See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/247434805 Let’s Be Realistic: When Grief Counseling Is Effective and When It’s Not Article in Professional Psychology Research and Practice · June 2008 DOI: 10.1037/0735-7028.39.3.377 CITATIONS READS 30 1,599 2 authors, including: George A Bonanno Teachers College 256 PUBLICATIONS 24,843 CITATIONS SEE PROFILE Some of the authors of this publication are also working on these related projects: Drive to Thrive Theory: Sustaining Everyday Life Fabrics and Structure View project Article View project All content following this page was uploaded by George A Bonanno on 31 March 2016. The user has requested enhancement of the downloaded file. The Forum methodological and statistical experts” (p. boast about and pales in comparison with 349) commissioned by Gary R. Vanden- the average effect size for psychotherapy. Let’s Be Realistic: When Grief Bos but do not inform readers (a) what, if Indeed, given that Allumbaugh and Hoyt Counseling Is Effective and anything, these reviewers were told about (1999) acknowledged that an effect size of When It’s Not the reason for peer review, (b) what spe- .43 is “small relative to the .80 effect size cific flaws reviewers identified regarding of psychotherapy for a variety of problems the TIDE statistic, or (c) whether the re- found by previous meta-analyses” (p. George A. Bonanno viewers concluded that Fortner’s asser- 377), Larson’s and Hoyt’s assertion that Teachers College, Columbia University tions concerning the potential iatrogenic “there is not even any strong evidence that Scott O. Lilienfeld effects of grief counseling for normal be- grief counseling, as typically practiced, is Emory University reavement are themselves erroneous. At less efficacious than other forms of coun- the very least, it seems imprudent to dis- seling and psychotherapy” (p. 354) is be- Larson and Hoyt (2007) argued that be- regard the serious concerns raised by Fort- wildering. reavement researchers have erroneously and ner’s findings until they are subjected to Taken together, the extant meta- unscientifically advocated the pessimistic more rigorous and extensive peer review. analyses point toward a crucial conclusion conclusion that grief counseling is ineffec- Moreover, Larson and Hoyt seem surpris- dismissed by Larson and Hoyt (2007): tive and perhaps even harmful. They pro- ingly cavalier about the potential harm Most bereaved people do not need and will claimed that the news is actually good: Grief posed by certain forms of grief counseling. not benefit from clinical intervention. counseling is not harmful but is as effective For example, they neglected to note that Moreover, when treatment is focused ap- as other forms of psychotherapy. Therefore, two studies of grief counseling that in- propriately on bereaved people who do they concluded, most or all bereaved people cluded social activities interventions seek or need professional help and when should be considered candidates for treat- yielded an average negative (harmful) ef- interventions are appropriately tailored ment. This kind of unwarranted optimism is fect size (across four measures) of Ϫ.35 treatment effects will be comparable with as dangerous, if not more so, than an overly (Kato & Mann, 1999). those of other forms of psychotherapy. cautious pessimism. In addition to Fortner (1999), there are Several sources of evidence support this Warning of the “perils of unwarranted three published meta-analyses on adult conclusion. First, longitudinal and prospec- certainty” (p. 347), Larson and Hoyt (Allumbaugh & Hoyt, 1999; Kato & tive studies of grief course have established (2007) proceeded to reassure readers that Mann, 1999) and child (Currier, Holland, that the vast majority of bereaved people do “there is no empirical or statistical foun- & Neimeyer, 2007) samples. The average not evidence long-term difficulties (Bon- dation” (p. 354) for pessimism regarding weighted effect sizes for grief counseling anno, 2004). Indeed, approximately half of the effects of grief counseling and that across these studies were .43, .11, and .14, bereaved people show genuine resilience; “there is no evidence that bereaved clients respectively. Larson and Hoyt (2007) that is, they evidence little or no grief or are harmed by counseling” (p. 354). They elected to favor the former result and dis- depression and score just as highly as even suggested that journals “publish re- count the latter two. There is not sufficient matched nonbereaved samples on measures tractions in print from proponents of the space here to quibble about methodologi- of positive aspects of adjustment, even in the unfounded conclusions” (p. 354). Yet, it is cal details, as each of these studies differs early months after a loss (Bonanno, Mos- Larson and Hoyt who appear guilty of in its inclusion criteria and has its own kowitz, Papa, & Folkman, 2005; Bonanno et unwarranted certainty. methodological limitations. Indeed, as al., 2002; Bonanno, Wortman, & Nesse, The bulk of Larson and Hoyt’s (2007) Larson and Hoyt noted, there is minimal 2004). We see no