Relationship marketing focuses on retaining existing customers by creating and preserving mutually beneficial, long-term relationships (Christopher, Payne & Ballantyne 2008). The significance of focusing on relationship marketing tactics stems from the notion that building strong customer relationships leads to increased satisfaction, loyalty and customer referrals (Adjei, Griffith & Noble 2009:493; Wong & Sohal 2006:244). Retailers fostering long-term relationships with customers also benefit financially from the lower cost of acquiring customers and increasing their purchases (Ashley et al. 2011:749; Mark et al. 2013:233). Considering these benefits, it is not surprising that more retailers are pursuing long-term relationships with customers to gain a strategic and competitive advantage (De Cannière et al. 2010:87).

Relationships between retailers and customers come about when retailers respond promptly to, and satisfy, their customers’ needs (Adjei et al. 2009:494; Johnson & Ross 2014:207). This, in turn, increases the customers’ willingness to exchange information that allows the retailers to gain a deeper understanding of those needs and how to satisfy them better than their competitors can do (Ashley et al. 2011:749). Sheth and Parvatiyar argue that customers fundamentally engage in relationships with organisations for the purposes of choice reduction, which results in more efficient information processing and decision making. Particularly in the retail context, choice reduction is demonstrated in customers’ inclination to patronise the same retailer over a period of time (Sheth & Parvatiyar 1995:256).

Notwithstanding the fact that relationship marketing offers mutual benefits for retailers and their customers, marketing academics have criticised the application of a relationship marketing approach in consumer markets on account of the size of the market and the limited customer interactions (Leahy 2011:651). Despite this valid criticism, retailers continue to invest in relationship building as a customer retention strategy (Bojei et al. 2013:171; Grewal & Levy 2007:449). It thus becomes essential to understand the dynamics of relationship marketing in consumer markets, including how relationships between retailers and customers develop (Adjei & Clark 2010:73; Mark et al. 2013:233). As customers’ desires to enter into a relationship ultimately affect the effectiveness of retailers’ relationship marketing efforts (Palmatier et al. 2007:210), customers’ intentions to engage in relationships with retailers should be scrutinised (Raciti et al. 2013:616).

Some customers may be indifferent to retailers’ relationship-building efforts (Adjei & Clark 2010:73) by focusing more on transactional exchanges (Dalziel, Harris & Laing 2011:399). Consequently, retailers’ efforts would be better directed at customers who are likely to respond to these efforts (Palmatier et al. 2007:210; Parish & Holloway 2010:63), that is, those who have relationship intentions. Relationship intention represents customers’ conscious and planned desire to engage in a relationship with an organisation whilst buying products or services attributed to the organisation itself, a brand or a channel (Kumar et al. 2003:667; Raciti et al. 2013:615). Kumar et al. (2003:667), who first introduced the relationship intention construct, suggest that customers’ relationship intentions consist of five sub-constructs: involvement, feedback, expectations, forgiveness and fear of loss of relationship.

Baker, Cronin and Hopkins (2009:116) explain that involvement is a function of the personal relevance that an object (for example, a product, a service or a relationship with an organisation) has for customers, along with their willingness to engage in the activities associated with that object. Furthermore, involvement plays a role in customers’ decision making, as involved customers demonstrate greater motivation to collect and process marketing information (Dagger & David 2012:450; Kinard & Capella 2006:365). Information processing and customer decision making influences customers’ interest in relationships with organisations (Varki & Wong 2003:89). Involvement should therefore be viewed as fundamental to the understanding of customer-organisational relationships, as involvement influences customers’ receptivity to organisations’ relationship marketing efforts (Kinard & Capella 2006:365; O’ Cass 2000:551).

Building on the premise that involvement influences customer decision making and that involved customers engage willingly in activities associated with objects of personal relevance (Dagger & David 2012:450), it can be deduced that customers’ involvement indicates their relationship intentions. Kumar et al. (2003:670) concur, arguing that involved customers engage in relationship activities with organisations without being forced to do so. Customers with relationship intentions not only display greater involvement with a retailer, but also value regular feedback to, and from, the retailer (Ashley et al. 2011:754; Baker et al. 2009:117).

