An empirical study of the reward preferences of South African employees

Research purpose: The aim of the study was to identify certain categories of rewards that employees consider to be most important, including base pay, contingency pay, benefits, performance and career management, quality work environment, and work–home integration. The impact of these reward categories on an organisation’s ability to attract, motivate and retain employees was explored, together with the influence of demographic variables on reward preferences.


Introduction Key focus of the study
The rewards that employees receive should be important to them and address their needs.Rewards refer to the compensation that an employee receives from an organisation in exchange for his or her services (Jiang, Xiao, Qi & Xiao, 2009).When companies structure their reward systems correctly, employees inherently perform well to achieve the organisation's goals as well as their own (Pfau & Kay, 2002).Milkovich, Newman and Gerhart (2011) also specified that rewards can play a significant role in influencing employees' attitudes and perceptions of work.Owing to the fact that rewards represent anything that is valuable and meaningful to the recipient, skilled and talented employees are less likely to be motivated by rewards that are not aligned with their preferences and values (Chiang & Birtch, 2007).In the past, the primary focus of reward systems was on compensating employees by financial means (e.g. through a basic salary), but with a rapidly changing workforce, a more holistic and integrated approach to reward management is now required (Zingheim & Schuster, 2001).
During the past several years, the concept of total rewards has become a popular topic of discussion.According to Heneman (2007), the shift from the phrase 'compensation and benefits' to 'total rewards' has resulted from the inability of many organisations to offer pay increases and added benefits because of intense competition in the marketplace.Employers therefore have to find alternate, less costly forms of rewards that will still attract, retain and motivate employees (Heneman, 2007).'Total rewards' thus describes a reward strategy that combines compensation, employee benefits, work-life balance, personal recognition and career development in the benefits package (Armstrong & Brown, 2001;Costello, 2010;WorldatWork, 2007).These elements represent an inventory from which organisations can design reward packages that create value for the organisation and its employees.A total rewards framework allows employers to consider employee preferences in accordance with individual needs and be more flexible when designing reward packages (Bussin, 2011).It also enables organisations to identify the correct combination of rewards for its particular workforce (WorldatWork & Towers Watson, 2012).This understanding of total rewards forms the key focus of this study.

Background
In the past few years, companies have faced a difficult economic environment that has challenged traditional reward practices and strategies (Deloitte Development LLP, 2009).Having experienced the effects of a worldwide economic recession since 2008, human resource managers within South Africa face increasing pressure to contribute more significantly towards the success of their organisations (Grobler et al., 2011).Remuneration accounts for over half the operating costs of most South African organisations, which points to the importance of maximising the return on these costs (Sutherland, 2011), especially in trying economic times.Yet, unfortunately, according to the Africa competitiveness report 2011, South Africa received a score of only 3.2 out of possible 7.0 within the sub-pillar of 'pay and productivity', coming 112th out of 139 participating countries (World Economic Forum, World Bank & African Development Bank, 2011).This implies that pay is not strongly related to employee productivity in this country.
In addition, growing concern regarding the retention of high performers has resulted in employers rethinking the way in which they reward their employees (Deloitte Consulting LLP, 2008), particularly because of the shortage of skilled and executive employees in Africa (Sutherland, 2011).Businesses in Africa are struggling to find the skilled employees they require, with many skilled African graduates emigrating to developed countries (World Economic Forum et al., 2011).Meyer and Kirsten (2012) agreed with this, stating that there has been a steady loss of skilled and qualified individuals from South Africa as a result of emigration, which poses a problem because of the need for skills to promote production in the country.Furthermore, Generation X and Y employees now comprise a large part of the South African workforce, resulting in new strategies needing to be formulated to successfully retain and motivate these younger employees (Grobler et al., 2011).In today's turbulent working environment where competition is rife, companies are moreover under pressure to add value and demonstrate flexibility in dealing with new challenges and opportunities (Armstrong & Brown, 2001).South Africa in particular is characterised by strong market competition (World Economic Forum et al., 2011), especially because globalisation has driven the need to be competitive (Swanepoel, Erasmus & Schenk, 2008).Demographic and sociological changes (Vandenberghe, St-Onge & Robineau, 2008), as well as differences in employee preferences as a result of generational differences and life stages (Giancola, 2008) and cultural influences (Chiang & Birtch, 2007), all present employers with major challenges when attracting, retaining and motivating employees.Companies are now required to adopt innovative ways of management in order to differentiate themselves in the labour market (Heneman, 2007), particularly with regard to how their reward systems are developed and managed.Organisations have responded to these pressures by adopting more flexible structures such as total reward packages, which better position organisations to attract and retain high quality employees (Bussin, 2011).

