knowledge related to CI's importance to travel behaviour to the potential intervening role of PBs in tourist travel motivation. Second, globalisation and the associated competitive forces have elevated the influence of CIs on tourists' decision-making when considering destination attributes as pull factors and their subsequent conative travel behaviour. Thus, the study provides tourism practitioners with critical insight into recreational tourist behaviour based on the stimulus of CI, potentially mediated by PB dimensions as heuristic cues. Third, the influence of both CI and PBs on the travel motives of tourists is the subject of limited academic inquiry within the African tourism context. Therefore, this study addresses the apparent lack of interest in studies associated with CI and tourist behaviour from an African recreational tourism viewpoint by exploring the under-researched subject in African tourism. Fourth, considering the COVID-19 pandemic, this study's findings provide tourism practitioners with retrospective empirical evidence of the potential utility of CI and PBs in managing tourist perception in the short- to medium-term post the crisis.

1.1 Theoretical framework

CI is a multi-dimensional construct associated with the "[…] tangible information, biases, imaginations, impressions, and emotional musings" that consumers utilise as heuristic cues in their consumptive decision-making concerning products of a country such as tourism offerings (Hwang, Asif and Lee 2020:4). Grounded in international marketing theory (Hwang et al.  2020), CI may be classified into two distinct typologies. Stereotypical CI is based on organic influences that manifest in the minds of consumers - which countries themselves cannot control (Motsi and Park 2020). At the same time, Reverse country-of-origin CI is the perception-based bias induced by planned activities and actions, including marketing and exports (Buhmann,2016). The ability of the CI to either positively or negatively influence consumer behaviour may be attributed to the halo effect and how the subsequent cognitive and affective inferences impact the evaluation of products associated with the country on a macro level (Septianto, Chiewb and Thaic 2020; Stepchenkova et al. 2018). The literature (Lala, Allred and Chakraborty 2008; Nadeau et al.  2008; Zhang et al. 2016) acknowledges that the formation of CI in the minds of consumers is highly individualised and subjective; hence there is generally no consensus prevails with regard to the definition of CI and its dimensions. Despite this, global competitive tourism market forces have resulted in CI becoming recognised as a crucial heuristic cue in marketing tourism destinations – with positive CI being harnessed as a tool for comparative advantage and the competitive brand positioning of tourism destination countries (Chaulagain et al. 2019).  

The PB may be characterised as a "[…] network of associations in the consumers' mind based on the visual, verbal, and behavioural expression of a place, which is embodied through the aims, communication, values, and the general culture of the place's stakeholders and the overall place design," (Zenker and Braun 2010:3). PB is rooted in the broader place marketing discourse, which O'Leary and Iredale (1976) first posited as a strategic approach to positively influencing consumer behaviour towards places and their 'products'. Place marketing has evolved into various discourses, including PB. PB theory and practice have conventionally been the preserve of the tourism discourse (Oliveira 2015) and have developed as a micro construct akin to destination branding (Hanna, Rowley and Keegan 2021; Matiza and Slabbert 2020). As a communicative platform in tourism, the PB is a multi-dimensional construct comprising generic intangible attributes (tourism, investment and immigration, governance, exports, people culture and heritage) that tourists utilise as heuristic cues to inform their subjective preference for a specific tourism destination (Dinnie 2008; Hanna et al.  2021).  

Travel motivation is a critical antecedent to tourist behaviour, informing essential decisions such as purchase behaviour, subjective preferences and destination choice (Xie and Ritchie 2019). The Push-Pull Framework (PPF) is the seminal explanatory construct for how tourists are 'pushed' to engage in tourism activity by intrinsic needs and 'pulled' towards specific tourism destinations with the attributes they perceive would best satisfy their needs (Chen and Chen 2015; Crompton 1979; Xu and Chan 2016). However, as a subjective construct, the literature acknowledges the dichotomous nature (push versus pull motives) of travel motivation and the subsequent array of travel motivation factors (Phau, Lee and Quintal 2013; Xu and Chan 2016). Contemporary studies have generally established the influence of push and pull travel motivations on tourism activity and destination choice respectively in the cases of countries that include New Zealand (Fieger, Prayag and Bruwer 2019), Taiwan (Chen and Chen 2015), Turkey (Caber and Albayrak 2016) and Australia (Michael, Wien and Reisinger 2017).

