However, a meta analysis of writing studies by Smyth (1998) indicated that men may
benefit more from writing than women and those who do not naturally talk openly about
their emotion may benefit more than those who do.Personality and mood: Pennebaker (1997) also concludes that anxiety, inhibition
or constraint do not influence the effectiveness of writing. However, Christensen et al.
(1996) concluded that individuals high on hostility scores benefited more from writing
than those low on hostility.Use of language: To explain the effectiveness of writing Pennebaker and colleagues
(2001) developed a computer programme to analyse the content of what people were
writing during the task. They coded the transcripts in terms of the types of words used:
negative emotion words (sad, angry), positive emotion words (happy, laugh), causal
words (because, reason) and insight words (understand, realize). The results from this
analysis showed that greater improvement in health was associated with a high number
of positive emotion words and a moderate number of negative emotion words. More
interestingly, they also found that those who showed a shift towards more causal and
insight words also showed greater improvement (Pennebaker et al. 1997). They con-
cluded from this that this shift in language use reflected a shift from poorly organized
descriptions towards a coherent story and that a coherent story was associated with
better health status. However, in contrast to this Graybeal et al. (2002) directly assessed
story-making and found no relationship with health outcomes.Stress
Stress can cause illness through physiological changes such as raised heart rate, blood
pressure, heart beat irregularities and an increase in fatty deposits (see above). It can also
result in changes in immune function. Research on rats showed that stressors such as tail
pinching, a loud noise and electric shocks could produce immunosuppression (Moynihan
and Ader 1996). Research in humans shows a similar picture. One area of research
which has received much attention relates to the impact of caregiver stress. In an early
study, Kiecolt-Glaser et al. (1995) explored differences in wound healing between people
who were caring for a person with Alzheimers and a control group. Using a punch biopsy
which involves removing a small area of skin and tissue they explored the relationship
between caregiver stress and the wound healing process. The results showed that wound
healing was slower in the caregivers than the control group. The wound healing para-
digm has also been used to show links between stress and slower healing in students
during an exam period (Marucha et al. 1998) and slower healing using high resolution
ultrasound scanning which is more accurate than the more traditional measurement
strategies involving photography (Ebrecht et al. in press). Herbert and Cohen (1993)
carried out a meta analysis of 38 studies which had explored the stress–immune system
link. They concluded that stress consistently resulted in changes in immune function in
terms of proliferative response to mitogens, NK cell activity, and was related to greater
numbers of circulating white blood cells, immunoglobulin levels, and antibody titers to
herpes viruses. They also concluded that greater changes in immune response were264 HEALTH PSYCHOLOGY