Customer feedback, whether positive or negative, is a valuable source of managerial information (Voss et al. 2004:212), as it is an opportunity for organisations to learn from customer interactions (Caemmerer & Wilson 2010:289), thereby highlighting organisational strengths and generating new ideas for service improvements (Wirtz, Tambyah & Mattila 2010:363). Feedback identifies problem areas and is an opportunity for an organisation to rectify problems that may occur during service delivery, thus preventing dissatisfied customers from defecting to competitors and spreading negative word-of-mouth remarks to others. Consequently, customer feedback is appreciated as a customer relationship driver (Lacey 2012:137).

Previous research suggests that customers who have strong relationships with organisations are more at ease in providing feedback (Rothenberger, Grewal & Iyer 2008:359), and are motivated to communicate with organisations about correcting problems, as opposed to terminating a relationship when service transgressions occur (Lacey 2012:138). Subsequently, customers’ willingness to give an organisation feedback is indicative of strong customer-organisational relationships (Blodgett, Wakefield & Barnes 1995:31), and also of their relationship intentions towards that organisation (Kumar et al. 2003:670). Kumar et al. (2003:670) add that customers with relationship intentions are motivated by altruism to give the organisation both positive and negative feedback on their expectations.

Expectations denote customers’ beliefs about a product, service or organisation, derived from previous experiences (Wilson et al. 2012:51). These beliefs, in turn, become reference points or standards against which to judge performances (Zeithaml, Berry & Parasuraman 1993:1). Consequently, expectations significantly influence customer satisfaction (Oliver 1980:460), and, ultimately, their behavioural intentions (Choy, Lam & Lee 2012:14).

Customer expectations differ, depending on the type of relationship they have with an organisation (Mason & Simmons 2012:231). For example, customers in established relationships with organisations often have higher expectations of the organisations. To sustain customer-organisational relationships requires considerable investment of irrecoverable resources, including time and effort (De Wulf et al. 2001:34; Liang & Wang 2006:120–121). Kumar et al. (2003:670), therefore advocate that customers who expect more from organisations are concerned with the enhancement of organisational products and services and are thus more intent on building relationships.

Although the strongest customer-organisation relationships are susceptible to disruption and problems when customer expectations are not met (Tsarenko & Tojib 2011:383), research has shown that strong customer relationships give organisations some protection after a service transgression has occurred (Priluck 2003:37; Yu & Xie 2011:1). Kim, Ok and Canter (2012:60–61) and Singh and Sirdeshmukh (2000:150) explain that customers in strong organisational relationships display greater tolerance of service failures, and are therefore more likely to forgive organisations for their poor product or service experiences. Such customers are also more likely to forgive service transgressions, as they expect to maintain their relationships with the organisation in the future (Priluck 2003:37; Yu & Xie 2011:1). Kumar et al. (2003:670) accordingly suggest that customers’ willingness to forgive an organisation when their expectations are not met, or when transgressions occur, reveals their relationship intentions towards the organisations.

Customers’ wishes to maintain organisational relationships are embedded in the relational benefits arising from such relationships, including confidence and social and special treatment benefits (Bojei et al. 2013:171; Henning-Thurau, Gwinner & Gremler 2002:234). Dagger, David and Ng (2011:273) posit that customers in organisational relationships experience less anxiety, as they know what to expect from an organisation (confidence benefits).They receive personal recognition, familiarity and friendship from the organisation's employees (social benefits), as well as discounted prices and customised product or service offerings (special treatment benefits).