Research purpose
Employee rewards has thus become an important topic in today's economy as companies strive to find a balance between reward components, as well as identify those components that provide for employees' personal needs (Costello, 2010).As competition in the workplace increases, employees are becoming more demanding when it comes to rewards, expecting organisations to make exceptions based on their individual preferences and needs (Herman & Gioia, 2000).Peoples' needs and preferences change over time (Gross & Friedman, 2004) and there are many factors that can affect employees' reward preferences, including age, values, religion, marital status, number of dependents and culture (Meyer & Kirsten, 2012).Rewards packages that fit well with such characteristics and preferences of employees can result in increased motivation and, ultimately, improved performance (Lawton & Chernyshenko, 2008).For example, flexible benefits and rewards packages allow employees to tailor their packages to suit their personal motivations, work and leisure interests, career stage and domestic circumstances (Armstrong & Brown, 2009).Thus, it is useful to provide employees with flexibility in choice regarding the different reward categories available, allowing for the design of 'different, employee-initiated reward profiles' (Nienaber & Bussin, 2011, p. 2).This can also be referred to as the individualisation of rewards packages (Corby, Palmer & Lindop, 2009), where choices are built into the reward system (Swanepoel et al., 2008).
However, Nienaber and Bussin (2011) explained that it is virtually impossible for large-sized and medium-sized companies to administer rewards based on the unique preferences of individuals, because of the difficulty such companies would experience in administering and governing thousands of unique reward profiles.One manageable way to address this issue is to design reward packages according to different employee groups or segments.This would entail clustering various types of rewards depending on the needs of certain employee segments.For example, according to Mercer's seventh global total rewards survey (Mercer, 2008), different rewards are often offered for different employee groups specifically according to job level, job family, business unit, product line or lifecycle, as well as geographic location.Further examples of ways by which to segment a workforce include categorising it by generation, age, gender, family size, occupation, educational level or religion (Du Toit, Erasmus & Strydom, 2007).Segmentation is one of three key principles that, once applied to organisations' reward and talent management models, will result in significant improvements to both talent return on investment and human capital risk management (WorldatWork & Towers Watson, 2012).

Research objectives
Based on the above, the aim of this study is to identify categories of rewards that are preferred by South African employees within the corporate world, as well as how satisfied these employees are with these rewards as offered by their current employers.Furthermore, this study will explore the influence of demographic variables on reward preferences, as well as the impact that reward categories have on an organisation's ability to attract, motivate and retain employees.

What will follow
Numerous means of segmenting employees with regard to reward preferences are examined in this paper, including the influence of demographic factors and generational differences.Thereafter, a discussion on the use of rewards in attracting, retaining and motivating employees will be provided.

Demographic influences on reward preferences
Research has shown that individuals' reward preferences are influenced by their personal demographic characteristics.For example, Nienaber, Bussin and Henn (2011) found that reward preferences differ according to a number of demographic factors, such as number of children, race, age, job level, educational qualifications, marital status, years of service, as well as gender.In terms of job level and type, Lawton and Chernyshenko (2008) found that full-time employees with dependants prefer family-supportive rewards, such as medical and life insurance, and that lower level employees feel that additional education and training would increase their chances for career advancement.Employees in lower level jobs, such as administrative and junior management levels, have also been observed to hold significantly higher preferences for remuneration and benefits when compared to senior and executive management (Nienaber et al., 2011).It is thus possible that job level be used to categorise a workforce so that employee groups can be more effectively rewarded (Mercer, 2008).Chiang and Birtch (2006) stated that demographic variables such as gender affect reward preferences.Nienaber et al.'s (2011) research showed that women have a stronger preference for remuneration and benefits as well as for a conducive working environment.Additionally, Konrad, Ritchie, Lieb and Corrigall (2000) revealed that men place greater importance on career advancement and promotions as well as increased responsibility without the interference from managers.However, Fisher and Yuan (1998) found no gender differences in reward preferences such as wages, working conditions, interesting work and growth and promotion opportunities.Similarly, Chow and Ngo (2002) established that a high salary and good working conditions are considered to be important by both male and female employees.
In terms of age, it has been found that needs change as employees get older; for example, employees who are older value rewards such as flexible work arrangements and skill development rather than cash compensation (Hedge, Borman & Lammlein, 2006).Therefore, it can be suggested that a higher age is related to a preference for rewards that offer work flexibility and training and development opportunities.This is in contrast with research by Cennamo and Gardner (2008), who discovered that extrinsic rewards such as pay and benefits are valued more highly by older employees than younger employees.Lawton and Chernyshenko (2008) concluded that young employees with simpler jobs value training and development opportunities more than their older peers, a finding that is also contrary to that of Hedge et al. (2006).Nienaber et al. (2011) discovered that reward categories, such as remuneration and benefits, as well as a conducive working environment, are valued more by younger employees, reducing in need as employees age.Younger employees also may prefer more tangible rewards, such as medical plans or maternity or paternity leave, whereas older employees may value stock options or retirement plan contributions (Mehta, Anderson & Dubinsky, 2000).Furthermore, Fisher and Yuan (1998) found that younger employees place more importance on promotion, growth in skills, as well as interesting work, whereas older employees value job security and sympathetic help with personal problems more than their younger peers.
Based on the above, it is suggested that if employers have an understanding of the demographic characteristics of employees, they will be able to design rewards packages such as personal benefit schemes that will create unique value for their employees and ultimately lead to a competitive advantage in the labour market (Lawton & Chernyshenko, 2008).

