Development of a Destination Image Recovery Model for Enhancing the Performance of the Tourism Sector in the Developing World

Abstract

This chapter is based on a doctoral thesis on the development of a destination image (DI) recovery model for enhancing the performance of the tourism sector in Zimbabwe. The study was prompted by the failure of African destinations to develop DI image recovery models. A pragmatist paradigm, a convergent parallel mixed methodology research approach and a cross sectional survey were adopted. A sample of three hundred and nineteen comprising international tourists, service providers and key informants was used. A structured, semi-structured questionnaire and semi-structured interview guide were used respectively. Quantitative data was analyzed using the Statistical Package for Social Sciences (SPSS) and AMOS version 25 while qualitative data was analyzed using NVivo version 12. Tests were conducted using descriptive statistics, exploratory factor analysis, and confirmatory factor analysis. Structural Equation Modeling (SEM) was used to analyze the multiple independent variables. The major findings were that price, ancillary services and amenities significantly influenced affective image while ancillary services significantly influenced destination performance. The study recommended that the Ministry of Environment, Climate, Tourism and Hospitality Industry trains tourism stakeholders including the host community in order to achieve sustainable destination image recovery.

1. Introduction

Travel and tourism has become the world’s largest and fastest growing industry, and its growth shows a consistent year to year increase [1]. The sector contributes directly to 5% of the world’s GDP, one in 12 jobs globally, and is a major export sector for many countries, both in the developing and developed world [1]. The increase in global tourism numbers (1 billion in 2012) compared to 710 million in 2000 [2] has resulted in intense competition between destinations to grow their market shares. According to [3], international tourist arrivals rose by 6% in 2018 to hit the 1.4 billion mark from 1.3 billion in 2017. The [3’s] tourism forecast which was published in 2010 suggested that the 1.4 billion arrivals would be attained in 2020, yet the rapid tourism growth on the international scale has seen that target being attained two years ahead of time [4]. This growth has seen tourist destinations wrestle fiercely not only for the tourist’s expenditure but also for their voice and mind.

The intangibility of tourism products means that their image is the only way which potential tourists have of comparing destinations and choosing between them and therefore it is important to create and transmit favorable images to potential tourists in target markets [4]. As tourism services are intangible, images become more important than reality [5]. This makes the tourist’s perceptions of the product not only a fundamental component of the decision-making process, but also a key determinant of the performance of the tourist destination. The idea of DI was introduced into tourism studies in the early 1970s by [5, 6, 7] and Pike [4], and has since become one of the most researched topics in tourism-related research [8] due to its association with tourism performance [9]. However, there is a less marked mention of DI recovery and performance in literature, that is, the DI recovery-performance correlation has been marginalized. Some studies have focused on DI and tourist loyalty [10, 11]. Others have examined DI and technology (film, Internet and others) [11, 12, 13]. There is therefore an interstice in research on the DI recovery and performance nexus.

Globally, France, Switzerland, Austria, Germany, the United States, Spain and China continue to top the rankings in terms of both international arrivals and receipts in 2018 [14]. The sound performance of these tourist destinations tended to suggest a strong DI. The European continent continued to lead in terms of arrivals in 2017 (713 million, plus six percent) [15]. Africa’s weak image was generally attributed to political upheavals, disease, a poor infrastructure, poverty, and frequent droughts [16]. These factors negatively impacted the economies of the African destinations and specifically, the tourism economies. [16] identified Africa’s ‘unfortunate’ image as an obstacle to the region’s competitiveness in the global tourism market, ascertaining that there is overwhelming evidence to suggest that Africa faces a huge challenge in counteracting the continent’s prolonged negative image and perceived risks as a tourist destination. This was part of the reason why in 2018 Africa attracted only five percent of the international overnight visitors, accounting for 67 million international tourists [3] against a global total of 1.403 billion international tourists. Tourist arrivals in sub-Saharan Africa grew by 6% with the island destinations, namely Cabo Verde, Reunion and Mauritius registering strong growth [3]. What seems to be emerging from the above discussion is that many African destinations, including Zimbabwe, are faced with a challenge of a weak DI and an equally weak performance of the tourism sector. There is a dearth of research on destination performance [17]. Scholars who have explored tourism performance include [17, 18, 19, 20, 21, 22, 23]. However, these researchers did not explore DI and the performance of the tourism sector jointly. This study sought to fill this interstice.

The [14] has made efforts to react to the tarnished image through various promotions focusing on rebranding as an exceptional ingredient in order to give the country’s tourism a facelift [24]. This has seen Zimbabwe as a tourist destination rebranding three times between 1980 and 2011 [24]. However, it appears that the negative image has remained in place, well after the hosting of the highly touted UNWTO General Assembly. Zimbabwe’s tourism arrivals and receipts indicate that the sector has been on an unstable path in the last decades, with fluctuating performances in tandem with the deteriorating local economic conditions and the global economic crisis in 2008 exacerbated by the global economic crisis/credit crisis which affected mostly developed world tourism markets and led to many traditional tourists cutting back on their travel and leisure expenditure [25]. In Zimbabwe, tourism is one of the four pillars anchoring economic growth after Agriculture, Mining and Manufacturing [26].

1.1 Statement of the problem

Zimbabwe is grappling with a negative tourist DI and a decline in the performance of the tourism sector. Despite several studies, for example [23, 27, 28, 29] which have been carried out to improve the performance of the country’s tourism sector, image and performance remain problematic. Zimbabwe’s travel and tourism competitiveness index (ranking) has not been impressive. In 2015, Zimbabwe was ranked 115 out of 141 tourist destinations across the world and an equally low 114 out of 136 destination in 2017 [30]. In terms of prioritization of travel and tourism, Zimbabwe was at 105 out of 136 while it scored a very low 134 out of 136 destinations for its business environment in 2017 [30]. In terms of international arrivals, the 2011 figure of 2423 20 was marginally better than the 2017 figure of 2,422,930 [14]. This suggested a lack of tangible growth in terms of arrivals. The country’s image in the source markets is still associated with political instability, policy inconsistency, and disease outbreaks [30]. There is a strong market perception that the destination is not price competitive and that the overall product is tired [14]. In [25], it is noted that [14] has been promoting tourism through beauty pageants, carnivals and sporting events such as soccer tournaments. The ZTA website has also served as a promotional tool [14]. However, as indicated by [26], Zimbabwe is still failing to gain its previous position as a destination of choice. Furthermore, the goal of a middle income economy for Zimbabwe by 2030 may remain a pipe dream unless there is an improvement in the economic, social and political environments [26]. Although Zimbabwe’s tourist figures have increased here and there since 2008, as a destination, it is still struggling to restore itself to its former glory as a competitive force in southern Africa. Zimbabwe’s negative perception hinders its visibility in the international markets as a tourist destination which in turn is reflected in weak demand among international tour operators and travel agencies [4]. Negative perceptions of tourist destinations lead to the poor performance of the industry [31]. The highest number of tourists the country has received (2579974) in 2018 almost equals that of 2007 (2505988), that is, twelve years ago. Unless this problem of a weak image is resolved, Zimbabwe’s negative perception in the source markets will remain, and the performance of the tourism sector will remain depressed resulting in a low tourism multiplier effect. The study will benefit tourism and hospitality stakeholders such as tourists, the Zimbabwe Tourism Authority, tourism and hospitality researchers, planners, policy formulators, tourism and hospitality business operators and local communities.

