executives in a company who are likely to experience stress, it is in fact the case that
those lower down an organization suffer the most heart disease (e.g. Marmot et al.,
1991). The JDC model suggests that this is because they have low levels of control
and may therefore be unable to moderate their job demands.
The JDC model also describes two further job types. Active jobs are those with
high levels of demand and high control. These are hypothesized to be less stressful
than high strain jobs because they offer the individual opportunities to develop
protective behaviours, encourage active learning and motivate engagement in new
behaviours (Karasek, 1979). Thus, for example, a high level manager or senior exec -
utive might have high job demands, but may also be in a position to delegate some
of the work. By contrast, passive jobs, that is, those with low demands and low control
(for example, an assistant in a seaside cafe out of season) are essentially boring and
are suggested to result in learned helplessness and reduced activity. The original
conceptualization of the model (Karasek, 1979) emphasized the importance of the
interaction between the two variables in predicting strain. In a later publication, Karasek
(1989) argued that this interaction is not of central importance. The model is,
therefore, frequently considered supported if there is an additive effect of demands and
control. However, van der Doef and Maes (1999) make a clear distinction between
two alternative hypotheses, which are tested in different studies:
1 The strain hypothesis(the additive hypothesis), which states that greater psycho -
logical strain and physical illness will be suffered by those in the high strain quadrant
of the model (see Figure 4.2).
2 The buffer hypothesis, which states that there is an interaction between demand
and control (i.e. control moderates, or buffers, the impact of demand). This can
be tested by entering a multiplicative interaction term into a multiple regression
equation predicting strain outcomes after the main effects terms (of demand and
control).
The model has subsequently been expanded by the addition of social support to
form the job demand–control–support model (Johnson and Hall, 1988) and both forms
of the model have stimulated considerable research looking at a wide range of physical
and psychological outcomes (for reviews see van der Doef and Maes, 1998, 1999; de
Lange et al., 2003; Hausser et al., 2010). This has been facilitated by the availability of
measures of the core variables. The ‘job content questionnaire’ (Karasek, 1985) aims
to measure demand and control as objectively as possible using a self-report question -
naire to tap the existence of particular stressors using items such as ‘my job requires
working very fast’ or ‘my job allows me to make a lot of decisions on my own’. People
are asked to respond to these items by ticking one of four options ranging from
‘strongly agree’ to ‘strongly disagree’.
The JDC model has been successful in predicting cardiovascular disease (Kivimaki
et al., 2012) and associated risk factors (e.g. ambulatory blood pressure levels) as well
as psychological well-being (van der Doef and Maes, 1998, 1999; O’Connor et al.,
2000a) although there is less support for the buffer hypothesis. There is also evidence
that it predicts health behaviours such as physical activity (e.g. Kouvonen et al., 2007;
Fransson et al., 2012). However, evidence is often complex and inconsistent (Jones,
Kinman and Payne, 2006). Indeed, a study of daily health behaviours suggested that,
STRESS AND HEALTH IN CONTEXT 75