Generational influences on rewards
A generation can be interpreted in terms of cohorts, life experiences or historical experiences.A cohort is a group of people who share a similar experience or characteristic within a specific time frame (Dencker, Joshi & Martocchio, 2007).The generational school of thought argues that values have been influenced by the defining moments or major historical events that occur as individuals grow older (Giancola, 2008).Each group originates from a distinct and unique era and therefore has its own view on business issues.
Presently, four distinct generations of people are interacting in the workplace (Murphy, 2007).Generational experts typically place individuals into the following generations according to their birth date: Veterans, the Silent Generation, Traditionalists or the GI Generation (1925-1942), Baby Boomers (1943-1960), Generation X (1961-1981) and Generation Y or Millennials (1982-Present) (Giancola, 2008).According to Bussin (2011), each generation has a different view on rewards.The most valued rewards for Veterans are flexible schedules, part-time hours and recognition of expertise and experience.Baby Boomers prefer recognition of hard work, healthcare and retirement benefits and autonomy.Generation X values flexible work arrangements, worklife balance and rewards that are tangible.Lastly, learning and development opportunities, immediate performance feedback, working in teams and having positive role models are important to Generation Y (Hewlett, Sherbin & Sumberg, 2009;Reynolds, 2005;Zingheim & Schuster, 2008).Cennamo and Gardner (2008) also discovered that Generation Y employees value work-life balance, freedom and flexibility in their jobs.
Despite the above research, Jorgensen (2003) questioned whether there is enough empirical evidence to base workplace strategies and practices on the differing values, likes and dislikes of the generations.According to findings of research conducted by Noble and Schewe (2003), there was no empirical evidence to support the validity of generational segmentation in the workforce.In South Africa, research conducted by Moore (2009) and Nienaber et al. (2011) revealed that different generations do not have different reward preferences.The implication therefore is that structuring reward packages around generational differences can lead to dissatisfaction amongst employees.
However, Smola and Sutton (2002) argued that companies do need to examine their reward systems in relation to the needs and expectations of generational groups, because failure to address these differences can result in lower levels of employee productivity, conflict in the workplace and miscommunication.This is especially because employees in different generational groups expect the rewards they receive to meet their individual needs (Dencker et al., 2007).Yet, a survey conducted in 2008 to gain insight into how employers are handling the challenge of rewarding a multigenerational workforce revealed that 56% of organisations do not even consider generational differences when designing and implementing total reward programmes (WorldatWork, 2008).This means that although there is an awareness of generational differences, organisations may be underestimating the importance of addressing the needs of each generation and rewarding them accordingly (WorldatWork, 2008).

Culture and personality effects on reward preferences
Although not empirically investigated in this study, it is worth briefly noting the influence that culture and personality play on individuals' preferences for certain rewards.Multinational organisations can improve their competitive position by changing their compensation practices to match the cultural preferences of employees across countries (Westerman, Beekun, Daly & Vanka, 2009).Research conducted by Chiang and Birtch (2005, 2006, 2007) revealed that culture does influence certain reward preferences and thus understanding how reward preferences differ across cultures is key to managing multinational reward systems.Ensuring that compensation practices are in line with a country's culture is important because it shows cultural awareness and sensitivity and results in increased financial performance (Schuler & Rogovsky, 1998).Furthermore, in terms of personality, Nienaber et al.'s (2011) South African study confirmed that individuals with certain personality types and preferences experience different reward preferences.Vandenberghe et al. (2008) found that individuals who are open to experience are attracted to rewards such as flexible work schedules, variable pay, bonuses, relationships with co-workers and development and career opportunities, whilst Roberts, Chernyshenko, Stark and Goldberg (2005) proposed that conscientious individuals value family-orientated benefits, as they feel the duty to support their families and make sure they provide for their needs.

The use of rewards in the attraction, motivation and retention of employees
Remuneration is a driver of retention, job satisfaction, as well as employee commitment in South Africa (Sutherland, 2011).The objectives of a South African compensation system should thus include attracting highly qualified employees, as well as motivating and retaining these employees through incentivising desired behaviour and rewarding good performance (Grobler et al., 2011;Meyer & Kirsten, 2012).Highlighting this finding, Swanepoel et al. (2008) and Armstrong, Brown and Reilly (2010) mentioned that remuneration impacts the attraction and retention of employees, whilst Phillips and Gully (2012) stated that competitive wages and benefits have, time and again, been listed as a means of attracting and retaining employees.According to Grobler et al. (2011), job applicants compare different job offers and pay scales, putting more weight on the comparative salaries being offered to them as opposed to benefits and intrinsic rewards on offer because it is easier to compare monetary offers.
In terms of retention, Grobler et al. (2011) noted that inadequate compensation is often the cause of turnover.This is confirmed in the South African context by Mohlala, Goldman and Goosen (2012), who found that monetary rewards were lacking in their organisation under study.This negatively affected retention, especially for those who had been employed there for longer periods of time.Respondents confirmed that employees had left the organisation to move to where they had been offered better salaries.Good benefits can also be used to increase retention, but Phillips and Gully (2012) mentioned that this will only affect retention positively if employees understand and appreciate the benefits provided.Furthermore, these authors suggest that workplace flexibility, in the form of alternate job arrangements, could assist in retaining employees, because arrangements such as job sharing, flextime and telecommuting could help employees to better balance their work and home lives.Recognition of performance is also stated by Rose (2011) as helping to reduce employee turnover, as well as improve the engagement and motivation of employees.Armstrong and Brown (2009, p. 53) stated that strategic rewards can assist an organisation in establishing itself as a 'great place to work', which will result in attracting highperforming individuals to work there and thereafter remain working there.The ability of rewards to recruit and retain can thus be incorporated as evaluation criteria for assessing the effectiveness of rewards, such as by measuring rates of involuntary turnover and the retention of high performance or key skill staff (Armstrong & Cummins, 2011).
Rewards and recognition have also been shown to lead to engagement, which comprises motivation, commitment and organisational citizenship and, in turn, leads to improved performance (Armstrong et al., 2010).Motivation can be improved by rewarding good performance and offering incentives (Grobler et al., 2011).These authors elaborate that organisations can make use of rewards such as promotions and merit increases, as well as intrinsic rewards such as goal accomplishment, to encourage continued high levels of performance from employees.

