THE VALIDATION OF TWO BURNOUT MEASURES IN THE 
SOUTH AFRICAN EARTHMOVING EQUIPMENT INDUSTRY 
A.M. le Roux, Hons.B.Com. 
Mini-dissertation submitted in partial fulfilment of the requirements for the degree 
Magister Commercii in Industrial Psychology at the 
North-West University, Potchefstroom Campus 
Supervisor: Dr. K. Mostert 
November 2004 
Potchefstroom 
COMMENTS 
The reader should keep the following in mind: 
The editorial style as well as the references referred to in this mini-dissertation follow the 
format prescribed by the Publication Manual (4'h edition) of the American Psychological 
Association (MA). This practice is in line with the policy of the Programme in Industrial 
Psychology of the North-West University, Potchefstroom Campus, to use the APA style in 
all scientific documents as h m  January 1999. 
The mini-dissertation is submitted in the form of a research article. The editorial style 
specified by the South African Journal of Industrial Psychology (which agrees largely with 
the APA style) is used, but the M A  guidelines were followed in constructing tables. 
ACKNOWLEDGEMENTS 
'Thank you" is a very small word, but with tremendous meaning. Allow me to thank all the 
people who supported, guided and motivated me. Without the support of all these people, I 
would not have accomplished my dream to write a mini-dissertation. Thank you, from the 
bottom of my heart, to.. . 
My Lord Jesus, for leading my path to this great experience and blessing me with the 
ability to complete this project. 
Dr. Karina Mostert, my mentor and supervisor, who was never too busy to lend a helping 
hand. Thank you for your guidance, motivation, inspiration and patience. I have learned 
so much from you. Thank you for being the wonderful person you are. 
A very special person in my life, Neil. Thank you for all your love, support and 
encouragement the past three years. Thank you for always being there. and believing in 
me. 
Mom and Dad, for always being there for me and supporting me through everythmg in 
my life thus far. Thank you that I know you are just a phone call away when I need you. 
Thank you for all your love and guidance. Also, to my brother and sister for all your love 
and support. 
All of my fiiends who listened to all the struggles and who supported me. A special 
thanks to Liana, Michkl and Sherril. Thank you for being the greatest fiiends there are. 
Izel and Michkl, for their help and hard work in collecting and capturing the data. 
All the employees of the eaxthmoving equipment industry who cooperated and set time 
aside to participate in this study. 
Dr. K. Mostert, for the careful work she did in preparing my statistical processing. 
Mr. W. Cloete, for the professional manner in which he conducted the language editing. 
The financial assistance of the National Research Foundation (NRF) towards this research is 
hereby acknowledged. Opinions expressed and conclusions arrived at are those of the author and 
are not necessarily to be attributed to the National Research Foundation. 
TABLE OF CONTENTS 
List of Tables 
Abstract 
Opsomming 
CHAPTER 1: INTRODUCTION 
hoblem statement 
Research objectives 
General objective 
Specific objectives 
Research method 
Research design 
Participants 
Measuring battery 
Statistical analysis 
Overview of chapters 
Chapter summary 
References 
CHAPTER 2: RESEARCH ARTICLE 
CHAPTER 3: CONCLUSIONS, LIMlTATIONS AND 
RECOMMENDATIONS 
6.1 Conclusions 
6.2 Litations of this research 
6.3 Recommendations 
6.3.1 Recommendations for the organisation 
6.3.2 Recommendations for future research 
vi 
vii 
References 
LIST OF TABLES 
Table Description 
Research Article 1 
Table 1 Characteristics of the Participants 
Table 2 Item-bias analysis for the MBI-GS 
Table 3 Item-bias analysis for the OLBI 
Table 4 Pattern matrix of the 15-item MBI-GS for the African group and the 
Engllsh group 
Table 5 Pattern matrix of the 14-item MBI-GS for the African group and the 
English group 
Table 6 Pattern matrix of the 16-item OLBI for the African group and the English 
group 
Table 7 Pattern matrix of the 16-item OLBI for the African group and the English 
group 
Table 8 Descriptive Statistics and Alpha Coefficients of the MBI-GS and the OLBI 
(n = 326) 
Table 9 Correlation Coefficients between the MBI-GS and the OLBI 
Table 10 MANOVAS - Differences in Burnout Levels of Demographic Groups 
Table 11 Differences between Employees in an Earthmoving Equipment Industry, 
based on Gender 
Page 
ABSTRACT 
Title: The validation of the Maslach Burnout Inventory - General S w e y  and the 
Oldenburg Burnout Inventory in the South &can Earthmoving Equipment Industry 
Kw terms: Burnout, Maslach Burnout Inventory - General S w e y  (MBI-GS), Oldenburg 
Burnout Inventory (OLBI), validity, reliability, construct equivalence, item bias, 
earthmoving equipment industry, construction workers, construction industry. 
The objectives of this study were to validate the Maslach Burnout Inventory - General Survey 
(MBI-GS) and the Oldenburg Burnout Inventory (OLBI), to determine the construct equivalence 
and item bias of these instruments for employees in an earthmoving equipment industry in South 
&ca of different language groups, to determine the correlation between the MBI-GS and the 
OLBI, and to investigate the relationship between burnout and various demographic 
characteristics. A random sample of 326 employees in an earthmoving equipment industry of 
eight provinces in South Africa was taken. The MBI-GS, the OLBI and a biographical 
questionnaire were used as measuring instruments. Cronbach alpha coefficients, inter-item 
correlation coefficients, Pearson product-moment correlation coefficients and exploratory factor 
analysis were used to analyse the data 
The results showed that three factors of the MBI-GS and two factors of the OLBI exist for 
different language groups. Exploratory factor analysis with target rotations confirmed the 
construct equivalence of the three factors of the MBI-GS and the two factors for the OLBI for 
different language groups. No evidence was found for uniform or nonuniform bias of the items 
of the MBI-GS or the OLBI for different language groups. Results indicated that no differences 
with respect to levels of burnout exist between various demographic characteristics. 
Recommendations were made for further research. 
Titel: Die validering van die Maslach-Uitbrandingmaelys - Algemene Opname en die 
Oldenburg-Uitbrandingsvraelys in die Suid-Afrikaanse Grondverskuiwings- 
toerustingindustrie 
Sleutelterme: Uitbranding, MaslacbUitbrandingsvraelys - Algemene Opname, Oldenburg- 
Uitbrandingsvraelys (OLBI), betroubaarheid, geldigheid, konstrukekwivalensie, 
itemsydigheid, grondverskuiwingstoerustingindustrie, konstruksiewerkers, 
konstruksie-industrie. 
Die doelstellings van hierdie studie was om die Maslach-Uitbrandingsvraelys - Algemene 
Opname (MI3I-GS) en die Oldenburg-Uitbrandingsvraelys (OLBI) te valideer vir werknemers in 
'n grondverskuiwingstoerustingindustrie, om die konstrukekwivalensie en itemsydigheid daarvan 
vir verskillende taalgroepe te bepaal, die korrelasie tussen die MBI-GS en die OLBI te bepaal, 
amok om die verhouding tussen uitbranding en verskillende demografiese groepe te bepaal. 'n 
Ewekansige steekproef (n= 326) is geneem onder werknemers in 'n 
grondverskuiwingstoerustingindustrie in agt provinsies. Die MBI-GS, die OLBI en 'n 
biografiese vraelys is as meetinstrumente gebruik. Cronbach alfakoEffisihte, interitem- 
korrelasiekoeffisihte, Pearson-produkmoment-korrelasiekoeffisihte en eksploratiewe 
faktoranalise is gebruik om die data te ontleed. 
Die resultate het getoon dat drie faktore van die MBI-GS en twee faktore van die OLBI vir 
verskillende taalgroepe bestaan. Geen bewyse is gevind vir uniforme of nie-uniforme sydigheid 
van die items van die MBI-GS of van die OLBI vir verskillende taalgroepe nie. Bevindinge het 
getoon dat daar gem verskille ten opsigte van vlakke van uitbranding tussen verskillende 
demografiese groepe voorkom nie. 
Aanbevelings vir toekomstige navorsing is am die hand gedoen. 
vii 
CHAPTER 1 
INTRODUCTION 
This mini-dissertation focuses on the validation of the Maslach Burnout Inventory - General 
Survey and the Oldenburg Burnout Inventory in a South African earthmoving equipment 
industry. 
This chapter contains the problem statement and a discussion of the research objectives, in which 
the general objective and specific objectives are set out The research method is explained and 
the division of chapters given. 
1.1 PROBLEM STATEMENT 
Burnout is recognised as a meaningful phenomenon of the modem age and the term began to 
appear frequently in the 1970s (Maslach, Schaufeli & Leiter, 2001). Maslach and Jackson 
(1986), defined burnout as a syndrome of emotional exhaustion, depersonalisation and reduced 
personal accomplishment, which can occur among individuals who do ''people work" of some 
kind. However, a "working" definition of burnout was given by Schaufeli and Enzmann (1998, 
p. 36), in which the burnout phenomenon was described as "a persistent, negative, work-related 
state of mind in 'normal' individuals, which is primarily characterised by distress, a sense of 
reduced effectiveness, decreased motivation, and the development of dysfunctional attitudes and 
behaviours at work". 
Many experts regard burnout as the h a l  step of a range of unsuccessful attempts to cope with 
stressful conditions (Gold, 1984). According to Schaufeli and Enzmann (1998), burnout can be 
considered as a particular kind of prolonged stress. Thus, burnout can be seen as a final stage in 
a breakdown in adaptation, which results from the long-term imbalance of demands and 
resources and is accompanied by chronic malfunctioning at work (Lazarus & Folkman, 1984). 
The concept of burnout has been researched and studied in countries all over the world (Maslach 
et al., 2001). Burnout is also the subject of numerous books (e.g. Edelwich & Brodsky, 1980), 
magazine and newspaper reports (e.g. Longcope, 1982), and professional articles (e.g. Thornton, 
1992), indicating the importance of this phenomenon. 
The initial work on burnout developed out of the occupational sector of human services and 
education (Iviaslach et al., 2001). Thereafter, burnout was studied in various occupational fields 
- both nationally and internationally. International research on burnout has been carried out in 
various occupations, including policing (Biggam, Power, MacDonald, Carcary & Moodie, 1997; 
Kop, Euwema & Schaufeli, 1999), human services (Anderson, 2000) and the teaching profession 
@learns & Cain, 2003; Weisberg & Sagie, 1999). 
In South Africa, extensive research has been carried out on burnout of police officers (Storm, 
2002), emergency workers (Naudt & Rothmann, 2004), pharmacists (Rothmann, Malan & 
Rothmann, 2001), psychiatric nurses (Levert, Lucas & Ortlepp, 2000), managers in the 
manufacturing industry (Jansen van Vuuren & Rothmann, 2002), pharmacist's assistance 
(Iviostert, Van Rensburg & Rothmann, 2004), academic staff in Higher Education Institutions 
(Barkhuizen & Rothmann, 2004), student leaders in Higher Education Institutions (Sieberhagen 
& Pienaar, 2004) as well as tertiary students (Jackson, Mostert & Pienaar, 2004). However, it 
seems that not much research has been done into the experience of burnout of employees in the 
earthmoving equipment industry. 
The earthmoving equipment industry includes the building of schools, office buildings, shopping 
centres, dams, bridges, sewer treatment facilities, industrial plants, roads and highways 
(Applebaum, 1981). This industry has also become progressively larger over the last four 
decades (Singh, 1997). Earthmoving plays an important role in industries such as construction 
and mining that are constantly under pressure to improve productivity, efficiency and safety 
(Singh, 1997). Furthermore, construction work can lead to physical fatigue, which, in turn, may 
lead to reduced productivity and motivation, as well as job dissatisfaction and safety risks, 
accidents and injuries (National, 2000). The construction industry is also a very demanding and 
stressful work environment and many job-related stressors are likely to be pertinent to the work 
of manager, supervisors and professionals engaged in the industry (Lingard, 2003; Smallwood, 
1997). Many worksites are hazardous for humans - unstructured and outdoor environments, 
poor visibility conditions and inclement weather conditions (Singh, 1997). According to Gold 
(1984), work-related stressors such as these can cause burnout. 
It is important to use valid and reliable instruments when burnout is measured in the earthmoving 
equipment industry. In most of the research that has been done on burnout, the Maslach Burnout 
Inventory (MBI) (Maslach & Jackson, 1986) was used (Schaufeli & Emmann, 1998). The MBI 
encompassed all three of the core dimensions of burnout w l a c h  et al., 2001). However, with 
the growing body of empirical research, adaptive versions of the original MBI were developed. 
According to Maslach et al. (2001), the MBI - Human Services Survey (MBI-HSS) was 
developed for use by people working in the human services and health care. A second version 
was developed for use by people working in educational settings, namely the MBI - Educators 
Survey (MBI-ES). However, research showed that burnout is no longer restricted to the caring 
professions - it also exists outside the human services (Maslach & Leiter, 1997). The apparent 
need for an instrument that measures burnout in non-contactual professional context was met by 
the introduction of a third version of the MBI, the MBI - General Survey (Schaufeli, Leiter, 
Maslach & Jackson, 1996). The MBI-GS can be used in my occupational context and assesses 
parallel dimensions to those contained in the on@ MBI, except that the items do not explicitly 
refer to working with people (Schaufeli & Enzmann, 1998). 