These relational benefits facilitate the formation of relational bonds, when customers interact satisfactorily with organisations over consecutive time periods (Homburg, Giering & Menon 2003:44; Liang & Wang 2007:339). Relational bonds include financial items (price attractions and discounts to stimulate customer consumption), social aspects (a sense of affiliation through the provision of support, advice and empathy during interactions), and structural bonds (rules, policies, procedures, infrastructure or agreements) that give formal structures for relationships and interactions) (Liang & Wang 2007:339–341; Wilson et al. 2012:154). Relational bonds also increase customers’ perceptions of switching cost, that is, the cost of lost resources invested when they transfer to another organisation (Jones et al. 2007:337).

As the establishment of customer-organisational relationships requires the investment of customer effort and time (Vázquez-Casielles et al. 2009:2293), any perceived loss of relational benefits, termination of relational bonds, and possible switching costs give rise to customer concern about the consequences that may come about should their relationship with an organisation end (Jones et al. 2007:337; Kumar et al. 2003:670). Therefore, customers who fear losing their relationship with an organisation demonstrate relationship intentions (Kumar et al. 2003:670).

A loyalty programme is an integrated system of marketing activities aimed at increasing customer spending, loyalty, and ultimately retention by rewarding customers for repeated patronage (Meyer-Waarden 2008:89; Vesel & Zabkar 2010:1339). If they want to join a retailer's loyalty programme, customers usually have to apply to become a member, after which they are issued with a loyalty card to validate their membership so that they can take advantage of the rewards offered (Leenheer & Bijmolt 2008:429). Rewards, such as price discounts and gifts, are obtained either immediately when customers present their loyalty card, or when their points, accumulated over a period of time, are redeemed (Liu 2007:20).

It is believed that loyalty programmes facilitate customer relationship development in a number of ways. Firstly, loyalty programmes allow retailers to obtain information about their customers’ product and retail mix preferences (Mauri 2003:13), which enables the retailers to better satisfy customer needs, a necessary condition for relationship formation (Adjei et al. 2009:494). Secondly, as loyalty programmes require the investment of considerable irretrievable resources, it signals the retailers’ intention to sustain the customer relationships (Leenheer & Bijmolt 2008:429; Liu 2007:19). Lastly, loyalty programmes focus on encouraging customer loyalty by rewarding customers for their continued patronage (De Wulf & Odekerken-Schröder 2003:97; Leenheer & Bijmolt 2008:429; Liu 2007:19). It is thus understandable that clothing retailers increasingly use loyalty programmes as one of their relationship marketing strategies for establishing and strengthening customer relationships and stimulating customer loyalty (Huang 2015:1318; Leenheer & Bijmolt 2008:429).

South African examples of clothing retailers using loyalty programmes include the WRewards card from Woolworths, the TFG Rewards & More card from Foschini, the Thank U Rewards card from Edgars and the PEPclub card from Pep Stores (Pep Stores 2014; Thank U Rewards 2014; The Foschini Group 2014; Woolworths 2014). The WRewards card focuses on dividing customers into tiers according to their spending, and rewards customers with instant price discounts and benefits (Woolworths 2014), whilst the Edgars Thank U Rewards card allows the customers to earn and accumulate points based on the monetary value of their purchases, which, in turn, can be redeemed for cash discounts (Thank U Rewards 2014). Foschini attempts to foster relationships with their customers by providing instant, unique rewards based on the customers’ specific needs (The Foschini Group 2014), whilst PEP customers earn luckies (competition entries into a lucky draw for a variety of prizes) every time their PEP club card is swiped in-store (PEP Stores 2014).

Whilst clothing retailers use loyalty programmes to enhance customer relationships, Bridson, Evans and Hickman (2008:364) caution that some customers may view loyalty programmes as mere promotional tools and may switch from one clothing retailer to the next to take advantage of the best offers. For this reason, the customers’ loyalty programme membership may not necessarily indicate that they are willing to respond to the clothing retailers’ relationship efforts.