1.2 Research objectives

The major objective of the study was to develop a destination image recovery model to enhance tourism performance in Zimbabwe. The specific objectives of the study were to assess the current situation with regards to destination image and performance of the Tourism sector in Zimbabwe, examine the determinants of destination image and performance of the tourism sector in Zimbabwe, investigate the extent to which destination image affects performance of the tourism sector in Zimbabwe and develop a destination image recovery model for enhancing performance of the Tourism sector in Zimbabwe.

1.3 Research hypotheses

H1: Price is significantly positively related to affective image.

H2: There is a significant and positive relationship between amenities and affective image.

H3: Ancillary services have a significant relationship with affective image.

H4: Accessibility has a significant positive influence on affective image.

H5: Price significantly influences performance.

H6: Amenities significantly influence performance.

H7: Ancillary services significantly influence performance.

H8: Accessibility significantly influences performance.

2. Literature review

Literature review entailed describing, comparing, contrasting and evaluating the major theories, arguments, themes, approaches and controversies in the scholarly literature on DI and performance of the tourism sector. Literature review was also conducted to identify gaps in literature with regards to DI and performance of the tourism sector. This was done in order to fill such gap or gaps with new knowledge thereby contributing towards extending the frontiers of knowledge in terms of DI recovery and tourism performance.

2.1 Theoretical framework

The study on which this chapter is based was premised on the stakeholder theory [32] and the [33] tourism performance model. Its tenets are that organizations depend on a wide range of audiences or groups of stakeholders in order to realize their objectives [34]. Modern life and tourism in particular is affected by a wide range of variables which include technology, social dimensions, political developments, environmental factors and others. The stakeholder groups cited in this theory include clients, end users (the other theories above do not make this distinction), employees, suppliers, pressure groups, local communities and the media and each stakeholder makes a decisive role in the organization’s future. This theory is currently popularly used in tourism development and in destination image recovery and in the enhancement of tourism performance.

According to [33], there are eight drivers of tourism performance which are indicated here in their order of importance: (1) tourism and related infrastructure; (2) economic conditions; (3) security, safety, and health; (4) tourism price competitiveness; (5) government policies; (6) environmental sustainability; (7) labor skills and training; and (8) natural and cultural resources. This theory links quite well with destination image recovery in that it focuses on attributes which are central to destination image recovery. The [30] noted that these destination attributes are important in generating a destination’s appeal. However, the limitations of this theory are that it assumes that any factor outside these eight may not as critical as those included on this list.

2.2 Perceptual images of a tourist destination

Most frequently, the concept of DI has been operationalized as consisting of two components: A perceptual-cognitive component that captures knowledge and beliefs about a destination’s attributes and an affective component that describes feelings towards a destination [35]. The cognitive component of the image refers to a person’s beliefs and knowledge about a destination and its attributes, which together help to form an internally accepted mental picture of the place [36]. It also includes a set of attributes that mainly correspond to the resources of a tourist destination [37, 38]. Those resource attributes generally involve the natural environment (scenic beauty, weather, beaches); Amenities (hotels, restaurants, service quality, shops); Attractions (water sports, well - known attractions, a variety of tourist activities); Accessibility (convenient transportation, developed infrastructure, ease of access, Social Environment (personal safety - security, friendly local people, good value for money, a clean environment) [37]. All these can induce an individual to visit a specific destination. The affective component refers to the evaluation stage, concerning the feelings that the individual associates with the place of visit [38]. The affective component generally covers a number of categories: distressing -relaxing, unpleasant-pleasant, boring-exciting, sleepy-lively [39]. The destination should conjure the right emotions in the potential visitor for it to earn a visit [40]. The Conative component (behavioral intention) of DI has been considered by several researchers in DI formation [31, 41, 42]. For these researchers conation is part of the image formation process which is “analogous to behavior evolving from cognitive and affective images” [43] denoting the “intent or action component” [44]. Understanding tourists’ intention or the likelihood of visiting a destination is crucial for destination marketing managers.

Destination image comprises functional characteristics, psychological characteristics, common and unique dimensions [45]. Common psychological attributes refer to the friendliness of the locals or beauty of the landscape, whereas unique psychological factors include feelings associated with places of religious pilgrimage or some historic event. [46] indicates that functional characteristics can be easily measured while psychological characteristics, on the contrary, cannot be easily measured. However, together they influence the formation of DI explaining why the use of mixed methods in DI studies has gained prominence [46, 47, 48].

2.3 Determinants of destination image and performance of the tourism sector

The determinants of DI include natural resources; general infrastructure; tourist infrastructure; tourist and leisure recreation; culture, history and art; political and economic factors; the social environment and the atmosphere of the place [49]. However, this view tends to marginalize the role of the tourist’s reasoned and emotional interpretation in DI formation. Most studies [50, 51] tend to consider image to be a concept formed by the consumer’s reasoned and emotional interpretation as the consequence of two closely inter-related concepts: perceptive/cognitive evaluations referring to an individual’s own knowledge and beliefs about the object, and affective appraisals related to the individual’s feelings towards the object. The combination of these two factors produces an overall, or compound, image related to the positive or negative evaluation of the product or brand [52].

2.4 Demand (tourist) factors as determinants of DI and performance

The demand side of determinants of DI and performance relates to issues which pertain to tourists’ socio-demographic factors [53], tourists’ nationality [52, 54] tourists’ level of awareness or familiarity with a particular destination [55]. In [56], internal factors influencing the image construct include socio-demographic factors. Specifically, the social and cultural environment relate to socio-demographic aspects of a human being [57]. It is postulated that today’s tourists play a leading role in image projection [58]. They have become an active agent who use Web 2.0 tools to disclose their opinion, experiences and feelings about the destination visited [59]. Figure 1 depicts factors which influence DI recovery.

Figure 1.

Factors influencing DI recovery. Source: Adapted from Harahsheh (2009, p. 78).

2.5 Supply factors as determinants of destination image and performance

According to [60], the natural resources are the main attraction of the tourism destination. Thus, they influence destination perception and performance [33]. Scenery for example, constitutes one of the dimensions used by researchers to measure DI [61]. The natural environment is one of the three dimensions of DI and performance [33]. The first dimension comprises the socio-cultural amenities such as wonderful cultural traditions, interesting local arts and interesting cultural diversity. Second, natural amenities such as: beautiful mountains, outstanding natural wonders, wonderful sightseeing opportunities, and appealing opportunities for exploring wilderness and nature. And third, climate attributes interrelation: appealing winter climate, appealing summer climate. Nature tends to feature prominently in the classifications of the destination-image management dimensions by different scholars. In [62], for example, came up with nine dimensions/attributes that determine the perceived DI of an individual. These include natural resources such as weather and its variations, beaches and their variations, wealth of countryside such as protected nature reserves and variety and uniqueness of flora and fona.