In the helping professions (including education), the three dimensions of burnout are 
distinguished as emotional exhaustion, depmnalisation md low personal accomplishment 
(Maslach, et al., 1996). Since the employees of the earthmoving equipment industry are 
functioning in a non-helping profession, the study will focus on exhaustion, cynicism and 
professional efficacy (as it is measured with the MBI-GS): 
Exhaustion refers to feelings of fatigue, without direct reference to people as the source 
of those feelings. 
Cynicism can been seen as an indifference or an aloof attitude towards one's work in 
general, where the items refer to work itself rather than to recipients of one's services or 
to personal relationships at work. 
Professional Efficacy encompasses both social and non-social accomplishments at work. 
The three dimensions of the MBI-GS are interrelated. Cynicism is highly related to Exhaustion 
(0,44 < r < 0,61), and also strongly related to Professional Efficacy (-0,38 < r < -0,57) (Schaufeli 
et al., 1996). In light of the above, it appears that burnout can validly be measured across a range 
of different occupations, despite the difference between human service occupations and non- 
contactual professions. 
The MBI-GS has evidenced satisfactory internal consistencies ranging fiom 0,73 (Cynicism) to 
0,91 (Exhaustion) (Leiter & Schaufeli, 1996). Reliability analyses showed that the Exhaustion 
and Professional Efficacy subscales were sufficiently internally consistent, but that one Cynicism 
item should be removed in order to increase internal consistency beyond the criteria of 0,70 
(Klopper & Mostert, 2004; Schutte, Toppinen, Kalimo & Schaufeli, 2000; Storm, 2002). 
Various studies were found in South Atiican samples where the MBI-GS was used. In a sample 
of senior managers in a manufacturing industry, Rothmann and Jansen van Vuuren (2002) found 
satisfactory alpha coefficients (Exhaustion = 0,79; Cynicism = 0,84 (after item 13 has been 
omitted); Professional Efficacy = 0,84). Rothmann and Malan (2003) reported higher alpha 
coefficients (Exhaustion = 0,89; Cyrucism = 0,76; Professional Efficacy = 0,85), while 
Rothmann, Jackson and Kruger (2003) found lower alpha coefficients for Cynicism ( a  = 0,72, 
&er item 13 has been omitted) and Professional Efficacy (a = 0,69). Storm and Rothmam 
(2003) found alpha coefficients of 0,88 for Exhaushushon, 0 78 for Cyncism and 0,79 for 
Professional Efficacy. With this taken into account, it can be expected that the MBI-GS will be 
sufficiently internally consistent. 
Although thepsychometric properties of the MBI-GS were tested .in South AiZcan samples (e.g. 
police officers, Stom, 2002; and teachers, Jackson & Rothmann, 2004), the Maslach Burnout 
Inventory - General Survey (MBI-GS) is not yet validated for employees in a South Atiican 
earthmoving industry. Consequently, it is also difficult to identifi burnout at an early stage and 
due to lack of South AiZcan norms for the MBI-GS, the implementation of intervention 
programmes is hampered. 
Although the Maslach Burnout Inventoryremains themost popular, and almost universally u s 4  
instrument to assess burnout (Schaufeli & Enzmam, 1998), an alternative measure of burnout 
that can be used is the Oldenburg Burnout Inventory (OLBI) (Demerouti, Bakker, Vardakou & 
Kantas, 2003). The OLBI was developed in Germany, whereas the MBI-GS was initially 
developed for English-speaking populations (Demerouti et al., 2003). The OLBI was 
constructed and validated among different occupational groups (Demerouti & Nachreiner, 1998) 
at the same time that the MBI-GS was developed. The OLBI is based on the theoretical work of 
Chemiss (1980) and Hall (1976) and can be used among occupations within and outside human 
service profRssons with consistent central elements of the burnout construct (Demerouti, 
Bakker, Nachreiner & Ebbinghaus, 2002). 
The OLBI conceives burnout as a syndrome of work-related negative experiences, conqxising 
two dimensions: feelings of exhaustion and disengagement fhm work (Demerouti et al., 2002, 
2003). E&amtioen is dehed as a consequence ofprolonged and intense physical, affective and 
cognitive strain as the result ofprolonged exposure to specific working conditions (or stressors). 
This conceptualisation corresponds to definitions of exhaustion as proposed by Aronson, Pines 
and Ka&y (1983), Lee and Ashfort (1993), and Shirom (1989). In contrast to exhaustion as it is 
operationalisedin the MBI-GS, the OLBI covers not only affective aspects (e.g. feelings of being 
emotionally drained), but also physical and cognitive aspects of exhaustion (e.g. the need for 
long resting time). This make the OLBI more applicable both to those workers who perform 
physical work and to those whose job is mainly aboutprocessing information, especially because 
there is something more related to fdings of exhaustion and not just emotions (Demerouti et al., 
2002,2003). 
The OLBI's disengagement scale refers to distancing oneself fiom one's emotions regarding the 
work task (e.g. uninteresting, no longer challenging, "disgusting'> and to a devaluation and 
mechanical execution o f  one's work (Demerouti et al., 2002). Demerouti et al. (2003) refer to 
disengagement as distancing oneself h m  one's work and experiencing negative attitudes toward 
the work object, work content, or one's work in general, whereas depersonalisation in the 
original MBI refers to distancing oneself emotionally fiom sem'ce recipients. Furthermore, 
disengagement items concern the relationship between the employee and his or her job, 
especially regarding engagement and identification @emerouti et al., 2003). Thus, 
disengagement and depersonalisation differ in their content and the objects to be valued 
(Demerouti et al., 2003). 
Professional Efficacy is not considered as a separate burnout dimension (Demerouti et al., 2003). 
According to Leiter (1993), emotional exhaustion leads to depersonalisation, whereas feelings of 
reducedpersonal accomplishment - that parallel professional efficacy - develop independently. 
Green, Walkey and Taylor (1991) argue that emotional exhaustion and depersonalisation 
constitute the core dimensions o f  burnout, whereas feelings of reduced personal 
accomplishments may also be intqreted as a possible consequence of bumout (Shirom, 1989). 
A meta-analysis of the correlates of the burnout dimensions confinus the independent role of 
professional efficacy (Lee & Ashforf 1996). 
From a psychometric p o d  of view, Guilford (1954) argues that one-sided scales (such as the 
MI-GS, where all the exhaustion and cynicism items are phrased negatively and all the 
professional efficacy items are phrased positively) are inferior to scales that include both 
positively and negatively worded items. The OLBI includes both negatively and positively 
worded items. 
The OLBI has been constructed and validated in an independent study among 293 Geman 
employees h m  different occupational fields, including human services and blue-collar workers 
(Ebbinghaus, 1996). A factor analysis confirmed the two-factor structure of the bumout 
construct (Ebbinghaus, 1996). Discriminate and convergent validity for both scales were 
examined by assessing their relationships with the subscales of the BMS (in German: BeZmhing 
Monotonic SitngUng), a reliable and valid German questionnaire for measuring short-term stress 
reactions at work. Furthermore, a series of first and second order factor analyses supported the 
OLBI's discriminate validity. Ebbinghaus' (1996) study supported the convergent validity of  the 
OLBI by showing that both burnout dimensions were only related to the conceptually most 
relevant constructs. Exhaustion was significantly related to psychological fatigue (r = 0,53,p < 
0,05), but not to satisfaction (r = 0,0Q), whereas disengagement was si&ificantly related to 
satisfaction (r = 0,53,p < 0,05), but not topsycholog!'cal fatigue (r = -0,lO). 
Apart h m  a study in Greece (Demerouti et al., 2003), both burnout instruments (the MBI-GS 
and the OLBI) have not yet been validated on the same population. Also, no other studies were 
found in the literature where burnout was measured by means of the OLBI. The OLBI has not 
yet been validated in other countries or in an earthmoving equipment industry. This is therefore 
the first study in South Aiiica to validate the OLBI and to investigate whether both the MBI-GS 
and the OLBImeasure the same bumout construct. 
When bumout measures are applied to different cultural groups, issues of measurement 
equivalence and bias become important (Van de Vijver & Tanzer, 1997). Equivalence and bias 
of  measuring instruments should be computed in all studies that take place in a multicultural or 
cross-cultural context (Van de Vijver & Leung, 1998). 
Van de Vijver and Leung (1997) make a hierarchical distinction of three types of equivalence. 
The first type, namely construct equivalence, indicates the extent to which the same construct is 
measured across all cultural groups studied. When an instrument measures different constructs 
in different cultures, i.e. when cultural equivalence exists, no cornparion can be made. The 
same construct is measured in case of  construct equivalence (also labelled structural 
equivalence). The second type of equivalence is called measurement unit equivalence and can 
be obtained when metric measures have the same measurement unit but have different origins. 
The third type of equivalence is called scalar equivalence and can be obtained when two metric 
measures have the same measurement unit and the same origin. Equivalence cannot be assumed 
but should be established and nported in each study (Van de Vijver & Leung, 1997). Construct 
equivalence is the most &qua fly studied type of  equivalence. 
Item bias should also be computed. An item is an unbiased measure of a theoretical construct 
(for example burnout) ifpersons h m  different cultural groups who are equally burned out have 
the same average score on the item (Van de Vijver & Leung, 1997). Persons with an equal 
standing on the theoretical construct underlying the instrument should have the same expected 
score on the item, irrespective of group membership. The definition of bias does not stipulate 
that the averages of cultural groups should be identical, but only that these averages should be 
identical across culturalgroups forpe~sons who are equally burned out. 
Item bias can be produced by sources such as incidental differences in appropriateness of the 
item content and inadequate item formulation. Bias will lower the equivalence of a measuring 
inshument. Two types of item bias are distinguished, namely uni'fom bias and nonuni'form bias 
(Van de Vijver & Leung, 1997). Uni'form bias refers to influences of bias on scores that are 
more or less the same for all score levels, whereas influences that are not identical for all score 
levels are referred to as nonuni'fom bias. 
Finally, the influence of demographic variables on burnout is examined The relationship 
between burnout on the one hand and gender, race, age, language, educational level, business 
unit and household situation on the other hand is concentrated on in this study. 
The following research questions emerge h m  the above-mentionedproblem statement: 
0 What are the reliability and validity of the MBI-GS for employees in the earthmoving 
equipment industry in South Atiica? 
What are the reliability and validity of the OLBI for employees in the earthmoving 
equipment industry? 
0 What are the construct equivalence and item bias of the MBI-GS? 
0 What are the construct equivalence and item bias of the OLBI? 
How do the MBI-GS and the OLBI correlate? 
What is the relationship between burnout and various demographic characteristics? 
What recommendations could be made for future research? 
1.2 RESEARCH OBJECTIVES 
The research objectives consist of a general objective and specific objectives. 
1.2.1 General objectives 
The general objective of this research is to validate the MBI-GS and the OLBI and to determine 
the construct equivalence and item bias of these instruments for employees in an earthmoving 
equipment industry in South Africa. 
1.2.2 Specific objectives 
The specific objectives of the research are the following: 
To determine the reliability and validity of the MBI-GS for employees in the 
earthmoving equipment industry in South Afiica. 
To determine the reliability and validity of the OLBI for employees in the earthmoving 
equipment industry in South Africa 
To determine the construct equivalence and item bias of the MBI-GS. 
0 To determine the construct equivalence and item bias of the OLBI. 
To determine the correlation between the MBI-GS and the OLBI. 
0 To investigate the relationship between burnout and various demographic characteristics 
To make recommendations for future research. 
1.3 RESEARCH METHOD 
The research method consists of a literature review and an empirical study. The results obtained 
h m  the research will be presented in an article format. Because separate chapters were not 
targeted for literature reviews, this paragraph focuses on aspects relevant to the empirical study 
that was conducted. The reader should note that a brief literature review was compiled for the 
purpose of the article. 
1.3.1 Research design 
A cross-sectional survey design is used to reach the objectives of this research. Cross-sectional 
designs are used to examine groups of subjects in various stages of development simultaneously, 
while the survey describe a technique of date collection in which questionnaires are used to 
gather data about an identified population (Bums & Grove, 1993). This design is well suited to 
the descriptive and predictive functions associated with correctional research, whereby 
relationships between variables are examined (Shaughnessy & Zechmeister, 1997). 
1.3.2 Participants 
Random samples (n = 326) were taken h m  earthmoving companies in Limpopo, Gauteng, 
Mpumalanga, Northern Cape, Western Cape, Eastern Cape, KwaZulu-Natal and North West. 
The following formula proposed by Kerlinger and Lee (2000) is used to determine the sample 
size for this study: 
and 
where n' = estimated sample sue; n = the estimated sample size using the formula; N = the sue 
of the population; z = standard score corresponding to the specified probability of risk, SD = the 
standard deviation of the population, and d = the specified deviation. 
The values for z, SD and d were determined, based on previous studies of bumout in South 
Aliica (Storm & Rothmann, 2003). 
1.3.3 Measuring battery 
Two questionnaires are used in the empirical study, namely the Maslach Burnout Inventory - 
General Survey (Ml3I-GS) (Schaufeli et al., 1996) and the Oldenburg Burnout Inventory (OLBI) 
Pemerouti et al., 2000). 
