Content uploaded by Gerhard (Gert) Roodt
Author content
All content in this area was uploaded by Gerhard (Gert) Roodt on Oct 24, 2014
Content may be subject to copyright.
Original Research
doi:10.4102/sajhrm.v11i1.507
hp://www.sajhrm.co.za
The validaon of the turnover intenon scale
Authors:
Chris F.C. Bothma1
Gert Roodt1
Aliaons:
1Department of Industrial
Psychology and People
Management, University of
Johannesburg, South Africa
Correspondence to:
Gert Roodt
Email:
groodt@uj.ac.za
Postal address:
PO Box 524, Auckland Park
2006, South Africa
Dates:
Received: 01 Nov. 2012
Accepted: 13 Dec. 2012
Published: 15 Apr. 2013
How to cite this arcle:
Bothma, C.F.C., & Roodt, G.
(2013). The validaon of the
turnover intenon scale. SA
Journal of Human Resource
Management/SA Tydskrif vir
Menslikehulpbronbestuur,
11(1), Art. #507, 12 pages.
hp://dx.doi.org/10.4102/
sajhrm.v11i1.507
Copyright:
© 2013. The Authors.
Licensee: AOSIS
OpenJournals. This work
is licensed under the
Creave Commons
Aribuon License.
Orientation: Turnover intention as a construct has attracted increased research attention in
the recent past, but there are seemingly not many valid and reliable scales around to measure
turnover intention.
Research purpose: This study focused on the validation of a shortened, six-item version of the
turnover intention scale (TIS-6).
Motivation for the study: The research question of whether the TIS-6 is a reliable and a valid
scale for measuring turnover intention and for predicting actual turnover was addressed in
this study.
Research design, approach and method: The study was based on a census-based sample
(n = 2429) of employees in an information, communication and technology (ICT) sector
company (N = 23 134) where the TIS-6 was used as one of the criterion variables. The leavers
(those who left the company) in this sample were compared with the stayers (those who
remained in the employ of the company) in this sample in respect of different variables used
in the study.
Main ndings: It was established that the TIS-6 could measure turnover intentions reliably
(α = 0.80). The TIS-6 could signicantly distinguish between leavers and stayers (actual
turnover), thereby conrming its criterion-predictive validity. The scale also established
statistically signicant differences between leavers and stayers in respect of a number of the
remaining theoretical variables used in the study, thereby also conrming its differential
validity. These comparisons were conducted for both the 4-month and the 4-year period after
the survey was conducted.
Practical/managerial implications: Turnover intention is related to a number of variables
in the study which necessitates a reappraisal and a reconceptualisation of existing turnover
intention models.
Contribution/value-add: The TIS-6 can be used as a reliable and valid scale to assess turnover
intentions and can therefore be used in research to validly and reliably assess turnover
intentions or to predict actual turnover.
Introducon
The retention of staff is considered to be a pressing people issue and consequently much has
been published about it (cf. Bothma & Roodt, 2012; Du Plooy & Roodt, 2010; Greyling & Stanz,
2010; Griffeth, Horn, & Gaertner, 2000; Kotzé & Roodt, 2005; Mendes & Stander, 2011). Turnover
intentions (intentions to stay or leave the organisation) is an important criterion variable in similar
types of studies, but such studies seldom publish any additional validation information on these
criterion measures. The challenge and importance of this study therefore is to develop a scale that
can serve as a valid and reliable criterion variable in future turnover or retention studies.
Although turnover intention is covered well in the literature, the need remains to validate turnover
cognition scales (Sager, Griffeth & Horn, 1998). The motivation for validating the shortened
version of the turnover intention scale (TIS-6) is that most other scales use only a limited number
of scale items. Martin (2007) observed that various researchers have only used single item scales
(Guimaraes, 1997; Lambert, Hogan, & Barton, 2001), with obvious metric limitations. According
to Martin (2007), only a limited number of other studies have used more than three items in their
instruments (Becker, 1992; Fox & Fallon, 2003; Lum, Kervin, Clark, Reid & Sirola, 1998). It seems
that information on the metric properties of such instruments is lacking and that no validation
research is reported specically on the TIS-6 (the studies by Jacobs [2005] and Martin [2007] report
on the longer TIS versions).
The main research question of this study is therefore as follows: is the shortened TIS-6 a reliable and
valid scale for measuring turnover intention and for predicting actual turnover? The objectives of
Page 1 of 12
Scan this QR
code with your
smart phone or
mobile device
to read online.
Read online:
Original Research
doi:10.4102/sajhrm.v11i1.507
hp://www.sajhrm.co.za
the study are to investigate, (1) the reliability, (2) the construct
(factorial) validity, (3) the criterion-predictive validity and
(4) the differential validity of the TIS-6 within the context
of a large South African information, communication and
technology (ICT) sector company. The contribution of this
study would be that a valid and reliable turnover intention
scale is developed for future use as a criterion or predictor
variable. It is also important to use valid and reliable scales
as a proxy for predicting actual turnover.
Literature review
Dening turnover intenon
Bester (2012) noted that turnover intention is seldom
precisely dened in reported studies. He concluded that
this practice is probably attributable to the assumption that
people perceive the term to be self-explanatory. Bester (2012)
further argued that many researchers (Horn, Griffeth &
Salaro, 1984; Mobley, 1982; Mowday, Steers, & Porter, 1979;
Steers, 1977) viewed turnover intention as the nal step in
the decision-making process before a person actually leaves
a workplace. Turnover intention can therefore be described
as an individual’s behavioural intention or conation,
in Fishbein and Ajzen’s (1975) framework of planned
behaviour, to leave the employ of the organisation. Lacity,
Lyer and Rudramuniyaiah (2008, p. 228) dened turnover
intention as ‘… the extent to which an employee plans to
leave the organisation’. For the purpose of this study, the
denition of Tett and Meyer (1993, p. 262) is used, who
aptly dened turnover intention as: ‘… the conscious and
deliberate wilfulness to leave the organisation’. The TIS-
6 was developed as a conation (intention) to distinguish it
from the affective (emotion) and the cognitive (knowledge)
components of psychological activities as conceptualised by
Fishbein and Ajzen (1975).
Against the background of Fishbein and Ajzen’s (1975)
theoretical framework, behavioural intention is a reliable
determinant of actual behaviour (Jaros, Jermier, Koehler &
Sincich, 1993; Muliawan, Green & Robb, 2009). It has also been
empirically established that turnover intention (conation) has
a positive relationship with actual turnover (actual behaviour)
(Byrne, 2005; Hendrix, Robbins, Miller & Summers, 1998;
Steensma, Van Breukelen & Sturm, 2004). Several authors
argued that turnover intention can be used as a valid proxy
for actual labour turnover (Jaros et al., 1993; Muliawan et al.,
2009; Tett & Meyer, 1993). Turnover is the result of a coping
strategy used by employees to escape the current situation
(cf. Petriglieri, 2011). Turnover can be permanent, when
employees leave the employment institution, or it can be
characterised by horizontal mobility when employees seek
and accept transfers to other departments (Kirpal, 2004). Tett
and Meyer (1993, p. 262) referred to turnover as ‘… the last
in a sequence of withdrawal cognitions …’, a practice that
Petriglieri (2011, p. 648) named an ‘identity exit’.
Theorecal models that explain turnover intenons
Several authors (Bakker & Demerouti, 2006; Jacobs, 2005;
Lee & Mitchell, 1994; Mobley, 1982; Morrell, Loan-Clarke,
Arnold & Wilkinson, 2008; Petriglieri, 2011; Schaufeli &
Bakker, 2004; Zeffane, 1994) have developed and tested
models in an attempt to explain turnover intentions and
related constructs. Perhaps the most prominent of these
is the job resources-demands (JD-R) model (Bakker &
Demerouti, 2006; Bakker, Demerouti & Verbeke, 2004),
which provides plausible explanations as to why individuals
may choose to leave an organisation. In most studies that
used the JD-R model, the path to turnover intention is the
result of job demands that cause burnout. An indirect
relationship between job demands and turnover intention
is therefore proposed. Bester (2012) also suggested that
this idea is based upon studies which have found that
job demands, especially when there are less resources,
stimulate exhaustion (the opposite of engagement) and, in
turn, cause turnover intentions (Bakker & Demerouti, 2006;
Bakker et al., 2004; Demerouti, Bakker, Nachreiner &
Schaufeli, 2000; Schaufeli & Bakker, 2001; Schaufeli &
Bakker, 2004). According to Schaufeli and Bakker (2004),
the link between work engagement, burnout and turnover
intention is well established. The abovementioned studies
also indicated that the absence of job resources was related
to disengagement, which increased turnover intention. A
possible limitation of the JD-R model may be that it mostly
emphasises contextual and/or organisational resources and
demands and, to a lesser extent, personal resources or the
role of personal agency. Sweetman and Luthans (2010), on
the other hand, introduced the concept of ‘psychological
capital’ (personal resources), which includes facets such as
efcacy, optimism, hope and resiliency that may act as a
buffer between contextual demands and turnover intention.
The nding that the absence of job resources stimulated
turnover intention was also supported in a study that did
not use the JD-R model (Agarwal, Ferrat & De, 2007). Along
a similar line, Du Plooy and Roodt (2010) indicated on a
bivariate level that both work engagement and organisational
citizenship behaviour are negatively related to turnover
intention, whilst work alienation and burnout are positively
related. In a stepwise multiple regression, however, work
alienation explains the largest amount of variance (54%) in
turnover intention, whilst the beta weight of organisational
citizenship behaviour (β = 0.064) in the prediction model
turned positive. These ndings therefore suggest that work
engagement and work alienation should rather be viewed as
polar opposites with organisational citizenship behaviours
and burnout, respectively, as resulting consequences. It
seems that in tight economic or labour market conditions,
individuals do not wish to ‘burn bridges’, which may explain
the positive relationship between organisational citizenship
behaviours and turnover intention.