Research question
The above review of research literature points towards the need for the present study.With the exception of Nienaber et al.'s (2011) research, limited South African research exists that quantitatively explores the relationship between demographic variables and reward preferences.Research in the South African context is also lacking in terms of how employees perceive different reward categories to affect their levels of attraction, motivation and retention.This must be investigated so that South African reward packages can be tailored accordingly.
Therefore, based on the above, the research question for this study is: 'What preferences do South African employees hold in terms of rewards?'

Research approach
This study used a quantitative research approach.Primary data on the reward preferences of employees were obtained by using a structured questionnaire, based on Nienaber et al.'s (2011) reward preferences questionnaire (RPQ).The research was primarily descriptive in nature, as it aimed to understand and interpret the reward categories most preferred by employees, as well as which categories are most likely to attract, retain and motivate employees.

Research participants
Respondents consisted of employees from 10 medium-sized to large-sized Port Elizabeth-based companies and one Johannesburg-based company.The researchers made contact with the human resource managers, owners or managers of these companies and requested that they forward the questionnaire to employees within their companies via electronic mail.No particular industry was focused on, as non-probability convenience and snowball sampling was relied upon to collect the data.Convenience sampling was utilised because the researchers were in direct contact with the managers in the various firms and this was followed by snowball sampling when these managers passed the questionnaires on to further respondents.As stated by Malhotra (2010), subsequent respondents are selected in this way, based on referrals.A total of 250 respondents were obtained for this study in this manner.As the researchers relied on the managers of the various companies to disseminate the questionnaires, they were unable to ascertain the exact number of questionnaires distributed.The response rate was, for this reason, difficult to estimate, although it could be estimated that the questionnaire reached over 4500 potential respondents based on the sizes of the companies.
The biographical profile of the sample is summarised in Table 1.It is evident from this table that there was a fairly equal distribution of genders and that the majority of respondents earned over R20 000.00 per month, were White, married, had over four people in their household, held a postgraduate qualification and had been working for their current employer for over 10 years.Nienaber et al.'s (2011) RPQ was used as a base from which to achieve the objectives of this study.According to these authors, the RPQ was informed by the theory of the total rewards model and was originally tested on two pilot groups before being finalised.For the purposes and objectives of this study, the RPQ was modified with permission from Nienaber et al.Although the rewards categories remained the same as in Nienaber et al.'s questionnaire, the phrasing of the items was changed and the Likert scale was altered from a seven-point scale to a five-point scale.The modified questionnaire had two distinct sections that clearly measured both reward preferences and respondents' levels of satisfaction with their current rewards received, which differs from the original RPQ.Furthermore, whilst the original RPQ asked respondents to indicate one of the six reward categories that had the greatest impact on them with regard to attraction, retention and motivation, the modified questionnaire allowed respondents to rate, on a five-point Likert scale, the extent to which each of the six reward categories impacted on an organisation's ability to attract, retain and motivate them.

Measuring instrument
As a result of these modifications to the original RPQ, the questionnaire was piloted with 15 randomly selected respondents in industry.After completing the pilot study, suggestions from the respondents were noted and the questionnaire was amended accordingly.Based on the above insight into the development of the questionnaire, an explanation of the questionnaire's two sections will now be provided.
Section A -Demographics: Section A contained demographic variables in order to determine their influence on reward preferences.The items included gender, age, racial group, marital status, level of education, number of members in the household, household income per month, job level, job category and years of service.The response format consisted of a range of options for each item and respondents were asked to select the box which best represented their own personal characteristics.
Section B -Reward preferences: Section B consisted of three sub-sections that collected responses with regard to different reward categories.Section B1 aimed to determine the importance of the reward categories included in the total rewards mix model, as well as the levels of satisfaction that the respondents experienced with the rewards that they currently receive from their employers.This section was divided into six categories or dimensions of rewards, with a total of 42 items.Examples of items in each category are: • Base pay: 'A market-related salary'.A five-point Likert scale was used to measure both reward preferences (importance) and satisfaction.The scale ranged from 1 = 'Not important at all' to 5 = 'Extremely important' for reward preferences and 1 = 'Very dissatisfied' to 5 = 'Very satisfied' for satisfaction.
In Section B2, respondents were requested to indicate which of the six reward categories listed above would have an impact on an organisation's ability to attract, retain and motivate them.The aim of this sub-section was to understand the key elements in the total rewards model that would attract, retain and motivate talented and high performing employees.
Section B3 consisted of two open-ended, ranking questions.
The first question asked the respondents to list two rewards in order of importance, excluding salary, which they would include in their total rewards package.The second question asked the respondents to list the two most important factors, such as lifestyle, economic situation or family, which would affect their reward choices.The aim of these questions was to enable the respondents to express their attitudes and opinions and to assist the researchers in interpreting the answers to the structured questions in the previous sub-sections.

