How Firm Is Our Foundation? Current Play Therapy Research Roger D. Phillips Alburtis, Pennsylvania

The present article examines play therapy research since Phillips’s (1985) review. Play therapy’s evidence base remains largely inadequate using specific scientific/ methodological criteria. The most compelling evidence for play therapy’s effectiveness is found for children facing medical procedures, although alternative explanations of the same data cannot be disconfirmed. The present conclusions are considered relative to findings from recent meta-analyses of play therapy research. Suggestions are made for improving play therapy research as well as broad questions to guide such research. Keywords: play therapy, empirical research, evidence base, literature review

Pick up the 10th anniversary special issue of the Journal of Clinical Child and Adolescent Psychology (JCCAP; Silverman & Hinshaw, 2008) on evidence-based psychosocial treatments for children and adolescents and feel its considerable professional heft. Then look through its pages for a discussion of play therapy (PT). There is none. One would have ended in the same predicament with the inaugural compilation (Journal of Clinical Child Psychology, Lonigan, Elbert, & Johnson, 1998) or with any of the recent general volumes on evidence-based treatments for children (e.g., Kazdin & Weisz, 2003). How can one of the most widely used and historically illustrious treatments for children warrant only the briefest mention, if that much, within these load-bearing, forward-looking professional publications? The sobering answer is that a body of credible scientific evidence for most of PT still does not exist. Unfortunately, this is not a new conclusion (e.g., Lebo, 1953; Phillips, 1985; Pumfrey & Elliott, 1970), but was reached again only after a thorough review of the published, empirical PT research (PTR) since Phillips (1985), including the two recent corner-turning meta-analyses (Bratton, Ray, Rhine, & Jones, 2005; LeBlanc & Ritchie, 2001), although excluding the literatures on filial therapy and parent– child interaction therapy (see below). It is important to note that others have read the PTR literature quite differently and would disagree with the present conclusion (Bratton & Ray, 2000; Ray, 2006; Ray, Bratton, Rhine & Jones, 2001); indeed, the title of a recent edited volume confidently asserts precisely the opposite conclusion (Reddy, Files-Hall, & Schaefer, 2005). Thus, it is important to consider how, if at all, these divergent views can be reconciled.

Roger D. Phillips, Alburtis, Pennsylvania. Correspondence concerning this article should be addressed to Roger D. Phillips, 120 West Second Street, P.O. Box 100, Alburtis, PA 18011. E-mail: [email protected] 13 International Journal of Play Therapy 2010, Vol. 19, No. 1, 13–25

© 2010 Association for Play Therapy 1555-6824/10/$12.00 DOI: 10.1037/a0017340

14

Phillips

CREDIBLE SCIENTIFIC EVIDENCE Nathan and Gorman (2002) proposed a typology of studies based on methodological rigor and it is used routinely in determinations of sufficient scientific evidence (e.g., see JCCAP, Silverman & Hinshaw, 2008, issue noted above). Type I studies are randomized, prospectively designed, clinical trials (RCTs) using randomly assigned comparison groups, blind assessments, clear inclusion/exclusion criteria, state-of-the-art diagnosis, measurement of treatment fidelity, and adequate samples sizes to power clearly described statistical analyses. Type II studies, according to Nathan and Gorman, are clinical trials using a comparison group to test an intervention but may contain weaknesses limiting conclusions, although no critical flaws that would undermine the integrity of the data for answering the study’s question(s). Type III studies have significant methodological flaws that make their findings suspect, and include simple pre–post designs or studies using retrospective designs. (Nathan & Gorman, 2002, noted three additional study types, but those are irrelevant to the present work.) How does this typology of scientific rigor inform conclusions about current PTR? Are there areas of PT that might legitimately warrant the appellation of “evidence based” using these demanding requirements? The following three sections review one promising domain of PT (i.e., children facing medical procedures), raise questions that belie the inherent complexity even in this promising area, and then assess the remaining PTR using research related to anxiety/fear as an illustrative exemplar.