2.6 Effect of destination image on performance of the tourism sector

Literature does not directly point out the direct effect which DI has on performance of the tourism sector. It spells out the effect of DI on value, satisfaction and loyalty of the tourists [63, 64] and not that of DI and destination performance. However, given that a direct relationship between DI, satisfaction and revisit intentions [65], it follows that there is a relationship between DI and destination performance but it appears that it is more of a derived effect than a direct one. In [64], it is found out that the tourism image is a direct antecedent of perceived quality, satisfaction, intention to return, and willingness to recommend the destination. In [53], it is conducted a study on DI, perceived value, tourist satisfaction and loyalty focusing on Mauritius. They reported that their results supported the proposed destination loyalty model, which advocated that DI directly influenced attribute satisfaction; DI and attribute satisfaction were both direct antecedents of overall satisfaction; and overall satisfaction and attribute satisfaction in turn had direct and positive impact on destination loyalty. The implication of these relationships seems to that it is important to develop positive images of a tourist destination in order to increase the number of tourists and tourist receipts [53].

2.7 Strategies to improve destination image

In [16], it is proposed two broad categories of DI recovery: the cosmetic and strategic approaches both of which emphasize the role of the media. Media strategies in the cosmetic approach try to change the destination’s image without really changing the reality behind it; the destination’s problems are not solved or managed but the local decision-makers try to potray it in a positive light, by using advertising or public relations campaigns [66]. Strategies within this category include ignoring the image crisis problem, disassociation from the problematic location, association with prestige locations, acknowledging a negative DI, delivering a counter-message to the negative stereotype, spinning the negative characteristic to positive and ridiculing the stereotype [16]. However, these strategies are associated with a low level of change in the destination’s characteristics including its performance while those which use the strategic approach tend to be associated with a high level of change (Figure 2).

Figure 2.

The strategic vs. cosmetic approach for altering prolonged negative images. Source: Avraham and Ketter (2013).

Destinations host major events to attract visitors, gain positive attention from the media and improve their image [16]. Zimbabwe has used special events which include hallmark and mega events to improve DI and ultimately destination performance. Special events describe specific rituals, presentations or anniversaries specifically planned or designed to mark a specific occasion, cultural or organizational goals [67]. Special events can include national days and celebrations, important civic occasions, unique cultural performances, major sporting fixtures, corporate functions, trade promotions and product launches [68, 69]. However, it appears that very little has been achieved by way of improving image and performance of Zimbabwe as a tourist destination.

2.8 Conceptual framework

The conceptual framework depicts the study hypotheses, that the components of the cognitive image, in this case price, amenities, accessibility and ancillary services impact both affective image and destination performance (Figure 3). Affective image which derives from the potential tourists’ feelings towards the destination contributes to the improvement in overall destination image especially after visiting the destination [70]. This will ultimately result in; enhancing the performance of the tourism sector (destination performance) as the tourists spend money in the destination.

Figure 3.

Conceptual framework. Source: Author’s compilation (2018).

3. Research methodology

This study adopted the pragmatic research philosophy. This philosophy is a position that contends that the research question is the most important determinant of the research philosophy adopted for the study [71]. A mixed research methodology was used. Literature shows that the use of qualitative and quantitative research methodologies separately would not yield the best results for this study. Mixing quantitative and qualitative approaches is increasingly popular in DI research, although it appears that there is insufficient theoretical rationale for doing so [72]. However, [73] highlighted that the mixed methodology adds value in terms of increasing confidence in the research findings.

3.1 Research design

The research design was more of quantitative than qualitative. The large amounts of quantitative data came from the tourists who far outnumbered the service providers and key informants who together provided mostly qualitative data. Research designs can be classified into three broad categories, namely quantitative, qualitative and mixed methods research designs [74]. Creswell [74] describes this design as concurrent procedures, in which the researcher converges quantitative and qualitative data in order to provide a comprehensive analysis of the research problem. This view is corroborated by [75] who highlighted that the purpose of doing this is to best understand or develop a more complete understanding of the research problem by obtaining different but complementary data. Both forms of data are collected simultaneously and the information is integrated in the interpretation of the overall results. Data analysis is kept independent and there is need to look for convergence, divergence, contradictions, or relationships of the two sources of data [76]. The convergent parallel mixed methods design supported the research requirements. It was the most appropriate research design in that allows for the collection and analysis of both qualitative and quantitative data separately [76]. This was consistent with what the research sought, that is, to collect qualitative data from service providers in the tourism sector and quantitative data from tourists. These two groups were mutually exclusive thus facilitating the independent collection and analysis of data. This would be followed by a comparison of the results to see if the results confirmed or disconfirmed each other. The result would be used to develop a DI recovery model which enhances and improves tourism performance in Zimbabwe.

A structured questionnaire was used to collect data from international tourists and a semi-structured one was applied on the service providers. For both tourists and service providers, the closed questions included a Five-point Likert scale: (For tourists) 5-Very Good, 4-Good, 3-Unsure, 2-Poor and 1-Very Poor and another one, 5-Very Important, 4-Important, 3-Unsure, 2-Somewhat Important and 1-Not Important. For service providers: 5-Strongly Agree, 4-Agree, 3-Neutral, 2-Disagree and 1-Strongly Disagree. Semi-structured interviews were conducted on key informants. Semi-structured interview is a term that typically refers to a context in which the interviewer has a series of questions that are in the general form of an interview guide but is able to vary the sequence of the questions [77]. This type of interview is used to find out what is happening, it seeks new insights, identifies general patterns and helps to understand the relationship between variables [78]. It uses a combination of open and close-ended questions and hence it is consistent with the mixed methodology research design. Data was collected from 240 international tourists, 62 service providers and 17 key informants. Figure 4 shows how the study was conducted. Price, amenities, accessibility and ancillary services are some of the determinants of cognitive image.

Figure 4.

Research design adopted in the study. Source: Author’s compilation (2018).

3.2 Data analysis

Descriptive analysis was applied on demographic data and on interval-scaled (Likert scale) data. Frequency table analysis and proportion percentage analysis was used to transform raw data into a form that would facilitate easy understanding and interpretation. Descriptive statistics were thus used to analyze and profile the perceptions (attitudes) and future intentions of the sampled international tourists. Quantitative data which was obtained from the tourists and some from service providers was analyzed using the Statistical Package for Social Sciences (SPSS) and AMOS version 25. The study used inferential statistics in order to analyze the multiple independent variables. Structural Equation Modeling (SEM) was used to analyze the multiple independent variables which included accessibility, amenities ancillary services and prices as well as dependent variables such as affective image and performance. Kaiser-Meyer-Olkin Measure of Sampling Adequacy (KMO) and Bartlett’s test were used to test data for validity. Cronbach’s Alpha was used to test for data reliability. Overall, the Alpha value was 0.7 and above. SEM is one of the most often used statistical techniques used by researchers to test complex models which involve a number of dependent and independent variables [79]. Similar studies have used SEM [35, 46, 53]. Multivariate analysis was used to test hypotheses because it is optimal for analyzing multiple relationships [75]. Factor analysis was applied on the thirty eight destination image attributes which tourists rated on a Likert scale. These destination attributes were classified into constructs which included price, accessibility, amenities, ancillary services, affective image and performance. This was done in order to facilitate data analysis. Quantitative data was presented using tables and graphs.