0 The Maslach Burnout inventory - General Survey (MBI-GS) (Schaufeli et al., 1996) has 
three subscales: Exhaustion (Ex) (five items, e.g. "I feel used up at the end of the 
workday"), Cynicism (Cy) (five items, e.g. "I have become less enthusiastic about my 
work") and Professional Efficacy (PE) (six items, e.g. "In my opinion, I am good at my 
job"). Together, the subscales of the MBI-GS provide a three-dimensional perspective on 
burnout. Internal consistencies (Cronbach coefficient alphas) reported by Schaufeli et al. 
(1996) varied from 0,87 to 0,89 for Exhaustion, 0,73 to 0,84 for Cynicism and 0,76 to 
0,84 for Professional Efficacy. Test-retest reliabilities after one year were 0,65 
(Exhaustion), 0,60 (Cynicism) and 0,67 (Professional Efficacy) (Schaufeli et al., 1996). 
AU items are scored on a 7-point muency-rating scale, ranging from 0 to 6. High scores 
on Exhaustion and Cynicism, and low scores on Professional Efficacy are indicative of 
burnout. Stom (2002) confirmed the 3-factor structure of the MBI-GS in a sample of 2 
396 members of the South a c a n  Police Service (SAPS), but recommended that Item 13 
be dropped from the questionnaire. She confirmed the structural equivalence of the MBI- 
GS for different race groups in the SAPS. The following Cronbach alpha coefficients 
were obtained for the MBI-GS: Exhaustion: 0,88; Cynicism: 0,79; Professional Efficacy 
0,78 (Storm, 2002). 
The Oldenburg Burnout Inventory (OLBI) (Demerouti et al., 2000) was constructed and 
validated in an independent study among 293 employees from different occupational 
fields, including human services and blue-collar workers (Ebbinghaus, 1996). The 
inventory measures burnout independent of vocational aspects on two dimensions, 
namely exhaustion and disengagement. The seven items of the Exhaustion subscales are 
generic and refer to general feelings of emptiness, overtaxing from work, a strong need 
for rest, and a state of emotional exhaustion. Examples are: 'After my work, I regularly 
feel worn out and weary', and 'After my work, I regularly feel totally fit for my free-time 
activities' (reversed). Disengagement refers to distancing oneself from one's work (work 
object and content), and to negative, cynical attitudes and behaviours towards one's work 
in general. This subscale encompasses eighteen items, including: 'I fkquently talk about 
my work in a negative way', and 'I get more and more engaged in my work' (reversed). 
Furthermore, a biographic questionnaire was developed to gather information about demographic 
characteristics. Information that was gathered included gender, race, age, language, educational 
level, business unit and household situation. 
1.3.4 Statistical Analysis 
The statistical analysis is carried out with the help of the SPSS Program (SPSS Inc., 2003) as 
well as the SAS Program (SAS Institute, 2000). Cronbach alpha coefficients and exploratory 
factor analyses are used to assess the reliability and validity of the measuring instruments (Clark 
& Watson, 1995). Descriptive statistics (e.g. means, standard deviations, skewness and kurtosis) 
are used to analyse the data 
Item-level analysis (item-bias analysis) is performed by using analysis of variance (ANOVA) for 
the instruments (yielding interval-level scores). An item is unbiased if persons from different 
language groups, with an equal standing on the theoretical construct underlying the instrument, 
have the same expected score on the item (e.g. Van de Vijver & Leung, 1997). Even though 
several statistical techniques are available for analysing item bias, analysis of variance has the 
advantage of computational simplicity, robustness and the possibility to study both uniform and 
nonuniform bias @lellenbergh, 1982). Therefore, analysis of variance is used in this study. The 
dependent variable is the item score, while language and score levels are the independent 
variables. A significant main effect of language groups is taken to point to uniform bias and a 
significant interaction of score level and language groups points to nonuniform bias. 
Construct equivalence of the instruments is also performed. According to Van de Vijver and 
hung  (1997), construct equivalence can be investigated with several techniques, such as factor 
analysis, cluster analysis, and multidimensional scaling or other dimensionality-reducing 
techniques. The basic idea behind the application of these techniques is to obtain a structure in 
each culture - which can then be compared across all cultures involved. The most fkequently 
employed technique for studying construct equivalence is factor analysis. Exploratory factor 
analysis is used to examine construct equivalence. A principal components analysis is conducted 
to determine the number of factors of the MBI-GS and the OLBI in the total sample. In order to 
determine the solution for each language group, a direct oblimin rotation is used. Factors 
obtained in each group are compared (after target rotation). The agreement is evaluated by a 
factor congruence coefficient, Tucker's phi (Van de Vijver & Leung, 1997). Values above 0,90 
are taken to point to essential agreement between cultural groups, while values above 0,95 point 
to very good agreement. A high agreement implies that the factor loadings of the lower and 
higher level are equal up to a multiplying constant. The latter is needed to accommodate 
possible differences in eigenvalues of factors for the different language groups. 
Multivariate analysis of variance (MANOVA) is used to determine the significance of 
differences between the burnout of demographic groups. MANOVA tests whether mean 
differences among groups on a combination of dependent variables are likely to have o c c ~ ~ ~ e d  
by chance (Tabachnick & Fidell, 2001). In MANOVA a new dependent variable that maximises 
group differences is created £torn the set of dependent variables. One-way analysis is then 
performed on the newly created dependent variable. Wilk's lambda is used to test the likeliiood 
of the data under the assumption of equal population mean vectors for all groups against the 
likelihood under the assumption that the population mean vectors are identical to those of the 
sample mean vectors for the different groups. When an effect is significant in MANOVA, one- 
way analysis of variance (ANOVA) is used to discover which dependent variables had been 
affected. Because multiple ANOVAs are used, a Bonfemoni-type adjustment is made for 
inflated Type I error. Tukey tests are done to indicate which group differed si@cantly when 
ANOVAs are done. 
T-tests are used to determine differences between the groups in the sample. Effect size (Cohen, 
1998; Steyn, 1999) is used in addition to statistical significance to determine the significance of 
relationships. Effect sizes indicated whether obtained results are practically significant. A cut- 
off point of 0,50 (medium effect) (Cohen, 1998) is set for the practical significance of 
differences between means. 
1.4 OVERVIEW OF CHAPTERS 
In Chapter 2, the Maslach Burnout Inventory - General Survey and the Oldenburg Burnout 
Inventory are discussed, as well as the differences of burnout between various groups. This 
chapter also deals with the empirical study. Chapter 3 deals with the discussion, limitations, and 
recommendations of this study. 
1.5 CHAPTER SUMMARY 
This chapter discussed the problem statement and research objectives. The measuring 
instruments and research method used in this research were explained, followed by a brief 
overview of the chapters that follow. 
REFERENCES 
Anderson, D.G. (2000). Coping strategies and burnout among veteran child protection workers. 
Child Abuse and Neglect, 24, 839-848. 
Applebaum, H.A. (1981). Royal Blue: The culture of construction workers. New York: 
Rinehart and Winston. 
Arbuckle, J.L. (1999). Amos 4.0. Chicago, IL: Smallwaters. 
Aromon, E., Pines, A.M. & Kafiy, D. (1983). Ausgebrannt. Vom &erd@ zur 
Selbstenfaltung [Burnout: From tedium to personal growth]. Stuttgart, Germany: Klett- 
Cotta 
Barkhuizen, E.N. & Rothmann, S. (2004, March). Burnout of Academic staff in a higher 
education institution. Paper presented at the 2nd South African Work Wellness Conference, 
Potchefstroom. 
Biggam, F.H., Power, KG., MacDonald, R.R., Carcary, W.B. & Moodie, E. (1997). Self- 
perceived occupational stress distress in a Scottish police force. Work and Stress, 11, 118- 
133. 
Burns, N. & Grove, S.K (1993). The practice of nursing research, conduct, critique, and 
utilization (2d ed.). Philadelphia, PA: W.B. Saunders. 
Byrne, B.M. (1994). Burnout: Testing for the validity, replication and invariance of causal 
structure across elementary, intermediate and secondary teachers. American Educational 
Research Journal, 31, 645-673. 
Cherniss, C. (1980). Professional burnout in human service organizations. New York: Praeger. 
Clark, L.A. & Watson, D. (1995). Constructing validity: Basic issues in objective scale 
development. Psychological Assessment, 7, 309-3 19. 
Cohen, J. (1998). Statisticalpower analysis of the behavioural sciences (2nd 4.). Orlando, CA: 
Academic Press. 
Demerouti, E., Bakker, A.B., Nachreiner, F. & Ebbinghaus, M. (2002). From mental strain to 
burnout. European Journal of Work and Organizational Psychology, 11(1), 423-441. 
Demerouti, E., Bakker, A.B., Vardakoy I. & Kantas, A. (2003). The convergent validity of two 
burnout instruments: A multitrait-multimethod analysis. European Journal of Psychological 
Assessment, 19, 12-23. 
Demerouti, E. & Nachreiner, F. (1998). Zur Spezifit5t von Burnout fiir Dienstleistungsberufe: 
Fakt oder Artefakt [On the specificity of burnout for human services: Fact or artefact]. 
ZeitschnBfitr Arbeitswissenschaft, 52, 82-89. 
Ebbinghaus, M. (1996). Erfassung von Burnout. Entwicklung und aerpriifung eines 
MeBinstrumentes liir die Anwendung in verschiedenen Berufsbereichen [Conceptualising 
burnout: development and validation of a measurement instrument for the application in 
various working domains]. Unpublished manuscript, Department of Work and 
Organizational Psychology, University of Oldenburg, Oldenburg, Germany. 
Edelwich, J. & Brodsky, A. (1980). Burnout: Stages of disillusionment in the helping 
professions. New York : Human Sciences Press. 
Freudenberger, H.J. (1974). Staff burnout. Journal of Social Issues, 30, 159-165. 
Gold, Y. (1984). Burnout: A Major problem for the teaching profession. Education, 104(3), 
271-275. 
Green, D.E., Walkey, F.H. & Taylor, A.J.W. (1991). The threefactor structure of the Maslach 
Burnout Inventory. Journal of Social Behavior and Personality 6,453472. 
Guilford, J.P. (1954). Psychometric methods. New York: McGraw-Hill. 
Hall, D.T. (1976). Careers in organizations. Pacific Palisades, CA: Goodyear. 
Jackson, L.T.B. & Rothmann, S. (2004, March). The validation of the Maslach Burnout 
Inventory for South Afican teachers. Paper presented at the 2* South Afiican Work 
Wellness Conference, Potchefstroom. 
Jackson, L.T.B., Mostert, K. & Pienaar, J. (2004, March). The measurement of burnout and 
engagement in a sample of tertiary students. Paper presented at the 2nd South African Work 
Wehess Conference, Potchefstroom. 
Jansen van Vuuren, J.A. & Rothmann, S. (2002, September). The relationship between 
psychological strengths, coping and burnout of senior managers in manufacturing industry. 
Paper presented at the 14& Conference of the South &can Institute for Management 
Scientist, Pretoria. 
Hopper, J. & Mostert, K. (2004, March). Coping, job stress and burnout in the South AfLican 
Police Service in the Free State. Poster presented at the 2nd South Afiican Work Wellness 
Conference, Potchefstroom, South Afiica 
Kop, N., Euwema, M. & Schaufeli, W. (1999). Burnout, job stress and violent behavior among 
Dutch police. Work and Stress, 13,326-340. 
Lazarus, R.S. & Folkman, S. (1984). Coping and adaptation. In W.D. Gentry m.), Handbook 
of behavioral medicine (pp.282-325). New York: Guilford. 
Lee, R.T., & Ashfort, B.E. (1996). A M e r  examination of managerial burnout: Toward an 
integrated model. Journal of Organizational Behavior, 14, 3-20. 
Leiter, M.P. (1993). Burnout as a developmental process: Consideration of models. In W.B. 
Schaufeli, C. Maslach, & T. Marek (eds.), Professional burnout: Recent developments in 
theory and research (pp.237-250). Washington, DC: Taylor & Francis. 
Levert, T., Lucas, M. & Ortlepp, K. (2000). Burnout in psychiatric nurses: Contributions of the 
work environment and sense of coherence. South Afncan Journal of Psychology, 30(2), 36- 
43. 
Lingard, H. (2003). The impact of individual and job characteristics on 'burnout" among civil 
engineers in Australia and the implications of employee turnover. Construction Management 
and Economics, 21(1), 69-80. 
Longcope, K. (1982). 'Burnout' syndrome seems to be spreading like wildfire. The Harrikburg 
(Pa.) Evening News, September, 15, p. B6. 
Maslach, C. & Jackson, S.E. (1986). MBI: Maslach Burnout Inventory: Manual research 
edition. Palo Alto, CA: Consulting Psychologists Press. 
Maslach, C., Jackson, S.E. & Leiter, M. (1996). Maslach Burnout Inventory: Manual (3d ed.). 
Palo Alto, CA: Consulting Psychologists Press. 
Maslach, C. & Leiter, M.P. (1997). The truth about burnout. San Francisco, CA: Jossey-Bass. 
Maslach, C., Schaufeli, W.B. & I~5ter, M.P. (2001). Job burnout. Annual Review of 
Psychology, 52, 397-422. 
Meams, J. & Cain, J.E. (2003). Relationship between teachers' occupational stress and their 
burnout and distress: Roles of coping and negative mood regulation. Anxiety, Stress and 
Coping, 16, 71-82. 