CHILDREN FACING MEDICAL PROCEDURES A reasonably solid evidence base undergirds PT interventions for children facing medical procedures, which represents an evidentiary improvement over the years (see Phillips, 1988). Recent work, mostly Type II studies, has added to the promising early findings (Burstein & Meichenbaum, 1979; Cassell, 1965; Johnson & Stockdale, 1975) described originally. For example, in Rae’s widely cited study (Rae, Worchel, Upchurch, Sanner, & Daniel, 1989) children who received a twosession nondirective PT intervention reported improvements in hospital-related fears compared to children in nondirective verbal support, “diversionary play” (board games, puzzles, etc.) and no-treatment control groups; no other outcomes, however, revealed group differences, including objective measures provided by parents or nurses. Young and Fu (1988) looked at the impact of a “needle play” intervention on children’s self-reports and objective reports of pain during and after a blood test. Children who received the intervention (three different groups) had lower pulse rates 5 min after the blood test when compared to the control group, and two intervention groups were rated as more physically still and comfortable during the blood test; however, no group differences emerged immediately after the blood test nor were there any differences at any time on self-reported pain. Zahr (1998) evaluated a therapeutic puppet show depicting preoperative and surgical events as intervention for Lebanese preschoolers. Children in the puppet-show group were significantly less upset behaviorally and physiologically at the time of a

Current Play Therapy Research

15

preoperative injection than were no-treatment children and they urinated significantly sooner after surgery (a medical marker for recovery); however there were no group differences in children’s cooperation, and significant differences in parentreported, posthospitalization, behavioral adjustment were irretrievably compromised by notable and systematic subject attrition. Of particular note, Li and his colleagues (Li, 2007; Li & Lopez, 2008; Li, Lopez, & Lee, 2007a, 2007b) evaluated a therapeutic play intervention to prepare Chinese children for outpatient surgery in a Type I, double-blind RCT. Both treatment and control groups received the standard informational presentation by the hospital staff 1 week before the surgery, but treatment-group children also received a doll demonstration of the upcoming surgical procedures within the operating room and were encouraged to repeat the same procedures on the doll themselves. This intervention was conducted in groups of five children, with one accompanying parent for each child. Children in the treatment group reported significantly less anxiety after the intervention (both before and after surgery) and evinced fewer emotional behaviors at anesthesia induction, but there were no differences in self-reported postoperative pain. Parents in the treatment group also reported significantly less anxiety and greater satisfaction compared to those in the control group. Parents’ participation during the intervention is a potential confound in specifying the true child-related treatment effect, although their presence accurately represents both reality as well as the good and right thing to do during such events. Thus, the available scientific evidence is reasonably persuasive for this PT context. One can have considerable confidence in Li’s findings and the promise of his group-based, therapeutic-play intervention as a method for preparing children for outpatient surgery, although conclusions beyond the parameters of this RCT should await replication, preferably by an independent research team. It is worth noting, too, that the content of Li’s intervention is quite comparable to other medical-preparation play interventions, but differs mostly in the quality of the intervention’s evaluation (i.e., research design, controls, execution).

QUESTIONS OF COMPLEXITY IN MEDICAL-RELATED PTR This PT domain is certainly closer to evidentiary “terra firma” than other PT domains, but also not quite there yet. Specifying this domain as “children facing medical procedures” narrows the focus substantially, but leaves quite a bit still unknown and some of that unknown might explain the current results just as plausibly. For example, with the exception of Rae et al. (1989), virtually all of these findings were based on cognitive– behavioral PT strategies (e.g., providing information, modeling, rehearsal), even if they were not specifically identified as such by the researchers. Thus, the interventions are confounded with the population/context, and the evidence base might instead reflect a cognitive– behavioral approach to PT (see Knell, 1998). Studies that disentangle these factors within one design will be necessary. The outcomes for this PT context also seem more circumscribed than initially thought. Treatment-group differences were found for specific measures of anxiety or fears, but usually not found for other experiences (e.g., generalized anxiety,

16

Phillips

self-reported pain, cooperation with procedures). Differences were found in selfreports, but not always via the observations of others (e.g., parents, nurses). Effects may be time dependent in unknown ways, as some studies reported immediate impacts but not longer lasting differences, whereas others found slightly delayed effects but nothing at the actual time of the procedure. Although children facing medical procedures is more specific than the larger universe of children, greater precision may be required, especially in relating the medical context to specific outcomes. The reviewed studies, for example, included children facing outpatient surgery, blood draws, and hospitalization for acute conditions. It seems possible that there would be different effects among these various medical contexts as well as others not addressed in these studies (e.g., extended or repeated hospitalizations, children with chronic conditions). Last, some of these investigations included the presence of parents and others did not; the possibility exists that PT’s effects for young children are mediated through or moderated by parents.

