A critique of expressive writing experiment in the cancer population: Focus on construct validity

摘要

Expressive writing was proposed to facilitate cognitive processing and disinhibition of emotions in people confronted with traumatic events. Contrary to the findings found in the general population, recent meta-analyses failed to support its effect in cancer. Given the discrepancy in evidence concerning the effect of expressive writing experiment for the cancer population, this paper endeavored to reexamine the cancer literature on expressive writing focusing on assessment of threats to construct validity. A systematic search was performed in the PubMed and PsycINFO databases up to September 2018. Studies were limited to randomized controlled trials in the cancer population published in English language with written disclosure as the only source of treatment effect. Construct validity was assessed with respect to all four study features - persons, settings, treatments and outcomes. Threats to construct validity arising in the primary research include underrepresentation of the newly diagnosed cancer patients, debatable study design concerning manipulation at the home setting, the use of factual writing as attention control and study operation of treatment reflecting more than one construct, as well as missing outcomes central to the construct of interest. This review highlights the need for both primary research and research synthesis to attend to construct validity when design or evaluate experiments.