Documentary analysis was adopted to establish the trends which were emerging from international aircraft and passenger movements provided by CAAZ from 2016 and 2018 and the international arrivals provided by ZTA during the same period. Data was first captured on a template before it was cleaned (edited). Qualitative data from key informants and service providers was analyzed using NVivo version 12, thematic coding. The goal was to identify, analyze and describe patterns, or themes, across a data set [77]. Word cloud, Word tree, Hierarchical charts, Word popularity (Word frequencies) including Word query were used to analyze data. Thematic analysis was used because it allows for the classification into themes of various and divergent views from the respondents and it is a flexible research technique which is not tied to a specific philosophical orientation [77]. This approach was therefore quite consistent with the pragmatism research philosophy which informed this study. Results which emerged from the qualitative analysis were compared and contrasted with those which were obtained from quantitative analysis.

3.3 Ethical considerations

Respecting the respondent’s rights, needs, values and desires is emphasized when collecting research data (Creswell, 2014). Research which includes human input should ensure that they are well informed and consent sought from relevant authorities. Permission was sought from international tourists, service providers and key informants to carry out research. Various organizations which included the ZTA, CAAZ, Zimbabwe Parks and Wildlife Management Authority, the Ministry of Tourism and the Ministry of Tourism and Hospitality Industry issued letters to the researcher granting him permission to conduct the research at Robert Gabriel Mugabe International Airport. Research assistants were trained to ensure that they behaved ethically as they went about administering research instruments.

4. Results and discussion

4.1 Response rate

Response rate refers to the total number of responses divided by the total number in the sample after ineligible respondents have been excluded [78]. A total of 397 respondents comprising 293 international tourists, 90 service providers and 17 key informants was targeted. However, the actual tally was 319 giving a response rate of 80% which was quite commendable [78]. This total of respondents consisted of 240 international tourists, 62 tourism and hospitality service providers and 17 key informants (Table 1).

Narration	Targeted respondents	Actual respondents	Response percentage	rate
International tourists	293	240	82
Service providers	90	62	69
Key informants	17	17	100
Total	400	319	80

Table 1.

Response rate.

Source: Field Survey (2018).

4.2 Demographic characteristics of respondents

In a study sample of 319, fifty three percent were males while forty-seven were females. The slight dominance of males could be due to the fact that men traveled more for tourism than their female counterparts and feel more motivated to meet their need for sport and adventure experiences than females [80]. The [80] further noted that there were more men than women in the business world and a lot of business travel occurs across the world and as a result, men tended to travel more than females.

The variation suggests that gender can influence perceptions of destination’s appeal. This is in line with [81] who asserted that females tended to engage in long-haul travel more than their male counterparts. Respondents aged between 25 and 35 years old formed the largest group (25.2%) followed by those aged between 35 and 44 years old (18.1), 17.2% of the respondents were in age group 45–54 years old, 15.1% of the respondents were in age group 55–65 years old, age group 66 or older constituted 13.4% of respondents, and age group 18-24 years old were 10.9%. The results showed that most of the tourists ranged from young to middle aged. The study findings resonate with those by [82]. A Visitor Exit Survey which was conducted at Zimbabwe’s ports of entry by [82] revealed that the majority of visitors to Zimbabwe were young (35–39) years (16.4%) and middle-aged (40–49) years (13.9%). A study which was conducted in Egypt by [83] focusing on cultural dimensions, demographics, and information sources as antecedents to cognitive and affective DI found out that tourists in the age ranges 26–35 and 36–50 were more likely to use the Internet, while younger (aged 18–25) and older (51–65) were less likely to use it. This finding in terms of age was similar to that of tourists in that most of the international tourists were fairly young. This may create a scenario whereby the young tourists are served by young service providers. This can help to create telepathy and rapport between the tourist and the server. This may enhance both employee and customer satisfaction leading to improved firm and destination performance. These results show that most of the respondents were well educated indicating that their responses were given from a position of enlightenment and knowledge.

Most of the tourists (37.5%) received an income of US$50000 and more before tax per annum followed by those who were earning between US$10001 and US$20000 (17.9%), and those who earned between US$20001 and US$30000 (11.9%), those who earned between US$30001 and US$40000 (10%) and those who earned US$ 40,001 to US$50000 (10%). There is limited research which has directly examined the relationship between destination attractiveness and income of the tourists. In [84], it is noted that in a study conducted in Taiwan, it was found that income was an influencer of tourist behavior. Tourists with a higher income tended to travel internationally more and were likely to stay in luxury hotels. On the other hand, travelers with less income tended to be associated with domestic trips rather than international vacations. In that regard, income was found to be an important determinant of destination choice [84].

4.3 Reliability analysis

Reliability analysis is used to determine the extent of internal consistency that is represented by a set of items in a construct [85]. For this study, reliability analysis was used to determine the extent to which the items within each and every construct were consistent. According to [86], the optimal minimum alpha statistic is 0.7. However, other scholars such as [87] argue that even alpha statistics of 0.6 are still reliable. The reliability tests for each and every construct will be presented.

The Cronbach’s Alpha statistic was 0.900, and being greater than 0.7, it follows that the construct price was internally consistent and reliable (Table 2). Further, assessing the corrected item-total correlation, none of the items had a coefficient less than 0.3 as recommended by [88] and this means that all the items extracted using PCA were reliable. For affective image, the Cronbach’s alpha statistic was 0.881. This was greater than the threshold of 0.7, and thus validates that affective image was internally consistent. On the other hand, none of the items had a corrected item-total correlation that was less than 0.3. Effectively, this meant that all the items were internally consistent. The Cronbach’s alpha for amenities was computed to be 0.842 and this was greater than 0.7. These results validate that the construct amenities were reliable. Regarding the corrected item to total correlation, the minimum observed was 0.533. This again, does fall below the 0.3 threshold set by scholars. In this regard, the researcher confirmed that amenities as a construct was reliable. With respect to ancillary services, the construct was internally consistent since the alpha statistic was 0.759, which is greater than the minimum expected 0.7. With respect to the corrected item-total correlation, the minimum was 0.431 and being greater than 0.3, none of the items were to be dropped.