WHAT ABOUT THE REST OF PTR? What should be made of the rest of PTR vis-a´-vis the conclusion about a limited scientific evidence base for PT? One feature of PTR previously noted by Phillips (1985) remains true today: The field is characterized by a disparate array of studies that often do not build incrementally or conceptually on previous work. For instance, it is very difficult to reach a critical mass of findings necessary for trustworthy conclusions when widely varying samples of children are used, even when studying, in principle, the same population. Unfortunately, even if one tries to work around this shortcoming by organizing research along other dimensions, most PTR is too methodologically compromised (i.e., Type 3 studies) to allow for the credible accumulation of findings. To illustrate these broader points about PTR, consider PT for children’s anxieties/fears, which seems at least plausible based on the reasonably good evidence from the medical-procedures PTR and promising earlier work (e.g., Milos & Reiss, 1982), and that has been proffered by others (Russ, 2004). However one scans the recent PTR in vain for studies that focus fully and carefully on children’s anxieties/fears. Baggerly (2004) found improvement on only one (of several) measures of anxiety in a study of children in a homeless shelter; however, this work was compromised by significant attrition (48%) without a follow-up intent-to-treat analysis and widely varying treatment implementation. Fourth- and fifth-graders referred for classroom behavioral problems were unchanged in teacher ratings of their internalizing behaviors after receiving a 10-week sandtray intervention while their wait-list peers increased such behaviors (Flahive & Ray, 2007), although it appears that teachers knew the children’s treatment status. Post (1999) found no differences in anxiety among fourth- to sixth-grade children from an “at risk” school after an extremely variable, uncontrolled course of individual PT. Shen (2002) reported significant improvements in children who received a short-term group PT intervention after an earthquake in China compared to those not receiving treatment; however, the measures used in this study were not valid for Chinese populations, the statistical analyses were uncorrected for the large number of

Current Play Therapy Research

17

separate analyses, and a placebo effect cannot be ruled out. Ray, Schottelkorb, and Tsai (2007) found that children with attention deficit/hyperactivity disorder (ADHD) symptoms who received 16 sessions of individual PT reduced their teacher ratings of comorbid anxious behavior significantly more than those children who received a “reading mentoring” intervention. However, it is difficult to situate these results because reading mentoring seems untenable as a comparison group for this population given that there are known effective treatments for ADD/ADHD (see Pelham & Fabiano, 2008; Pelham, Wheeler, & Chronis, 1998). In a carefully designed study (Constantino, Malgady, & Rogler, 1986), two forms of “cuento therapy” (based on Puerto Rican folktales) were compared to PT and no treatment in reducing young Puerto Rican children’s anxiety (among other behaviors). One of the two versions of cuento therapy yielded significantly less anxiety than all other groups (including PT) in first-grade children (but not in older children) immediately after intervention; this effect was largely maintained at follow-up 1 year later. Thus, PT’s value in treating anxiety, outside the context of medical procedures, seems unclear and inconclusive—some negative findings, some null, and some positive but with significant design factors that vitiate their integrity, and all of them indirect tests. The PTR field would benefit from carefully controlled experiments that address directly the effects of PT on clearly defined anxieties/fears in children. This inconclusive and methodologically compromised state of affairs regarding anxieties/fears exemplifies the current status of most PTR (Danger & Landreth, 2005; Dougherty & Ray, 2007; Fall, Balvanz, Johnson, & Nelson, 1999; Fall, Navelski, & Welch, 2002; Gariepy & Howe, 2003; Garza & Bratton, 2005; Kaduson & Finnerty, 1995; Karcher & Lewis, 2002; Nicol, Smith, Kay, Hall, Barlow, & Williams, 1988; Reams & Friedrich, 1994; Reyes & Asbrand, 2005; Scott, Burlingame, Starling, Porter, & Lilly, 2003; Trostle, 1988; Tyndall-Lind, Landreth, & Giordano, 2001). Perhaps the only exception to the preceding conclusion is the Type I RCT from the Oklahoma group (Bonner, Walker, & Berliner, 1993, 1999) working with a clearly and empirically defined sample of young children with sexual problem behaviors. Bonner and her colleagues developed a 12-session, cognitive– behavioral, group intervention that they compared to a 12-session intervention of group dynamic/client-centered PT; both interventions also included structured collateral parent groups. At the end of treatment, there were statistically significant reductions in parent-reported sexual behaviors for children in both treatment groups; similar findings persisted through 1- and 2-year follow-ups using only a subsample of the total children. In a recent 10-year follow-up using the entire sample and public-sector databases, they found that children from both treatment groups were no more likely than individuals in the general population to perpetrate any sexual abuse or other sexual offenses as adolescents and young adults (Carpentier, Silovsky, & Chaffin, 2006). Statistical survival analyses found that if abuse/offenses did occur they were more likely among those who received the group PT compared to the cognitive– behavioral group; however, these researchers regarded this finding as the likely result of the enhanced statistical power associated with using their entire sample as opposed to genuine treatment differences. Thus, dynamic/clientcentered group PT, as manualized by the Oklahoma researchers, seems equally effective in reducing the sexual behaviors among young children diagnosed with such problems. Interestingly, it is not clear if this conclusion would still be true