Research procedure
The questionnaire was distributed via electronic mail to companies in both Port Elizabeth and Johannesburg.A covering letter accompanied the questionnaire, emphasising that participation was voluntary and that the respondents would remain anonymous.The human resource managers, owners or managers of these companies were asked to distribute the questionnaire electronically to their members of staff.
Data for this study were collected in two ways.Originally, electronic questionnaires in Microsoft Word format were sent to respondents (Format 1).However, when the response rate for these questionnaires proved low because respondents had to email the questionnaires back to the researchers, it was decided to develop an online version of the questionnaire using an online survey tool to assist in increasing the response rate (Format 2).The advantages of using online questionnaires are that the response time is shorter and the data are collected into a central database, therefore saving time and resources associated with the data capturing process (Ilieva, Baron & Healey, 2001).
Respondents were given 2 weeks to complete the questionnaire and either return it to the researchers via electronic email (Format 1) or complete the questionnaire via the online survey tool (Format 2).A total of 124 responses were received via electronic mail from the respondents (Format 1) and a total of 126 responses were received from the online questionnaire (Format 2), bringing the total number of respondents to 250.Once the completed questionnaires were received in either of these formats, the responses were captured electronically in one central database and sent for statistical analysis.

Statistical analysis
Both descriptive and inferential statistics were used to analyse the data.Descriptive statistics in the form of frequency tables were used to calculate the reward categories' summated scores.In terms of inferential statistics, item analysis, analysis of variance (ANOVA) and multivariate analysis of variance (MANOVA) were used.Item analysis was conducted in two ways.Firstly, it was established whether the summated scores in the questionnaire had adequate internal consistency by calculating Cronbach's alpha coefficients.Secondly, a series of factor analyses with principal component analysis were used as the factor extraction technique for each set of items corresponding to a particular summated score individually.This was conducted to determine the significance of the factor loadings.MANOVA was used to determine biographical related variance to the importance of rewards.Univariate ANOVA was conducted to determine relationships between biographical variables that were found to be significantly related to the importance of rewards.Cohen's d statistic was then calculated where statistically significant differences were found, to determine practical significance.Content analysis was also conducted for the last two open-ended questions in the questionnaire, in order to extract the themes from responses.

Results
Item analysis determined the reliability of the summated scores for the six reward categories derived from the 42 items in Section B1 of the questionnaire for the importance and satisfaction scores, as well as the six items in Section B2 for the consolidated attract, retain and motivate scores.The first part of the item analysis consisted of establishing whether the summated scores demonstrated adequate internal consistency by calculating Cronbach's alpha coefficients, as previously explained.The values are reflected in Tables 2.
From Table 2, it can be seen that the Cronbach's alpha coefficient scores were almost all greater than 0.60, the recommended minimum value for reliability (Malhotra, 2010).
One exception was the importance score for contingency pay, with a value of 0.52.However, with this exception, the observed Cronbach's alpha coefficients confirmed the reliability of the reward categories' summated scores.
The second part of the item analysis consisted of a series of factor analyses, using principal component analysis as the factor extraction technique, for each set of items corresponding to a particular summated score individually. 1Almost all factor loadings were greater than 0.350, the minimum value deemed statistically significant according to the guidelines of Hair, Black, Babin, Anderson and Tatham (2006) for a sample of size n = 250.Only three loadings less than the threshold value were observed: 'A dedicated parking bay at work' (0.308) for benefits importance, 'Total control over work methods without a manager's interference' (0.284) for performance and career management importance and 'The ability to work flexible working hours' (0.323) for work-home integration satisfaction.Given the acceptable Cronbach's alpha coefficients recorded for all summated scores and the importance of not jeopardising content validity, the three items with loadings slightly less than the recommended minimum value were retained in calculating the summated scores.
The extent to which each of the six reward categories had an impact on an organisation's ability to attract, retain and motivate respondents is reflected in Table 3, in which the reward categories are ranked in order of importance based on mean scores.It is seen from this table that base pay (monthly salary or remuneration) is the reward that most strongly attracts and retains employees, whereas performance recognition and career management is the reward that most strongly motivates employees.
1.The table depicting the results of the factor analyses was not included in this article because of its large size.It is however available from the authors upon request.Descriptive statistics for the reward category summated scores are depicted in Table 4.It is important to note that the mean scores are in the range 1-5, corresponding to the Likert-scale used in the questionnaire but standardised so that higher (or lower) scores depict positive (or negative) perceptions.In Table 4, the factors are categorised into groups based on the significance of mean score differences as determined by a series of one-sample t-tests at the alpha = 0.05 significance level (Bonferroni adjusted) for statistical significance and Cohen's d statistics with a threshold value of 0.20 for practical significance.
Five of the importance scores obtained mean values greater than the lower limit of the positive interval of the scale (3.4 -5.0), with the value of 3.39 for work-home integration being only marginally below this level.The satisfaction mean values were all in the neutral interval (2.6 -3.4), with the exception of quality work environment obtaining a mean score of 3.41, only slightly higher than the upper limit of the interval.All the reward categories obtained mean scores in the positive interval in terms of their ability to attract, retain and motivate.
Inspection of the groups based on the significance of betweenfactor mean score differences, revealed the following with regard to the ranking of reward categories as perceived by respondents: • Importance: base pay is the most important reward factor, whilst contingency pay, performance and career management and quality work environment are the second most important, followed by benefits in third place and, finally, work-home integration.• Satisfaction: respondents were most and second-most satisfied with quality work environment and benefits, respectively.Grouped together in third place were performance and career management, work-home integration and base pay.The least satisfactory reward category was contingency pay.• Attract, retain and motivate: there were no significant differences amongst the top four rewards, that is, base pay, performance and career management, contingency pay and work-home integration.Benefits and quality work environment were single factor groups placed second and third.
It is worth noting from these results that whilst base pay was deemed the most important reward category and also placed first with regard to its impact to attract, reward and motivate employees, it came second last in terms of respondents being satisfied with this reward.
Multivariate ANOVA was conducted to determine which, if any, of the biographical variables are related to the importance of rewards construct.The biographical variables are listed in Table 5, a summary of the results of the first MANOVA iteration.It is evident that gender, age and job level have a statistically significant relationship with importance, because their p-values were lower than the significance value of 0.05.
The univariate ANOVA results for these three biographical variables that were found to be significantly related to importance are reflected in Table 6.Statistically significant relationships can be seen between gender and base pay, quality work environment and work-home integration, between age and base pay, contingency pay and work-home integration and between job level and benefits.
Table 7 indicates the statistics for those factors related to biographical variables according to the significant results shown in Table 6.A small practically significant difference can be seen when Cohen's d statistic falls in the interval of 0.20 < |d| < 0.50 and a moderate practically significant difference is found when Cohen's d statistic falls in the interval of 0.50 < |d| < 0.80.The following observations can be made based on the results depicted in Table 7: • Gender: men perceive base pay, quality work environment and work-home integration as less important than women.• Age: employees in the age group 18 -29 years regard base pay and contingency pay as less important than older employees.The younger the employee is, the more important he or she deems work-home integration.• Job level: the job levels ranked according to the perceived importance of benefits for the group are: administrative or clerical, followed by junior management, specialist or professional, general management and, finally, senior management.The differences amongst groups are significant, except for those between junior management, specialist or professional and general management.
Respondents were asked in the last two open-ended questions to list, in order of importance, the two most important rewards (excluding salary or base pay) that they would include in their total reward package and the two most important factors that would affect their reward choices.The results of the content analysis are depicted in Tables 8 and 9, ranked in order of importance based on the number of first-placed responses.It is evident from Table 8 that contingency pay, benefits and performance and career management are the top three rewards mentioned in terms of importance (similar to findings in Table 4).Family needs, lifestyle and financial security are the most important factors influencing the choice of rewards, according to Table 9.