		Scale mean if item deleted	Scale variance if item deleted	Corrected item-total correlation	Cronbach’s alpha if item deleted
Price	0.900
Lodging Prices		12.42	6.011	0.582	0.938
Prices of Restaurant Food		12.53	4.969	0.892	0.826
Prices of Restaurant Beverages		12.51	5.004	0.883	0.830
Prices of Goods and Services		12.41	5.542	0.767	0.874
Affective Image	0.881
Destination’s capacity to Relieve Stress		16.22	10.162	0.729	0.852
Destination’s Capacity to Provide Relaxation		16.16	10.223	0.793	0.838
Destination as a Pleasant Place		16.04	11.278	0.682	0.865
Destination as an Arousing Place		16.46	9.779	0.696	0.862
Destination as a Provider of Excitement		16.41	9.985	0.703	0.859
Amenities	0.842
Conference Facilities		11.08	7.261	0.726	0.776
Facilities for Young Children		10.88	7.433	0.729	0.775
Facilities for People living with Disabilities		10.65	7.841	0.724	0.779
Shopping Facilities		10.67	8.766	0.533	0.857
Ancillary Services	0.759
Cleanliness		16.80	6.030	0.569	0.702
Tourist Information		16.97	5.629	0.602	0.688
Quietness		17.15	5.580	0.548	0.709
Friendliness of Local People		16.72	6.631	0.514	0.725
ICT Readiness		17.14	6.100	0.431	0.753
Accessibility	0.801
Zimbabwe’s Accessibility as a Destination		12.17	5.384	0.524	0.793
Infrastructure at the entry point		12.29	4.295	0.736	0.685
Service at Immigration		12.20	4.679	0.685	0.714
Accessibility Destinations		12.00	5.926	0.533	0.789
Value	0.854
Value as a Vacation Destination		7.24	2.248	0.633	0.612
Value as a Business Destination		7.44	2.329	0.553	0.711
Overall Quality of the Destination		6.97	2.615	0.571	0.688
Attractions	0.636
Natural Landscape		4.32	0.591	0.472	.
Climate		4.48	0.433	0.472	.

Table 2.

Reliability analysis.

From the results above, the Cronbach’s alpha for accessibility was 0.801 and being greater than 0.7, it follows, therefore, that the construct was internally consistent and reliable. With respect to the corrected item-total correlation, the lowest observed was 0.524 and because this was greater than the minimum 0.3, the researcher confirms that all the items making up the construct.

accessibility were reliable. The next construct that was tested was value/performance. The corresponding Cronbach’s alpha for value/performance was 0.754 and being greater than 0.7, we can confirm that the construct was reliable and internally consistent. With respect to the corrected item-total correlation coefficient, the lowest observed was 0.553 and being greater than 0.3, it followed that all the items were very reliable. The construct attractions had a Cronbach alpha statistic of 0.636, this was less than the expected minimum of 0.7 and effectively, this meant that the construct was not so reliable. This is, however, despite that the corrected item-total correlation coefficients were greater than 0.3. Overall, from the reliability analysis, it was confirmed that the reliable constructs were: Price, Affective Image, Amenities, Ancillary Services, Accessibility and Value/Performance.

4.4 KMO and Bartlett’s test-destination image recovery and performance

In order to ensure that the conditions for the use of factor analysis were satisfied, [89], argue that the Kaiser-Meyer-Olkin Measure of Sampling Adequacy (KMO) and Bartlett’s tests ought to be tested. With respect to the KMO test, which measures the adequacy of the sample, the lower expected threshold should be 0.5, with higher values being more desirable [90]. With respect to the Bartlett’s test, which is a measure of multivariate normality, the p-value ought to be significant at p < 0.05 [88]. These tests were computed and the results are summarized in Table 3.

Kaiser-Meyer-Olkin measure of sampling adequacy	.843
Bartlett’s test of sphericity Approx. Chi-Square	4170.258
Df	703
Sig.	.000

Table 3.

KMO and Bartlett’s test.

Source: Data Survey.

The results above show that the KMO statistic was 0.843, and being greater than the minimum 0.5, it follows that the sample adequacy condition was satisfied. On the other hand, the Bartlett’s test was significant at p < 0.01 and this confirms the assumption of multivariate normality was met. The results above show that the KMO statistic was 0.843, and being greater than the minimum 0.5, it follows that the sample adequacy condition was satisfied. On the other hand, the Bartlett’s test was significant at p < 0.01 and this confirms the assumption of multivariate normality was met.

4.5 Modeling process

The modeling process below looked at price, amenities, conducive environment, affective image, accessibility and performance. The research instrument comprised of 38 items that measured the determinants of destination image recovery and performance of the tourism sector in Zimbabwe. With a view to establishing the principal factors behind these determinants of destination image recovery and destination performance, [91] recommend the use of exploratory factor analysis (EFA) dimension reduction techniques. According to [92], these dimension reduction techniques help in the classification of items that share a common underlying structure into a set of similar items collectively known as components [88]. One of the major dimension reduction methods recommended by scholars is factor analysis and this was considered in this study to be the optimal dimensionality reduction technique as prescribed by [93]. To achieve this dimensionality reduction, the principal component analysis (PCA) was used as the factor analysis component extraction method.

4.5.1 Factor extraction

Because the normality assumption was met, the Principal Component Analysis (PCA) was used in this study as the component extraction method, instead of the principal axis factoring, which works best when the normality assumption is not met [92]. With a view to simplifying the factors extracted, rotation was used. The components were assumed to be uncorrelated and o this effect, orthogonal rotation was done instead of oblimin rotation [90]. For the orthogonal rotation, Varimax was selected and this was done with Kaiser Normalization as prescribed by [89].

4.5.2 Communalities matrix

Having run PCA, the communalities that emerged are presented in Table 4. Generally, the communalities inform us on the extent of correlation between one item and the rest of the other items [88]. The higher the common variance, the higher is the validity of the item, and [85] recommend communalities to be at least 0.5 in magnitude.

	Initial	Extraction
Zimbabwe’s Accessibility as a Destination	1.000	.606
Infrastructure at the Country’s Immigration (entry point used)	1.000	.809
Service at Immigration (entry point used)	1.000	.717
Accessibility of Tourist Destinations within Zimbabwe	1.000	.697
Road Condition	1.000	.600
Inland Transportation/Taxi/Bus	1.000	.526
Natural Landscape	1.000	.730
Climate	1.000	.506
Tourist Attractions	1.000	.653
Opportunities for Learning Ethnic Customs	1.000	.751
Local Cuisine	1.000	.600
Outdoor Activities	1.000	.520
Cleanliness	1.000	.639
Tourist Information	1.000	.665
Quietness (Noise Pollution)	1.000	.565
Friendliness of Local People	1.000	.649
Nightlife/Entertainment	1.000	.644
Attitude of Service Personnel	1.000	.685
Safety and Security	1.000	.476
ICT Readiness	1.000	.624
Conference Facilities	1.000	.660
Facilities for Young Children	1.000	.800
Facilities for People living with Disabilities	1.000	.795
Shopping Facilities	1.000	.617
Lodging Facilities	1.000	.726
Restaurants	1.000	.721
Lodging Prices	1.000	.757
Prices of Restaurant Food	1.000	.864
Prices of Restaurant Beverages	1.000	.863
Prices of Goods and Services	1.000	.735
Destination’s capacity to Relieve Stress	1.000	.753
Destination’s Capacity to Provide Relaxation	1.000	.798
Destination as a Pleasant Place	1.000	.724
Destination as an Arousing Place	1.000	.712
Destination as a Provider of Excitement	1.000	.734
Value as a Vacation Destination	1.000	.751
Value as a Business Destination	1.000	.689
Overall Quality of the Destination	1.000	.727
Extraction Method: Principal Component Analysis.