18

Phillips

today given that more recent work by this research team has moved solidly in the direction of its cognitive– behavioral intervention as the evidence (usually in RCTs) for its effectiveness has mounted.

META-ANALYSES OF PTR Two recent meta-analyses of PT outcomes (Bratton et al., 2005; LeBlanc & Ritchie, 2001) have caused considerable excitement in the field and are often cited now as support for PT’s evidence base. These are ambitious and impressive achievements for PTR, providing much-needed organization of far-flung research activities in the field. They offer a chance to look across the entire PTR field, allowing for the emergence of some expectable and some unexpected findings. Reading them, one has the sense of the field being lit as never before, which affords the opportunity to see some things that were once only dimly viewed. In particular, the central finding of the average effect sizes for PT, hovering near .70, is impressive and likely to be a real finding given the close range of the average effect sizes across several meta-analyses looking at somewhat different slices of the child psychotherapy and/or PT literatures (Casey & Berman, 1985; Kazdin, Bass, Ayers, & Rodgers, 1990; Weisz, Weiss, Alicke, & Klotz, 1987; Weisz, Weiss, Han, Granger, & Morton, 1995). These findings indicate that PT is considerably better than nothing, and that it is better to about the same degree as most other forms of adult or child psychotherapy. In and of itself, this quantification of value in PT and its relative stature among the repertoire of interventions is noteworthy. However meta-analyses are not without their limitations, and if the PTR field now seems better lit there are also shadows that remain. The grand view that is gained in the compiling and combining of individual studies sometimes sacrifices the smaller picture (i.e., specificity in effects or contributing factors), even though these two research teams coded a number of study features. “The devil is in the details” of the coding categories used to evaluate the individual studies and then the actual codes assigned to the studies (which typically are invisible to readers). A number of these coding details, as well as some not undertaken (e.g., theoretical investment, institutional affiliation), were puzzling, raising uncertainty about some of the major conclusions and their value as an evidence base. For example, “parent involvement” in both meta-analyses yielded larger average effect sizes. Bratton et al. (2005) noted that most of their data regarding parents were derived from studies of filial therapy (22 of 26 studies), and LeBlanc and Ritchie (2001) included studies of filial therapy and parent– child interaction therapy (PCIT). It seems likely, however, that the range of parental involvement in PT is much broader than just these two specific parent-training programs, so the parent-involvement findings may be much narrower than described. And, on a conceptual level are filial therapy and PCIT truly PT? PCIT (Eyberg, 1988) contains as much non-PT as PT (i.e., Patterson-ian parent-directed intervention and nondirective-PT child-directed intervention, respectively), and there is evidence that the child-directed PT-portion of PCIT is insufficient for producing significant change in children’s disruptive behavior (Eisenstadt, Eyberg, McNeil, Newcomb, &