Ethical considerations
The researchers were not required to process this research through the ethical clearance committee of their university, because of the fact that the research was not of a sensitive nature; respondents were not part of a vulnerable group such as children, nor were the respondents students or staff of the university.

Potential benefits and hazards
The measurement instrument utilised in this study posed no risk or danger to respondents.The data collected were stored securely, with only the researchers having access to it.Benefits will arise from this study if organisations around South Africa take note of the recommendations made by the researchers.

Recruitment procedure
Respondents were recruited for this research study via electronic mail, which included a description of the study.All respondents were aware that participation in this research study was voluntary and that they could withdraw from the research at any stage if they so wished.It was emphasised to respondents that all responses would be kept confidential.

Informed consent
The researchers clearly outlined the purpose of the study to respondents.Respondents provided their informed consent by means of completing the questionnaire, whether in Formats 1 or 2, as explained previously.

Data protection
All responses to the questionnaires were electronically captured and stored in a password-protected file that was only accessible to the researchers.

Trustworthiness Reliability
Internal consistency reliability assesses the reliability of a set of items when several items are summed to form a total score for a scale (Malhotra, 2010).In this line, Cronbach's coefficient alpha was calculated in order to determine the internal consistency of the modified questionnaire, as discussed previously.

Validity
Construct validity was confirmed as a result of the initial RPQ being developed out of theory pertaining to the total rewards model, which ensured that the items in the questionnaire were linked with literature pertaining to rewards.Furthermore, content validation was attained by means of expert analysis of the content of the modified questionnaire.The questionnaire was analysed by a specialist in the field of statistics, as well as two rewards professionals who are nationally renowned in this field.All three individuals approved the modified questionnaire.