Table 4.

Communalities-destination image recovery and performance.

Source: Data Survey (2018).

From the results, only one item had a communality that was less than 0.5 and this was safety and security and the respective correlation coefficient was 0.476. Effectively, this was discarded off from the results. The rest of the other coefficients were considered to be significant for accurate factor extraction, with the highest communalities being 0.864 and 0.863 for prices of restaurant food and prices of restaurant beverages respectively. The resultant model is presented in Figure 5 below.

Figure 5.

Structural equation model. Source: Data survey (2018).

The corresponding table with the detailed results is presented in Table 5 below.

			Estimate	S.E.	C.R.	P	Standardized
AF	<−--	PR	.237	.051	4.681	.000	.320
VA	<−--	PR	−.114	.065	−1.750	.080	−.128
AF	<−--	AM	.089	.044	1.995	.046	.135
VA	<−--	AM	.072	.062	1.173	.241	.091
AF	<−--	AC	.025	.046	.543	.587	.036
VA	<−--	AC	−.070	.066	−1.071	.284	−.083
AF	<−--	AN	.586	.146	4.003	.000	.345
VA	<−--	AN	.356	.172	2.066	.039	.175

Table 5.

Structural equation model - regression weights.

Source: Data Survey (2018).

From the results above, the strongest relationship was found to exist between ancillary and affective image, whose standardized coefficient was 0.345 and this was seconded by price and affective image, with a standardized coefficient of 0.320. The p-value was less than 0.05 for the relationship between Price and Affective Image (p < 0.01), amenities and affective image (p < 0.05), ancillary services and affective image (p < 0.01) as well as ancillary and value. It should be noted that only one of the four hypotheses linking performance was significant. The conclusions to the research hypotheses are indicated below:

4.5.3 Dependent variable: Affective image

With respect to the dependent variable, affective image, the key hypothesis decisions are summarized below:

H₁: Price is significantly positively related to affective image.

SIGNIFICANT (CR = 4.681; p = 0.000 < 0.05).

The hypothesis is therefore accepted.

H₂: There is a significant and positive relationship between amenities and affective image.

SIGNIFICANT (CR = 1.995; p = 0.046 < 0.05).

The hypothesis is therefore accepted.

H₃: Ancillary services have a significant relationship with affective image.

SIGNIFICANT (CR = 4.003; p = 0.000 < 0.05).

The hypothesis is therefore accepted.

H₄: Accessibility has a significant positive influence on affective image.

NOT SIGNIFICANT (CR = 0.543; p = 0.578 > 0.05).

The hypothesis is therefore not accepted.

Table 6 presents hypothesis testing results.

		CR	p	Result	Decision
H1	Price is significantly positively related to affective image	4681	0,000	p < 0,05	H1►S
H2	There is a significant and positive relationship between amenities and affective image	1995	0,046	p < 0,05	S
H3	Ancillary services have a significant relationship with affective image	4003	0,000	p < 0,05	S
H4	Accessibility has a significant positive influence on affective image	0,543	0,578	p > <0,05	H4►R
H5	Price significantly influences performance	1759	0,080	p > <0,05	R
H6	Amenities significantly influence performance	1173	0,241	p > <0,05	R
H7	Ancillary services significantly influence performance	1066	0,039	p < 0,05	S
H8	Accessibility significantly influences performance	1071	0,284	p > <0,05	R

Table 6.

Hypothesis testing.

Key: S: Hypothesis Supported. R: Hypothesis Rejected.

From these findings, it was established that the significant factors affecting the affective image were price, amenities and ancillary services. Further review into the respective magnitudes, using the critical ratios, the findings above do confirm that the most significant of the three is the issue of price. In other words, lodging prices, prices of restaurant food, prices of restaurant beverages and prices of goods and services play the most significant role towards improving the affective image. On the other hand, ancillary services such as cleanliness, tourist information, quietness, friendliness of local people as well as ICT readiness were found to be the second most important factor that has a significant positive influence on affective image. Amenities, while significant, was not so important, comparing with the above two that is price and ancillary services.

4.5.4 Performance

With regards to the second dependent variable, that is, value/performance, it emerged that there was only one significant determinant and this was ancillary services as shown below.

H₅: Price significantly influences performance.

NOT SIGNIFICANT (CR = -1.759; p = 0.080 > 0.05).

The hypothesis is therefore not accepted.

H₆: Amenities significantly influence performance.

NOT SIGNIFICANT (CR = 1.173; p = 0.241 > 0.05).

The hypothesis is therefore not accepted.

H₇: Ancillary services significantly influence performance.

SIGNIFICANT (CR = 1.066; p = 0.039 < 0.05).

The hypothesis is therefore accepted.

H₈: Accessibility significantly influences performance.

NOT SIGNIFICANT (CR = -1.071; p = 0.284 > 0.05).

The hypothesis is therefore not accepted.

From the outcome above, accessibility, amenities and price were not significant determinants of performance. However, ancillary services were. One of the key aspects in the ancillary services category was the friendliness of local people. In this regard, it follows that the value of tourists was shaped more buy ancillary sub-factors such as friendliness of local people, more than traditionally known factors such as accommodation, amenities and price. The lack of significance of tourism resources such as amenities could be an indication of the evolving nature of the type of tourists now visiting Zimbabwe. Generally, the friendliness of local people is a known attribute that is valued by drifters and explorers, or rather allocentric and near allocentric tourists [94]. The lack of significance of amenities could mean that the nature of the tourists visiting Zimbabwe has drifted from being mass tourists, who from the literature, are divorced from the local people, to being drifters and explorers, who tend to interact with the local people, and will try to blend with the host community. This is further validated by the fact that attractions such as the natural landscape and climate had been dropped as not being valid, again, another indication of the evolving interests of tourists, from focusing on the attractions to showing interest in mixing with the host community. This tends to suggest the need to develop community and cultural tourism. Cultural tourism entails interacting with the local people in order to understand their history, present and future [95].

4.5.5 Squared multiple correlations

The researcher went on to evaluate the overall squared multiple correlations for the two dependent variables, that is, affective image and value. The corresponding results are presented in Table 7.

	Estimate
VA	.204
AF	.467

Table 7.

Squared multiple correlations.

Source: Data Survey (2018).

From the results above, the r-square for value was 0.204 while that for affective image was 0.467. It follows from the above finding that the independent variables price, amenities, ancillary services, accessibility and attractions explained the greatest variance in affective image (46.7%) than in value (20.4%). What this means is that the independent variables determined more of the destination’s capacity to relieve stress, the destination’s capacity to provide relaxation, the destination as a pleasant place, the destination as an arousing place as well as the destination as a provider of excitement than they defined the value of the destination.