Jacobs (2005) proposed a different turnover intention model,
where positive or negative perceptions of organisational
culture (predictors) were related to turnover intentions
(criterion). A number of variables mediated this said
relationship, such as job satisfaction, organisational
citizenship behaviour, organisational commitment and
knowledge sharing (cf. Boshoff, Van Wyk, Hoole & Owen,
Page 2 of 12
Original Research
doi:10.4102/sajhrm.v11i1.507
hp://www.sajhrm.co.za
Page 3 of 12
2002; Wasti, 2003 for similar types of models). Individuals’
perception of organisational culture may therefore trigger
key mediating variables, which may, again in turn, lead to
decisions to leave or stay with the organisation.
Another theoretical framework which may shed light on an
individual’s decision to exit an organisation is Petriglieri’s
(2011) theory of identity threat responses. In a nutshell, this
theory argues that individuals assess the identity threat and
possible coping responses against the threat strength and the
background of existing social support. This results in two
broad coping strategy categories of either identity protection
responses or identity restructuring responses. Identity exit
is one of the identity restructuring responses which will
eliminate the identity threat. This model has particular
relevance for turnover within the conceptual framework of
work identity.
Implicaons of turnover intenons
Bothma (2011) argued that leaving a job may not always
be an option for an individual. The decision to leave is
inuenced by many personal and contextual factors such as
employability and labour market conditions. An individual’s
turnover intention is dependent on perceived chances and
the ease of nding another job (especially in tough economic
conditions), the role of mobility cognitions, as well as
individual differences in search behaviour. Alternative
employment opportunities therefore inuence actual labour
turnover behaviour (Agarwal et al., 2007; Akgün & Lynn,
2002; Allen & Meyer, 1996; Bellou, 2008; Boies & Rothstein,
2002; Brown, 1996; Carmeli & Gefen, 2005; Chen, Chu, Wang
& Lin, 2008; Jaros et al., 1993; Lee & Mitchell, 1994; Martin &
Roodt, 2008; Mobley, 1982; Senter & Martin, 2007; Wheeler,
Gallagher, Brouer & Sablynski, 2007).
Bothma (2011) concluded that the turnover phenomenon
has signicant cost and other negative consequences for
any organisation (Bluedorn, 1982; Greyling & Stanz, 2010;
Mobley, 1982). Losing employees that are highly skilled
may have disruptive implications for organisations, such as
impaired organisational functioning, service delivery and
administration. It may also contribute to increased costs
of re-hiring and re-training employees (Roodt & Bothma,
1997; Sulu, Ceylan & Kaynak, 2010). These mentioned
consequences provide a sound rationale for this validation
study of the TIS-6.
In summary, this study will investigate and make comparisons
of those that leave the employ of the organisation (leavers)
versus those that stay in service (stayers) in respect of
turnover intentions, work-based identity scores, the three
dimensions of work engagement, the three dimensions of
burnout, organisational citizenship behaviour, personal
alienation and task performance.
Research design
Research approach
The research approach followed in this study is empirical
and quantitative, where a cross-sectional eld survey
generated the primary research data for this study. For data
analyses, correlational statistical procedures were applied
for generating plausible, ex post facto explanations for
relationships between variables.
Research method
The research method used in this study will be explained
under the headings that follow. A more detailed explication
of the research method can also be found in Bothma and
Roodt (2012, pp. 6–8), whilst the discussion of the measuring
instruments are also detailed in the rst author’s thesis
(Bothma, 2011).
Research parcipants
A census-based sampling approach1 was used to survey the
target population below middle management (N = 23 134),
in the service of a South African ICT sector company. The
survey was conducted over a 1-month period with a Web-
based questionnaire application. An invitation to participate
in the survey was sent to the entire target population via
e-mail, with the universal resource locator (URL) address
attached for ease of responding. Responses on the Web-
based questionnaire were anonymous. A response rate of
about 11% yielded a sample of 2429 research participants.
This sample was used for the comparisons of leavers and
stayers over both the 4-month and the 4-year period.
Table 1 reects that most participants were men (63.2%). The
majority of the participants (44.1%) were White, followed
by Black (26.3%), Coloured (16.3%) and Asian or Indian
(13.3%). These ethnic proportions reect the heterogeneity of
the company’s work force. The mean age of the participants
was about 40 years, which reects a mature labour force. The
majority of the respondents were from operational levels
(55.0%) and were stationed in the corporate region (25.0%) of
the company. About 41.0% of the participants had a Matric
or lower qualication, followed by 27.0% that possessed a
National or National Higher Diploma.
Measuring instruments
A number of established measuring instruments with
known reliabilities and validities were used in this study.
Owing to the lack of space, not all the validity and reliability
coefcients as reported in previous studies can be reported
here. Only brief reference will be made to Cronbach alpha
reliabilities reported by the original authors and those found
in this study.
Turnover intention scale: Turnover intention (the
intention to leave or stay) was measured with a six-item scale
1.Bothma and Roodt (2012) described a census-based sampling approach as follows:
‘Before the term census-based sampling can be understood, the terms census and
random sample need to be explained. In a census the whole target populaon is
surveyed and parcipaon is compulsory. A random sample on the other hand is
a randomly selected poron of the target populaon; they can choose whether to
parcipate in the survey or not. A census-based sampling approach enumerates all
members of the target populaon (similar to a census) with the choice to parcipate
in the survey or not. Self-selecon bias (which falls outside the control of the
researcher) equally aects response rates of census-based as well as other random
sampling strategies. Because a census-based sampling approach enumerates the
complete populaon as a sample, it is a more accurate sampling strategy compared
to normal random sampling strategies where only small porons of the populaon
are sampled’.
Original Research
doi:10.4102/sajhrm.v11i1.507
hp://www.sajhrm.co.za
Page 4 of 12
adapted from the 15-item scale initially developed by Roodt
(2004). To enhance the reliability of responses, behaviour
intention should be measured within a reasonable timeframe
after accepting a position within a company. Based on
recommendations from literature (Muliawan et al., 2009), this
study used a 6-month period.
Examples of items included in the TIS-6 are: ‘How often have
you considered leaving your job?’ and ‘How often do you
look forward to another day at work?’ Jacobs (2005) reported
a Cronbach alpha coefcient of 0.91 for the 15-item version
of the TI scale. Martin (2007) and Martin and Roodt (2008) in
their study reported a Cronbach alpha coefcient of 0.90 for
a 13-item version of the scale. The reliability of the TIS-6 will
be reported in the ‘Results’ section.
Alienation scale: The ve-item alienation scale (AL) of Banai,
Reisel and Probst (2004) is based on the personal alienation
scale of Korman, Wittig-Berman and Lang (1981) and was
also later used by Banai and Reisel (2007) to measure work
alienation in a cross-national study.
Examples of the selected items are: ‘To what extent do you
feel that your daily activities don’t reect your real interests
and values?’ and ‘How likely is it that you would prefer to
live a different life than you are currently doing?’ Banai and
Reisel (2007) reported a Cronbach alpha coefcient of 0.80 for
the AL. The Cronbach alpha reliability for the scale in this
study was 0.81.
Helping behaviour: According to Podsakoff, MacKenzie,
Paine and Bachrach (2000), helping behaviour includes
various conceptualisations such as altruism, peace-making,
cheerleading and interpersonal helping. Helping behaviour
was measured with a nine-item scale of which ve items
were from the helping behaviour scale (Van Dyne & LePine,
1998) and four items from the altruism dimension of the
citizenship behaviour scale (Smith, Organ & Near, 1983).
Examples of the selected items are: ‘How often do you
volunteer to do things in your work group?’ and ‘How
often do help others who have heavy workloads?’ Van Dyne
and LePine (1998) reported Cronbach alpha coefcients
for the helping behaviour scale in a range from 0.88 to
0.95. This study reported a Cronbach alpha coefcient of
0.86 for helping behaviour measured by the combined and
adjusted scale.
Maslach burnout inventory – human services survey: For
the purpose of this study, the Maslach burnout inventory
– human services survey (MBI-HSS-20) (Maslach &
Jackson, 1981) was used to measure burnout at work. The
20-item instrument is composed of three dimensions, namely
emotional exhaustion (EE) (eight items) with Cronbach alpha
coefcients ranging from 0.85 to 0.90, depersonalisation (DP)
(ve items) with Cronbach alpha coefcients ranging from
0.58 to 0.79 and reduced personal accomplishment (PA)
(seven items) with Cronbach alpha coefcients ranging
from 0.70 to 0.71 (Gil-Monte, 2005; Maslach, Jackson &
Leiter, 1996).
Examples of the selected items are: ‘I feel emotionally drained
from my work’ and ‘I feel used up at the end of the work day.’
A seven-point Likert-type frequency rating scale ranging
between extreme values of 0 (never) and 6 (always) was used
to rate job burnout items (Maslach, Jackson & Leiter, 1986,
1996). This study found Cronbach alpha coefcients of 0.89
for emotional exhaustion, 0.70 for depersonalisation and 0.71
for reduced personal accomplishment.
Utrecht work engagement scale: The Utrecht work
engagement scale (UWES-17) was used to measure work
engagement in this study. The 17-item version consists
of three dimensions, namely vigour (VI) (six items) with
Cronbach alpha coefcients ranging from 0.75 to 0.82,
dedication (DE) (ve items) with Cronbach alpha coefcients
ranging from 0.88 to 0.90 and absorption (AB) (six items)
with Cronbach alpha coefcients ranging from 0.70 to 0.77
(Schaufeli, Salanova, Conźalez-Romá & Bakker, 2002).