Mostert, K., Van Rensburg, S. & Rothmann, J.C. (2004, March). Apsychometric analysis of the 
Maslach Burnout Inventory -Human Services Survey (MBI-HSs) in auxiliary and enrolled 
nurses. Paper presented at the 2"* South African Work Wellness Conference, Potchefstroom. 
National Safety Council. (2000). Injury facts, Itasca, 111. 
Naudi, J.L.D. & Rothmann, S. (2004, March). Burnout of emergency workers in South Ajiaca. 
Paper presented at the 2nd South Afiican Work Wellness Conference, Potchefstroom. 
Rothmann, S. (2003). Burnout and engagement: A South Afiican perspective. South Afncan 
Journal of Industrial Psychologv, 29(4), 16-25. 
Rothmann, S. & Jansen van Vuuren, J.A. (in press). Sense of coherence, self-efficacy, locus of 
control, coping and burnout of senior managers in a manufacturing industry. South Ajiacan 
Journal of Industrial Psychology. 
Rothmann, S., Malan, A.M. & Rothmann, J.C. (2001). Sense of coherence, coping and burnout 
in a coiporate pharmacy group. Paper presented at the 7" annual PsySSA Conference, 
Johannesburg. 
SAS Institute. (2000). The SAS System for Windows: Release 8.01. Cary, NC: SAS Institute Inc. 
Schaufeli, W.B. & Enanann, D. (1998). The burnout companion to study andpractice: A 
critical analysis. London: Taylor & Francis. 
Schaufeli, W.B., Leiter, M.P., Maslach, C. & Jackson, S.E. (1996). MBI - General Survey. In 
C. Maslach, S.E. Jackson & M.P. Leiter (Eds.), Maslach Burnout Inventory Manual (3* 4.). 
Palo Alto, CA: Consulting Psychologists Press. 
Schutte, N., Toppinen, S., Kalimo, R. & Schaufeli, W.B. (2000). The factorial validity of the 
Maslach Burnout Inventory - General Survey (MBI-GS) across occupational groups and 
nations. Journal of Occupational and Organizational Psychology, 73, 53-66. 
Shaughnessy, J.J. & Zechmeister, E.B. (1997). Research methorls in psychology (4'h 4 . ) .  New 
York: McGraw-Hill. 
Shirom, A. (1989). Burnout in Work organizations. In C.L. Cooper & I.T. Robertson (Eds.), 
International review of industrial and organizationalpsychology (pp.25-48). NY: Wiley. 
Sieberhagen, C. & Pienaar, J. (2004, March). Burnout and engagement of student leaders in a 
higher education institution. Paper presented at the 2* South Afiican Work Wellness 
Conference, Potchefstroom. 
Singh, S. (1997). State of the art in automation of earthmoving. Journal of Aerospace 
Engineering, 10(4), 179-1 88. 
Smallwood, J. (1997). Construction is a stressful industry: management. SA Builder, 889,30. 
SPSS Inc. (2003). SPSS 12.0 for Windows. Chicago, IL: Author. 
Step,  H.S. (1999). Praktiese betekenisvolheid: Die gebruik van effekgroottes. Wetenskaplike 
bydraes -Reeks B: Natuunvetenskappe Nr. I1 7. Potchefstroom: PU vir CHO. 
Storm, K. (2001). The relationship between personality dimensions, coping and burnout of 
pharmacists in a corporate pharmacy group. Unpublished dissertation, PU for CHE, 
Potchefstroom. 
Storm, K. (2002). Burnout and work engagement in the South Afn'an Police Service. 
Unpublished doctoral thesis, PU for CHE, Potchefstroom. 
Storm, K. & Rothmann, S. (2003). A psychometric analysis of the Maslach Bumout Inventory 
- General Survey in the South African Police Services. South Afiican Journal of 
Psychology, 39(4), 219-266. 
Tabachnick B.G. & Fidell, L.S. (2001). Using multivariate statistics (4& 4.). Boston, MA: 
Allyn & Bacon. 
Thornton, P. (1992). The relation of coping, appraisal and bumout in mental health workers. 
Journal ofPsychology, 126.261-271. 
Van de Vijver, F. & Leung, K. (1997). Methods and data analysis for cross-cultural research. 
Thousand Oaks, CA: SAGE. 
Van de Vijver, F. & Tamer, N.K. (1997). Bias and equivalence in crosscultural assessment: 
An overview. European Review ofApplied P d o l o g y ,  47, 263-279. 
Weisberg, J. & Sagie, A. (1999). Teachers' physical, mental and emotional burnout: Impact on 
intention to quit. Journal ofpsychology, 133(3), 333-340. 
Chapter 2 
Research Article 
THE VALIDATION OF TWO BURNOUT MEASURES IN THE SOUTH 
AFRICAN EARTHMOVING EQUIPMENT INDUSTRY* 
A.M. LE ROUX 
K. MOSTERT 
Work Well: Research Unit for People, Policy and Pe$ormance, Faculty of Economic & 
Management Sciences, North-West University, Potchefstroom Campus 
ABSTRACT 
The objectives of this study were to validate the Maslach Bumout Inventoty - General Survey (MBI-GS) 
and the Oldenburg Burnout Inventory (OLBI) for employees in an earthmoving equipment industry in 
South Aftica, to determine the construct equivalence and bias for different language groups, and to 
investigate the relationship between bumout and various demographic characteristics. A cross-sectional 
survey design was used. Random samples (n = 326) were taken of employees in an eartbmoving 
equipment industry of eight provinces in South Aftica The MEI-GS, the OLBI and a biographical 
questionnaire were administered. Exploratory factor analysis with target rotations confirmed the 
wnshc t  equivalence of the three factors of the MEI-GS and the two factors for the OLBI for different 
language groups. No evidence was found for uniform or nonuniform bias of the items of the MBI-GS or 
the OLBI for different language groups. Results indicated that no differenca with respect to levels of 
bumout exist between various demographic chamcteristics. 
OPSOMMING 
Die doelstellings van hierdie studie was om die Maslach Uitbrandingsmelys - Algemene Opname 
(MBI-GS) en die Oldenburg Uitbmdingsvraelys (OLBI) te valideer vir wehemem in 'n 
grondverskuiwingstoerustingindustrie, om die konstrukekwavilensie en sydigheid d a a ~ ~ a n  vir 
verskillende taalgroepe te bepaal, en om die vahouding tussen uitbmding en verskillende demopfiese 
gmepe te bepaal. 'n Dwarsdeursneeontwerp is gebluik Gestratifkeerde ewekamige steelqnoewe (n = 
326) is geneem van werknemers in 'n grondv+wingstoemtingindustrie. Eksplo~atiewe faktoranalise 
met teikemotasie het die konstrukekwivalensie van die drie faktore vir die MBI-GS en twee faktore vir 
die OLBI vir verskillende taalgroepe bevestig. Bewyse is nie gevind vir uniforme of nie-uniforme 
sydigheid van die items van die MBI-GS of van die OLBI vir verskillende taalgroepe nie. Bevindinge 
het getoon dat daar geen verskille ten opsigte van vlakke van uitbranding tussen versldllende 
demografiesekaraktertrekke voorkom nie 
- 
This m a t e d  is based u p  wok supported by the National Research Foundation under Grant number 2054383. 
Burnout was originally almost exclusively studied as an occupational issue for people providing 
human services (Schaufeli & Emmam, 1998). The concept of burnout was linked to human 
services such as health care, education and social work, where employees do "people work" of 
some kind (Maslach & Leiter, 1997). However, nearly a quarter of a century of research and 
practice made it clear that burnout also exists outside the human services. The need for an 
instrument that measures bumout in non-contactual professional contexts was met by the 
introduction of the Maslach Burnout Inventory - General Survey (MBI-GS) (Schaufeli, Leiter, 
Maslach & Jackson, 1996). The MBI-GS assesses parallel dimensions to those contained in the 
o r i d  Maslach Burnout Inventory, except that the items do not explicitly refer to working with 
people. 
Although the Maslach Burnout Inventory remains the most popular, and almost universally used, 
instrument to assess burnout (Schaufeli & Enzmaun, 1998), an alternative measure of burnout 
that can be used is the Oldenburg Burnout Inventory (OLBI) (Demerouti, Bakker, Nachreiner & 
Schaufeli, 2001). The OLBI was constructed and validated among different occupational groups 
(Demerouti & Nachreiner, 1998) at the same time that the MBI-GS was developed. 
A great deal of research has been done on bumout in various occupational fields in South A k a  
(e.g. police officers, Storm, 2002; emergency workers, Naud6 & Rothmaun, 2004; pharmacists, 
Rothmann, Malan & Rothmann, 2001; academic staff in Higher Education Institutions, 
Barkhuizen & Rothmann, 2004, managers in the manufacturing industry, Jansen van Vuuren & 
Rothmam, 2002). However, it seems that not much research has been done into the experience 
of burnout of employees in the earthmoving equipment industry. 
According to Applebaum (1981), the earthmoving equipment industry includes the building of 
schools, office buildings, shopping centres, dams, bridges, sewer treatment facilities, industrial 
plants, roads and highways. In industries such as construction and mining, earthmoving plays a 
fundamental role (Singh, 1997). The construction industry evolves around a very demanding 
and stressful work environment (Lingard, 2003). Many worksites are hazardous for humans - 
unstructured and outdoor environments, poor visibility conditions and inclement weather 
conditions (Singh, 1997). According to Gold (1984), such work-related stress experienced by an 
individual could cause burnout. 
Apart fiom a study done in Greece (Demerouti et al., 2003), both burnout instruments (the MBI- 
GS and the OLBI) have not yet been validated on the same population. Also, no other studies 
were found in the literature where burnout was measured by means of the OLBI. The OLBI has 
not yet been validated in other countries, or in the earthmoving equipment industry. This is 
therefore the first study in South &ca to validate the OLBI and to investigate whether both the 
MBI-GS and the OLBI measure the same burnout construct. 
The objectives of this study are to validate the MBI-GS and the OLBI, to determine the construct 
equivalence and item bias of these instruments for employees in an earthmoving equipment 
industry in South Afiica, to determine the correlation between the MBI-GS and the OLBI and to 
investigate the relationship between burnout and various demographic characteristics. 
Burnout and its different ways of measurement 
Herbert Freudenberger (1974) is generally considered to be the 'founding father' of the burnout 
syndrome. He described burnout as a state of chronic emotional fatigue. More recently, 
Schaufeli & Enzmann (1998, p. 36) defined burnout as "a persistent, negative, work-related state 
of mind in 'normal' individuals, which is primarily characterised by distress, a sense of reduced 
effectiveness, decreased motivation, and the development of dysfunctional attitudes and 
behaviours at work". 
Initially research concentrated on caregiving and service occupatio~ls (Freudenberger, 1974; 
Maslach, 1976). However, Schaufeli, Martinez, Pinto, Salovana and Bakker (2002) explain that 
burnout is no longer restricted to the caring professions and that all types of professions and 
occupational groups can experience burnout. Thus, an adapted version of the Maslach Burnout 
Inventory (Maslach & Jackson, 1986) was developed, namely the Maslach Burnout Inventory - 
General Survey (Schaufeli et al., 1996). The MBI-GS assesses parallel dimensions to those 
contained in the on@ MBI, except that the items do not explicitly refer to working with 
people (Schaufeli et al., 1996). The MBI-GS comprises three subscales: Exhaustion (EX), 
Cynicism (CY) and Professional Efficacy (F'E). Exhaustion refers to fatigue, but without direct 
reference to people as the source of those feelings. Cynicism can been seen as an indifference or 
an aloof attitude towards one's work in general, where the items refer to work itself rather than to 
recipients of one's services or to personal relationships at work 
The three dimensions of the MBI-GS are interrelated Cynicism is highly related to Exhaustion 
(0,44 < r < 0,61), and also strongly related to Professional Efficacy (-0,38 < r < -0,57) 
(Schaufeli et al., 1996). Thus it appears that burnout can be validly measured across a different 
range of occupations. Internal consistencies (Cronbach coefficient alphas) reported by Schaufeli 
et al. (1996) varied from 037 to 0,89 for Exhaustion, 0,73 to 0,84 for Cymcism and 0,76 to 484 
for Professional Efficacy. Test-retest reliabilities after one year were 0,65 (Exhaustion), 0,60 
(Cynicism) and 0,67 (l'rofessional Efficacy) (Schaufeli et al., 1996). 
Although the psychometric properties of the MBI-GS were tested in samples in South Africa, 
(police officers, Stonn, 2002; teachers, Jackson & Rothmann, 2004), the Maslach Burnout 
Inventory - General Survey (MBI-GS) has not yet been validated for employees in a South 
scan earthmoving industry. 
The MBI-GS still suffers fkom one of the same methodological deficiencies as the original MBI: 
the items in each subscale are phrased in the same direction @ememuti, Bakker, Nachreiner & 
Ebbinghaus, 2002). All Exhaustion and Cynicism items are phrased negatively, while all 
Professional Efficacy items are phrased positively. According to Guildord (1954), such one- 
sided scales are inferior to scales that include both positively and negatively phrased items. 
