Instructional manipulations in writtten emotional disclosure interventions: An experimental test of content effects on cortisol, mood, and linguistic dimensions

Date of Award


Degree Type


Degree Name

Doctor of Philosophy (PhD)




Joshua Smyth


Expressive writing, Theoretical mechanisms, Written emotional disclosure, Intervention, Content effects, Writing instructions

Subject Categories

Clinical Psychology | Psychology | Social and Behavioral Sciences


Expressive writing (EW) interventions encourage individuals to express their thoughts and emotions through writing about a stressful/traumatic event and are associated with various physical and psychological benefits. Recently, EW interventions have tested a range of altered intervention content, such as focusing on perceived benefits of past stress/trauma, or trying to facilitate emotional processing. No studies to date have experimentally tested altered intervention content to establish if such instructions systematically and reliably produce the response intended. The purpose of this study was thus to manipulate EW intervention instructions to investigate if content alterations change the process of writing and participant responses. Community and student participants ( n =204) were randomized to 1 of 6 groups representing different EW instructions that are commonly used in experimental disclosure studies. These include instructions thought to elicit greater processes of cognitive-processing, exposure, self-regulation, and benefit-finding. Each modified group was compared with a standard EW and an emotionally-neutral control group. Participants wrote for 20 minutes, once a week for 3 weeks. Short-term outcomes included writing content and pre-to-post writing changes in salivary cortisol and mood. The experimental groups differed from the control group on immediate outcomes in expected ways; however, altered instructions produced both intended and unintended responses compared to a standard EW group. Some intervention manipulations produced responses that differed from standard EW instructions as expected: cognitive-processing based instructions lead to more cognitive-processing ( p <.0001) and instructions that promoted self-regulation and benefit-finding lead to more positive affect ( p 's<.003). In contrast, some intervention manipulations produced unanticipated changes on processes thought to be associated with other writing instructions. For example, exposure-based instructions lead to more extensive emotional habituation over time ( p =.04) but also promoted the unanticipated response of greater cognitive-processing ( p =.04). These results suggest that altering EW instructions to presumptively target specific processes "works" since most intended processes were shifted as intended. Importantly, however, altered instructions also elicited processes inconsistent with theoretical/clinical predictions which suggest that altering EW instructions to tailor intervention content seems to activate multiple processes. Thus, modified intervention instructions may elicit a much wider range of mechanisms than those thought to be unique to one instructional set.