For the purpose of this study, the duration of customers’ support refers to the time period during which customers have repeatedly purchased from the same clothing retailer (Rust & Williams 1994:108). The length of time during which customers have supported a particular retailer is viewed as significant, as it is believed that longer periods of support are directly associated with retailers’ profitability (Ward & Dagger 2007:283). Further, scholars argue that the impact of relationship marketing tactics on relationship investment improves over time, suggesting that customer relationships strengthen the longer their support lasts (De Wulf et al. 2001:47; Ward & Dagger 2007:282). Clothing retailers may assume that the length of time during which customers have supported them is indicative of their successful relationship marketing strategies. However, postulating that relationships with customers automatically strengthen over time erroneously distorts the dynamics of relationship marketing (Ward & Dagger 2007:287). Likewise, the assumption that the duration of their support indicates customers’ relationship intentions is incorrect (Parish & Holloway 2010:69). Customers may, therefore, buy from an organisation over an extended period of time, yet lack the emotional attachment or intention to develop a relationship with that organisation (Kumar et al. 2003:670).

The target population for this study was clothing retail customers aged 18 years and older, residing in the greater Pretoria metropolitan area. In the absence of a sample frame, non-probability convenience sampling was used to collect data from the respondents in the target population (Iacobucci & Churchill 2010:287). This approach is similar to some recent South African studies (Beneke et al. 2012; Roberts-Lombard & Immelman 2015) as well as other South African studies on relationship intention (Kruger & Mostert 2013:339, 2015:57).

An interviewer-administered survey approach using a structured questionnaire complemented the descriptive research design used for the study (Zikmund & Babin 2013:49, 164). Trained fieldworkers approached respondents on the basis of convenience. Participation in the study was voluntary. The respondents were not asked to divulge any personal or identifiable information, thereby ensuring their anonymity.

The questionnaire commenced with a set of screening questions which would identify the respondents as part of the target population. As previous studies examining relationship marketing in the retail context emphasise that the respondents should have had regular contact with a retailer (Bettencourt 1997:393; De Wulf & Odekerken-Schröder 2003:101), a screening question was included to ensure that they had purchased items from a clothing retailer within the past three months. A second screening question ensured that those participating in the study were the main decision makers in terms of choosing a clothing retailer from which to purchase clothing.

The first section of the questionnaire focused on determining the respondents’ clothing retail patronage habits, including their loyalty programme membership and how long they had supported the clothing retailer where they shopped most frequently. Similarly to the approach followed by De Wulf and Odekerken-Schröder (2003:101), the next section focused on measuring the respondents’ relationship intentions towards the clothing retailer where they shopped most often.

Previous studies have measured relationship intentions using 26 items (Kruger & Mostert 2012:45, 2013:350). This has resulted in fielding administration issues concerning the length of the questionnaire and the time needed to complete it (Fricker et al. 2012:3). The relationship intention measurement scale proposed by Kruger and Mostert (2012:45) was therefore adapted and shortened to 15 items for the present study. Specifically, a 5-point unlabelled Likert scale, where 1 = strongly disagree and 5 = strongly agree, was used to measure relationship intention.

The questionnaire was pre-tested amongst 60 respondents from the target population. After minor wording changes, the data was collected by trained fieldworkers, who approached prospective respondents on the basis of convenience. In total, 511 usable questionnaires were collected from the respondents for analysis.

The Statistical Package for Social Sciences (SPSS) (Version 22) was used to capture, clean and analyse the data. With reference to the normality of distribution relating to each scaled item included in the questionnaire, Curran, West and Finch (1996:16) indicate that the distribution of results can be deemed normal if it displays skewness and kurtosis absolute values of less than 2.00 and 7.00 respectively. As all the scaled items in the questionnaire were within these parameters (ranging between -0.027 to -1.425 for Skewness and 0.061 to 1.925 for Kurtosis) (Curran et al. 1996:16), as well as the fact that the sample size was relatively large, parametric tests were considered suitable for testing the hypotheses.