Discussion
This research aimed to identify employees' rewards preferences, the degree of satisfaction experienced with   4 indicate that base pay is considered to be the most important reward category amongst the respondents, whilst work-home integration the least important reward category.This finding is not surprising, as pay has been cited as the main reason why individuals work (Price, 2011).Research conducted by Chiu, Luk and Tang (2002) also showed that base salary was listed as one of the five most popular reward components offered to employees, together with annual leave, paid sick leave, year-end bonuses and maternity leave.It is interesting to note though that when respondents were asked to list their two most important rewards that they would include in their total reward package excluding salary or base pay, contingency or variable pay and benefits were the most frequent responses (Table 8) and contingency or variable pay was also listed as second-most important in Table 4. Based on large amounts of research on contingency or variable pay plans, Lawler (2000) concluded that providing employees with such incentives increases the work behaviours that are rewarded.Yet, more recently, Armstrong and Brown (2009) have stated that academic studies frequently demonstrate that contingent pay, or pay for performance, actually demotivates rather than motivates employees.This is because most employees are average performers as opposed to top performers, which results in the majority of employees feeling hard done by when they do not receive salary increases based on performance.The researchers note that this is one of the problems with extrinsic forms of motivation, which will be discussed in greater detail in the 'Practical implications' section of this article.
The results in Table 4 also reveal that whilst base pay and contingency or variable pay were deemed to be important reward categories and were also placed first with regard to their impact on attracting, rewarding and motivating employees (in line with findings by WorldatWork, Scott, McMullen & Royal, 2012), respondents were the most dissatisfied with both of these categories.This suggests that organisations are not providing their employees with sufficient financial rewards to fund their needs outside of work, or to ensure that their chosen or desired lifestyles can be maintained.Interestingly though, quality work   environment was show to be the least important factor in attracting, retaining and motivating employees (Table 3); yet, respondents were most satisfied with this element of their total reward package (Table 4).This indicates that employers are not focusing their efforts on ensuring that employees are satisfied with those rewards that hold the greatest importance in attracting, motivating and retaining employees, such as base and contingency pay.
Table 3 showed that performance and career management was rated the most important reward in terms of motivating employees and rated the second-most important factor in retaining staff.This supports research conducted by WorldatWork ( 2010) that showed that performance and development rewards were most important in motivating and retaining employees.Therefore, although pay is one of the most important factors when it comes to attracting and retaining employees, organisations should not only use this element of rewards to attract, retain and motivate employees but rather non-financial elements as well (Hill & Tande, 2006).As stated by Chen and Hsieh (2006), high performing employees with valuable skills work for more than just money, looking for a job that can offer them an optimistic future, an opportunity for professional and personal growth, a positive working environment, as well as recognition.
The study also aimed to determine whether certain demographic variables were related to the importance of rewards.Table 5 indicates that gender, age and job level were found to be related to reward preferences, whereas marital status, education level and number of household members held no relationship with the importance of rewards.In terms of gender, the results in Table 7 revealed that women prefer base pay, quality work environment and work-home integration more than men.The fact that women place more importance on pay and a conducive working environment is in line with South African findings by Nienaber et al. (2011), yet is in contrast to Chow and Ngo's (2002) study that showed that both male and female employees consider high salary and good working conditions to be important in their jobs.The finding that women place more importance on integrating their work and home lives is also in contrast to findings by Paddey and Rousseau (2011), who found no gender differences in South Africa with regards to restructuring work to accommodate one's home life such as family responsibilities.Work-family conflict can have serious implications for organisations, such as poor performance, lower levels of commitment and loyalty and an increase in employee turnover rates (Fu & Shaffer, 2001;Thornthwaite, 2004), no matter whether experienced by men or women.This was confirmed in the present study, with work-home integration being found to be a key factor in the attraction, retention and motivation of employees.Additionally, in a 2009 job satisfaction survey, 46% of employees rated flexibility to balance life and work as very important (Society for Human Resource Management, 2009), regardless of gender.Thus, organisations cannot downplay the importance of workhome integration both to male and female employees.
Job level was found in this study to influence employees' preference for benefits, with administrative and junior management level employees exhibiting the highest preference for benefits.This finding correlates with Nienaber et al.'s (2011) South African study, in which it was discovered that employees occupying lower level jobs, such as administrative and junior management levels, had significantly higher preferences for remuneration and benefits when compared to senior and executive management.The researchers suggest that this could be because employees first need to be satisfied with pay and benefits before intrinsic rewards such as career management or workhome integration can be effective in motivating individuals.This is in line with Tang and West (1997), who stated that when people have enough money, they may then turn their attention to satisfying intrinsic needs such as recognition, achievement and career advancement.
With regard to age, the study showed that respondents in the age group 18-29 years (Generation Y) place less importance on base pay and contingency or variable pay than older respondents.This is in contrast to Nienaber et al. (2011), who found no significant differences in reward preferences amongst generations in South Africa, with all generations stating that salary and benefits are important aspects of rewards.Yet, Cennamo and Gardner (2008) found that older employees value extrinsic rewards, such as pay and benefits, more than younger employees, which is in line with the present study's findings.They argued that this may be as a result of the career stage of employees; that is, older employees may be at a stage in their working career where they receive higher salaries and benefits, therefore placing more importance on these types of rewards.
Younger employees, on the other hand, place more emphasis on rewards such as work-home integration, as revealed in Table 7.In support of this finding, Cennamo and Gardner (2008) discovered that employees who fall into the Generation Y category value freedom in their jobs, such as work-life balance, and prefer rewards that allow them to have a more flexible lifestyle.Owing to the changing demographics of the workforce, companies are becoming more aware of the importance of these types of benefits, especially for the younger generation (Milkovich et al., 2011).
The present findings do not agree with Giancola (2008) and Cennamo and Gardner's (2008) statements that employees prefer to select their benefits according to their life stage (for example, retirement benefits being less important to employees in their twenties but becoming more important as they age), because no significant age differences were found for benefits.However, based on the other generational results discovered, this study does support certain generational reward differences.
These results adequately answer the research question posed earlier, as they have demonstrated the preferences that South African employees hold in terms of rewards.