An instrument that does include both positively and negatively worded items, as requested by 
conventional psychometric standards, is the Oldenburg Burnout Inventory (OLBI) (Demerouti et 
al., 2002). The OLBI conceives burnout as "a syndrome of work-related negative experiences, 
including feelings of exhaustion and disengagement at work" @emmuti et al., 2002) and can be 
used among occupations within and outside human service professions. The OLBI includes two 
dimensions, namely exhaustion and disengagement from work. Exhaustion is deiined as "a 
consequence of intensive physical, affective and cognitive strain, i.e. as a long-term consequence 
of prolonged exposure to certain job demands'' (Demerouti, Bakker, Vardakou & Kantas, 2003). 
According to Demerouti et al. (2003), the disengagement scale refers to distancing oneself from 
one's work and experiencing negative attitudes towards the work object, work content, or one's 
work in general. 
The OLBI was constructed and validated in Gennany among different occupational groups 
(Demerouti, 1999; Demerouti & Nachreiner, 1998). Exploratory and confirmatory factor 
analysis confirmed its two-dimensional factorial structure (Ebbinghaus, 1996). Demerouti et al., 
(2002) examined the factorial validity of burnout as measured by the OLBI. Results confirmed 
that burnout has a two factor structure, with exhaustion and disengagement as separate but 
correlated dimensions. This structure was also invariant across human services and industrial 
production work. This consistency of the factorial structure of the OLBI across different 
occupational groups confirms the generalisability of the burnout construct to other occupations 
and suggests that human service burnout represents only one specific manifestation of burnout 
(Demerouti & Nachreiner, 1998). 
Comparing the different measurements of burnout 
When looking at the two different means of measurement, it can be seen that the scales differ in 
content. The OLBI's exhaustion scale refers not only to affective aspects of exhaustion, as is the 
case with the MBI-GS, but includes physical and cognitive aspects as well (Demerouti et al., 
2002, 2003). The OLBI's disengagement scale refers to emotions toward the work task as well 
as to devaluation and mechanical executions of one's work, while the MBI-GS's cynicism scale 
restricts itself to measuring mainly subjective job meaninglessness and the lack of interest 
employees have in their jobs (Demerouti et al., 2002). 
The third dimension of burnout, feelings of reduced personal accomplishment - or professional 
efficacy - in more general terms are not considered as a separate burnout dimension, when using 
the OLBI. Several reasons are given for this argument. Firstly, Leiter (1993) argued and 
showed that emotional exhaustion leads to depersonalisation, whereas professional efficacy 
develops independently. Secondly, the independent role of personal accomplishment has been 
cordirmed through a meta-analysis of the correlates of the burnout dimensions (Lee & Ashfort, 
1996), and thirdly, personal accomplishment is the weakest burnout dimension in terms of 
significant relationships with other variables (Lee & M o r t ,  1996; Schaufeli & Enzmann, 
1998). 
The convergent validity of the OLBI and the MBI-GS was confirmed in a study among 232 
Greek employees h m  different occupations @ememuti et al., 2003). Results of multitrait- 
multimethod analysis included both the OLBI and the MBI-GS and showed that the burnout 
components and the measurement instruments could explain the responses to the various items. 
The bivariate correlation between OLBI exhaustion and MBI-GS exhaustion was 0,60 and the 
same correlation was found between OLBI disengagement and the MBI-GS cynicism. 
From an empirical point of view it is not only important to obtain a valid and reliable 
measurement of burnout but also to test measurement equivalence and bias in contexts where 
differences in scores could be attributed to cultural influences in terms of item meaning and 
understanding. Invalid conclusions regarding the constructs under study could be made if 
cultural influences are not accounted for. 
Construct equivalence, according to Van de Vijver and Leung (1997), indicates the extent to 
which the same construct is measured across all cultural groups studied. No comparison can be 
made if an instrument measures different constructs in different cultures. Item bias concerns 
aspects of measurement validity in intercultural group comparison (Van de Vijver & Leung, 
1997). An unbiased item would provide the same average score on an item if two people h m  
different cultural groups are similar in terms of the construct measured by this item. Two types 
of item bias are distinguished, namely uniform bias (influences of bias on scores that are more or 
less the same for all score levels) and nonuniform bias (influences that are not identical for all 
score levels). 
Finally, the influence of demographic variables on burnout, as measured with the MBI-GS and 
the OLBI, was examined. The relationship between burnout on the one hand and gender, race, 
age, language, educational level, business unit and household situation on the other hand, will be 
concentrated on in this study. 
The above discussion leads to the following hypotheses: 
Burnout, as measured by the MBI-GS, is a three-dimensional construct (consisting of 
Exhaustion, Cynicism and hofessional Efficacy) and shows high internal consistency. 
Burnout, as measured by the OLBI, is a two-dimensional construct (consisting of 
Exhaustion and Disengagement) and shows high internal consistency. 
Burnout, as measured by the MBI-GS and the OLBI, is an equivalent and unbiased 
construct for different language groups within the earthmoving equipment industry. 
No differences exist between burnout and various demographic characteristics. 
METHOD 
Research design 
A cross-sectional s w e y  design was used to reach the objectives of this research. Cross- 
sectional designs are used to examine groups of subjects in various stages of development 
simultaneously, while the s w e y  describes a technique of data collection in which questionnaires 
were used to gather data about an identified population @urns & Grove, 1993). This design is 
well suited to the descriptive and predictive functions associated with correctional research, 
whereby relationships between variables are examined (Shaughnessy & Zechmeister, 1997). 
Study population 
The participants used in the research were selected randomly h m  the population (n = 326). 
Samples were taken h m  eaahmoving companies in Limpopo, Gauteng, Mpumalanga, Northern 
Cape, Western Cape, Eastern Cape, KwaZulu-Natal and North West. Table 1 presents some of 
the characteristics of the participants. 
Table 1 
Characteristics of the Participants 
- - - - - - - - - - 
Item Category Percentage 
Gender Male 74,s 
Race 
Language 
Business unit 
Education level 
Female 
White 
AfIican 
Colomd 
Indian 
Other 
Eoglish 
Sepedi 
Sesotbo 
s e t s ~ ~ n a  1,s 
Tshivenda 0,9 
isiNdebele 1,2 
isiXhosa 2,1 
isizulu 3,9 
Other 0 3  
Construction 52,7 
Shared Services 14,s 
Mi- 243 
Rental 2,4 
Handliig 4,8 
Other 1,2 
Grade 10 (Std 8) 16,4 
W e  11 (Std 9) 3,6 
Grade 12 (Std 10) 38,2 
Technical CoUege Diploma 23,6 
Technikon Diploma 9,4 
University Degree 4 2  
Postgraduate Degree 2,4 
According to Table 1, the participants were predominantly male (74,8%), and 61,5% was white. 
Just over half of the participants (52,7%) worked in the construction unit, while 20,3% worked in 
the mining unit. The majority of the participants spoke Aftham (40,9%) and Enghsh (38,5%). 
A total of 38,2% of the participants possessed a Grade 12 certificate while 23,6% of the 
participants possessed a Technical College Diploma 
Measuring battery 
The following questionnaires were utilised in the empirical study. 
The Maslach Burnout Inventory - General Survey (MBI-GS) (Schaufeli et al., 1996) was 
used to measure burnout. The MBI-GS has tbree subscales: Exhaustion (Ex) (five items, 
e.g. "I feel used up at the end of the workday"), Cynicism (Cy) (five items, e.g. "I have 
become less enthusiastic about my work") and Professional Efficacy (PE) (six items, e.g. 
'%I my opinion, I am good at my job"). Together, the subscales of the MBI-GS provide a 
threedimensional perspective on burnout. Internal consistencies (Cronbach coefficient 
alphas) reported by Schaufeli et al. (1996) varied from 0,87 to 0,89 for Exhaustion, 0,73 
to 484 for Cyucism, and 476 to 484 for Professional Efficacy. Test-retest reliabilities 
after one year were 0,65 (Exhaustion), 0,60 (Cynicism) and 0,67 (F'rofessional Efficacy) 
(Schaufeli et al., 1996). All items are scored on a 7-point frequency-rating scale, ranging 
from 0 to 6. High scores on Exhaustion and Cynicism, and low scores on Professional 
Efficacy are indicative of burnout. Storm (2002) confirmed the 3-factor structure of the 
MBI-GS in a sample of 2 396 members of the South African Police Service (SAPS), but 
recommended that Item 13 be dropped from the questionnaire. She confirmed the 
structural equivalence of the MBI-GS for different race groups in the SAPS. The 
following Cronbach alpha coefficients were obtained for the MBI-GS: Exhaustion: 488; 
Cynicism: 0,79; Professional Efficacy: 0,78 (Storm, 2002). 
The Oldenburg Burnout Inventory (OLBI) (Demerouti et al., 2000) has been constructed 
and validated in an independent study among 293 employees from different occupational 
fields, including human services and blue-collar workers (Ebbinghaus, 1996). The 
inventory measures burnout independent of vocational aspects on two dimensions, 
namely exhaustion and disengagement. The seven items of the Exhaustion subscales are 
generic and refer to general feelings of emptiness, overtaxing h m  work, a strong need 
for rest, and a state of emotional exhaustion. Examples are: 'After my work, I regularly 
feel worn out and weary', and 'After my work, I regularly feel totally fit for my free-time 
activities' (reversed). Disengagement refers to distancing oneself h m  one's work (work 
object and content), and to negative, cynical attitudes and behaviours towards one's work 
in general. This subscale encompasses eighteen items, including: 'I h p e n t l y  talk about 
my work in a negative way', and 'I get more and more engaged in my work' (reversed). 
Furthermore, a biograpbic questionnaire was developed to gather information about demographic 
characteristics. Information that was gathered included gender, race, age, language, educational 
level, business unit and household situation. 
Statistical analysis 
The statistical analysis was carried out with the help of the SPSS Program (SPSS Inc., 2003) as 
well as the SAS Program (SAS Institute, 2000). Cronbach alpha coefficients and exploratory 
factor analyses were used to assess the reliability and validity of the measuring instruments 
(Clark & Watson, 1995). Descriptive statistics (e.g. means, standard deviations, skewness and 
kurtosis) were used to analyse the data Product-moment correlation was used to determine the 
correlation between the MBI-GS and the OLBI. 
Item-level analysis (item-bias analysis) was performed by using analysis of variance (ANOVA) 
for the instruments (yielding interval-level scores). An item is unbiased if persons h m  different 
language groups, with an equal standing on the theoretical construct underlying the instrument, 
have the same expected score on the item (e.g. Van de Vijver & Leung, 1997). Even though 
several statistical techniques are available for analysing item bias, analysis of variance has the 
advantage of computational simplicity, robustness and the possibility to study both uniform and 
nonuniform bias (Mellenbergh, 1982). Therefore, analysis of variance was used in this study. 
The dependent variable is the item score, while language and score levels are the independent 
variables. A significant main effect of language groups was taken to point to uniform bias and a 
significant interaction of score level and language groups point to nonuniform bias. 
Construct equivalence of the instruments was also performed. According to Van de Vijver and 
Leung (1997), construct equivalence can be investigated with several techniques, such as factor 
analysis, cluster analysis, and multidimensional scaling or other dimensionality-reducing 
techniques. The basic idea behind the application of these techniques is to obtain a structure in 
each culture - which can then be compared across a l l  cultures involved. The most kequently 
employed technique for studying construct equivalence is factor analysis. Exploratory factor 
analysis was used to examine construct equivalence. A principal components analysis was 
conducted to determine the number of factors of the MBI-GS and the OLBI in the total sample. 
In order to determine the solution for each language group, a direct oblimin rotation was used. 
Factors obtained in each group were compared (after target rotation). The agreement was 
evaluated by a factor congruence coefficient, Tucker's phi (Van de Vijver & Leung, 1997). 
Values above 0,90 were taken to point to essential agreement between cultural groups, while 
values above 0,95 point to very good agreement. A high agreement implies that the factor 
loadings of the lower and higher level are equal up to a multiplying constant. The latter was 
needed to accommodate possible differences in eigenvalues of factors for the different language 
groups. 
Multivariate analysis of variance (MANOVA) was used to determine the significance of 
differences between the burnout of demographic groups. MANOVA tests whether mean 
differences among groups on a combination of dependent variables are likely to have occurred 
by chance (Tabachnick & Fidell, 2001). In MANOVA a new dependent variable that maxirnises 
group differences was created from the set of dependent variables. Oneway analysis was then 
performed on the newly-created dependent variable. Wilk's lambda was used to test the 
likelihood of the data under the assumption of equal population mean vectors for all groups 
against the likelihood under the assumption that the population mean vectors were identical to 
those of the sample mean vectors for the different groups. When an effect was significant in 
MANOVA, oneway analysis of variance (ANOVA) was used to discover which dependent 
variables had been affected Because multiple ANOVAs were used, a Bonferroni-type 
adjustment was made for inflated Type I error. Tukey tests were done to indicate which group 
differed significantly when ANOVAs were. done. 
T-tests were used to determine differences between the groups in the sample. Effect size 
(Cohen, 1998; Steyn, 1999) was used in addition to statistical significance to determine the 
significance of relationships. Effect sizes indicated whether obtained results were practically 
significant. A cut-off point of 0,50 (medium effect) (Cohen, 1998) was set for the practical 
significance of differences between means. 