An exploratory factor analysis was performed to reduce the dimensionality of the data (i.e. the items used to measure relationship intentions) into a smaller set of more manageable factors (Field 2013:628; Pallant 2013:188). The exploratory factor analysis also helped the researchers understand the underlying structure of the latent variable (i.e. factors that might underlie relationship intentions) and test the validity of the shortened relationship intention measurement scale (Field 2013:628). Cronbach's alpha coefficients were calculated to assess the reliability of the scale measuring the respondents’ relationship intentions (Pallant 2013:6). Field (2013:679) recommends that Cronbach's alpha coefficients should ideally be greater than 0.7 to indicate the acceptable internal consistency and reliability of the measurement scale.

In order to test the hypotheses formulated in this study, the researchers relied on a 95% confidence level, meaning that p-values of less than 0.05 were interpreted as being statistically significant (Hair et al. 2013:281). An independent sample t-test and one-way Anovas were carried out in order to determine whether there were statistically significant differences between the means of different groups, whereas significant associations between constructs were determined by means of Pearson's chi-square tests for independence by cross-tabulating the variables (Field 2013:722; Pallant 2013:247, 258).

Although statistical significance demonstrates whether statistical differences or associations exist between variables, it does not indicate the strength of the significance (Field 2013:79). Ellis (2010:4) therefore suggests that effect sizes should be calculated to enable the researcher to judge the practical importance of an effect or a result. In this study, Cohen's d-values (effect size for differences) and w-value (effect size for associations) were calculated (Bagozzi 1994:248; Cohen 1988:25–26). Cohen (1988:25–26) explains that d-values can be considered small at 0.2, medium at 0.5, and large (practically significant) at a value greater than or equal to 0.8. All the d-values were rounded off to one decimal point. In terms of the effect sizes for associations, Steyn (1999:8) suggests that w-values should be interpreted according to the following guidelines: w = 0.1 indicates a small effect, w = 0.3 indicates a medium effect and w = 0.5 indicates a large effect, and a practically significant, association between variables. Similar to previous studies (Farrington 2014; Kruger & Mostert 2013:349), this study used effect sizes that were large and practically significant (specifically w-values ≥ 0.5 and d-values ≥ 0.8) when deciding whether or not to support hypotheses.

Table 1 provides an overview of the sample profile and clothing retail patronage habits of the respondents who participated in the study.

From Table 1 it can be derived that, in terms of gender, more females (61.8%) than males (38.2%) participated in the study. Furthermore, most of the respondents were white people (50.9%) and black people (32.0%), the remainder being either Indian/Asian (10.6%) or those of mixed race (6.1%). The majority of the respondents were single (65.8%), married or living with a partner (30.5%). Regarding the highest level of education, the majority of the respondents had completed matric/grade 12 (56.4%), a degree (18.4%) or a diploma (15.4%).

Most of the respondents shopped most frequently at Mr Price (27.2%), followed by Edgars (22.7%), Woolworths (15.1%) and Markhams (5.9%). Slightly more than seven per cent of the respondents indicated that they most frequently purchased clothing from ‘other’ clothing retailers, including lesser known clothing retailers like Jay Jays, PQ Clothing and Top Shop. How long the respondents had patronised the clothing retailer they shopped at most frequently ranged from a period of 1 year or more, but less than 5 years (42.7%), 10 years and longer (29.2%) or 5 years or more, but less than 10 years (23.1%). Finally, whilst the majority of the respondents (48.5%) indicated that they did not have a loyalty card from the clothing retailer in question, 32.7% did have one, whereas 10.4% indicated that the clothing store where they bought most often did not offer a loyalty card.

An exploratory factor analysis was conducted using maximum likelihood extraction with varimax rotation (Field 2013:642, 644; Hair et al. 2014:94) to identify the underlying factors constituting the clothing retail respondents’ relationship intentions. The data was considered appropriate for factor analysis, as the Bartlett's test of sphericity yielded a significant result (p < 0.0001) and the Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy (MSA) for the overall measure was 0.791 (which is greater than the recommended cut-off value 0.5) (Field 2013:647; Pallant 2013:199). Table 2 shows the rotated pattern matrix for the 15 items used to measure the respondents’ relationship intentions in the clothing retail context.