1.2 Theoretical framework

CI as a summative construct with a superseding halo effect influences the decision-making process of tourists based on the generic tourism and non-tourism stereotypes of a country. At the same time, the PB is a critical marketing axiom in the branding of tourism destinations representing the value proposition of a tourism destination in tourists' minds (Dinnie 2008; Matiza and Slabbert 2020; Zenker and Braun 2010). Previous studies have established the heterogeneity and subjective nature of CI and PB within the tourism context (Elliot, Papadopoulos and Kim 2011; Palau-Saumell et al. 2016). Moreover, both constructs influence the decision-making and conative behaviour of tourists, with the CI being a macro construct (often stereotypical) that supersedes the micro (often induced) PB dimensions (Chaulagain et al. 2019; Matiza and Slabbert 2020). Within the scope of decision-making and tourists" conative behaviour based on location-specific attributes, travel motivation is associated with factors that attract tourists to visit a specific country to satisfy their intrinsic needs (Fieger et al.  2019). To this end, prior tourism studies have generally established a causal relation between destination attributes and tourist behaviour (Caber and Albayrak 2016; Chen and Chen 2015; Prayag and Ryan 2011; Van Dyk et al.  2019). Contrastingly, only a handful of tourism studies (Hwang et al.  2020; Palau-Saumell et al.  2016; Prayag 2009) have modelled the general relation between CI and travel motives. Notwithstanding the notion that travel motives may be based on other intrinsic personal factors such as nostalgia (Leong et al. 2015) and ethnocentrism (Prayag and Ryan 2011), the CI  and PBs are emerging as critical formative and summative decision-making constructs.

1.3 Recreational tourism and travel motivation

Recreational tourism can be differentiated into two distinct typologies based on the resources utilised in recreational activities, namely natural and anthropogenic recreational tourism resources (Gjorgievski, Kozuharov and Nakovski 2013). Natural recreational tourism resources occur in natural geographical spaces and mainly satisfy the needs of nature-oriented tourists. In contrast, anthropogenic recreational tourism resources are man-made facilities intentionally developed to fulfil the needs of leisure tourists (Da Silva et al. 2013; Gjorgievski et al.  2013). Therefore, recreation may be associated with both leisure (cultural events, the arts, entertainment, sports events) and nature-based tourism (adventure excursions, wildlife viewing, beach-going) activities (Nafi and Ahmed 2018; Rudskii et al. 2019).  

Leisure tourism is synonymous with activities undertaken during tourists' free time to voluntarily partake in intrinsically pleasurable, relaxing and rewarding experiences based on psychological and physical motivators (Nafi and Ahmed 2018; Yousefi and Marzuki 2015). Previous studies have established heterogeneous variables as motivational factors for leisure tourism in the cases of Taiwan (Chen and Chen 2015), Tanzania (Kara and Mkwizu 2020) and Turkey (Nikjoo and Ketabi 2015). These may include access to museums, monuments, and historical locations and artefacts of interest (Nikjoo and Ketabi 2015); parks, conservancies and reserves (Fieger et al. 2019); entertainment activities such as mega sports events such as the Olympics, theme parks and casinos (Nafi and Ahmed 2018); festivals, the arts, music concerts (Sangeeta and Anandkumar 2016); as well as unique cultures and ways of life (Yousefi and Marzuki 2015).  

Nature-based tourism is an overarching tourism typology that encompasses various tourism activities, including adventure, eco- and sustainable tourism-oriented activities (Mehmetoglu and Normann 2013). Few studies (Mehmetoglu and Normann 2013; Meric and Hunt 1998; Xu and Chan 2016) have explored the motives associated with nature-based tourism as a generic tourism typology. Instead, the literature has focused on exploring tourist motivation based on specific tourist activities such as adventure tourism (Giddy and Webb 2018), bird watching (Chen and Chen 2015), and rock-climbing (Caber and Albayrak 2016). However, previous studies have recognised that generic location-specific nature tourism attributes (pull factors) motivate tourists. These include the climate of the location (Michael et al.  2017; Rice and Khanin 2019); flora and fauna found at the site (Chen and Chen 2015); the availability of and access to wildlife (Fieger et al.  2019); as well as available outdoor activities (Caber and Albayrak 2016; Fieger et al.  2019), respectively.