RESULTS 
The composition of the sample was of such nature that it was decided to conduct the analysis in 
this study on language groups. The sample was divided into an Afiican group (Afibam, 
Sepedi, Sesotho, Setswana, Tshivenda, isiNdebele, isiXhosa, isiZulu) and an English group. 
Firstly, items that show uniform and nonuniform bias for the two language groups were 
identified through a bias analysis. Based on the total score of the MBI-GS and the OLBI, score 
groups were compiled. Two effects were tested for si@cance in the subsequent variance 
analysis, namely the main effects of culture (uniform bias) and the interaction effects of culture 
and score level (nonuniform bias). If both the main effects of culture and the interaction of 
culture and score level were found to be non-significant, the item was taken to be unbiased. The 
results of the individual item analysis of the variance for the 15-item' MBI-GS are presented in 
Table 2. 
' Item 13 (''I just want to do my job and not be bothered")~ eliminated h m  the original MBI-GS, resulting in a 
15-item scale. The decision to eliminate this item is consistent with previous research (Schutte et al., 2000). 
Table 2 
Ztem-bias analysis for the MBZ-GS 
Item Uniform Bias Nonuniform Bias 
P P 
MBI 1 0,85 0,31 
MBI 2 O N  0,49 
MBI 3 0,09 0,70 
MBI 6 
MBI 7 
MBI 8 
MBI 10 0,67 09'  
MBI 11 0,09 0,49 
MBI 14 0,48 0,75 
MBI 15 0,88 0,45 
MBI 16 0,52 0,71 
' q2 > 0,06 - Practically significant (medium effect) 
According to Table 2, no sigdicant eta square ($) value was obtained for any item of the MBI- 
GS, indicating that none of the items showed uniform or nonuniform bias. 
The results for the individual item analysis of variance for the 16-item OLBI are represented in 
Table 3. 
Table 3 
Item-bias analysis for the OLBZ 
Item Uniform Bias Nonnniform Bias 
P P 
OLBI 1 0,51 0,71 
OLBI 2 
OLBI 3 
OLBI 4 
OLBI 5 %go 0,05 
OLBI 6 0,82 0,63 
OLBI 7 0,58 0,88 
OLBI 8 0,22 438 
OLBI 9 
OLBI 10 
OLBI 11 
OLBI 12 0,18 0,85 
OLBI 13 0,87 0,60 
OLBI 14 0,32 0,29 
OLBI 15 0,59 0,61 
OLBI 16 0,43 0,49 
' q2 > 0,06 - Practically significant (medium effect) 
According to Table 3, no significant eta square (qz) value was obtained for any item of the OLBI. 
A simple principal components analysis was conducted on the 15 items of the MBI-GS on the 
total sample of employees in an earthmoving equipment industry. Analysis of the eigenvalues 
(larger than 1) and scree plot indicated that three factors could be extracted. Next, principal 
component analysis with a direct oblimin rotation was used in canying out factor analysis per 
language group. The pattern matrices for the f i c a n  group and the English group are reported 
in Table 4. 
Table 4 
Pattern matrix of the 15-item MBI-GS for the African group and the English group 
AFRICAN GROUP ENGLISH GROUP 
Item Factor 1 
M B I  1 -0,06 
MBI 2 -0,91 
MBI 3 0,11 
MBI 4 0,12 
MBI 5 -0,32 
M B I  6 0,OS 
MBI 7 -0,25 
MBI 8 0,67 
MBI 9 0,64 
MBI 10 0,11 
M B I  11 -0,38 
MBI 12 0,W 
MBI 14 0,81 
M B I  15 0,67 
MBI 16 -0.02 
Factor 2 
0,85 
0,77 
0,74 
0,50 
0,11 
0,83 
0,02 
025 
0,16 
0,03 
0,02 
-0,06 
-0,lO 
0,16 
-0,09 
Factor 3 
0,11 
0,07 
-0,19 
-0,15 
0,34 
0,05 
0,61 
0,Ol 
-0,20 
0,77 
021 
0,69 
0,41 
-017 
0.76 
Item 
MBI 1 
MBI 2 
MBI 3 
MBI 4 
MBI 5 
MBI 6 
M B I  7 
M B I  8 
M B I  9 
MBI 10 
M B I  11 
M B I  12 
M B I  14 
MBI 15 
MBI 16 
Factor 1 
0,Ol 
-om 
0,52 
-1,ll 
-],I4 
0,13 
-0,35 
0,46 
O,@ 
0,39 
-0,W 
-0,15 
0,81 
0,49 
-0.15 
Factor 3 
-0,07 
-0,20 
0,05 
0,26 
-0,65 
0,05 
-0,51 
0,16 
0,15 
-0,SO 
479 
-0,69 
-0,03 
0,29 
-,069 
The pattern matrices of the three-factor solutions for the African group and the English group 
were then used as input for an exploratory factor analysis with target rotations. By rotating the 
one solution to the other, the three-factor structure was compared across groups. The following 
Tucker's phi coefficients were obtained after target rotation: Factor 1 = 0,92; Factor 2 = 0,98 
and Factor 3 = 0,89. Factor 1 and 2 compared favourably with the guideline of 0,90. However, 
it is clear that Factor 3 showed a lower equivalence for the two language groups. 
Inspection of Table 4 revealed that one item was complex and problematic. Item 11 ("I am 
enthusiastic about my job") loaded on different factors for the African group and the English 
group. This item loaded on Factor 1 and 3, while it loaded strongly (-0,79) on Factor 3 in the 
English group. 
Another simple factor analysis was conducted after Item 11 was removed. The eigenvalue and 
scree plot showed three factors. The pattern matrices for the African group and the English 
group are reported in Table 5. 
Table 5 
Pattern matrix of the 14-item MBZ-GS for the Afican group and the English group 
AFRICAN GROUP ENGLISH GROUP 
Item 
MBI 1 
MBI 2 
MBI 3 
MBI 4 
MBI 5 
MBI 6 
MBI 7 
MBI 8 
MBI 9 
MBI 10 
MBI 12 
MBI 14 
MBI 15 
Factor 1 
-0,07 
-0,08 
0.11 
0,12 
435 
0,11 
-0,25 
0,s 
0,61 
0,12 
0,w 
0,81 
0,69 
Factor 2 
0986 
0,77 
0,74 
0 9  
0,13 
0,82 
0,03 
0,26 
0,17 
0,04 
-0,07 
-0,lO 
0,15 
Factor 3 
0,12 
0,08 
-412 
-0J6 
Of 6 
O M  
o,a 
-0,03 
-0,24 
0,77 
0969 
0,37 
-0,22 
Item 
MBI 1 
MBI 2 
MBI 3 
MBI 4 
MBI 5 
M B I  6 
MBI 7 
MBI 8 
MBI 9 
MBI 10 
MBI 12 
MBI 14 
MBI 15 
Factor 1 
0,Ol 
0,02 
0,06 
-0J4 
-0,w 
0,12 
-0,29 
O M  
0,57 
0,39 
-0J4 
48J 
445 
M B I  16 -0,OO -0,08 0.77 MB I  16 -0.14 -0.11 
Factor 3 
-0,07 
-0,23 
0,05 
0,30 
4,71 
0,08 
- 0 9 6 0  
0,15 
0,22 
477 
-466 
404 
0,35 
-0.71 
The three factors were labelled as: Factor 1: Cynicism, Factor 2: Exhaustion, and Factor 3: 
Professional Efficacy. Tucker's phi coefficients of 0,96 (Cynicism) 0,98 (Exhaustion) and 0,90 
(Professional Efficacy) resulted after a target rotation was carried out. These coefficients can be 
regarded as acceptable. 
Another simple principal components analysis was conducted on the 16 items of the OLBI on the 
total sample of employees in an earthmoving equipment industry. Analysis of the eigenvalues 
(larger than 1) and scree plot indicated that two factors could be extracted. Next, principal 
component analysis with a direct oblimin rotation was used in carrying out factor analysis per 
language group. The pattern matrices for the Afirican group and the English group are reported 
in Table 6. 
Table 6 
Pattern matrix of the 16-item OLBZfor the Afican group and the English group 
- -  - 
AFRICAN GROUP ENGLISH GROUP 
Item Factor 1 Factor 2 
OLBI 1 
OLBI 2 
OLBI 3 
OLBI 4 
OLBI 5 
OLBI 6 
OLBI 7 
OLBI 8 
OLBI 9 
OLBI 10 
OLBI 11 
OLBI 12 
OLBI 13 
OBLI 14 
OLBI 15 
OLBI 16 
-414 
0,67 
0,58 
0,65 
0 3  
066 
0,Ol 
0,55 
0,70 
-0,09 
opa 
451 
-0,oo 
-0,15 
-0,06 
0.01 
Item 
OLBI 1 
OLBI 2 
OLBI 3 
OLBI 4 
OLBI 5 
OLBI 6 
OLBI 7 
OLBI 8 
OLBI 9 
OLBI 10 
OLBI 11 
OLBI 12 
OLBI 13 
OBLI 14 
OLBI 15 
Factor 1 
0,65 
0,05 
-0,24 
0,13 
0,23 
-0,2 1 
0960 
0,11 
-0,20 
0,09 
-0,18 
0,12 
0,54 
0,87 
Factor 2 
0,lO 
0,75 
0,415 
0,80 
-0,08 
0,58 
-0,IO 
0,76 
o,@J 
455 
0,m 
0,81 
-0,14 
-0J7 
0,12 
OLBI 16 0.80 0.10 
The pattern matrices of the two-factor solutions for the African group and the Enghsh group 
were then used as input for an exploratory factor analysis with target rotations. By rotating the 
one solution to the other, the two-factor structure was compared across groups. The following 
Tucker's phi coefficients were obtained after target rotation: Factor 1 = 0,93 and Factor 2 = 
0,85. Although Factor 1 compared favourably with the guideline of O,9O, it is clear that Factor 2 
showed an unacceptable low equivalence for the two language groups. 
Inspection of Table 6 revealed that one item was complex and problematic. Item 10 ("After my 
work, I usually feel still fit for my leisure activities"), loaded on different factors for the Afiican 
group and the Enghsh group. 
Another simple factor analysis was conducted after Item 10 was removed. The eigenvalue and 
scree plot showed two factors. The pattern matrices for the e c a n  group and the English group 
are reported in Table 7. 
Table 7 
Pattern matrix of the 16-item OLBI for the Afncan group and the English group 
AFRICAN GROUP ENGLISH GROUP 
Item Factor 1 Factor 2 Item Factor 1 Faetor 2 
OLBI 1 0964 -0,14 OLBI 1 096 0,09 
OLBI 2 -0,09 049 OLBI 2 0,35 0,75 
OLBI 3 4,13 03s OLBI 3 -0,24 0,50 
OLBI 4 0.15 0964 OLBI 4 0,12 0,80 
OLBI 5 0952 0,11 OLBI 5 op -0,08 
OLBI 6 0,21 46 OLBI 6 -021 0,58 
OLBI 7 0,7l 0,Oo OLBI 7 039 -0,ll 
OLBI 8 -0,17 0,61 OLBI 8 0,lO 0,76 
OLBI 9 -0,08 046 OLBI 9 -0,ZO 0,63 
OLBI 11 -0,38 0 9  OLBI 11 -0,18 0,61 
OLBI 12 -0,13 0 9  OLBI 12 0,lO 0,80 
OLBI 13 0,43 0,02 OLBI 13 044 -0,14 
OBLI 14 0,45 -0,21 OBLI 14 036 -0,14 
OLBI 15 0,71 -0,09 OLBI 15 047 0,13 
OLBI 16 0969 -0,Ol OLBI 16 O F  0,lO 
The two factors were labelled as: Factor 1: Engagement (all the positive items clustered together) 
and Factor 2: Disengagement (all the negative items clustered together). Tucker's phi 
coefficients of 0,96 (Engagement) and 0,91 (Disengagement) resulted after a target rotation was 
carried out. These coefficients can be regarded as acceptable. 
Above tindings confirmed hypothesis 3, that Burnout, as measured by the MBI-GS and OLBI, is 
an equivalent and unbiased construct for different language groups within the earthmoving 
equipment industry. 
The descriptive statistics and alpha coefficients o f  the three factors of the MBI-GS and the two 
factors of the OLBI are given in Table 8. 
Table 8 
Descriptive Statistics and AIpha Coefficients of the MBI-GS and the OLBI (n = 326) 
Item Mean SD Skewness Kuttosis a 
MBI-GS 
I 
Emotional Exhaustion 12,16 7,25 0,35 -472 o,@ 
Cynicism 635 5,45 477 -456 472 
Professional Efficacy 25,42 4,60 -1,33' 2,II' 0,71 
Engagement 1297 333 0,34 -0,57 0,71 
Disengagement 21,60 5,08 0,25 -464 48.2 
According to Table 8, the scores on the MBI-GS are normally distributed for Emotional 
Exhaustion and Cynicism. However, Professional Efficacy is w t  normally distn'buted. Tbe 
scores and the OLBI are normally distributed for both subscales. Tbe Cronbach alpha 
coefficients o f  the subscales of both measuring instruments are considered to be acceptable 
compared to the gui'deline of a > 0,70 (NunnaLly & Bemstein, 1994). 
Above findings confirm hypotheses 1 and 2. 
The correlation coefficients between the MBI-GS and the OLBI are reported in Table 9. 