Examples of scale items are: ‘At my work, I feel bursting with
energy’ and ‘Time ies when I’m working’. A seven-point
Likert-type frequency rating scale ranging between extreme
TABLE 1: Biographical and demographical prole of the respondents (n = 2429).
Variable Category Frequency (f) Percentage (%)
Age (years) 20–29 292 12.0
30–39 960 39.5
40–49 877 36.1
50+ 300 12.4
Gender Female 893 36.8
Male 1536 63.2
Race Black 640 26.3
White 1070 44.1
Coloured 395 16.3
Asian or Indian 324 13.3
Job tenure (years) 0–1 205 8.4
2–5 433 17.8
6–10 700 28.8
11–15 303 12.5
16–20 226 9.3
20+ 562 23.1
Educaon Grade 12 or less 988 40.7
Post-school cercate or
diploma
479 19.7
Naonal Diploma or
Naonal Higher Diploma
653 26.9
Bachelor’s degree or
equivalent, or more
309 12.7
Locaon Central 119 4.9
Corporate 605 24.9
Eastern 318 13.1
Gauteng 450 18.5
North-eastern 336 13.8
Southern 159 6.5
Western 442 18.2
Marital status Single 511 21.0
Married or cohabing 1678 69.1
Divorced or separated 214 8.8
Widowed 26 1.1
Level Management 446 18.4
Operaonal 1334 54.9
Specialist 649 26.7
Original Research
doi:10.4102/sajhrm.v11i1.507
hp://www.sajhrm.co.za
Page 5 of 12
values of 0 (never) and 6 (always) was used to rate work
engagement items (Schaufeli & Bakker, 2003). This study
found a Cronbach alpha coefcient of 0.91 for the UWES-17.
Task performance scale: Task performance assessment
was independently conducted by participants’ supervisors.
These assessments were measured with an adaptation of
a nine-item scale (Rotenberry & Moberg, 2007; Williams &
Anderson, 1991).
Examples of the scale items are: ‘How often does this
employee perform the tasks that are expected from him or
her?’ and ‘How frequently does this employee fail to perform
essential duties?’ Care was taken that the scale did not
overlap with items related to contextual, helping behaviour
performance as discussed above. Rotenberry and Moberg
(2007) reported a Cronbach alpha coefcient of 0.91 for the
task performance scale. This study found a Cronbach alpha
coefcient of 0.94 for task performance.
Work-based identity scale: Previous attempts were made
to measure work-based identity (Aryee & Luk, 1996; Buche,
2003, 2006, 2008; Walsh & Gordon, 2007; Wayne, Randel &
Stevens, 2006), but no suitable measuring instrument was
found that complied with the theoretical denition of work-
based identity. Different scales that measure different facets of
work-based identity as dened in the work-based identity
prototype (refer to Bothma, 2011), such as work role centrality,
person–environment t, organisational identication, job
involvement, occupational and/or professional identity
and career identity were sourced, adapted and combined to
measure work-based identity (Lauver & Kristof-Brown, 2001;
Lodahl & Kejner, 1965; Mael & Ashforth, 1992; Roodt, 1997;
Roodt, De Braine, Bothma & Jansen, 2009; Serani, Maitland
& Adams, 2006).
A proposed work-based identity scale was compiled,
consisting of 36 items representing the different facets of
work-based identity, selected from a number of various
pre-existing scales. Firstly, items were selected from the
organisational-related commitment scale of Roodt (1997).
Examples of selected items are: ‘To what extent do you
regard work as the most important aspect in your life?’ and
‘To what extent does your job allow for the achievement of
personal goals?’ Secondly, job involvement was measured
with items that were selected from Lodahl and Kejner’s
(1965) job involvement scale, such as: ‘How likely are you
to regard your work as only a small part of who you are?’
Thirdly, items were selected from three subscales from the
functions of identity scale of Serani et al. (2006). The items
were selected from the subscales: ‘structure’ – dened as ‘…
the structure of understanding of who one is’ (p. 1), ‘goals’ –
dened as ‘… meaning and direction through commitments,
values and goals’ (p. 1) and ‘future’ – dened as ‘… meaning
and direction through commitments, values and goals and
sense of future’ (p. 1). Fourthly, organisational identication
was measured with the scale of Mael and Ashforth (1992).
Examples of the adapted items are: ‘How often do you say
“we” rather than “they” when you talk about the organisation
that you work for?’ and ‘How interested are you in what
others think about the organisation that you work for?’
Finally, person-organisation t was measured with items
selected from the scale of Lauver and Kristof-Brown (2001).
Examples of the selected items are: ‘To what degree do your
values match or t the values of the organisation that you
work for?’ and ‘To what degree are you able to maintain your
values at the organisation that you work for?’
Reliability and validity of the instrument was determined
by submitting the 36-item questionnaire to rst-level and
second-level factor analyses to determine the factor structure
(Figure 1). Three columns can be identied in Figure 1. The
left-hand column shows the theoretical sub-constructs as
explained above with their respective reliabilities. The middle
column shows the results of the rst-level factor analysis
based on the six postulated factors and their respective
reliabilities. The right-hand column shows the results of the
second-level factor analysis and the respective reliabilities
of the two postulated factors. The factor analyses yielded a
28-item, one-dimensional work-based identity scale with a
Cronbach alpha coefcient of 0.95 (Roodt et al., 2009). The
second factor, labelled ‘future’, was omitted because it was
based on experimental items.
Research procedure
An electronic invitation to participate in the survey was
sent to all ICT sector company employees up to middle
management. The survey instrument was designed in such a
way that it allowed for ‘one-at-a-time completion’ of separate
components of the survey. The successful completion of the
questionnaire by the participants activated the last survey
instrument, the task performance scale, to be completed
by their immediate supervisor. Electronic reminders were
sent out on a weekly basis to all participants requesting
and reminding them to participate (or thanking those that
participated already). Participation was voluntary, responses
were treated as condential and no incentives were provided
to enhance participation.
In the 4-month period after the initial survey was conducted
it was established that 84 respondents left the service of the
company (identied by means of the company’s PERSAL
system) – this group was labelled the ‘leavers’. A random
sub-sample of 88 was drawn from the remainder of the initial
sample – and this group was labelled the ‘stayers’. The leavers
and stayers were compared in terms of the different variable
mean scores. The same procedure was then repeated after a
4-year period, where 405 leavers of the same initial sample
were compared with 405 randomly selected stayers from the
remaining sample. By only using their PERSAL numbers, all
the participants remained anonymous to the researchers.
Stascal analysis
All statistical analyses were conducted with the standard
SPSS (version 18.0) software program (Pallant, 2007) by the
Statistical Consultation Service (Statcon) of the University
of Johannesburg. These analyses were conducted in two
Original Research
doi:10.4102/sajhrm.v11i1.507
hp://www.sajhrm.co.za
Page 6 of 12
phases. In the rst phase, descriptive statistical analyses,
factor and iterative item reliability analyses and correlations
between all the variables were conducted. In the second
phase, inferential statistical analyses were conducted.
Results
A summary of the factor analysis procedure and results on
the TIS-6 is presented in Table 2. The second column in Table 2
refers to the item loadings (ranging between 0.73 and 0.81) on
the single extracted factor and the third column to the scale
internal consistency reliability (item GQ2 was reected). A
single factor was extracted (principal axis factoring with
varimax rotation) with a Cronbach alpha reliability coefcient
(α = 0.80) for the TIS-6. These ndings conrm the factorial
validity as well as the reliability of the TIS-6.
Similar factor analytic procedures were repeated for the
other variables used in the study (which are not repeated
here), but their reliabilities were reported individually
earlier under the sub-heading ‘Measuring instruments’.
The intercorrelations between the different variables are
presented in Table 3, in which it is evident that the different
variables are all signicantly related. More specically, the
correlations between turnover intentions and other variables
range between r(2428) = -0.11, p = 0.050 for helping behaviour
and r(2428) = 0.73, p = 0.001 for alienation. In the rst case,
turnover intention would decrease if helping behaviour
Theorecal sub-constructs First-level factor analysis Second-level factor analysis
Item per
dimension
Item – total
correlaon
Dimension
reliability
Item Item – total
correlaon
Factor reliability Item Item – total
correlaon
Construct
reliability
DQ1 0.72 Work
α = 0.82
DQ18 0.78 WI 1
α = 0.94
DQ18 0.74 Work-based
identy
α = 0.95
DQ2 0.69 DQ17 0.77 DQ17 0.73
DQ3 0.73 DQ7 0.76 DQ7 0.69
DQ4 0.54 DQ8 0.76 DQ8 0.72
DQ16 0.49 DQ10 0.76 DQ10 0.74
EQ1 0.42 DQ9 0.63 DQ9 0.53
DQ5 0.62 Job
α = 0.82
DQ5 0.74 DQ5 0.69
DQ6R0.40 DQ2 0.72 DQ2 0.66
DQ9 0.55 DQ19 0.71 DQ19 0.70
DQ13 0.41 DQ3 0.74 DQ3 0.69
DQ14 0.40 DQ12 0.75 DQ12 0.75
DQ15 0.50 EQ1 0.51 EQ1 0.48
DQ17 0.71 DQ6R0.45 DQ11 0.73
DQ19 0.69 DQ11 0.67 DQ1 0.69
DQ20 0.46 DQ1 0.67 DQ4 0.55
DQ7 0.72 Career or occupaon
α = 0.85
DQ4 0.54 DQ20 0.48
DQ8 0.76 DQ20 0.47 GQ19 0.55
DQ18 0.70 EQ4 0.44 GQ20 0.48
DQ10 0.58 Organisaonal identy
α = 0.87
GQ19 0.76 WI 2
α = 0.87
GQ17 0.56
DQ11 0.63 GQ20 0.69 GQ18 0.63
DQ12 0.61 GQ17 0.71 GQ15 0.55
GQ15 0.66 GQ18 0.72 GQ16 0.36
GQ16 0.44 GQ15 0.66 DQ15 0.55
GQ17 0.68 GQ16 0.47 DQ14 0.42
GQ18 0.72 DQ15 0.58 WI 3
α = 0.74
DQ16 0.57
GQ19 0.70 DQ14 0.53 DQ13 0.47
GQ20 0.62 DQ16 0.58 EQ7 0.57
GQ21R0.20 DQ13 0.41 EQ8 0.50
EQ2 0.94 Future
α = 0.72
EQ2 0.65 WI 4
α = 0.78
EQ2 0.46 Future
α = 0.74
EQ3 0.56 EQ3 0.65 EQ3 0.52
EQ4 0.30 EQ7 0.69 WI 5
α =0 .82
EQ5 0.70
EQ5 0.66 EQ8 0.69 EQ6 0.56
EQ6 0.51 EQ5 0.67 WI 6
α = 0.80
EQ7 0.52 Person–environment t
α = 0.60
EQ6 0.67
EQ8 0.57
EQ9R0.20
Source: Adapted from Bothma, F.C. (2011). The consequences of employees’ work-based identy. Unpublished DCom thesis, University of Johannesburg, Johannesburg.