1.4 Hypothesis formulation

The literature (Dinnie 2008; Hanna et al. 2021; Hwang et al. 2020) regard tourist behaviour and decision-making as being complex processes susceptible to biases that may be informed by heuristic cues such as CI and PBs. Furthermore, considering the theoretical framework, it is reasonable to hypothesise that due to the CI’s overarching halo effect on both consumer perception and behaviour, South Africa's CI may be viewed as an exogenous variable that can influence PB perceptions and tourist’s travel motives alike (Matiza and Slabbert 2020). Moreover, the established notion that tourists’ travel motives are susceptible to the influence of both the CI (Chaulagain et al. 2019) and PB dimensions (Zenker and Braun 2010) supports the viability of modelling travel motives as an endogenous variable (Xie and Ritchie 2019) of which its relationship with the CI may be mediated by the PB. As a result, in line with the transmittal approach to mediation analysis (Memon et al. 2018) and some of the literature (Agler and De Boeck 2017; Gunzler et al. 2013), the null hypothesis was tested as follows (Figure 1),

H0: South Africa's place brand does not mediate the relation between the country's image and inbound recreational tourists' travel motivations.

H1: South Africa's place brand mediates the relationship between the country's image and inbound recreational tourists' travel motivations.

Latent Construct 1

Conceptual framework

Data were generated as part of a quantitative-deductive study that surveyed inbound tourist perceptions of South Africa as a tourism destination based on CI and PB dimensions and their travel motivations. The Aerial Cableway station at Table Mountain in Cape Town, South Africa, was the study site where 400 self-administered questionnaires were distributed to a convenient sample of international tourists between the 6th and 9th of November 2018. Three hundred ninety-five usable questionnaires were returned, of which n=233 were suitable for the present study's purposes.

2.1 Measuring instrument

A measuring instrument was developed for the study. The questionnaire consisted of four sections, socio-demographic; CI; PB dimensions; and travel motivations.

Independent variable

(exogenous): Country image (CI) as a summative construct of the stereotypical beliefs held by a country was measured by nine items which were drawn from the literature (see Dinnie 2008; Nadeau et al. 2008; Palau-Saumell et al. 2016; Zhang et al. 2016; Zhang et al. 2018). Respondents were asked how specific CI dimensions of South Africa"s image influenced their views of the country. Responses were recorded on a five-point Likert scale of influence between (1) Very negative influence, to (5) Very positive influence.

Mediator variable

: The influence of South Africa's PB as the value proposition of the country as a tourism destination in the minds of international recreational tourists was measured based on 44 items drawn from previous studies (see Avraham 2020; Han and Hyun 2013; Žugić and Konatar 2018). Respondents were asked about the extent of influence each aspect influenced their decision-making towards South Africa as a tourism destination. Responses were recorded on a five-point Likert scale of influence (1) = 'Not at all influential' and (5) = 'Extremely influential'.

Dependent variable

(endogenous): The outcome (dependent) variable in the mediation analysis was the recreational travel motives of tourists, and these were measured based on ten items drawn from the literature (see Caber and Albayrak 2016; Chen and Chen 2015; Nafi and Ahmed 2018; Nikjoo and Ketabi 2015; Sangeeta and Anandkumar 2016). Respondents were asked to rate the attributes that would motivate them to consider visiting South Africa for recreational tourism. Responses were recorded on a five-point Likert scale of agreement between (1) 'Strongly disagree' and (5) 'Strongly agree'.