4.5.6 Research model equation

The research model originally comprised of two endogenous variables as well as four main exogenous variables and these are presented in the quotations below:

Initial Eq. 1:

$$\begin{array}{l}\mathit{AF}={\mathit{PR}}_{\mathit{ai}1\dots \mathit{ai}4}+{\mathit{AM}}_{\mathit{bi}1\dots \mathit{bi}4}+{\mathit{AC}}_{\mathit{ci}1\dots \mathit{ci}4}+{\mathit{\text{AN}}}_{\mathit{di}1\dots \mathit{di}5}\\ +{\epsilon }_{\mathit{ai}1\dots \mathit{ai}4}+{\epsilon }_{\mathit{bi}1\dots \mathit{bi}4}+{\epsilon }_{\mathit{ci}1\dots \mathit{ci}4}+{\epsilon }_{\mathit{di}1\dots \mathit{di}5}\end{array}$$

Price, amenities, and accessibility had four items each, and hence i1-i4, while ancillary services had five items, and hence i1-i5. The equation in simple terms was,

$$\mathit{AF}=\mathit{\alpha PR}+\mathit{\phi AM}+\mathit{\vartheta AN}+\mathit{\eta AC}+{\kappa }_1+{\epsilon }_1\dots \left[1\right]$$

Where:

i: Items.

κ: intercept.

ε: Error term.

α, φ, ϑ, η: Path coefficients.

PR: Price.

AM: Amenities.

AN: Ancillary services.

AC: Accessibility.

AF: Affective image.

Initial Eq. 2:

Again, for Eq. 2, price, amenities, and accessibility had four items each, and hence i1-i4, while ancillary services had five items, and hence i1-i5.

$$\begin{array}{l}\mathit{VA}={\mathit{PR}}_{\mathit{ai}1\dots \mathit{ai}4}+{\mathit{AM}}_{\mathit{bi}1\dots \mathit{bi}4}+{\mathit{AC}}_{\mathit{ci}1\dots \mathit{ci}4}+{\mathit{\text{AN}}}_{\mathit{di}1\dots \mathit{di}5}\\ +{\epsilon }_{\mathit{ai}1\dots \mathit{ai}4}+{\epsilon }_{\mathit{bi}1\dots \mathit{bi}4}+{\epsilon }_{\mathit{ci}1\dots \mathit{ci}4}+{\epsilon }_{\mathit{di}1\dots \mathit{di}5}\end{array}$$

The equation in simple terms was:

$$\mathit{VA}=\mathit{\beta PR}+\mathit{\epsilon AM}+\mathit{\chi AN}+\mathit{\lambda AC}+k_2+{\epsilon }_2\dots \left[2\right]$$

Where:

i: Items

κ: Intercept.

ε: Error term.

ε, β, χ, λ: Path coefficients.

PR: Price.

AM: Amenities.

AN: Ancillary services.

AC: Accessibility.

VA: Performance.

From the above, ε,β, χ, λ, α, φ, ϑ, η were all weights of the exogenous variables that were used to predict the endogenous variables. κ was the intercept and ε was the error term, or residuals. Nevertheless, upon testing the structural equation model, some of the variables were dropped off after their p-values were found to be non-significant (p > 0.05). In this regard, the original equations were subsequently revised. Upon structural equation modeling, for Eq. 1, accessibility was dropped off as it did not have a significant effect on affective image and the subsequent equation comprised one endogenous variable and three exogenous variables as shown below:

Revised Eq. 1:

$$\mathit{AF}=\mathit{\alpha PR}+\mathit{\phi AM}+\mathit{\vartheta AN}+{\kappa }_1+{\epsilon }_1\dots$$E1

On the other hand, for Eq. 2, price, amenities and accessibility did not have a significant impact on value (performance), and in this regard, these were dropped off and the subsequent equation comprised one endogenous variable and one exogenous variable as shown below:

Revised Eq. 2:

$$\mathit{VA}=\mathit{\chi AN}+k_2+{\epsilon }_2\dots$$E2

Where:

κ: Intercept

ε,χ: Path coefficients.

AN: Ancillary services.

VA: Performance.

4.5.7 Model fit test

With a view to testing the validity of a structural equation model above, several goodness-of-fit tests are carried out as prescribed by [96]. There are three broad categories of model fitness tests, and these include absolute fit indices, the relative fit indices as well as the parsimonious fit indices [89]. For the absolute fit indices, the CMIN/DF is the most common, and the chi-square test p-value should be greater than 0.05, while the CMIN/DF ought to be less than 3.0. On the other hand, for the relative fit indices, Goodness-of-Fit Index (GFI), Comparative Fit Index (CFI), Incremental Fit Index (IFI) and Normed Fit Index (NFI) are the most common and this ought to be greater than 0.90. With respect to the parsimonious fit indices, the most common include the Parsimony Normed Fit Index (PNFI), Parsimony Comparative Fit Index (PCFI) as well as the Root Mean Square Error of Approximation (RMSEA) according to [97]. Nevertheless, the most common is RMSEA and according to [98], the maximum acceptable is 0.08. Satisfying the goodness-of-fit at these three levels qualifies the structural model being tested to be accurate and valid [89, 97]. The model fit indices from the study are presented from Table 5.24 to Table 5.27. From the results, with respect to the absolute fit indices, CMIN/DF = 1.730 and this was less than the prescribed maximum of 3.0, and this was the first validation of the model. Table 8 shows absolute fit.

Model	NPAR	CMIN	DF	P	CMIN/DF
Default model	70	487.929	282	.000	1.730
Saturated model	351	.000	0
Independence model	26	3752.085	325	.000	11.545

Table 8.

Model fit-absolute fit indices.

Source: Data Survey (2018).

Further validation was accomplished by the relative fit indices for which IFI and CFI were 0.941 and 0.940 respectively against the expected minimum threshold of 0.90. Table 9 shows relative fit.

Model	NFI Delta1	RFI rho1	IFI Delta2	TLI rho2	CFI
Default model	.870	.850	.941	.931	.940
Saturated model	1.000		1.000		1.000
Independence model	.000	.000	.000	.000	.000

Table 9.

Relative fit indices.

Source: Data Survey (2018).

Regarding the model parsimony, PNFI was 0.755 and PCFI was 0.816 > 0.50. Again, both parsimony measures were greater than the expected minimum 0.50 and this confirmed that the model parsimony was not violated. Table 10shows the parsimony measures.

Model	PRATIO	PNFI	PCFI
Default model	.868	.755	.816
Saturated model	.000	.000	.000
Independence model	1.000	.000	.000

Table 10.

Parsimony-adjusted measures.

Source: Data Survey (2018).

Lastly, with respect the RMSEA statistic, this was found to be 0.052. Because the observed statistic was less than the expected maximum of 0.08, if follows, therefore, the model was valid. Table 11 depicts the RMSEA statistic.