Boxes shaded in grey denote deleted items or factors, whilst dierent colours indicate items of dierent theorecal dimensions.
WI 1 – WI 6 denote items included in rst-level and/or second-level factors by using dierent colours
FIGURE 1: Factor analyses results of the work-based identy scale.
TABLE 2: Factor analysis results of the turnover intenon scale.
Scale reliability Scale Items Item loadings
Turnover intenon α = 0.80 GQ1 0.733
GQ2R 0.772
GQ3 0.815
GQ4 0.733
GQ5 0.767
GQ6 0.779
R, item score is reected.
Original Research
doi:10.4102/sajhrm.v11i1.507
hp://www.sajhrm.co.za
Page 7 of 12
increases. In the second case, turnover intention would
increase if alienation increases.
The data proles of the 84 employees who resigned from
the ICT company over the 4-month period after the survey
was conducted were compared with the data proles of the
88 employees drawn randomly from the remaining sample
(n = 2345) who stayed with the company. Independent sample
t-tests were conducted to compare the different variable
scores of those employees who resigned versus those who
stayed. The following analyses (displayed in Table 4) provide
evidence that the turnover intention score can be used as a
proxy for actual labour turnover. The guidelines of Cohen
(1988, pp. 284–287) were followed to calculate the effect
sizes for independent-sample t-tests, expressed as partial
eta-squared. The variance strength of partial eta-squared
is indicated as ranging between 0.01 ≤ ηp
2 ≤ 0.05 (small*),
0.06 ≤ ηp
2 ≤ 0.13 (moderate**) and ηp
2 ≥ 0.14 (large***) effect.
There was a signicant difference in the turnover intention
scores of those employees who resigned (M = 5.14,
SD = 1.26) compared to those who stayed (M = 4.13, SD = 1.28):
t(170) = 5.20, p ≤ 0.001 (two-tailed). The difference in the
means (mean difference = 1.01, 95% CI: 0.63 to 1.39) has a
large effect (ηp
2 = 0.14). This nding supports the criterion-
predictive validity of the TIS-6 to predict actual turnover.
There was a signicant difference in the work-based identity
scores of those who resigned (M = 4.16, SD = 1.22) compared
to those who stayed (M = 4.96, SD = 0.92): t(153.8) = -4.84,
p ≤ 0.001 (two-tailed). The difference in the means (mean
difference = -0.79, 95% CI: -1.12 to -0.47) has a moderate effect
(ηp
2 = 0.12).
TABLE 3: Intercorrelaon matrix (Pearson correlaons) of the dierent variables.
Variables used
in the study
M SD WI AL H-OCB PA EE DP WE TI TP
WI 136.80 29.0 [0.95] – – – – – – – –
AL 20.76 7.0 -0.56** [0.81] – – – – – – –
H-OCB 49.36 8.7 0.37** -0.13** [0.86] – – – – – –
PA 29.15 7.6 0.35** -0.22** 0.29** [.71] – – – – –
EE 19.82 12.0 -0.39** 0.51** -0.11** -0.05*[0.89] – – – –
DP 8.49 6.3 -0.26** 0.33** -0.07** -0.05*0.66** [0.70] – – –
WE 24.13 7.0 0.71** -0.62** 0.35** 0.42** -0.40** -0.27** [0.91] – –
TI 25.21 8.1 -0.56** 0.73** -0.11** -0.20** 0.56** 0.37** -0.58** [0.80] –
TP 51.75 8.8 0.08** -0.13** 0.11** 0.08** -0.06** -0.07** 0.09** -0.13** [0.94]
WI, work-based identy; AL, alienaon; H-OCB, helping behaviour; PA, reduced personal accomplishment; EE, emoonal exhauson; DP, depersonalisaon; WE, work engagement; TI, turnover
intenon; TP, task performance; M, mean; SD, standard deviaon.
Coecient alphas are presented in square brackets along the diagonal.
*p ≤ 0.050; **p ≤ 0.001
n = 2429
TABLE 4: Independent-samples t-tests comparing group means (4-month period).
Variable Equal variances
(assumed or
not assumed)
Levene’s test for
equality of variances
t-Test for equality of means
FSig. t df Sig.
(two-tailed)
Mean
dierence
SE
dierence
95% Condence interval
of the dierence
Eta
squared
Lower Upper
Work-based identy Assumed 7.110 0.018 -4.84 170.000 0.001 -0.79 0.16 -1.12 -0.47 –
Not assumed – – -4.81 153.849 0.001 -0.79 0.17 -1.12 -0.47 0.12**
Personal alienaon Assumed 0.020 0.889 3.43 170.000 0.001 0.73 0.21 0.31 1.14 0.07**
Not assumed – – 3.43 169.319 0.001 0.73 0.21 0.31 1.14 –
Emoonal exhauson (BO1) Assumed 0.730 0.392 -2.18 170.000 0.030 -0.34 0.16 -0.65 -0.03 0.03**
Not assumed – – -2.18 169.819 0.030 -0.34 0.16 -0.65 -0.03 –
Depersonalisaon (BO2) Assumed 1.720 0.191 4.06 170.000 0.001 0.86 0.21 0.45 1.28 0.09**
Not assumed – – 4.07 169.999 0.001 0.86 0.21 0.45 1.28 –
Reduced personal
accomplishment (BO3)
Assumed 1.310 0.253 2.32 170.000 0.021 0.45 0.19 0.07 0.83 0.04**
Not assumed – – 2.32 168.025 0.021 0.45 0.19 0.07 0.83 –
Turnover intenon Assumed 0.640 0.423 5.20 170.000 0.001 1.01 0.19 0.63 1.39 –
Not assumed – – 5.21 169.827 0.001 1.01 0.19 0.63 1.39 0.14***
Vigour (WE1) Assumed 14.010 0.000 -4.24 170.000 0.001 -0.78 0.18 -1.14 -0.42 –
Not assumed – – -4.21 150.503 0.001 -0.78 0.18 -1.14 -0.42 0.09**
Dedicaon (WE2) Assumed 12.540 0.001 -4.32 170.000 0.001 -1.00 0.23 -1.46 -0.54 –
Not assumed – – -4.28 145.060 0.001 -1.00 0.23 -1.46 -0.54 0.10**
Absorpon (WE3) Assumed 14.540 0.000 -3.93 170.000 0.001 -0.73 0.19 -1.10 -0.36 –
Not assumed – – -3.90 145.300 0.001 -0.73 0.19 -1.10 -0.36 0.08**
Helping behaviour (H-OCB) Assumed 2.604 0.108 -1.565 170.000 0.119 -0.25 0.16 -0.56 0.07 Not sig.
Not assumed – – -1.558 162.188 0.121 -0.25 0.16 -0.56 0.07 –
Task performance Assumed 0.934 0.335 -1.566 154.000 0.120 -0.25 0.16 -0.57 0.07 Not sig.
Not assumed – – -1.554 144.947 0.122 -0.25 0.16 -0.57 0.07 –
F, F-value; Sig., signicance; t, t-value; df, degrees of freedom; SE, standard error; Eta squared, where the paral eta-squared of the variance strength is indicated as 0.01 (small*), 0.06 (moderate**)
and 0.14 (large***) eects.
The researchers followed the guidelines of Cohen, J.W. (1988). Stascal power analysis for the behavioral sciences. (2nd edn.). Hillsdale: Lawrence Erlbaum, p. 284.
Original Research
doi:10.4102/sajhrm.v11i1.507
hp://www.sajhrm.co.za
Page 8 of 12
There was a signicant difference in the personal alienation
scores of those who resigned (M = 4.85, SD = 1.40) and those
who stayed (M = 4.12, SD = 1.37): t(170) = 3.43, p ≤ 0.001
(two-tailed). The difference in the means (mean difference =
0.73, 95% CI: 0.31 to 1.14) has a moderate effect (ηp
2 = 0.07).
There was a signicant difference in the emotional exhaustion
(BO1) scores of those who resigned (M = 3.06, SD = 1.36)
compared to those who stayed (M = 2.19, SD = 1.42): t(170)
= -2.18, p = 0.030 (two-tailed). The difference in the means
(mean difference = 0.86, 95% CI: -0.65 to -0.03) has a small
effect (ηp
2 = 0.03).
