Automated computer-based feedback in expressive writing

https://doi.org/10.1016/j.chb.2012.01.003Get rights and content

Abstract

Previous studies indicate that changing pronoun use can moderate the health benefits of expressive writing. Participants who change their pronoun use from essay to essay benefit more from the exercise. The current experiment attempted to improve the expressive writing paradigm by altering subjects’ focus (specifically self- and other-related focus) through text-based instructions. The automated language-based system was simple to implement, and it successfully changed participants’ pronoun use, but the intervention had no mood-related or subjective benefits. While the automated intervention system has promise in guiding people’s natural writing, future research must address the broader question of the inherent links between language use and substantive psychological change.

Highlights

► Writing about emotional upheavals can improve physical and psychological health. ► Owen et al. tried to improve this by changing subjects’ use of emotional words with 108 video messages. ► Pronoun words, which this study altered, have a bigger impact on writing outcomes. ► The manipulation was purely text-based, making it sensitive to change over time. ► The manipulation worked, but overall outcomes remained the same. Future directions are discussed.

Introduction

Almost 200 published studies on expressive writing suggest that writing about emotional upheavals can improve physical and psychological health. In a typical expressive writing study, participants write about a traumatic event 3–5 times for 10–30 min each time. Control participants are generally asked to write about superficial or non-emotional topics for comparable times. Meta-analyses indicate that the expressive writing method has modest health effects using objective behavioral and physiological measures over several weeks, especially for non-clinical populations (Frattaroli, 2006, Frisina et al., 2004, Smyth, 1998).

A recurring theme in the expressive writing literature concerns the underlying mechanism. That is, why does writing work and is it possible to identify the features of writing that account for the health improvements? The answers have proven to be complicated. The emerging consensus is that expressive writing is beneficial for a number of overlapping reasons. The labeling of an event along with its constituent emotions, changes in the ways the event is cognitively represented, possible habituation of the emotional reactions to the event, better sleep after writing, and even positive changes in subsequent social behaviors have all been implicated as mechanisms (Pennebaker & Chung, 2011).

Separate lines of inquiry have focused on the writing samples themselves. Are some ways of writing more likely to yield health improvements than others? Using a variety of text analysis methods, some promising findings suggest that the ways people write may be a key to health improvements. For example, one study relied on the Linguistic Inquiry and Word Count (LIWC, Pennebaker, Booth, & Francis, 2007) software in the analysis of several writing studies. Those people whose health improved the most evidenced greater use of positive emotions and an increasing use of cognitive words over the course of writing (Pennebaker, Mayne, & Francis, 1997; for conceptual replications see Klein and Boals (2001) and Boals and Perez (2009).

A later study relied on latent semantic analysis (LSA, Landauer & Foltz, 1998) and found that health improvements across three writing studies were linked to the ways participants used personal pronouns (Campbell & Pennebaker, 2003). Specifically, people who fluctuated in their use of first person singular pronouns (e.g., I, me, my) to other personal pronouns (e.g., you, he, she, they) from one writing essay to the next showed much greater health improvements than those who relied on a single pronoun orientation. The authors interpreted the findings as reflecting the ability to change perspectives in writing. That is, use of I-words typically involved writing about how the person felt or viewed themselves whereas essays with high rates of other-pronoun words suggested a more external focus.

An important question concerns the potential applicability of these language studies. If a method could be devised that influenced the ways people wrote, it may be possible to boost the effectiveness of the expressive writing intervention. In a creative test of this idea, Owen, Hanson, Preddy, and Bantum (2011) developed an intervention wherein participants’ emotion language was monitored using the LIWC program. In the No Feedback condition, participants wrote using the standard expressive writing paradigm. Those in the Simple Feedback condition were simply informed about their emotion language use after their first and second writing sessions. Those in the Directive Feedback condition learned about their emotion language and, through both text and video presentations, were encouraged to express their emotions at optimal rates based on their previous writings.