Table 9 
Correlation Coefficients between the MBI-GS and the OLBl 
Item 1 2 3 4 
1. Exhaustion 
2. Cynicism 0,46*+ 
3. Professional Efficacy -0,29' -0,55* 
4. Engagement 0,39- 0,52* -033% 
5. Disengagement -0,60* -0,49+ -037- -0,45" 
Inspection of Table 9 shows that Cynicism is statistically and practically significantly related to 
Exhaustion. Professional Efficacy is statistically significantly related to Exhaustion and 
practically significantly (negatively) related to Cylllcism (large effect). Engagement is practically 
significantly related to Exhaustion (medium effect), Cylllcism (large effect) as well as negatively 
related to Professional Efficacy (large effect). Disengagement is negatively and practically 
significantly related to Exhaustion (large effect), Cynicism (medium effect) Professional 
Efficacy (medium effect), and Engagement (medium effect). 
MANOVA analysis was conducted to determine the relationship between burnout and various 
demographic characteristics such as race, language, education level w o r m  in different 
business units, and household situations. The results of these comparisons are reported in Table 
10. 
Table 10 
UANOVAs - Diflerences in Burnout Levels ofDemographic Groups 
Variable Value F Df Den M D 
Race 0,94 1,55 12 321,06 0,IO 
Age 0,89 1,35 27 882,64 0,11 
Language o,go 1,27 27 917,68 0,16 
Educational Lcvel 0,94 1,05 18 888,68 0,m 
Business unit 0,95 1,ll 15 856,17 0,35 
Household situation 0,97 0,62 15 878,26 0,86 
After conducting an analysis of W i s  Lambda values, no statistical differences (p < 0,Ol) could 
be found regarding burnout levels of different race groups, language groups or education level. 
Also, no statistical significantly differences (p < 0,Ol) could be found regarding burnout levels of 
individuals at different business units or in different household situations. 
Table 11 indicates the differences between employees in an earthmoving equipment industry, 
based on gender. 
Table 1 1 
Dzfferences between Employees in an Earthmoving Equipment Industty, based on Gender 
Item Male Female P d 
Mean SD Mean SLJ 
MBI-GS 
Exhaustion 12,14 7,13 11,97 7,52 0,85 - 
Cynicism 4,OO -4,23 337 4,54 0,s 1 - 
Professional Efficacy 30,07 4 4 5  31,28 5,OO 0 ~ 7  - 
As can be seen in Table 11, there are no statistically (p < 0,OS) or practically (d  < 0,SO) 
significant differences with respect to the three dimensions of bumout and gender. These 
findings confirm hypothesis 4 .  
DISCUSSION 
The general aim of this research was to validate the Maslach Burnout Inventory - General 
Survey and the OLBI and to determine the construct equivalence and item bias of these 
instruments for employees in an earthmoving equipment industry in South Afiica 
The first objective was to determine the item bias of the MBI-GS and the OLBI. Neither 
uniform nor nonuniform bias was found for the items of the MBI-GS. This confirms the findings 
of Storm and Rotbmann (2003) who also found no uniform or nonuniform bias in a South 
Afiican Police Service sample. Also, no evidence of uniform or nonuniform bias was found for 
the items of the OLBI. 
Based on both empirical and conceptual grounds, item 13 was eliminated h m  the original 
adapted MBI-GS, resulting in a 15-item instrument. When testing the factorial validation of the 
MBI-GS, the deletion of item 13 ("I just want to do my job and not be bothered") is consistent 
with previous studies (e.g. Schutte et al., 2000; Storm & Rothmann, 2003). According to these 
autho~s, problems might be caused by the ambivalent nature of this item. On the one hand, a 
high score may indicate disengagement and social isolation by closing oneself off from contacts 
with others at work. It could, however, also be indicative of strong motivation and engagement. 
Although validation is needed in future studies and the deletion of item 13 was part of the post 
hoc analysis, the decision to eliminate this particular item is consistent with previous research 
(Schutte et al., 2000; Storm & Rothmann, 2003). 
The second objective was to determine the construct equivalence of the MBI-GS. By using 
exploratory factor analysis, three factors were extracted kom the 15-item MBI-GS, namely 
Exhaustion, Cynicism and Professional Efficacy. However, exploratory factor analysis with 
target rotations showed that the construct equivalence of the scales was not acceptable. 
Inspection of the factor loadings revealed that item 11 ("I am enthusiastic about my job") was 
problematic. Item 11 loaded on different factors for the Afiican group and the English group. 
This item loaded on Professional Efficacy in the English group (as it should), but loaded on 
Cynicism in the Afiican group. It is possible that people of different cultures add different 
values to the word "enthusiastic". After the removal of Item 11 from the analysis, target rotation 
resulted in acceptable construct equivalence of the three factors for both language groups in the 
earthmoving equipment industry. 
The third objective was to determine the construct equivalence of the OLBI. Two factors were 
extracted for the OLBI after conducting an exploratory factor analysis. These factors were 
labelled Engagement and Disengagement. Interestingly enough, all the positive items 
(Engagement) and all the negative items (Disengagement) clustered together and loaded on two 
factors respectively. This finding is in contrast with previous research (see Demerouti et al., 
2003). According to Demerouti et al. (2003), both subscales, Exhaustion and Disengagement, 
include positively and negatively worded items that refer to their opposites - vigour or energy, 
and drive or engagement (which have to be reversed coded). These authors found that two of the 
six items on the Exhaustion scale, and four of the seven items on the Disengagement scale, 
loaded positively. 
Furthermore, exploratory factor analysis with target rotations showed that the construct 
equivalence of the scales was not acceptable. Further inspection of the factor loadings revealed 
that item 10 was problematic. Item 10 ("After my work, I usually feel still fit for my leisure 
activities") loaded on different factors for the f i c a n  group and the En&& group. This item 
loaded on Engagement in the African group (as it should), but loaded on Disengagement in the 
English group. It is possible that the item might have been misunderstood, although explanations 
of items were provided and there were people present during the conducting of the inventory 
whom the participants could have asked for explanations. It is also possible that the item is 
sample specific. If the OLBI is conducted on another sample the item might not be problematic. 
However, after the removal of Item 10 h m  the analysis, target rotation resulted in acceptable 
construct equivalence of the two factors for both language groups. 
Another objective was to determine the relationship between the MBI-GS and the OLBI. A 
product-moment correlation was conducted. The results obtained indicated that the Engagement 
scale of the OLBI correlate negatively with all three subscales of the MBI-GS. The 
Disengagement scale correlated positively with all three subscales of the MBI-GS. 
The last objective was to investigate the relationship between burnout and various demographic 
characteristics. The results indicated that no differences exist between levels of burnout and 
race, language, educational level, business unit, household situation and gender. This finding 
confirms the findings of Demerouti et al. (2003). These authors also found no significant effects 
between the demographic variables and any of the burnout scales. 
In conclusion, based on the results obtained in this study it seems as if the MBI-GS is a valid, 
reliable and suitable instruments for measuring burnout of employees working in an earthmoving 
equipment industry. Therefore, the MBI-GS can be used to measure bumout in South Africa in 
other non-contactual occupations. The OLBI did not measure burnout as it was supposed to. A 
possible reason why all the positive items and all the negative items loaded together is that 
positive constructs are measured with positive items, while negative constructs are measured 
with negative items. Extended research need to be conducted in other samples in South Africa to 
determine the validity of the OLBI. 
This study had various limitations. Fi t ly ,  the study relied on self-report measures. Validation 
studies that use self-report measures exclusively increase the likelihood that at least some part of 
the shared variance between measures can be attributed to method variance (Schaufeli & 
Enzmann, 1998). Secondly, only 38,5% of the participants spoke English as their home 
language. This could have contributed to misunderstanding of items and incorrect interpretations 
during the answering of the questionnaires. Thirdly, a cross-sectional design was used, and as a 
result no casual inferences could be drawn. The present study should be considered as one of the 
first attempts to investigate both burnout measures for a South Aliican sample, and it offers 
suggestions for future improvements. 
RECOMMENDATIONS 
Despite the limitations of this study, the present findings may have important implications for 
future research and practice. Based on the results of this study, it is recommended that the MBI- 
GS be used to assess burnout in the earthmoving equipment industry. However, item 13 and 11 
should be left out when administering the MBI-GS. Further research is needed regarding the 
psychometric properties of the OLBI. 
It is suggested that additional research be conducted to determine the validity and reliability of 
both instruments in other South Atiican samples. Especially additional research into the OLBI is 
recommended, seeing as the factors clustered together as negative and positive factors. In future 
studies, larger samples should be used. Large sample sizes might provide increased confidence 
that study findings would be consistent across other similar groups. 
Furthermore, the majority of the respondents have English as a second, or even third, language. 
Therefore translation of the questionnaires into a language other than English can be considered 
in future research. 
REFERENCES 
Applebaum, H.A. (1981). Royal Blue: The culture of construction workers. New York: 
Rinehart and Winston. 
Barkhuizen, E.N. & Rothmann, S. (2004, March). Burnout of Academic staff in a higher 
education institution. Paper presented at the 2"d South f i c a n  Work Wellness Conference, 
Potchefstroom 
Bums, N. & Grove, S.K. (1993). The practice of nursing research, conduct, critique, and 
utilization (2nd 4 . ) .  Philadelphia, PA: W.B. Saunders. 
Clark, L.A. & Watson, D. (1995). Constructing validity: Basic issues in objective scale 
development. Psychological Assessment, 7, 309-3 19. 
Cohen, J. (1998). Statisticalpower analysis of the behavioural sciences (2nd ed.). Orlando, CA: 
Academic Press. 
Demerouti, E. (1999). Burnout: Eine Folge konlcreter Arbeitsbedingungen bei Dienstleistungs- 
und Produkstionstitigkeiten. pumout: A consequence of specific working conditions 
among human services, and production tasks]. Frrmkfurt/Main: Lang. 
Demerouti, E., Bakker, A.B., Nachreiner, F. & Ebbinghaus, M. (2002). From mental strain to 
burnout. European Journal of Work and Organizational Psychology, 11(1), 423-441. 
Demerouti, E., Bakker, A.B., Nachreiner, F & Schaufeli, W.B. (2001). The job demands- 
resource model of burnout. Journal ofApplied Psychology 86(3), 499-512. 
Demerouti, E., Bakker, A.B., Vardakou, I. & Kantas, A. (2003). The convergent validity of two 
burnout instruments: A multitrait-multimethod analysis. European Journal of Psychological 
Assessment, 19, 12-23. 
Demerouti, E. & Nachreiner, F. (1998). Zur Spezifitiit von Burnout fiir Dienstleistungsberufe: 
Fakt oder Artefakt [On the specificity of burnout for human services: Fact or artefact]. 
Zeitschnpfir Arbeitswissenschaft, 52, 82-89. 
Ebbinghaus, M. (1996). Erfassung von Burnout. Entwicklung und Iherpriifung eines 
MeDinstrumentes fiir die Anwendung in verschiedenen Berufsbereichen [Conceptualizating 
burnout: Development and validation of a measurement instrument for the application in 
various working domains]. Unpublished manuscript, Department of Work and 
Organizational Psychology, Univenity of Oldenburg, Oldenburg, Germany. 
Freudenberger, H.J. (1974). Staff bumout. Journal of Social Issues, 30, 159-165. 
Gold, Y. (1984). Bumout: A Major problem for the teaching profession. Education, 104(3), 
271-275. 
Guilford, J.P. (1954). Psychometric methods. New York: McGraw-Hill. 
Jackson, L.T.B. & Rothmann, S. (2004, March). The validation of the Maslach Burnout 
Inventory for South Afican teachers. Paper presented at the 2" South African Work 
Wellness Conference, Potchefstroom. 
Jansen van Vuuren, J.A. & Rothmann, S. (2002, September). The relationship between 
psychological strengths, coping and burnout of senior managers in a manufacturing 
industry. Paper presented at the 14& Conference of the South African Institute for 
Management Scientist, Pretoria 
Lee, R.T., & Ashfort, B.E. (1996). A meta-analytic examination of the correlates of the three 
dimensions of burnout, Journal of Applied Psychology, 81, 123-133. 
Leiter, M.P. (1993). Burnout as a developmental process: Consideration of models. In W.B. 
Schaufeli, C. Maslach, & T. Marek (eds.), Professional burnout: Recent developments in 
theory and research (pp.237-250). Washington, DC: Taylor & Francis. 
Lingard, H. (2003). The impact of individual and job characteristics on 'burnout' among civil 
engineers in Australia and the implications of employee turnover. Construction Management 
and Economics, 21(1), 69-80. 
Maslach, C. (1976). Bumed-out. Human Behavior, 5, 16-22. 
Maslach, C. & Jackson, S.E. (1986). MBI: Maslach Burnout Inventory: Manual research 
edition. Palo Alto, CA: Consulting Psychologists Press. 
Maslach, C. & Leiter, M.P. (1997). The truth about burnout. San Francisco, C k  Jossey-Bass. 
Mellenbergh, G.J. (1982). Contingency table models for assessing item bias. Journal of 
Educational Statistics, 52,397-422. 
Naudk, J.L.D. & Rothmann, S. (2004, March). Burnout of emergency workers in South Afica. 
Paper presented at the 2"d South African Work Wellness Conference, Potchefstroom. 