Practical implications
Base pay and contingency or variable pay have been found to be classified as extremely important by respondents, as well as rewards that influence the retention of employees.This finding is in line with Rose (2011), who stated that individuals naturally indicate a preference for cash (in this case, base and contingency pay) over a non-cash award, such as work-home integration.Yet, respondents in this study were most dissatisfied with these rewards; and many managers in South Africa become reluctant to provide wage increases during periods of poor economic growth (Meyer & Kirsten, 2012), such as the after-effects of the recession presently being experienced.It can thus be recommended that employers place more emphasis on the design of their base and contingency pay, in order both to satisfy the financial needs of their employees and retain high performing employees.Organisations can make use of wage surveys to determine the 'going rate' in the South African labour market (Grobler et al., 2011) to achieve this.
However, Rose (2011) explained that non-cash awards, such as retail shop vouchers, magazine subscriptions or tickets to sporting events, should in fact be emphasised in the workplace for their motivational ability, instead of cash incentives.This is because the tangible effect of such a non-cash award will last longer than a cash award and such awards can be tailored to the interests and needs of employees, thus making them more personal.Furthermore, contingent pay in the form of incentives often results in the majority of employees not receiving the incentive and thus being demotivated instead of motivated.Recognition of achievement should thus be used as a motivator instead, provided it is differentiated from an employee's base pay and the message conveyed is one of celebrating success.This will reinforce current effective performance as opposed to incentivising performance through a typical carrot-andstick incentive plan (Rose, 2011).As stated by this author, incentives extrinsically motivate employees, resulting in 'the danger of destroying some intrinsic motivation by putting a price on it through cash' (Rose, 2011, p. 109).Organisations need to bear this in mind when developing incentive and recognition programmes.
The results of this study provide useful guidelines for organisations to structure their reward packages according to the needs and preferences of their employees based on factors such as gender and age, for example by emphasising work-home integration to women and younger employees, or providing flexible or 'cafeteria' styles of benefit arrangements to employees of differing job levels (see Meyer & Kirsten, 2012).This emphasises the fact that traditional forms of rewards are no longer solely applicable in today's diverse workforce and employers need to understand that rewards now encompass the overall value that a company offers its employees.Furthermore, family needs, lifestyle, financial security and career ambitions were all found to be important factors affecting the reward choices of employees.
A rewards programme that does not take such examples of employees' needs into account may decrease the motivation levels of employees (Salie & Schlechter, 2012).

Implications for future research
Literature provides limited research on reward preferences and segmentation of workforces in terms of rewards, other than flexible benefit packages.It is also sparse concerning research on the impact that demographic variables have on reward choices.The researchers thus recommend that future studies focus on the role that race, job level and income level play on reward preferences in South Africa.South Africa is a multicultural country with employees of varying socio-economic status, which makes it difficult for employers to cater for the different needs and lifestyle of its workforce.Demographic variables that are of relevance to South Africa, such as those mentioned, will provide a better indication of how reward packages should be developed and implemented in this country.
Additionally, research that focuses more specifically on reward preferences amongst generations, such as Generation X and Y differences, would be useful in today's working environment where organisations are dealing with a workforce composed of four different generations.This is in line with Giancola (2006), who specified that generational research is lacking in academic journals, which incorrectly indicate to experts that the concept lacks long-term value.
This study did not investigate two other important influences on the attractiveness of elements in total rewards packages, namely cultural and personality influences, which were briefly discussed in the literature review.These aspects have been empirically measured in previous studies internationally but have been largely neglected within South African research.Culture and personality thus indicate avenues for future research within a multicultural country such as South Africa.

Limitations of the study
As the researchers had to rely on managers in various companies to disseminate the questionnaires, they had little control over the composition of the sample or the industry to which the respondents belonged.They were also unable to compute an accurate response rate, as it was unable to be determined how many questionnaires were distributed.
Although the respondents were provided with written instructions, misunderstandings may also have taken place.
Categories of generations differ from culture to culture and nation to nation and the use of the existing cohort model may be inadequate for a South African analysis.Further research is thus required into appropriate South African categories.Additionally, the influences of culture and personality on reward preferences was introduced in the literature review, but were not investigated empirically in this study.Finally, a range of variables were analysed in relation to reward preferences, but a more focused analysis of specific variables in relation to reward preferences is necessary in future research.

Conclusion
Creating a holistic reward strategy is a far greater challenge than in the past, where a company's only concern was to offer a market-related salary (Gross & Friedman, 2004).
Today, a 'one-size-fits all' or 'one-dimensional' approach with regard to rewards is no longer effective (see Chiang & Birtch, 2005;Giancola, 2008;Mercer, 2008).Organisations need to take information about their employees into consideration in order to develop effective reward strategies.
Providing employees with the opportunity to vary their remuneration package to meet their specific needs can have a motivational impact and positively influence the company's competitive position in the labour market (Marchington & Wilkinson, 2005).Similarly, an understanding of individual reward preferences is needed if companies are to build effective reward systems that will attract, retain and motivate a skilled and high performing workforce (Lawler, 2000).Organisations must begin to focus on allocating their budgets in ways that not only increase return on investment and provide high levels of value to employees, but that also drive the required behaviours in employees (WorldatWork & Towers Watson, 2012).Thus, organisations should develop a total rewards approach in which both pay and nonpay reward elements are brought together to engage staff, whilst simultaneously segmenting this total reward approach to take key employee differences into account (Armstrong et al., 2010).

TABLE 1 :
Biological profile of the study sample.

TABLE 2 :
Internal consistency statistics for reward categories' summated scores.

TABLE 3 :
The impact of the reward categories on organisations' ability to attract, retain and motivate employees.

TABLE 4 :
Descriptive statistics for the reward category summated scores.

TABLE 5 :
Multivariate analysis of variance for the relationship between biographical variables and the importance of reward -Iteration 1.

TABLE 6 :
Univariate analysis of variance results.
df, degrees of freedom; F, variance of the group means; p, level of significance.Bold figures indicate p < 0.05.

TABLE 7 :
Statistics for factors related to biographical variables.

TABLE 8 :
Importance of rewards in total reward package.

TABLE 9 :
Importance of factors affecting employees' reward choices.