Model	RMSEA	LO 90	HI 90	PCLOSE
Default model	.052	.044	.060	.313
Independence model	.198	.193	.204	.000

Table 11.

RMSEA.

Source: Data Survey (2018).

To test for the sampling adequacy for the model, the researcher considered the use of the Hoelter’s statistics as prescribed by [99], Barrett (2007) and [100]. Table 12 shows sampling adequacy. [89, 100] argue that a critical N of 200 or higher indicates a satisfactory fit. From the results above, both the independence model and the default model had Ns greater than 200, and thus confirming the adequacy of the samples used for this study. Overall, the above tests confirmed the validity of the model as well as the model results. Table 12 shows sampling adequacy.

Model	HOELTER; 0,05; 0,01
Default model Independence model	397; 404; 245; 248

Table 12.

Sampling adequacy.

Source: Data Survey (2018).

Minimization: 0.703.

Miscellaneous: 3.219.

Bootstrap: 0.000.

Total: 3.922.

The conceptual framework suggested the relationships between four components of the cognitive image, namely price, amenities, accessibility and ancillary services and affective image and destination performance. In order to improve destination image, the study found that ancillary services were more significant than accessibility. This was because ancillary services had a bigger influence on affective image than accessibility. Also, price did not significantly influence tourism performance. Amenities tended to influence tourism performance much more than price. This tended to contradict general perceptions among tourism and hospitality stakeholders. This also contradicted conventional wisdom. Price, amenities and ancillary services had a notable influence on affective image while price did not significantly impact tourism performance. This suggested that the conceptual framework was logical and did not deviate much from what the proposed destination image recovery model revealed. In summary, the conceptual framework was supported by the model with regards to the significant influence of ancillary services on affective image and the significant influence of ancillary services on performance. Affective image is known in literature to be a significant factor in image formation [64].

5. Conclusions and implications

The conclusions of the study were derived from the study findings. In terms of the research objective to do with the Current situation with regards to DI and performance of the tourism sector in Zimbabwe, Zimbabwe was mainly visited by tourists traveling for purposes of visiting friends and relatives (VFR). Most of the tourists traveled alone and some in groups followed by those who traveled as couples. The VFR market is known to stay in private homesteads, avoiding hotel accommodation. Africa and Europe contributed most of the tourists who visited Zimbabwe and these were mostly educated males, highly educated with an annual income of at least US$50000 per annum. However, they spent very little in the destination (at most US$1000). This was not surprising given that the destination mainly hosted the VFR market. The national airline lacked capacity to adequately fly tourists into the country and to various tourist destinations in Zimbabwe. Thus the destination’s accessibility was compromised.

The second research objective covered determinants of DI and performance of the tourism sector in Zimbabwe. The most important factor which influenced image and performance of the tourism sector was the price charged by lodging facilities. It was followed by overall quality of the destination and the value tourists attached to Zimbabwe as a vacation destination. Immigration infrastructure and facilities for young children were rated highly. The ZTA and the Tourism Business Council of Zimbabwe (TBCZ), representing the government and the private organizations respectively in tourism and hospitality, were well positioned to influence DI recovery and tourism performance in Zimbabwe. However, both lacked funding to conduct image recovery activities. This implied that DI recovery could take long.

The third research objective looked at the extent to which DI affected performance of the tourism sector in Zimbabwe. Most of the service providers and key informants indicated that they had been affected by Zimbabwe’s unfavorable image to a large extent. Most of them were considering relocating their businesses to neighboring countries. Tourists spent more on food and beverages than on accommodation supporting the prevalence of the VFR market or transit business. The small expenditure by tourists in the destination also indicated the huge effect which DI had on performance of the tourism sector.

The fourth research objective dealt with developing a proposed DI recovery model for enhancing performance of the tourism sector in Zimbabwe. According to the proposed model, price, amenities and ancillary services had a significant influence on affective image. Ancillary services had a significant effect on tourism performance. Accessibility of Zimbabwe as a destination was found not to be significantly affecting destination performance. It can be derived from this that accessing the destination on its own is not the panacea for tourism firms to grow sales and profitability. This is because the tourist could still be constrained by prices when they are in the destination. From the study, the strongest relationship was found to exist between ancillary services and affective image. This suggests that a destination’s support services could influence a tourist’s feelings towards a place. In literature, a lot of attention tends to be put on tourist attractions-both natural and man-made and their capacity to draw tourists to the destination. It appears that the role of ancillary services in shaping DI is underrated. In view of the high prices of goods and services in Zimbabwe, accessibility becomes more of a hygiene factor than a key determinant of destination image and tourism performance. This finding suggested that accessibility would only be relevant in Zimbabwe’s tourism matrix only if the more important drivers of image and performance such as prices, amenities and ancillary services were right. The study showed that in terms of improving the affective image and value of Zimbabwe as a destination, the first thing which needed to be reviewed were the ancillary services then the price.

5.1 Implications for theory

The model has implications for theory. Past destination image recovery models assume that image recovery is synonymous with tourism performance. There was no attempt to isolate factors which influence image and the extent to which they do so and to identify factors which influence performance of the tourism sector and establish the extent to which they influence performance. This study has contributed to knowledge in that it identified specific components which form the cognitive image, measured them and established the extent to which they influence destination image. The challenge with using spinning as suggested by [16] is that there is an assumption that the tourists and potential tourists are not quite informed about the source of the problem at hand. The proliferation of modern technology makes it very difficult for destination marketers to depend on spinning nowadays. It is noted that mobile technologies which include smartphones, mobile applications and tablets have become the main devices for users to access the Internet.

5.2 Implications for policy and practice

The model has a number of implications for policy. It was established that price is a key factor in terms of the formation of the affective image. This implies that in order for tourists to have a favorable view of Zimbabwe as a tourist destination, more attention should be given to pricing. The stakeholder approach which informed this study as indicated in the theoretical framework, needs to be adopted and utilized. Affective image influences potential tourists to consider the destination among many and influences destination choice. Also, it was established that the friendliness of local people played a critical role in the performance of the tourism sector in Zimbabwe. However, the host community needed motivation. In [14], it is noted that eco-tourism could be used to motivate the host community since it increases employment opportunities and it enhances the tourism economy. Evidence is there to show the importance of the host community in tourism performance. The study revealed a need to attach more importance to the logistics and transport sector represented by the struggling national airline. This sector was key in increasing international tourist arrivals and generally enhancing the quality of inbound tourism. Another implication was the enforcement by government of green practices in logistical and transport-related operations. This was vital in enhancing environmental sustainability, reducing criminal activity and in attracting international tourists [34].

5.3 Implications for further research

This study focused on tourists and those employed in the tourism and hospitality industry. Further research could include the ordinary person and also explore strategic public-private partnerships and destination image recovery in Zimbabwean tourism. Respondents were selected from major cities and resorts such as Harare, Victoria Falls, Kariba and others. Possibly, if data had been collected from more areas, the research quality could have been better. Also, future research could explore the value attached by the tourism and hospitality industry on research.