Low, Stanton, Bower & Gyllenhammer (2010)
All women at Stave IV breast cancer, allows to moderate for stage level differences in psychological processing
Random assignment to experimental group
Control group also writes, just not expressively and about emotions
Based on social constraint theory
Used standardized scales for measuring depression and somatic symptoms
Took into account time since diagnosis
Short and infrequent writing (4 x 20 minute intervals over a period of 3 weeks)
Moderating variables (perceived emotional support; time since diagnosis) do not seem related to the writing
Perceived emotional support treated as an ancillary variable in the design, and yet was a defining feature of the hypothesis (weak internal validity)
Women received $80 for participating, which would have influenced the decision to participate
The participants had already been proactive in their psychological recovery
Homogenous experimental group.
The researchers hypothesized that especially among women who receive low emotional support and who were more recently diagnosed, the writing would reveal distinct improvements. However, the researchers did not control for these variables at the onset. Instead, they retroactively examined whether there was an effect based on perceived emotional support and time since diagnosis. This is a flaw in the research design. The researchers discuss social constraint theory, but the research design was not established to test specifically for social support as a variable. When the participants were asked about social support, they were asked to self-report rather than to objectively measure time spent with loved ones and other concrete measures for social support.
Other weaknesses in this study include the homogenous experimental group, which consisted mainly of well-educated whites in a specific geographic area. This means the results of the study cannot be extended necessarily to other populations and the results are not necessarily valid beyond this population. Writing for such short an infrequent times might have also led to the weak results and null hypothesis. Finally, the provision of a financial reward might also have skewed the motivations for participating.
Source: Henry, et al. (2010)
Narrowed down the population to focus on specific demographic variables like living in rural areas and the subsequent social isolation and lack of access to support resources.
Measures of depressive symptomology, mood states, and physical symptoms assessed at four different intervals, including immediately before the writing intervention as well as three and nine months after it.
Design was feasible, ie.…