2.2. Data analysis

The Statistical Package for Social Sciences (SPSS 27.0) and IBM SPSS AMOS 27 software were used to analyse the data. A three-step process was applied to analyse the data. Step one was the Kaiser-Meyer-Olkin (KMO >.50) and Bartlett's Test of Sphericity (p< .000) to determine the factorability of the data (Hair et al. 2010), followed by Principal Components Analysis (PCA) and Exploratory Factor Analysis (EFA) to determine its dimensionality. Scale reliability of the factors extracted by the PCA/EFA was assessed based on the Cronbach Alpha (α>.60) statistic (Ursachi, Horodnic and Zait 2015). The approach to the EFA and CFA of the total sample before mediation analysis was in line with the procedures by Cheung, Keung and Tam 2022, He and Song 2009 and Wang, Chen and Lawler 2022. Step two was the Confirmatory Factor Analysis (CFA) employed to establish the reliability and validity of the factors established by the EFA (Chen, Huang and Hou 2020). The goodness of fit of the model was determined based on the parameters recommended by the literature (Gaskin and Lim 2016; Pavlov, Maydeu-Olivares and Shi 2021), Comparative Fit Index (CFI ≥ 0.95); Tucker Lewis Index (TLI ≥ 0.95); Root Mean Square Error of Approximation (RMSEA < 0.06 to 0.08); Standardised Root Mean Square Residual (SRMR <0.08); and (CMIN/DF - between 1 and 3). Reliability was determined by a Composite Reliability statistic of CR > 0.7. In contrast, convergent validity was determined by the Average Variance Extracted of AVE > 0.5, and discriminant validity was indicated by the square root of AVE greater than inter-construct correlations (Hair et al. 2010). Step three was direct effect testing based on linear and multiple regressions of the CFA's validated factors, followed by mediation analysis using Model 4 in PROCESS macro (v3.5) SPSS (Hayes 2013). Mediation analysis was the most suitable data analysis approach as it is a regression-based statistical analysis procedure that can model the effect of exogenous variables (CI) on endogenous variables (travel motives) while considering the potential effect of one or more underlying mechanisms as intervening explanatory variables (place brand dimensions). Thus, mediation analysis facilitates understanding of the interplay between South Africa’s image as a country, its place brand, and international tourists' recreational travel motives.

3.1 Respondent profile

Table 1 presents the demographic profile of the inbound tourists surveyed.

Table 1:

Demographic profile	Descriptive statistics
Gender	Male (55.1%); Female (44.9%)
Average age	Between 20 and 40 years old (57%)
Level of education	Bachelor’s (39.4%); post-graduate (22.6%)
Employment status	Employed in the private or public sector (79.4%)
Country of residence	UK (19.4%), Germany (15.4%) the USA (11.6%), Netherlands (7.1%)
Number of visits to South Africa	First-time visitors (75%)

As is summarised in Table 1, there was generally an almost even split between the genders. The surveyed tourists were generally young to middle-aged, educated, formally employed, resided in Europe, and primarily first-time visitors to South Africa. Notably, the sample profile is consistent with Statistics South Africa"s (2017) data on inbound tourist source markets.

3.2 Dimension reduction

The KMO (>.50) and Bartlett's Test of Sphericity (p< .000) statistics confirmed the adequacy of the sample and the factorability of the data (Hair et al. 2010). The PCA and EFA [Varimax with Kaizer Normalisation] reduced the data into discernible factors with eigenvalues (EV>1) (Hair et al. 2010). As it emerged (Table 2), the CI dimension comprised two factors, Stereotypical CI and Reverse Country of Origin Image, which accounted for 54% of the variance in the data. The EFA for the PB dimension extracted seven discernible factors, Socio-Cultural; Governance; Marketing; People; Negative Events, Competitive Advantages, and Tourism, which accounted for a cumulative 59% of the variance in the data. The recreational travel motives PCA/EFA extracted two recreational tourism factors, Nature- and Leisure-oriented Tourism, accounting for 60% of the variance in the data. All extracted factors were deemed reliable at α > .65 (Ursachi et al. 2015).

3.3 Validity and reliability of the scale

CFA established the validity and reliability of the priori measurement scale from the PCA/EFA (Chen et al. 2020). The CFA [maximum likelihood estimation method] reported acceptable goodness of fit statistics. The CI–travel motives model (χ² [120] =184,051; CFI=.957; TLI=.801; RMSEA=.048; SRMR= .058; CMIN/DF=1.534) and the PB dimensions’ model (χ² [471] =1135,185; CFI=.832; TLI=.946; RMSEA=.078; SRMR=.080; CMIN/DF=2.410) each showed a reasonable fit for the data (Hu and Bentler 1999), including the CFI value for the PB model, which is close to the minimum cut-off of 0.9 (Bentler 1990; Rizcky, Maulana and Rufaidah 2014). Notably, after the CFA (Table 2), CI emerged as a second-order construct for Stereotypical CI and Reverse Country of Origin Image. The CR statistics for all the constructs were >.70 indicating good internal consistency (Nunnally 1978). The AVE statistics were either above the recommended .50 thresholds (Hu and Bentler 1999) or were within acceptable parameters of AVE < .50, given that their associated CR was above the discretional threshold of .60 (Fornell and Larcker 1981; Lam 2012). The AVEs of the respective models indicated satisfactory convergent validity. The correlations between constructs in both models were statistically significant. The square roots of the AVEs for the constructs ranged between .569 and .937. They were greater than the inter-construct correlations (except in People and Socio-Cultural in the PB dimensions), indicating suitable discriminate validity (Hair et al. 2010).