Rothmann, S., Malan, A.M. & Rothmann, J.C. (2001). Sense of coherence, coping and burnout 
in a corporate pharmacy group. Paper presented at the 7& annual PsySSA Conference, 
Johannesburg. 
SAS Institute. (2000). The SAS System for Windows: Release 8.01. Cary, NC: SAS Institute Inc 
Schaufeli, W.B. & Enzmann, D. (1998). Ike burnout companion to study and practice: A 
critical analysis. London: Taylor & Francis. 
Schaufeli, W.B., Leiter, M.P., Maslach, C. & Jackson, S.E. (1996). MBI-General Survey. In C. 
Maslach, S.E. Jackson & M.P. Leiter (Eds.), Maslach Burnout Znvento~y Manual (3" 4.). 
Palo Alto, CA: Consulting Psychologists Press. 
Schaufeli, W.B., Martinez, I.M., Pinto, A.M., Salanova, M. & Bakker, A.B. (2002). Burnout 
and engagement in university students: A cross-sectional study. Journal of Cross-Cultural 
Psychology, 33,464-48 1. 
Schutte, N., Toppinen, S., Kalimo, R. & Schaufeli, W.B. (2000). The factorial validity of the 
Maslach Burnout Inventory - General Survey (MBI-GS) across occupational groups and 
nations. Journal of Occupational and Organizational Psychology, 73, 53-66. 
Shaughnessy, J.J. & Zechmeister, E.B. (1997). Research methodr in psychology (4m ed.). New 
York: McGraw-Hill. 
Sin& S. (1997). State of the art in automation of earthmoving. Journal of Aerospace 
Engineering, 10(4), 179- 188. 
SPSS Inc. (2003). SPSS 12.0 for Windows. Chicago, IL: Author. 
Step, H.S. (1999). Praktiese betekenisvolheid: Die gebruik van effekgroottes. Wetenskaplike 
bydraes- Reeks B: Natuurwetenskappe Nr. I 17. Potchefstroom: Potchefstroome 
Universiteit vir Christelike Ho& Onderwys. 
Storm, K. (2002). Burnout and work engagement in the South Afncan Police Service. 
Unpublished doctoral thesis, PU for CHE, Potchefskoom. 
Storm, K. & Rothmann, S. (2003). A psychometric analysis of the Maslach Burnout Inventory 
- General Survey in the South African Police Services. South Afncan Journal of 
Psychology, 39(4), 219-266. 
Tabaclmick B.G. & Fidell, L.S. (2001). Using multivariate statistics (4& 4.) .  Boston, MA: 
A U p  & Bacon. 
Van de Vijver, F. & Leung, K. (1997). Methodr and data analysis for cross-cultural research. 
Thousand Oaks, CA: SAGE. 
CHAPTER 3 
CONCLUSIONS, LIMITATIONS AND RECOMMENDATIONS 
This chapter comprises conclusions regarding the literature review and the empirical study 
according to the specific objectives. The limitations of the research are discussed, followed by 
recommendations for the research problem in the organisation. Lastly, suggestions are made for 
future research. 
3.1 CONCLUSIONS 
The fust objective of this study was to determine the reliability and validity of the MBI-GS for 
employees in the earthmoving equipment industry in South Atiica Based on the results obtained 
in this study, it seems that the MBI-GS is a valid, reliable and suitable instrument for measuring 
burnout of employees working in an earthmoving equipment industry. Therefore the MBI-GS 
can be used to determine burnout in South Atiica in other non-contactual occupations. 
The second objective was to determine the reliability and validity of the OLBI for employees in 
the earthmoving equipment industry in South Atiica Based on the results obtained in this study, 
it seems that the OLBI did not measure burnout as it should have. The reason why all the 
positive items and all the negative items loaded together could be of a statistical nature. A 
possible reason could be that negative constructs should be measured with negative items, and 
vice versa. Extended research needs to be conducted in other samples in South Atiica in order to 
determine the validity of the OLBI. 
The third objective was to determine the construct equivalence and item bias of the MBI-GS. 
Based on both empirical and conceptual grounds, item 13 was eliminated h m  the original 
adapted MBI-GS, resulting in a 15-item instrument. It seems as if problems might be caused by 
the ambivalent nature of this item. On the one hand, a high score may indicate disengagement 
and social isolation by closing oneself off h m  contacts with others at work. It could, however, 
also be indicative of strong motivation and engagement: one concentrates on the task and does 
not want to be interrupted. The deletion of item 13 ("I just want to do my job and not be 
bothered") is consistent with previous studies (e.g. Schutte, Toppinen, Kalimo & Schaufeli, 
2000; Storm & Rothmann, 2003). 
An item is unbiased if persons from different language groups, with an equal standing on the 
theoretical construct underlying the instrument, have the same expected score on the item (e.g. 
Van de Vijver & Leung, 1997). Neither uniform nor nonuniform bias was found for the items of 
the MBI-GS. This confirms the findings of Storm and Rothmann (2003) who also found no 
uniform or nonuniform bias in a South Afiican Police Service sample. 
By using exploratory factor analysis, three factors were extracted h m  the 15-item MBI-GS, 
namely Exhaustion, Cynicism and Professional Efficacy. The three-factor model represented the 
data quite well, albeit after removal of one unsound item (item 11). After the removal of this 
item from the analysis, target rotation resulted in acceptable construct equivalence of the three 
factors for both language groups in the earthmoving equipment industry. 
The fourth objective was to determine the construct equivalence and item bias of the OLBI. No 
evidence of uniform or nonuniform bias was found for the items of the OLBI. Since this is the 
first study to measure burnout with the OLBI, extended research on other samples in South 
Afiica is necessary to confirm the findings of the OLBI. 
Two factors were extracted from the 15-item OLBI after conducting an exploratory factor 
analysis. These factors were labelled Engagement and Disengagement. All the positive items 
(labelled "Engagement") and all the negative items (labelled "Disengagement") clustered 
together and loaded on two factors respectively. This finding is in contrast to previous research 
(see Demerouti et al., 2003). According to these authors, both subscales, Exhaustion and 
Disengagement, include positively and negatively worded items that refers to its opposites - 
vigour or energy and drive or engagement (which have to be reversed coded). These authors 
found that two of the six items on the Exhaustion scale, and four of the seven items on the 
Disengagement scale, loaded positively. 
The two-factor model represented the data quite well, albeit after removal of one unsound item 
(item 10). After the removal of this item from the analysis, target rotation resulted in acceptable 
construct equivalence of the three factors for both language groups in the earthmoving 
equipment industry. 
It is believed that this confusing state of affairs does not necessarily reflect weaknesses inherent 
in the OLBI - that it could also be due to more general factors. Firstly, the OLBI is a recently 
constructed measuring instrument. Therefore relatively few studies have critically reviewed its 
psychometric properties. Secondly, the OLBI is an instrument that was originally constructed 
from data based on samples of individuals in Germany (Demerouti & Nachreiner, 1999). Hence, 
valid research that compares levels of work engagement in South Atiica is lacking and a 
thorough psychometric evaluation of this instrument in our specific national context will be 
influenced by the specific culwe of the country (or, more specifically, the culture of the 
earthmoving equipment industry). 
The fifth objective was to determine the relationship between the MBI-GS and the OLBI. A 
product-moment correlation was conducted The results obtained indicated that the Engagement 
scale of the OLBI correlated negatively with all three subscales of the MBI-GS. The 
Disengagement scale correlated positively with all three subscales of the MBI-GS. 
The last objective was to investigate the relationship between burnout and various demographic 
characteristics. The results indicated that no differences with respect to levels of burnout exist 
between race, language, educational level, business unit, household situation and gender. This 
finding confirms the findings of Demerouti, Bakker, Vardakou & Kantas (2003). These authors 
also found no significant effects between the demographic variables and any of the burnout 
scales. 
3.2 LIMITATIONS 
The first limitation of this study is that the results were obtained solely by self-report 
questionnaires. According to Schaufeli and Enzmann (1998), this may lead to a problem called 
"method-variance" or "nuisance". However, there are not a lot of alternative methodologies to 
deal with the use of self-report measures. Research into more objective measures is still needed. 
A suggestion was made that a logical way to capture the complex interactions between the 
numerous variables associated with burnout would be through a conceptually multivariate 
approach represented by a mosaic of individual traits and joblenvironmental factors (Dolan, 
1995). 
A second liitation of this study is that the majority of the participants did not speak English as 
their home language. This could have influenced the way in which the respondents answered the 
questionnaires, as a result of misunderstandings and incorrect interpretations. 
Thirdly, the design used in this study was a cross-sectional design, and no causal inferences 
could be drawn. Therefore, the causal relationships between variables were interpreted rather 
than established, and more complex f o m  of non-recursive linkages could not be examined. To 
deal with the limitations set by using a cross-sectional design, prospective longitudinal studies 
and quasi-experimental research designs are needed to further validate the hypothesised causal 
relationships. 
Another limitation is that the research was conducted in a homogeneous sample consisting of 
individuals of a specific profession, namely the earthmoving equipment industry. It should be 
noted that unique characteristics probably exist within this profession - such as the specific 
organisational culture, which could have influenced the participants' responses. The implication 
is that the results could not be generalised to other contexts or professions. Consequently, there 
is still the need for replication in other occupational groups and in heterogeneous samples. 
3.3 RECOMMENDATIONS 
The following recommendations are made for the organisation as well as for future research. 
3.3.1 Recommendations for the organisation 
A clear and accurate understanding of burnout is necessary in order to implement effective 
individual, managerial and organisational practices to deal with burnout. All individuals 
working within the earthmoving equipment industry should become aware of the causes and 
symptoms of burnout in order to become aware of their own and other people's exhaustion, 
cynicism, and low professional efficacy. If they are aware of these symptoms, interventions can 
be conducted before the effects of burnout become too serious. It is also important for managers 
to know whether they are suffering from burnout, in view of the danger that they could spread it 
to their subordinates. 
Prior to employees entering the earthmoving equipment industry, training programmes focusing 
on the identification of symptoms of burnout (with ongoing supervision and knowledge of 
effective stress management techniques) could reduce burnout. Employees currently in the 
industry should also be offered programmes where they can receive education on identifying the 
signs and symptoms of burnout, and on how to reverse the situation. 
Planned interventions should be designed and implemented. According to Lee and Ashfort 
(1993), these interventions should be designed for the long term and should deal with the root 
cause rather than just the symptoms. Individual-based interventions to reduce burnout symptoms 
might be an avenue to pursue. These interventions include techniques such as self-monitoring, 
self-assessment, didactic stress management, promoting a healthy lifestyle and relaxation (see 
Schaufeli & Enzmann, 1998, pp. 146-168). Stress-management programmes that use a 
cognitivebehavioural approach are also effective in reducing stress reactions, including burnout. 
Organisational development interventions in general, as well as interventions to influence culture 
and values, should be implemented to contribute to healthier workplaces. 
33.2 Recommendations for future research 
Future research in South Afiica needs to focus on the relative prevalence of burnout in various 
occupations. The differences in levels of burnout found between occupational groups may help 
identify occupations that are most at risk of burnout. This will also enable researchers to 
determine the relationship between variables (so as to ascertain which variable causes burnout) 
as well as the relationship between this variable and the remaining variables. 
Based on the results of this study, it is recommended that the MBI-GS be used to assess burnout 
in the earthmoving equipment industry. However, item 13 and 11 should be left out when 
administering the MBI-GS. Further research is needed regarding the OLBI. 
It is suggested that additional research be done to determine the validity and reliability of both 
instruments in other South African samples. Especially additional research on the OLBI is 
recommended, seeing as the factors clustered together as negative and positive factors. In future 
studies, larger samples should be used. It is also recommended that larger samples with more 
powerful sampling methods be utilised to enable generalisation of the findings to other similar 
groups. 
Furthermore, the majority of the respondents have English as a second, or even third, language. 
Therefore translation of the questionnaires into a language other than English (for example 
Afrikaans or Setswana) can be considered in future research. 
REFERENCES 
Demerouti, E., Bakker, A.B., Vardakou, I. & Kantas, A. (2003). The convergent validity of two 
burnout instruments: A multitrait-multimethod analysis. European Journal of Psychological 
Assessment, 19, 12-23. 
Demerouti, E. & Nachreiier, F. (1998). Burnout as an outcome ofjob demands, job resources 
and specific short-term consequences of strain. Paper presented at the 5m European 
Conference on Organizational Psychology and Health Care, Belgium. 
Lee, R.T. & Ashfort, B.E. (1993). A further examination of managerial burnout: Toward an 
integrated model. Journal of Organizational Behavior, 14, 3-20. 
Schaufeli, W.B. & Enzmann, D. (1998). The burnout companion to study and practice: A 
critical analysis. London: Taylor & Francis. 
Schutte, N., Toppinen, S., Kalimo, R. & Schaufeli, W.B. (2000). The factorial validity of the 
Maslach Burnout Inventory - General Survey P I - G S )  across occupational groups and 
nations. Journal of Occupational and Organizational Psychology, 73, 53-66. 
Storm, K. & Rothmann, S. (2003). A psychometric analysis of the Maslach Burnout Inventory 
- General S w e y  in the South African Police S e ~ c e s .  South Afican Journal of 
Psychology. 39(4), 219-266. 
Van de Vijver, F. & Leung, K. (1997). Methods and data analysis for cross-cultural research. 
Thousand Oaks, CA: SAGE.