Owen et al. (2011) study was effective in getting participants in the Directive Feedback condition to increase their use of emotion words. Unfortunately, however, there were no additional improvements in health relative to the control conditions. The Owen study points to the potential of automated interventions in expressive writing, although many other studies highlight the effectiveness of computer-based treatment in general (Barak et al., 2008, Hanley and Reynolds, 2009, Kaltenthaler et al., 2004). In retrospect, there are at least two possible reasons why the Owen study did not affect health. The instructions for those in the Directive Feedback condition asked participants to increase in their use of both positive and negative emotions. Recall that the Pennebaker et al. (1997) study found that the relation between emotional expression and health was linear for positive emotion but curvilinear (an inverted U-shape) for negative emotion. Perhaps the intervention resulted in levels of emotion that were not healthy.

A second possibility is that the Owen study focused on emotions and not social or cognitive changes associated with expressive writing. The later Campbell and Pennebaker (2003) project found that perspective switching, as measured by pronoun use, accounted for far more variance in health improvements that were reported in the emotion findings from the Pennebaker, Booth, and Francis project (2007).

The present study sought to extend the language feedback idea by focusing on perspective switching rather than emotional expression. Drawing on the Campbell and Pennebaker project, participants were encouraged to switch their perspectives and, indirectly, their use of pronouns from writing session to writing session. It was predicted that those who were asked to alter their perspectives would evidence greater self-reported benefits from writing than those who were given to perspective writing instructions.

Section snippets

Participants

Approximately 757 people accessed the experimental site which sought volunteers to participate in an online experiment that was attempting to “discover which ways of writing may be the best.” Prospective participants learned about the project through referring links on Pennebaker’s homepage (www.psy.utexas.edu/Pennebaker) and his experimental survey page (www.utpsyc.org). Of the 757 people who visited the site, 236 individuals wrote for all three writing sessions of the experiment. Of these,

Results

After briefly reviewing the content of the writing samples (3.1), participant compliance is assessed (3.2). Finally, the main outcome measures, which included both the subjective benefit of the exercise and measures of affect, are explored (3.3).

Discussion

The study had two goals. The first was to determine if it was possible to change the ways people write using automated feedback. That is, could we change the ways people focus on themselves versus others by relying on simple pronoun counts. The second, more important goal, was to determine if the manipulations in perspective switching would enhance the value of expressive writing. That is, does induced perspective switching bring about changes in the ways people think about and emotionally

Acknowledgements

Preparation of this paper was aided by funding from the Army Research Institute (W91WAW-07-C-0029), NSF (NSF-NSCC-090482), and START (DHS Z934002).

References (17)

  • J.E. Owen et al.

    Linguistically-tailored video feedback increases total and positive emotional expression in a structured writing task

    Computers in Human Behavior

    (2011)
  • J.L. Baddeley et al.

    An expressive writing intervention for military couples reuniting after a deployment

    Journal of Traumatic Stress

    (2011)
  • A. Barak et al.

    A comprehensive review and a meta-analysis of the effectiveness of internet-based psychotherapeutic interventions

    Journal of Technology in Human Services

    (2008)
  • A. Boals et al.

    Language use predicts phenomenological properties of Holocaust memories and health

    Applied Cognitive Psychology

    (2009)
  • R.S. Campbell et al.

    The secret life of pronouns: Flexibility in writing style and physical health

    Psychological Science

    (2003)
  • J. Frattaroli

    Experimental disclosure and its moderators: A meta-analysis

    Psychological Bulletin

    (2006)
  • P.G. Frisina et al.

    A meta-analysis of the effects of written emotional disclosure on the health outcomes of clinical populations

    Journal of Nervous and Mental Disease

    (2004)
  • T. Hanley et al.

    Counseling psychology and the internet: A review of the quantitative research into online outcomes and alliances within text-based therapy

    Counseling Psychology Review

    (2009)
There are more references available in the full text version of this article.

Cited by (0)

View full text