Table 2:

		EFA							*CFA
Dimension	Item(s)	Factor loading		Mean value (x̄)	Eigenvalue (EV)	Variance (%)	Cronbach alpha (α)		Items	Std. Est		CR	AVE	√ AVE	Valid (Y/N)
		Min	Max							Min	Max
1Country Image												+.925	.878	.937	Y
Stereotypical CI	5	.544	.817	4.12	3.467	38.53	.748		5	.432	.794	+	+		-
Reverse Country of Origin Image	3	.648	.820	3.44	1.356	15.07	.683		3	.424	.825	+	+		-
2Place Brand
Socio-Cultural	9	.503	.828	3.34	12.297	27.95	.891		8	.621	.788	.886	.494	.703	Y
Governance	5	.591	.803	3.19	3.323	7.552	.834		5	.522	.831	.842	.522	.722	Y
Marketing	5	.539	.723	3.56	2.738	6.223	.856		5	.616	.814	.847	.528	.726	Y
People	6	.561	.809	3.51	2.142	5.482	.805		5	.581	.727	.796	.440	.663	Y
Negative Events	6	.504	.802	3.12	2.157	4.903	.854		6	.599	.815	.864	.518	.720	Y
Competitive Advantages	5	.524	.746	2.95	1.513	3.438	.755		4	.420	.759	.646	.324	.569	Y
Tourism	4	.593	.703	4.05	1.449	3.292	.692		-	-	-	-	-		N
3Travel motivation
Nature-oriented Tourism	5	.692	.822	4.23	1.669	16.690	.858		5	.532	.897	.860	.559	.748	Y
Leisure-oriented Tourism	5	.636	.775	3.85	4.282	42.820	.771		5	.428	.789	.730	.361	.601	Y

Notes: 1KMO = .815 and Bartlett"s test of Sphericity of (χ² (36) = 537.504, p = .000) * CFA: CR= Composite Reliability; AVE=Average Variance Extracted

2KMO = .853 and Bartlett’s test of Sphericity of (χ² (946) = 5793.513, p = .000)

3KMO = .799 and Bartlett’s test of Sphericity of (χ² (45) = 1036.523, p = .000)

+After the CFA Country Image emerged as a second order construct for Stereotypical CI and Reverse Country of Origin Image

3.4 Direct effect testing

In line with prior approaches (Kane and Ashbaugh 2017; Mascha et al. 2013), linear and multiple regressions were run to establish the direct path relationships – predicting path c [CI (X) on Travel Motives (Y1-2)], path a [CI (X) on PB dimensions (M1-6)] and path b [PB dimensions (M1-6) on Travel Motives (Y1-2)]. The absence of a direct effect on the respective path(s) would violate the hypothesised mediation relationship, rendering further analysis of the variables redundant (Mascha et al.  2013).

Table 3:

	Unstandardised Coeff		Stand. Coeff	t-value	Sig.	VIF	Tol.
	B	Std. Error	β
Country Image (X) – Nature Tourism (Y1)
R2=.051, F(1,231)13.409, p=.000
X – Y1: path c	.306	.084	.234	3.662	.000***	1.000	1.000
Country Image (X) – Leisure Tourism (Y2)
R2=.084, F(1,231)21.167, p=.000
X – Y2: path c	.374	.081	.290	4.601	.000***	1.000	1.000
Country Image (X) – Social-Cultural (M1)
R2=.071, F(1,231)18.769, p=.000
X – M1: path a	.419	.097	.274	4.336	.000***	1.000	1.000
Country Image (X) – Governance (M2)
R2=.021, F(1,231)5.937, p=.016
X – M2: path a	.280	.115	.158	2.437	.016*	1.000	1.000
Country Image (X) – Marketing (M3)
R2=.025, F(1,231)6.994, p=.009
X – M3: path a	.268	.101	.171	2.645	.009**	1.000	1.000
Country Image (X) – People (M4)
R2=.052, F(1,231)13.616, p=.000
X – M4: path a	.381	.103	.236	3.670	.000***	1.000	1.000
Country Image (X) – Negative Events (M5)
R2=.009, F(1,231)2.052, p=.153
X – M5: path a	.148	.103	.094	1.433	.153	1.000	1.000
Country Image (X) – Competitive Advantages (M6)
R2=.104, F(1,231)27.976, p=.000
X – M6: path a	.572	.108	.329	5.289	.000***	1.000	1.000
Place brand dimensions (M1-6) – Nature Tourism (Y1)
R2=.062, F(6,226)3.574, p=.002
M1 – Y1: path b	.070	.076	.082	.926	.355	.511	1.958
M2 – Y1: path b	-.093	.057	-.127	-1.648	.101	.684	1.462
M3 – Y1: path b	.115	.068	.137	1.686	.094	.605	1.654
M4 – Y1: path b	.140	.075	.174	1.882	.061	.475	2.107
M5 – Y1: path b	-.102	.070	-.123	-1.450	.148	.562	1.780
M6 – Y1: path b	.078	.056	.104	1.404	.162	.735	1.361
Place brand dimensions (M1-6) – Leisure Tourism (Y2)
R2=.176, F(6,226)8.043, p=.000
M1 – Y1: path b	.149	.071	.177	2.092	.038*	.511	1.958
M2 – Y1: path b	-.109	.053	-.180	-2.057	.041*	.684	1.462
M3 – Y1: path b	-.001	.064	-.001	-.010	.992	.605	1.654
M4 – Y1: path b	.091	.070	.113	1.295	.197	.475	2.107
M5 – Y1: path b	-.032	.016	-.040	-.491	.624	.562	1.780
M6 – Y1: path b	.251	.052	.339	4.807	.000***	.735	1.361

Statistically significant at *p < .05, **p < .01, ***p < .001

The direct effect testing (Table 3) revealed no statistical violations associated with the data (Field 2013; Hayes 2013; Kane and Ashbaugh 2017). Path c (X on Y1 and Y2) was confirmed in bot instances. Path a (X on M1-6) was also confirmed in all direct relationships except for the direct effect of CI on Negative Events (M5) [β = .094, p >.05]; hence eliminating the mediator from further analysis. No statistically significant direct effects could be established between all the PB dimensions (M1-6) and Nature Tourism as a recreational tourism motive for tourists to South Africa (Y1); thus eliminating Nature Tourism from the mediation analysis. Additionally, no statistically significant direct effect could be identified between the PB dimensions Marketing (M3: β = -.001, p >.05); People (M4: β = .113, p >.05) and Leisure Tourism. As a result of the direct effect tests, parallel mediation was viable, as illustrated in the conceptual framework (Figure 2), where X is CI; M1 is Socio-Cultural; M2 is Governance, M3 is Competitive Advantages; and Y is Leisure Tourism travel motivation. Subsequently the null hypothesis was reformulated and illustrated in Figure 2 as follows,  

H0:         South Africa's socio-cultural [H01], governance [H02], and competitive advantages [H03] place brand dimensions do not mediate the relationship between the country's image and the travel motivations of inbound recreational tourists. H1:         South Africa's socio-cultural [H11], governance [H12], and competitive advantages [H13] place brand dimensions mediate the relationship between the country's image and the travel motivations of inbound recreational tourists.

Latent Construct 2

Parallel mediation conceptual framework

3.5 Direct effect testing

Table 4 summarises the results of the parallel mediation of the relationship between country image and leisure tourism in South Africa through the country's PB dimensions.

Table 4:

Testing Path	B	SE	95% BootCI		β	t-value	Sig.
			Lower Limit (LL) CI	Upper Limit (UL) CI
Country Image – Leisure Tourism
Path c’: R2=.191, F(4,228)13.4388, p=.000
Leisure Tourism (DV) - Country Image (IV)	.207	.083	.043	.371	.161	2.492	.013*
Country Image – Socio-Cultural - Leisure Tourism
Path a1: R2=.075, F(1, 231)18.769, p=.000
Country Image (IV) - Socio-Cultural (DV)	.419	.097	.229	.610	.274	4.332	.000***
Path b1: R2=.191, F(4,228)13.4388, p=.000
Socio-Cultural (IV) - Leisure Tourism (DV)	.164	.057	.051	.277	.161	2.856	.005**
Effect: a1b1	.069	.039	.011	.163
Country Image – Governance - Leisure Tourism
Path a2: R2=.025, F(1, 231)5.9371, p=.017
Country Image (IV) – Governance (DV)	.280	.115	.054	.507	.158	2.437	.016*
Path b2: R2=.191, F(4,228)13.4388, p=.000
Governance (IV) - Leisure Tourism (DV)	-.109	.051	-.210	-.008	-.145	-2.126	.035*
Effect: a2b2	-.031	.022	-.083	.002
Country Image – Competitive Advantage - Leisure Tourism
Path a3: R2=.108, F(1, 231)27.9764, p=.000
Country Image (IV) - Competitive Advantage (DV)	.572	.108	.359	.785	.329	5.2893	.000***
Path b3: R2=.191, F(4,228)13.4388, p=.000
Competitive Advantage (IV) - Leisure Tourism (DV)	.225	.051	.124	.326	.303	4.3831	.000***
Effect: a3b3	.129	.039	.059	.212

Note: N=233, statistically significant at *p < .05, **p < .01, ***p < .001

In line with the recommendations from the literature (Hayes 2013; Preacher and Kelley 2011) unstandardized coefficients were reported for the mediation analyses. Table 4 summarises mediation effects [95% bias-corrected, confidence intervals (CI) based on 5 000 bootstrap samples] where CI had a positive and statistically significant direct effect on all the PB dimensions Socio-cultural: a1 = .419, p < .001; Governance: a2 = .280, p < .05; and Competitive Advantage: a3 = .572, p < .001. The PB dimensions reported (Figure 3) statistically significant positive Socio-Cultural: b1 = .164, p < .01; Competitive Advantage: b3 = .225, p < .001 and negative Governance b2 = -.109, p < .01 direct effects on Leisure Tourism.

Latent Construct 3

Parallel mediation of the country image – leisure tourism nexus through place brand dimensions

The mediation of the Country Image-Leisure Tourism travel motivation via Socio-cultural: effect a1b1 = .069 [95% bootstrap CI (LL = .011, UL = .163)] and Competitive Advantage:effect a3b3 = .129 [95% bootstrap CI (LL = .059, UL = .212)] is confirmed and the null hypotheses H01 and H03 could be rejected (Preacher and Hayes 2004; Tsaur, Yang, and Tsai 2021). Thus, the alternative hypotheses H1 and H3 were accepted. However, the null hypothesis for the mediation effect of Governance: effect a2b2 = -.031 [95% bootstrap CI (LL = -.083, UL = .002)] while negative and statistically significant could not be rejected, as the LL and ULpassed through zero; thus null hypothesis H02 was accepted (Kane and Ashbaugh 2017; Preacher and Hayes 2004; Tsaur et al. 2021). The effect of X on Y adjusted for M1-3 indicated partial mediation (c" = .207, p < .05).

4. DISCUSSION

The study explored South Africa's PB potential mediating effect in the CI – recreational tourism travel motivation nexus. While South Africa's CI directly influenced the leisure-oriented recreational travel motives of inbound tourists to South Africa, the country’s PB dimensions [socio-cultural and competitive advantage factors] partially intervened in the relationship. CI has a positive direct effect on travel motivation and PB dimensions, while South Africa’s PB dimensions have a predictive effect on travel motivation for recreational tourism. Some of the literature corroborates some of these direct relationships. Chaulagain et al. (2019) found that American citizens' intentions to visit Cuba as a tourist destination were influenced by the country's image. Furthermore, the influence of CI on tourist decision-making has been established in both developed and less developed countries (Campo Martínez and Alvarez 2010; Hahm et al.  2018; Zhang et al.  2018); thus supporting the findings in the case of an African tourism destination. Prior empirical evidence (Matiza and Slabbert 2020; Nadeau et al.  2008) also supports the influence of CI on PBs, whereby the CI can influence how PBs may be interpreted and framed as heuristic cues by tourists. For instance, in the case of China, Zhang et al. (2016) also conclude that CI affects how the country is perceived as a tourism destination and has a moderating effect on the conative behaviour of tourists. While Mussalam and Tajeddini (2016) corroborate the PB-travel motivation nexus, establishing that PB attributes influenced Switzerland's travel behaviour as a holiday destination. Moreover, the literature suggests that aspects of competitive advantages (see Kruger and Saayman 2010) and socio-cultural (see Park et al.  2019) have influenced recreational tourism-oriented travel motivation.  