There was a signicant difference in the depersonalisation
(BO2) scores of those who resigned (M = 1.93, SD = 1.31)
compared to those who stayed (M = 1.48, SD = 1.23):
t(170) = 4.06, p ≤ 0.001 (two-tailed). The difference in the
means (mean difference = 0.45, 95% CI: 0.46 to 1.28) has a
moderate effect (ηp
2 = 0.09).
There was a signicant difference in the reduced personal
accomplishment (BO3) scores of those who resigned
(M = 3.92, SD = 0.98) compared to those who stayed
(M = 4.26, SD = 1.06): t(170) = 2.32, p = 0.021 (two-tailed). The
difference in the means (mean difference = -0.34, 95% CI: 0.07
to 0.83) has a small effect (ηp
2 = 0.04).
There was a signicant difference in the vigour (WE1) scores
of those who resigned (M = 3.96, SD = 1.38) compared to
those who stayed (M = 4.73, SD = 1.00): t(151) = -4.24, p
≤ 0.001 (two-tailed). The difference in the means (mean
difference = -0.78, 95% CI: -1.14 to -0.42) has a moderate effect
(ηp
2 = 0.09).
There was a signicant difference in the dedication (WE2)
scores of those who resigned (M = 3.60, SD = 1.78) compared
to those who stayed (M = 4.60, SD = 1.21): t(145) = -4.32,
p ≤ 0.001 (two-tailed). The difference in the means (mean
difference = -1.00, 95% CI: -1.46 to -0.54) has a moderate effect
(ηp
2 = 0.10).
There was a signicant difference in the absorption (WE3)
scores of those who resigned (M = 3.86, SD = 1.43) compared
to those who stayed (M = 4.59, SD = 0.97): t(145.3) = -3.94,
p ≤ 0.001 (two-tailed). The difference in the means (mean
difference = -0.73, 95% CI: -1.10 to -0.36) has a moderate
effect (ηp
2 = 0.08). All the effect sizes in respect of individual
variables as reported above support the criterion-predictive
and the differential validity of the TIS-6 in the 4-month
period after the survey.
There was no signicant difference in the helping behaviour
scores of those who resigned (M = 5.27, SD = 1.13) compared
to those who stayed (M = 5.51, SD = 0.95): t(170) = -1.565,
p = 0.119 (two-tailed). The difference in the means (mean
difference = -0.25, 95% CI: -0.56 to 0.07) was insignicant.
There was no signicant difference in the task performance of
those who resigned (M = 5.51, SD = 1.07) compared to those
who stayed (M = 5.76, SD = 0.93): t(154) = -1.566, p = 0.120
(two-tailed). The difference in the means (mean difference =
-0.25, 95% CI: -0.57 to 0.07) was insignicant.
The data proles of the 405 employees who resigned from
the ICT company over the 4-year period after the survey
was conducted were compared with the data proles of 405
employees drawn randomly from the remaining sample (n
= 2024) who stayed with the company. Independent-sample
t-tests were conducted to compare the different variable
scores of those employees who resigned versus those
who stayed. The following analyses (displayed in Table 5)
provide evidence that turnover intention scores can be used
as a proxy for actual labour turnover.
There was a signicant difference in the turnover intention
scores of those employees who resigned (M = 4.41, SD = 1.42)
compared to those who stayed (M = 4.03, SD = 1.30): t(801)
= -4.10; p ≤ 0.001 (two-tailed). The difference in the means
(mean difference = -0.39, 95% CI: -0.58 to -0.20) has a small
effect (ηp
2 = 0.02).
There was a signicant difference in the work-based identity
scores of those who resigned (M = 4.71, SD = 1.13) and those
who stayed (M = 4.99, SD = 0.99): t(793) = 3.88; p ≤ 0.001
(two-tailed). The difference in the means (mean difference =
0.29, 95% CI: 0.14 to 0.43) has a small effect (ηp
2 = 0.02).
There was a signicant difference in the personal alienation
scores of those who resigned (M = 4.34, SD = 1.49) and
those who stayed (M = 3.89, SD = 1.34): t(798.9) = -4.55;
p ≤ 0.001 (two-tailed). The difference in the means (mean
difference = -0.45, 95% CI: -0.65 to -0.26) has a small effect
(ηp
2 = 0.02).
There was no signicant difference in the emotional
exhaustion (BO1) scores of those who resigned (M = 2.73,
SD = 1.57) and those who stayed (M = 2.56, SD = 1.49):
t(808) = -1.58; p = 0.113 (two-tailed). The difference in the
means (mean difference = -0.17, 95% CI: -0.38 to 0.04) was
insignicant.
There was a signicant difference in the depersonalisation
(BO2) scores of those who resigned (M = 1.74, SD = 1.29)
and those who stayed (M = 1.63, SD = 1.24): t(808) = -1.22;
p = 0.223 (two-tailed). The difference in the means (mean
difference = -0.11, 95% CI: -0.28 to -0.07) was insignicant.
There was no signicant difference in the reduced personal
accomplishment (BO3) scores of those who resigned (M =
1.87, SD = 1.11) and those who stayed (M = 1.81, SD = 1.00):
t(808) = -0.96; p = 0.339 (two-tailed). The difference in the
means (mean difference = -0.07, 95% CI: -0.22 to 0.07) was
insignicant.
There was a signicant difference in the vigour (WE1) scores
of those who resigned (M = 4.45, SD = 1.26) and those who
stayed (M = 4.71, SD = 1.16): t(802.5) = 3.07; p ≤ 0.001 (two-
tailed). The difference in the means (mean difference = 0.26,
95% CI: 0.09 to 0.43) has a small effect (ηp
2 = 0.01).
Original Research
doi:10.4102/sajhrm.v11i1.507
hp://www.sajhrm.co.za
Page 9 of 12
There was a signicant difference in the dedication (WE2)
scores of those who resigned (M = 4.15, SD = 1.60) and those
who stayed (M = 4.55, SD = 1.44): t(797.9) = 3.49; p ≤ 0.001
(two-tailed). The difference in the means (mean difference =
0.38, 95% CI: 0.17 to 0.59) has a small effect (ηp
2 = 0.01).
There was a signicant difference in the absorption (WE3)
scores of those who resigned (M = 3.53, SD = 1.31) and those
who stayed (M = 3.84, SD = 1.16): t(797) = 3.57; p ≤ 0.001
(two-tailed). The difference in the means (mean difference =
0.31, 95% CI: 0.14 to 0.48) has a small effect (ηp
2 = 0.02).
There was no signicant difference in the helping behaviour
scores of those who resigned (M = 5.12, SD = 0.75) and those
who stayed (M = 5.12, SD = 0.75): t(808) = 0.02; p = 0.983 (two-
tailed). The difference in the means (mean difference = 0.00,
95% CI: -0.10 to 0.11) was insignicant.
There was a signicant difference in the task performance
of those who resigned (M = 5.70, SD = 0.99) and those who
stayed (M = 5.87, SD = 0.93): t(808) = 2.69; p = 0.007 (two-
tailed). The difference in the means (mean difference = 0.18,
95% CI: 0.05 to 0.31) has a small effect (ηp
2 = 0.01). All the
effect sizes in respect of the individual variables reported
above are small, but still signicant (except where stated
as insignicant) and do therefore still support the criterion-
predictive and the differential validity of the TIS-6 in the
4-year period after the survey.
The independent sample t-tests that were conducted to
compare the different variable scores of those employees
who resigned and those who stayed differed signicantly in
nine of the 11 t-tests (over a 4-month period after the initial
survey) and seven of the 11 t-tests (after a 4-year period after
the initial survey). These results conrm the differential
validity of the turnover intention scale over these two
time periods, as well as its use as a proxy for actual labour
turnover.
Ethical consideraons
All ethical protocols of the institution were observed and
adhered to in conducting this research.
TABLE 5: Independent-samples t-tests comparing group means (4-year period).
Variable Equal variances
(assumed or
not assumed)
Levene’s test for
equality of variances
t-Test for equality of means
FSig. t df Sig.
(two-tailed)
Mean
dierence
SE
dierence
95% Condence interval
of the dierence
Eta
squared
Lower Upper
Work-based identy Assumed 7.64 0.006 3.88 808.00 0.001 0.29 0.07 0.14 0.43 –
Not assumed – – 3.88 793.03 0.001 0.29 0.07 0.14 0.43 0.02*
Personal alienaon Assumed 7.31 0.007 -4.55 808.00 0.001 -0.45 0.10 -0.65 -0.26 –
Not assumed – – -4.55 798.94 0.001 -0.45 0.10 -0.65 -0.26 0.02*
Emoonal exhauson (BO1) Assumed 1.66 0.198 -1.58 808.00 0.113 -0.17 0.11 -0.38 0.04 Not sig.
Not assumed – – -1.58 805.92 0.113 -0.17 0.11 -0.38 0.04 –
Depersonalisaon (BO2) Assumed 0.58 0.445 -1.22 808.00 0.223 -0.11 0.09 -0.28 0.07 Not sig.
Not assumed – – -1.22 806.62 0.223 -0.11 0.09 -0.28 0.07 –
Reduced personal
accomplishment (BO3)
Assumed 3.72 0.054 -0.96 808.00 0.339 -0.07 0.07 -0.22 0.07 Not sig.