reason to assume that thesis rests on a single point: namely, that overlap in the studies included in these resilient individuals would desire or benefit researchers have relied on a summary meta-analyses. The methodological credo from grief counseling. (Neimeyer, 2000) of an unpublished meta- of the heterogeneity of irrelevancies Second, although some bereaved people analysis that yielded an overall effect size (Shadish, Cook, & Campbell, 2002) re- experience a more enduring struggle over of .13 for grief counseling (Fortner, 1999). minds us that the most robust conclusions time, only a small subset, usually about We grant Larson and Hoyt’s point and in science typically derive from the con- 10%–15%, exhibit extreme or complicated acknowledge that the treatment-induced vergence of differing studies with largely grief reactions (Bonanno & Kaltman, 1999, deterioration effect (TIDE) statistic used nonoverlapping methodological flaws. 2001). The logic of our profession suggests by Fortner (1999) in his dissertation had Given that three of four meta-analyses on that these are the bereaved people most ap- not been subjected to formal peer review the effects of grief counseling have yielded propriate for and most likely to benefit from prior to its citation by Neimeyer (2000) or effect sizes of less than .15, the overall intervention. The available evidence, includ- by later authors who cited Neimeyer. Yet, picture for the efficacy of grief counseling ing Allumbaugh and Hoyt’s (1999) meta- we are troubled by Larson and Hoyt’s can hardly be described as encouraging. analysis, supports this point. Bereavement claim that their article is the first to subject Even if we disregard all other meta- outcome studies restricted to clients who the TIDE statistic to peer review. They analyses and accept only Larson and sought treatment for grief-related difficulties describe a “post hoc blind peer review” (p. Hoyt’s (2007) conveniently restricted evidenced “robust effect sizes” (p. 377) sim- 349) of the TIDE statistic by two “national choice, an effect size of .43 is nothing to ilar to other forms of therapy. Allumbaugh Professional Psychology: Research and Practice 377 Copyright 2008 by the American Psychological Association 2008, Vol. 39, No. 3, 377–380 0735-7028/08/$12.00 378 THE FORUM and Hoyt themselves suggested that these Currier, J. M., Holland, J. M., & Neimeyer, counseling. In the brief space allotted for this clients were probably more distressed than R. A. (2007). The effectiveness of bereave- response, we correct several mischaracter- other clients. ment interventions with children: A meta- izations of our conclusions and note points The development of new diagnostic analytic review of controlled outcome re- of agreement between our actual conclu- search. Journal of Clinical Child and tools for the assessment of complicated sions, including those partially quoted by Adolescent Psychology, 36(2), 253–259. grief (Prigerson et al., 1995; Prigerson, & Fortner, B. V. (1999). The effectiveness of grief Bonanno and Lilienfeld, and their own stated Maciejewski, in press) has made it possi- counseling and therapy: A quantitative re- views about grief counseling. We elaborate ble to tailor grief treatments toward the view. Unpublished doctoral dissertation, Uni- on our earlier discussion about how re- most appropriate clinical samples. In fact, versity of Memphis, Memphis, TN. searchers and practitioners can draw valid two recent studies that preselected clients Kato, P. M., & Mann, T. (1999). A synthesis of conclusions from the empirical literature and for complicated grief and tested manual- psychological interventions for the bereaved. what factors add to our confidence about ized treatments boasted impressively large Clinical Psychology Review, 19, 275–296. these conclusions. Larson, D., & Hoyt, W. (2007).What has be- treatment effects (Boelen, de Keijser, van In our earlier article, we raised concerns come of grief counseling? An evaluation of den Hout, & van den Bout, 2007; Shear, about the impact of claims about Frank, Houck, & Reynolds, 2005). This is the empirical foundations of the new pessi- mism. Professional Psychology: Research treatment-induced deterioration effects where the debate should be focused—not and Practice, 38, 347–355. (TIDE) in a dissertation (Fortner, 1999), on tiresome battles about whether inter- Neimeyer, R. A. (2000). Searching for the mean- which appear to have attained the status of vention is or is not good for everyone, but ing of meaning: Grief therapy and the process scientific fact, even though neither the on (a) fine tuning assessment instruments of reconstruction. Death Studies, 24, 541–558. findings themselves nor the statistical so that they can best identify those be- Prigerson, H. G., & Maciejewski, P. K. (in method on which they are based has ever reaved people in serious clinical need and press). Prolonged grief disorder: Inclusion in been subjected to peer review. Bonanno DSM–V. In R. O. H. M. Stroebe, H. Schut, & (b) further developing effective treatments and Lilienfeld (2008) criticized us for re- W.