Literature (Campo Martínez and Alvarez 2010; Hahm and Tasci 2020; Zhang et al.  2018) provides anecdotal evidence to corroborate the most significant finding of a CI-PB-travel motivation nexus in South Africa's recreational tourism. Zhang et al. (2018) establish the mediation effect of destination image and the influence of China's CI on how tourists evaluate the country as a potential tourism destination. In Turkey, Campo Martínez and Alvarez (2010) determined that CI and PB are not mutually exclusive constructs and may be correlated in their influence on tourists’ decision-making. Therefore, considering tourism as a country's product, the CI as a summative and inferential construct has a halo effect on the PB and generic evaluation of the tourism destination (Hahm and Tasci 2020). This halo effect can, directly and indirectly, influence tourists' travel motives. Hence the mediating effect of socio-cultural and competitive advantage as salient PB dimensions in the influence of South Africa's CI on the travel motivations of inbound recreational tourists to South Africa is both novel and rational.

5. CONCLUSIONS

It can be concluded that South Africa's CI mainly influences the extrinsic travel motives of inbound recreational tourists. However, this influence is partly mediated by the PB [comprising social-cultural factors and competitive advantages] associated with South Africa. Critically, the empirical evidence shows that increases in the influence of South Africa's CI correspond to an increase in the direct effect of CI on tourists’ travel motivation. Conversely, the findings imply that a CI deficit may have a corresponding adverse effect on the perceived attractiveness of a destination through its PB and tourist decision-making. The empirical results are particularly illuminating for tourism practitioners in African countries which are susceptible to subjective biases and stereotypical monikers such as Afro-pessimism. The findings provide the impetus for the CI and PB's dual management in tourism marketing to influence the extrinsic travel motivations of a significant segment of the global tourism market, recreational tourists. Significantly, the established influence of CI and PB on tourist travel motives indicates a paradigm shift in the antecedents of tourists’ conative behaviour, thereby positing CI and PB as potentially viable perceptual variables in the psychographic segmentation of international tourists. Whereby tourism practitioners can identify the potential heterogeneity in the influence of CI and PBs on tourists and utilise any discernible differences to establish homogenous segments to target specific recreational tourists with concerted marketing programs that will maximise return on investment tourism marketing dollars.

Moreover, considering the ongoing COVID-19 pandemic, the ability of tourism practitioners to differentiate between or assimilate CI and PB will be critical to the comprehensive management of tourism recovery in the short to medium term post the COVID-19 pandemic. For instance, while a country such as South Africa may be deemed to be unsafe by some tourists due to the perceptions associated with the prevalence of unique variants of the coronavirus (see Tegally et al. 2020), South Africa's attributes and competitive advantages including wide-open spaces in nature and along its coasts (see Kruger and Saayman 2010; Van Dyk et al. 2019) may be very appealing to tourists and motivate them to engage in tourism activity in the country, despite the heightened perceived risk that may be generically associated with the country. This implies that it may be prudent for tourism practitioners to further interrogate and harness the competitive PB dimensions of South Africa to intercede and mitigate the potentially adverse effects of the country's generic image. Hence the consolidation of CI and PBs post the pandemic will help tourism practitioners better manage destinations and predict tourist behaviour in conjunction with the attributes that motivate recreational tourists.

Three limitations are noted. First, while the empirical evidence has far-reaching theoretical and practical implications, the findings of the study can only be generalised within the South African context. Therefore, replication of the study is recommended on a broader multi-country scale to establish a more global perspective on PB's mediating effect in the nexus between CI and travel motivation in the recreational tourism context. Second, the study is cross-sectional; hence the findings are a snapshot of the period under investigation. However, this is a standard limitation of behavioural studies. In light of the COVID-19 pandemic and the potential importance of CI and PB in tourism recovery post-the-pandemic, further research is recommended to explore the possible paradigm shift in tourist behaviour due to the pandemic and how CI and PB may influence recreational tourism demand. Third, a unique measuring instrument was developed for the study. Hence the measuring instrument may have been subject to measurement misspecification [reflective versus formative] (Mikulić 2018; Mikulić and Ryan 2018). As a growing challenge in the social sciences, futurestudies may better specify the items and mitigate this challenge.