Not assumed – – -0.96 800.04 0.339 -0.07 0.07 -0.22 0.07 –
Turnover intenon Assumed 3.86 0.050 -4.10 808.00 0.001 -0.39 0.10 -0.58 -0.20 –
Not assumed – – -4.10 801.90 0.001 -0.39 0.10 -0.58 -0.20 0.02*
Vigour (WE1) Assumed 8.52 0.004 3.07 808.00 0.002 0.26 0.09 0.09 0.43 –
Not assumed – – 3.07 802.47 0.002 0.26 0.09 0.09 0.43 0.01*
Dedicaon (WE2) Assumed 10.51 0.001 3.49 808.00 0.001 0.38 0.11 0.17 0.59 –
Not assumed – – 3.49 797.92 0.001 0.38 0.11 0.17 0.59 0.01*
Absorpon (WE3) Assumed 15.19 0.000 3.57 808.00 0.001 0.31 0.09 0.14 0.48 –
Not assumed – – 3.57 797.04 0.001 0.31 0.09 0.14 0.48 0.02*
Helping behaviour (H-OCB) Assumed 0.13 0.724 0.02 808.00 0.983 0.00 0.05 -0.10 0.11 Not sig.
Not assumed – – 0.02 807.98 0.983 0.00 0.05 -0.10 0.11 –
Task performance
Assumed 0.52 0.471 2.69 808.00 0.007 0.18 0.07 0.05 0.31 0.01*
Not assumed – – 2.69 804.25 0.007 0.18 0.07 0.05 0.31 –
F, F-value; Sig., signicance; t, t-value; df, degrees of freedom; SE, standard error; Eta squared, where the paral eta-squared of the variance strength is indicated as 0.01 (small*), 0.06 (moderate**)
and 0.14 (large***) eects.
The researchers followed the guidelines of Cohen, J.W. (1988). Stascal power analysis for the behavioral sciences. (2nd edn.). Hillsdale: Lawrence Erlbaum, p. 284.
Original Research
doi:10.4102/sajhrm.v11i1.507
hp://www.sajhrm.co.za
Page 10 of 12
Discussion
Despite the fact that turnover intention scales are frequently
used as criterion variables, little is known about their metric
properties. No previous studies were conducted to assess the
reliability and the validity of the shortened TIS-6, besides the
studies of Jacobs (2005) and Martin (2007) that used a longer
version of the scale, but did not investigate the relationship
with actual turnover. The research objectives of the present
study were therefore to evaluate the reliability, the factorial,
criterion-predictive and differential validity of the TIS-6 in
measuring turnover intentions or predicting actual turnover.
This study will add to the validity and reliability information
of the TIS, in general, and the TIS-6, specically, and will
contribute towards establishing its credibility for future use
in the scientic community.
Summary of key ndings
An exploratory factor analysis (EFA) using principal axis
factoring and varimax rotation established that the TIS-
6 is a one-dimensional construct, thereby conrming the
construct (more specically the factorial validity – cf.
Allen & Yen, 1979) of the scale. The item loadings (ranging
between 0.73 and 0.81) on the single extracted factor and the
overall reliability (α = 0.80) of the TIS-6 is on an acceptable
level, thereby conrming the reliability of the scale. It was
also established that scores of the TIS signicantly relate to
all other variables in this study, namely work engagement,
work-based identity, burnout, helping behaviour, work
alienation and task performance. These ndings conrm
previous research conducted by Bakker and Demerouti
(2006), Bakker et al. (2004), Demerouti et al. (2000), Schaufeli
and Bakker (2001) and Schaufeli and Bakker (2004), mainly
within the JD-R framework that linked turnover intention to
work engagement and burnout. The relationships between
turnover intention and work alienation and work-based
identity have not been reported on previously.
In order to establish the criterion-predictive validity of
the scale the TIS-6, mean score differences for those who
resigned were compared to a randomly selected group
from the sample of those who stayed with the organisation
(respectively for the 4-month and 4-year periods after the
survey). The obtained TIS-6 mean score differences were
signicant and the effect size was large (for the 4-month
period), which suggest that the TIS-6 could effectively
predict actual turnover. These ndings conrm previous
research conducted by Byrne (2005), Hendrix et al. (1998)
and Steensma et al. (2004) that turnover intention and actual
turnover are positively related. It also conrms the research
by Jaros et al. (1993), Muliawan et al. (2009) and Tett and
Meyer (1993) that turnover intentions can be used as a proxy
for actual turnover. The criterion-predictive validity of actual
turnover of the TIS-6 was hereby established.
In order to establish the differential validity of the scale,
independent-sample t-tests were conducted in respect of
the other variables to establish whether the mean scores of
those employees who resigned (n = 84) and those who stayed
(n = 88) differed signicantly. More specically, signicant
mean score differences (independent-sample t-tests) were
found in eight of the 10 remaining variables (work-based
identity, personal alienation, three work engagement
dimensions and three burnout dimensions), with effect sizes
ranging between moderate and small (in the 4-month period
after the survey).
The same procedure was repeated on a data set in a 4-year
period after the survey where scores of leavers (n = 405)
and stayers (n = 405) were compared in respect of the
same variables. More specically, signicant mean score
differences were found in six of the 10 remaining variables
(work-based identity, personal alienation, three work
engagement dimensions and task performance), but in this
case all the effect sizes were small. No previous research
could shed light on these longitudinal ndings over both a
short term and a medium term.
These results (both on a 4-month and a 4-year period
after the survey) conrm the differential validity of the
TIS-6. These results show that the TIS-6 is a reliable and valid
measure to assess the construct turnover intention and to
validly predict actual turnover behaviour, as was suggested
by Jaros et al. (1993) and Muliawan et al. (2009). These results
also conrm the differential validity of the TIS-6 as well as its
use as a proxy for actual labour turnover.
Praccal implicaons and recommendaons
Turnover intention in this study signicantly relates to a
number of other variables outside the JD-R framework,
such as work-based identity, personal alienation, the three
dimensions of work engagement (vigour, dedication
and absorption) and the three dimensions of burnout
(emotional exhaustion, depersonalisation, reduced personal
accomplishment), and not only to the suggested chain
of resources and demands as suggested by Bakker and
associates (Bakker & Demerouti, 2006; Bakker et al., 2004;
Demerouti et al., 2000; Schaufeli & Bakker, 2001, 2004). These
ndings, in combination with the different models proposed
by the Du Plooy and Roodt (2010) study, Jacobs (2005), Lee
and Mitchell (1994), as well as Petriglieri’s (2011) ndings,
necessitate a reappraisal and a reconceptualisation of models
that portray individuals’ cognitive processes before leaving
or exiting the organisation. Petriglieri’s model, especially
within an identity framework, shows potential in this regard.
The potential buffering role that psychological capital facets
(Sweetman & Luthans, 2010) may play in these cognitive
processes also warrants further research.
The ndings of this study further suggest that the TIS-6 can
be used as a reliable and valid measure to assess turnover
intention. The TIS-6 can therefore be used for business
applications and academic research to validly and reliably
assess turnover intention or to predict actual turnover.
Original Research
doi:10.4102/sajhrm.v11i1.507
hp://www.sajhrm.co.za
Page 11 of 12
Limitaons and suggesons for future research
A possible limitation of the study is that the TIS-6 was
only applied in a single organisational setting. However,
a strength of the study was that it yielded a fairly large
sample representing most of the different cultural groups
in the South African work context. Besides the suggestion of
reappraising and reconceptualising the cognitive processes
involved before individuals are making the decision to leave
the organisation, another suggestion for future research
may be to compare the scores of the TIS-6 across different
cultural groups in order to test for possible differential
item functioning and for measurement invariance. There is
a possibility that cultural groups may respond differently
to TIS-items and to antecedents leading to turnover
decisions. A third suggestion may be to investigate the role
of psychological capital facets as possible buffers between
contextual job demands and turnover intentions.
Conclusion
This study set out to determine whether the TIS-6 is a reliable
and valid instrument to assess turnover intentions and to
predict actual turnover. The results of the study conrm the
scale’s reliability, as well as its factorial, criterion-predictive
and differential validity. The research objectives of the study
are hereby achieved.
Acknowledgements
The research reported in this article is the product of a
collaborative research project between the University of
Johannesburg and the Vrije University, Amsterdam. Financial
support from the South Africa Netherlands Research
Programme on Alternatives in Development (SANPAD)
for conducting this research is hereby acknowledged.
Conclusions drawn or opinions expressed in this article are
those of the authors and do not necessarily reect the views
of SANPAD. A basic, draft version of this research article
was presented as a paper at the XXIX Pan Pacic Business
Conference, 24–27 May 2012 and a reduced version was
subsequently published in the Proceedings of the Pan Pacic
Conference.
Compeng interests
The authors declare that they have no nancial or personal
relationships which may have inappropriately inuenced
them in writing this article.
Authors’ contribuons
F.C.B. (University of Johannesburg) conducted the research
as a part of this doctoral study. G.R. (University of
Johannesburg) was the supervisor for this study. G.R. wrote
the largest portion of this article, whilst F.C.B conducted the
statistical analyses on which this article reported.
References
Agarwal, R., Ferrat, T.W., & De, P. (2007). An experimental invesgaon of turnover
intenons among new entrants in IT. The Data Base for Advances in Informaon
Systems, 38(1), 8–28. hp://dx.doi.org/10.1145/1216218.1216222
Akgün, A.E., & Lynn, G.S. (2002). Antecedents and consequences of team stability on
new product development performance. Journal of Engineering and Technology
Management, 19, 263–286. hp://dx.doi.org/10.1016/S0923-4748(02)00021-8
Allen, M.J., & Yen, W.M. (1979). Introducon to measurement theory. Monterey:
Brooks/Cole Publishing Company.