In Table 2, it can be seen that five factors were extracted to measure the clothing retail respondents’ relationship intentions. The five factors explain 75.40% of the total variance in the data. Furthermore, all the 15 items included in the measuring instrument loaded onto the five factors, and no items cross-loaded onto other factors. Table 2 also shows that all the items yielded factor loadings ≥ 0.5, indicating that each item should be retained (Hair et al. 2014:116). The decision to retain all the items in the analysis was further supported by considering the MSAs for all the pairs of items reflected in the factor analysis. According to Field (2013:687), the MSA for all the pairs of items included in the factor analysis should be greater than 0.5, and where MSA values smaller than 0.5 are found, the pair of corresponding items should be deleted from the analysis. The MSAs realised for all the pairs of items reflected in the factor analysis ranged between 0.656 and 0.849, thereby supporting the decision not to delete any pair items from the analysis.

In Table 2, it can be seen that three items loaded onto Factor 1 relating to the respondents’ concern about losing services, special privileges, or their relationship with their clothing retailer. Factor 1 was subsequently labelled Fear of relationship loss, which coincides with the label of the sub-construct that Kumar et al. (2003:667) suggested. The three items that loaded onto Factor 2 focused on the respondents’ willingness to forgive their clothing retailer for bad service or for their expectations not being met. Factor 2 was consequently labelled Forgiveness, and therefore supported the ‘forgiveness’ sub construct proposed by Kumar et al. (2003:667). Three items loaded onto Factor 3, which focused on customers’ feedback on their clothing retailers’ service delivery. The items comprising Factor 3 led to its being labelled Feedback, thereby supporting the ‘feedback’ sub-construct proposed by Kumar et al. (2003:667). Three items loaded onto Factor 4 related to the customers’ concern for the clothing retailers’ image. It suggested their pride in being a customer of the clothing retailer and seeing the clothing retailer's name or advertising material. The items comprising Factor 4 led to its label Involvement, which coincided with the involvement sub-construct proposed by Kumar et al. (2003:670). Lastly, the three items loaded onto Factor 5, dealing with the respondents’ expectations of receiving value for money and acceptable service. Factor 5 was consequently labelled Expectations, echoing the ‘expectations’ sub-construct proposed by Kumar et al. (2003:667).

Table 2 also reflects the Cronbach's alpha coefficients calculated to determine the reliability of the scales used to measure the factors identified in the exploratory factor analysis. Cronbach's alpha coefficient values were greater than 0.70, suggesting that the factors comprising relationship intention in the clothing retail setting are reliable (Hair et al. 2014:166).

According to the results of the exploratory factor analysis and Cronbach's alpha coefficients values, it can be concluded that the 15-item relationship intention measurement scale was valid and reliable when it came to measuring the respondents’ relationship intentions towards their clothing retailers.

An overall mean score was calculated for each respondent's relationship intentions (Kumar et al. 2003:675) in order to categorise the respondents into three relationship intention groups (by using the 33.3 and 66.6 percentiles as cut-points) according to their relationship intention levels. One-way Anovas was accordingly performed to determine whether there were any significant differences between the mean scores for the three relationship intention groups. Table 3 shows the descriptive statistics, Tukey's comparison (statistically significant at the 0.05 level) and d-values (effect sizes) for the respondents’ overall relationship intentions.

Using the cut-points described, Table 3 shows that 154 respondents were classified as having low relationship intentions (mean = 2.56), 202 had moderate relationship intentions (mean = 3.19), and 155 respondents had high relationship intentions (mean = 3.86). Suwal (2013) explains that the presence of multiple identical values at a cut-off point means that they all formed part of the same group. Consequently, the number of respondents per relationship intention group differed. Table 3 shows further that there are statistically significant differences between the three relationship intention groups, which can be regarded as practically significant, as all the effect sizes were greater than 0.8. It can therefore be concluded that hypothesis 1, stating that clothing retail customers with different levels of relationship intention differ significantly in terms of their relationship intentions, is supported.