Allen, N.J., & Meyer, J.P. (1996). Aecve, connuance, and normave commitment
to the organisaon: An examinaon of construct validity. Journal of Vocaonal
Behavior, 49, 252–276. hp://dx.doi.org/10.1006/jvbe.1996.0043, PMid:8980084
Aryee, S., & Luk, V. (1996). Work and non-work inuences on the career sasfacon
of dual-earner couples. Journal of Vocaonal Behavior, 49, 38–52. hp://dx.doi.
org/10.1006/jvbe.1996.0032
Bakker, A.B., & Demerou, E. (2006). The job demands-resources model: State
of the art. Journal of Managerial Psychology, 22(3), 309–328. hp://dx.doi.
org/10.1108/02683940710733115
Bakker, A.B., Demerou, E., & Verbeke, W. (2004). Using the job demands-resources
model to predict burnout and performance. Human Resource Management,
43(1), 83–104. hp://dx.doi.org/10.1002/hrm.20004
Banai, M., & Reisel, W.D. (2007). The inuence of supporve leadership and job
characteriscs on work alienaon: A six-country invesgaon. Journal of World
Business, 42, 463–476. hp://dx.doi.org/10.1016/j.jwb.2007.06.007
Banai, M., Reisel, W.D., & Probst, T.M. (2004). A managerial and personal control
model: predicons of work alienaon and organisaonal commitment in Hungary.
Journal of Internaonal Management, 10, 375–392. hp://dx.doi.org/10.1016/j.
intman.2004.05.002
Becker, T.E. (1992). Foci and bases of commitment: Are they disncons worth
making? Academy of Management Journal, 35(1), 232–244. hp://dx.doi.
org/10.2307/256481
Bellou, V. (2008). Exploring civic virtue and turnover intenon during organisaonal
changes. Journal of Business Research, 61, 778–789. hp://dx.doi.org/10.1016/j.
jbusres.2007.09.001
Bester, F. (2012). A model of work identy in mulcultural work sengs. Unpublished
DPhil thesis, University of Johannesburg, Johannesburg.
Bluedorn, A.C. (1982). Managing turnover strategically. Business Horizons, March–
April, 6–12. hp://dx.doi.org/10.1016/0007-6813(82)90097-0
Boies, K., & Rothstein, M.G. (2002). Managers’ interest in internaonal assignments:
The role of work and career sasfacon. Internaonal Journal of Intercultural
Relaons, 26, 233–253. hp://dx.doi.org/10.1016/S0147-1767(02)00002-0
Bosho, A.B., Van Wyk, R., Hoole, C., & Owen, J.H. (2002). The predicon of intenon
to quit by means of biographic variables, work commitment, role strain and
psychological climate. Management Dynamics, 11(4), 14–28.
Bothma, F.C. (2011). The consequences of employees’ work-based identy.
Unpublished DCom thesis, University of Johannesburg, Johannesburg.
Bothma, F.C., & Roodt, G. (2012). Work-based identy and work engagement as
potenal antecedents of task performance and turnover intenon: Unravelling
a complex relaonship. SA Journal of Industrial Psychology, 38(1), Art. #893, 17
pages. hp://dx.doi.org/10.4102/sajip.v38i1.893
Brown, S.P. (1996). A meta-analysis and review of organisaonal research on
job involvement. Psychological Bullen, 120(2), 235–255. hp://dx.doi.
org/10.1037/0033-2909.120.2.235
Buche, M.W. (2003). IT professional work identy: Construct and outcomes.
Unpublished doctoral thesis, University of Kansas, Kansas City.
Buche, M.W. (2006). Gender and IT professional work identy. In E.M. Trauth (Ed.),
Encyclopedia of gender and informaon technology, (pp. 434–439). Hershey: Idea
Group Reference.
Buche, M.W. (2008). Inuence of gender on IT professional work identy: Outcomes
from a PLS study. Retrieved October 10, 2008, from hp://portal.acm.org/
citaon.cfm?id=1355272
Byrne, Z. (2005). Fairness reduces the negave eects of organisaonal polics on
turnover intenons, cizenship behaviour and job performance. Journal of
Business and Psychology, 20, 175–200. hp://dx.doi.org/10.1007/s10869-005-
8258-0
Carmeli, A., & Gefen, D. (2005). The relaonship between work commitment models
and employee withdrawal intenons. Journal of Managerial Psychology, 20(2),
63–85. hp://dx.doi.org/10.1108/02683940510579731
Chen, H.C., Chu, C.I., Wang, Y.H., & Lin, L.C. (2008). Turnover factors revisited:
A longitudinal study of Taiwan-based sta nurses. Internaonal Journal of
Nursing Studies, 45, 277–285. hp://dx.doi.org/10.1016/j.ijnurstu.2006.08.010,
PMid:17011564
Cohen, J.W. (1988). Stascal power analysis for the behavioral sciences. (2nd edn.).
Hillsdale: Lawrence Erlbaum.
Demerou, E., Bakker, A.B., Nachreiner, F., & Schaufeli, W.B. (2000). A model of
burnout and life sasfacon among nurses. Journal of Advanced Nursing, 32,
454–464. hp://dx.doi.org/10.1046/j.1365-2648.2000.01496.x, PMid:10964195
Du Plooy, J., & Roodt, G. (2010). Work engagement, burnout and related constructs as
predictors of turnover intenons. SA Journal of Industrial Psychology, 36(1), Art.
#910, 13 pages. hp://dx.doi.org/10.4102/sajip.v36i1.910
Fishbein, M., & Ajzen, I. (1975). Belief, atude, intenon and behavior: An introducon
to theory and research. Reading: Addison-Wesley Publishing Company.
Fox, S.R., & Fallon, B.J. (2003). Modeling the eect of work/life balance on job
sasfacon and turnover intenons. Symposium paper presented at the 5th
Australian Industrial and Organisaonal Psychology Conference, Melbourne,
Australia.
Gil-Monte, P.R. (2005). Validacao factorial de Maslach Burnout Inventory (MBI-HSS)
para prossionais espanhois [Factorial validity of the Maslach Burnout Inventory
(MBI-HSS) of Spanish professionals]. Revista Saúde Pública, 39(1), 1–8. hp://
dx.doi.org/10.1590/S0034-89102005000100001
Original Research
doi:10.4102/sajhrm.v11i1.507
hp://www.sajhrm.co.za
Page 12 of 12
Greyling, J., & Stanz, K. (2010). Turnover of nursing employees in a Gauteng hospital
group. SA Journal of Industrial Psychology, 36(1), Art. #850, 11 pages. hp://
dx.doi.org/10.4102/sajip.v36i1.850
Grieth, R.W., Horn, P.W., & Gaertner, S. (2000). A meta-analysis of antecedents
and correlaons of employee turnover: Update, moderator tests, and research
implicaons for the next millennium. Journal of Management, 26, 463–488.
hp://dx.doi.org/10.1177/014920630002600305
Guimaraes, T. (1997). Assessing employee turnover intenons before/aer TQM.
Internaonal Journal of Quality & Reliability Management, 14(1), 46–63. hp://
dx.doi.org/10.1108/02656719710156770
Hendrix, W., Robbins, T., Miller, J., & Summers, T. (1998). Eects of procedural and
distribuve jusce on factors predicve of turnover. Journal of Social Behavior
and Personality, 13(4), 611–632.
Horn, P.W., Grieth, R.W., & Sellaro, L. (1984). The validity of Mobley ’s (1977) turnover
model. Organizaonal Behavior and Human Performance, 34, 141–174. hp://
dx.doi.org/10.1016/0030-5073(84)90001-1
Jacobs, E.J. (2005). The development of a predicve model of turnover intenons
of professional nurses. Unpublished DCom thesis, University of Johannesburg,
Johannesburg.
Jaros, S.J., Jermier, J., Koehler, J., & Sincich, T. (1993). Eects of connuance, aecve
and moral commitment on the withdrawal process: An evaluaon of eight
structural equaon models. Academic Management Journal, 36(5), 951–995.
hp://dx.doi.org/10.2307/256642
Kirpal, S. (2004). Work idenes of nurses. Between caring and eciency
demands. Career Development Internaonal, 9(3), 274–304. hp://dx.doi.
org/10.1108/13620430410535850
Korman, A.K., Wig-Berman, U., & Lang, D. (1981). Career success and personal
failure: Alienaon in professionals and managers. Academy of Management
Journal, 24, 342–361. hp://dx.doi.org/10.2307/255846
Kotzé, K., & Roodt, G. (2005). Factors that aect the retenon of managerial and
specialist sta: An exploratory study of an employee commitment model. SA
Journal of Human Resource Management, 3, 48–55. hp://dx.doi.org/10.4102/
sajhrm.v3i2.65
Lacity, M.C., Lyer, V.V., & Rudramuniyaiah, P.S. (2008). Turnover intenons of Indian IS
professionals. Informaon Systems Froners on Outsourcing, 10, 225–241. hp://
dx.doi.org/10.1007/s10796-007-9062-3
Lambert, E.G., Hogan, N.L., & Barton, S.M. (2001). The impact of job sasfacon on
turnover intent: A test of a structural measurement model using a naonal sample
of workers. Social Science Journal, 38(2), 233–250. hp://dx.doi.org/10.1016/
S0362-3319(01)00110-0
Lauver, K.J., & Kristof-Brown, A. (2001). Disnguishing between employees’
percepons of person-job and person-organizaon t. Journal of Vocaonal
Behavior, 59, 454–470. hp://dx.doi.org/10.1006/jvbe.2001.1807
Lee, T.W., & Mitchell, T.R. (1994). An alternave approach: The unfolding model of
voluntary employee turnover. Academy of Management Review, 19(1), 51–89.
hp://dx.doi.org/10.5465/AMR.1994.9410122008
Lodahl, T.M., & Kejner, M. (1965). The denion and measurement of job involvement.
Journal of Applied Psychology, 49, 24–33. hp://dx.doi.org/10.1037/h0021692,
PMid:14279757
Lum, L., Kervin, J., Clark, K., Reid, F., & Sirola, W. (1998). Explaining nursing turnover
intent: Job sasfacon, pay sasfacon, or organizaonal commitment. Journal of
Organisaonal Behaviour, 19(3), 305–320. hp://dx.doi.org/10.1002/(SICI)1099-
1379(199805)19:3<305::AID-JOB843>3.0.CO;2-N
Mael, F., & Ashforth, B.E. (1992). Alumni and their alma mater: A paral test of the
reformulated model of organisaon. Journal of Organisaonal Behaviour, 13,
103–124. hp://dx.doi.org/10.1002/job.4030130202
Marn, A. (2007). Employee percepons of organisaonal commitment, job
sasfacon and turnover intenons in a post-merger instuon. Unpublished
MCom dissertaon, University of Johannesburg, Johannesburg.
Marn, A., & Roodt, G. (2008). Percepons of organisaonal commitment, job
sasfacon and turnover intenons in a post-merger terary instuon. SA
Journal of Industrial Psychology, 34(1), 23–31. hp://dx.doi.org/10.4102/sajip.
v34i1.415
Maslach, C., & Jackson, S.E. (1981). The measurement of experienced burnout.
Journal of Occupaonal Behaviour, 2(2), 99–113. hp://dx.doi.org/10.1002/
job.4030020205
Maslach, C., Jackson, S.E., & Leiter, M.P. (1986). Maslach burnout inventory: Manual.
(2nd edn.). Palo Alto: Consulng Psychologists Press.
Maslach, C., Jackson, S.E., & Leiter, M.P. (1996). Maslach burnout inventory: Manual.
(3rd edn.). Palo Alto: Consulng Psychologists Press.
Mendes, F., & Stander, M.W. (2011). Posive organisaon: The role of leader behaviour
in work engagement and retenon. SA Journal of Industrial Psychology, 37(1), Art.
#900, 13 pages, hp://dx.doi.org/10.4102/sajip.v37i1.900
Mobley, W.H. (1982). Some unanswered quesons in turnover and withdrawal
research. Academy of Management Review, 7, 111–116. hp://dx.doi.
org/10.5465/AMR.1982.4285493
Morrell, K., Loan-Clarke, J., Arnold, J., & Wilkinson, A. (2008). Mapping the decision to
quit: A renement and test of the unfolding model of voluntary turnover. Applied
Psychology: An Internaonal Review, 57(1), 128–150. hp://dx.doi.org /10.1111/
j.1464-0597.2007.00286.x
Mowday, R., Steers, R., & Porter, L. (1979). The measurement of organizaonal
commitment. Journal of Vocaonal Behaviour, 14, 224–247. hp://dx.doi.
org/10.1016/0001-8791(79)90072-1Muliawan, A.D., Green, P.F., & Robb, D.A.
(2009). The turnover intenons of informaon systems auditors. Internaonal
Journal of Accounng Informaon Systems, 10(3), 117–136. hp://dx.doi.
org/10.1016/j.accinf.2009.03.001
Pallant, J. (2007). SPSS survival manual: A step-by-step guide to data analysis using
SPSS version 15. (3rd edn.). New York: Open University Press.
Petriglieri, J.L. (2011). Under threat: Responses to and the consequences of threats
to individuals’ idenes. The Academy of Management Review, 36(4), 641–662.
hp://dx.doi.org/10.5465/amr.2009.0087
Podsako, P.M., MacKenzie, S.B., Paine, J.B., & Bachrach, D.G. (2000). Organisaonal
cizenship behaviours: A crical review of the theorecal and empirical literature
and suggesons for future research. Journal of Management, 26(3), 513–563.
hp://dx.doi.org/10.1177/014920630002600307
Roodt, G. (1997). Theorecal and empirical linkages between work-related
commitment foci. SA Journal of Industrial Psychology, 23(2), 6–13. hp://dx.doi.
org/10.4102/sajip.v23i2.624
Roodt, G. (2004). Turnover intenons. Unpublished document. Johannesburg:
University of Johannesburg.
Roodt, G., & Bothma, F.C. (1997). Die koste van vrywillige, beheerbare arbeidsomset
[The cost of voluntary, controlable labour turnover]. Journal of Industrial
Psychology, 23(1), 26–30. hp://dx.doi.org/10.4102/sajip.v23i1.619
Roodt, G., De Braine, R., Bothma, F.C. & Jansen, P.G. (2009). The work-based identy
quesonnaire (WI-28). Unpublished quesonnaire. Johannesburg: University of
Johannesburg.
Rotenberry, P.F., & Moberg, P.J. (2007). Assessing the impact of job involvement
on performance. Management Research News, 30(3), 203–215. hp://dx.doi.
org/10.1108/01409170710733278
Sager, J.K., Grieth, R.W., & Horn, P.W. (1998). A comparison of structural models
represenng turnover cognions. Journal of Vocaonal Behaviour, 53(2), 254–
273. hp://dx.doi.org/10.1006/jvbe.1997.1617
Schaufeli, W.B., & Bakker, A.B. (2001). Werk en welbevinden: Naar een posiewe
benadering in de arbeids- en gezondheidspsychologie [Work and well-being:
Towards a posive approach in the work and health psychology]. Gedrag en
Organisae, 14, 229–253.
Schaufeli, W.B., & Bakker, A.B. (2003). Utrecht work engagement scale, preliminary
manual. Utrecht: Occupaonal Health Psychology Unit, Utrecht University.
Schaufeli, W.B., & Bakker, A.B. (2004). Job demands, job resources and their
relaonship with burnout and engagement: A mul-sample study. Journal of
Organisaonal Behavior, 25, 293–315. hp://dx.doi.org/10.1002/job.248
Schaufeli, W.B., Salanova, M., Conźalez-Romá, V., & Bakker, A.B. (2002). The
measurement of engagement and burnout: A two sample conrmatory factor
analyc approach. Journal of Happiness Studies, 3, 71–92. hp://dx.doi.
org/10.1023/A:1015630930326
Senter, J.L., & Marn, J.E. (2007). Factors aecng the turnover of dierent groups
of part-me workers. Journal of Vocaonal Behavior, 71, 45–68. hp://dx.doi.
org/10.1016/j.jvb.2007.03.004
Serani, T.E., Maitland, S.B., & Adams, G.R. (2006). The funcons of identy scale:
Revisions, validaon and model tesng. Poster presented at the Biennial Meeng
of the Society for Research on Adolescence, San Francisco, California.
Smith, C.A., Organ, D.W., & Near, J.P. (1983). Organizaonal cizenship behavior: Its
nature and antecedents. Journal of Applied Psychology, 68, 653–663. hp://
dx.doi.org/10.1037/0021-9010.68.4.653
Steensma, H., Van Breukelen, W., & Sturm, M. (2004). Studying employee turnover by
spling up the usual comparison group. Journal of Individual Employment Rights,
11, 211–227. hp://dx.doi.org/10.2190/46U9-T06L-8M32-PEFM
Steers, R. (1977). Antecedents and outcomes of organizaonal commitment.
Administrave Science Quarterly, 22 (1), 46–56. hp://dx.doi.
org/10.2307/2391745
Sulu, S., Ceylan, A., & Kaynak, R. (2010). Work alienaon as a mediator of the
relaonship between organisaonal injusce and organisaonal commitment:
Implicaons for healthcare professionals. Internaonal Journal of Business and
Management, 5(8), 27–38.
Sweetman, D., & Luthans, F. (2010). The power of posive psychology: Psychological
capital and work engagement. In A.B. Bakker & M.P. Leiter (Eds.), Work
engagement: A handbook of essenal theory and research (pp. 54–68). New York:
Psychology Press.
Te, R.P., & Meyer, J.P. (1993). Job sasfacon, organisaonal commitment, turnover
intenon and turnover: Path analyses based on meta-analyc ndings. Personnel
Psychology, 46(2), 259–293. hp://dx.doi.org/10.1111/j.1744-6570.1993.
tb00874.x
Van Dyne, L., & LePine, J.A. (1998). Helping and voice extra-role behaviours: Evidence
of construct and predicve validity. Academy of Management Journal, 41(1),
108–120. hp://dx.doi.org/10.2307/256902
Walsh, K., & Gordon, J.R. (2007). Creang an individual work identy. Human
Resource Management Review, 18(1), 46–61. hp://dx.doi.org/10.1016/j.
hrmr.2007.09.001
Was, A.S. (2003). Organisaonal commitment, turnover intenons and the inuence
of cultural values. Journal of Occupaonal and Organisaonal Psychology, 76,
303–321. hp://dx.doi.org/10.1348/096317903769647193
Wayne, J.H., Randel, A.E., & Stevens, J. (2006). The role of identy and work-family
support in work-family enrichment and its work-related consequences. Journal of
Vocaonal Behavior, 69, 445–461. hp://dx.doi.org/10.1016/j.jvb.2006.07.002
Wheeler, A.R., Gallagher, V.C., Brouer, R.L., & Sablynski, C.J. (2007). When person-
organisaon (mis)t and (dis)sasfacon lead to turnover. The moderang role
of perceived job mobility. Journal of Management Psychology, 22(2), 203–219.
hp://dx.doi.org/10.1108/02683940710726447
Williams, L.J., & Anderson, S.E. (1991). Job sasfacon and organisaonal commitment
as predictors of organisaonal cizenship and in-role behaviours. Journal of
Management, 17(3), 601–617. hp://dx.doi.org/10.1177/014920639101700305
Zeane, R.M. (1994). Understanding employee turnover: The need for a conngency
approach. Internaonal Journal of Manpower, 15(9/10), 22−37. hp://dx.doi.
org/10.1108/01437729410074182