An independent sample t-test was performed to determine whether significant differences existed between clothing retail customers’ relationship intentions and whether they belonged to a loyalty programme offered by the clothing retailer where they shopped most often. It should be noted that respondents who indicated that their clothing retailer of choice did not offer a loyalty programme (n = 53) were excluded from further analysis of loyalty programme membership. Table 4 shows the descriptive statistics, the p-value yielded by the t-test and d-values (effect sizes) when comparing the respondents’ overall relationship intention means according to their loyalty programme membership.

Table 4 demonstrates that there was a statistically significant difference between the respondents’ overall relationship intentions and their loyalty programme membership (p = 0.039). However, a small effect size (d = 0.4) indicates that the difference is not practically significant. It can therefore be concluded that hypothesis 2 is not supported, because there were no significant differences between the respondents’ relationship intentions and their loyalty programme membership.

It was decided to determine whether there were any significant associations between respondents with different levels of relationship intentions, and whether or not they belonged to a loyalty programme. Table 5 shows the results from a cross-tabulation between the different relationship intention groups and whether they belonged to a loyalty programme offered by their clothing retailer.

The cross-tabulation set out in Table 5 allows for the observation that most respondents with low and moderate relationship intentions did not belong to the loyalty programme offered by their clothing retailer of choice, whereas the majority of the respondents with higher relationship intention levels had loyalty programme membership. In order to determine whether there was a statistically significant association between the respondents’ level of relationship intention and their loyalty programme membership, a chi-square test was performed. The test realised a p-value of 0.00, indicating that there were statistically significant associations between variables. However, the realised effect size (w = 0.19) suggests that there was no practically significant association between the respondents’ relationship intention levels and whether they belonged to the loyalty programme offered by their clothing retailer. Hypothesis 3, stating that there are significant associations between clothing retail customers with different relationship intention levels and their loyalty programme membership, is therefore not supported.

One-way Anovas was conducted to determine whether significant differences existed between clothing retail customers’ relationship intentions and the duration of their support for their clothing retailer. As only eight respondents indicated that they had supported their clothing retailer for less than one year, the category was collapsed, together with the category one year or more, but less than five years and labelled ‘less than 5 years’. Table 6 shows the descriptive statistics, the comparison with Tukey (statistically significant at the 0.05 level) and the d-values (effect sizes) for the respondents’ overall relationship intentions.

Table 6 shows that there were no statistically or practically significant differences between the respondents’ overall relationship intentions and how long they had been supporting their clothing retailer of choice. It can therefore be concluded that hypothesis 4 is not supported, as there were no significant differences between the respondents’ relationship intentions and the duration of their support.

Although there were no significant differences between the respondents’ overall relationship intentions and how long they had supported their clothing retailer, it was decided to determine whether there were any associations between the respondents with different levels of relationship intentions and the duration of their support for their clothing retailer. Table 7 shows the results of a cross-tabulation between the different relationship intention groups and the duration of their support for the clothing retailer.

Table 7 shows that most of the respondents with low and moderate relationship intentions had been supporting their clothing retailer for 5 years or more, but for fewer than 10 years, whereas the majority of the respondents with high relationship intentions had been supporting their clothing retailer for fewer than 5 years. In order to determine whether there was a statistically significant association between the respondents’ level of relationship intention and how long they had been supporting their clothing retailer, a chi-square test was performed. The test realised a p-value of 0.126, indicating that no statistically significant association existed between the variables. The realised effect size (w = 0.119) also suggests that there was no practically significant association between variables. Hypothesis 5, stating that there were significant associations between clothing retail customers with different relationship intention levels and the duration of support is therefore not supported.