Current Play Therapy Research

19

Funderbuck, 1993). One could argue as well that training a parent to provide a PT intervention to their child (i.e., filial therapy) is quite a bit different in its relationship formation, history, context, and trajectory than an unfamiliar person beginning PT. As a result, one can feel confident of the effects derived from these specific parent-training approaches, but unsure how to situate them precisely within the larger field of PT. The largest coding issue, casting a very long shadow throughout the findings and conclusions, was the lack of adequate coding for study quality. Both metaanalyses coded for study design, with categories corresponding to studies using no-treatment and/or comparison groups. Bratton et al. reported no differences in average effect sizes (.79 to .89 range) as a function of their three-category studydesign code. Although study design is important, it does not suffice as a proxy for study quality, which should be a multifactor rating based on criteria similar to Nathan and Gorman’s (2002) Type I studies. This omission is all the more striking given the evidence for significantly larger average effect sizes as a function of greater methodological rigor from a previous meta-analysis focusing solely on nonbehavioral child psychotherapy research (Shirk & Russell, 1992). The consequences of this omission of study quality are fairly important. For example, in discussing their unexpected finding of a larger average effect size associated with humanistic/nondirective approaches to PT, Bratton et al. (2005) acknowledged both the dearth of clearly articulated interventions and the mixing of approaches within the same treatment protocols. These are serious qualifications to consider in interpreting the findings, but it is unclear how these problems were handled in the coding given that this particular analysis omitted only eight of the 93 studies. Later, when discussing the superior effects obtained when parents were the treatment provider, Bratton et al. duly warned readers in several places to bear in mind the unfortunate confluence of parents-as-treatment-providers and parents-asoutcome-raters. Such a serious measurement artifact, however, should figure prominently in a sturdy measure of study quality, which then could be examined systematically. Finally, there is the tantalizing result about the length of PT. Both research teams reported that larger effect sizes were found when PT reached 30 to 35 sessions, with diminished effect sizes for both briefer (⬍10 sessions) and longer treatment (⬎35 sessions). This finding seemed straightforward if a bit perplexing, and both research teams attempted to explain it in terms of putative clinical events or processes in PT. For example, Bratton et al. (2005) suggested that a sufficient number of sessions were needed for establishing a relationship and producing “enduring intrinsic change” (p. 386), whereas LeBlanc and Ritchie (2001) speculated about the “increase in and intensification of problem behaviors during initial stages of therapy” (pp. 158 –159). However, additional data clouded the picture. For example, Bratton et al. (2005) reported that, contrary to their general trend of increasing effect sizes across longer treatments, crisis settings provided the fewest sessions and yet yielded nearly the largest average effect size. Curiously, LeBlanc and Ritchie (2001) reported that their average PT effect size (across all studies) of .66 corresponded to 13 PT sessions. Their data are real and believable mathematically, but 13 sessions is not all that far from their threshold for significantly diminishing effect sizes (i.e., 10 sessions) and certainly much closer to the latter than to the threshold for more

20

Phillips

effective PT (i.e., 30 to 35 sessions). As a result, one has to wonder about the distributions of effect sizes and treatment length (which are depicted in their Figure 2, p. 157) and what realities—statistical, clinical, or otherwise—are reflected in those distributions. And, just how solid can these findings be regarding treatment duration when none of the individual studies evaluated or controlled treatment length in their original designs. Lengths of PT in the original studies were often driven by convenience, arbitrary prestudy decisions, or the parameters of other events (e.g., the school-year calendar, stays in settings), and in many cases were highly variable even within individual studies. Studies that systematically manipulate or control length of PT and measure differences in outcomes would be necessary before making the kinds of treatment-related recommendations suggested by Bratton et al. (2005, p. 386). Thus, this is a very intriguing, suggestive finding about treatment length rather than a definitive conclusion, and one that might serve as the launching pad for a series of systematic investigations. Indeed, this conclusion—suggestive or formative findings instead of definitive and summative ones—seems appropriate for these two important PT meta-analyses. Unfortunately, meta-analyses cannot compensate for decisions made about experimental questions, designs, and controls in the original PT studies. However, very promising PTR activities could emerge from these two benchmarking investigations.

SUGGESTIONS FOR FUTURE PTR Define the Population(s) and Sample(s) Precisely Specify the population of interest and measure the critical characteristics of one’s sample in order to ensure that it represents accurately the population of interest. Those critical characteristics of the sample must be verified empirically with accepted measurement instruments. PT will remain stalled if it continues to use convenience or poorly defined samples, or samples that only approximate the populations of interest.

Measure What Matters and Measure It Carefully Focus measurement activities on the critical issues defined within one’s research question. Those measurement activities will include, minimally, (a) verification of one’s sample (even if randomization is used), (b) treatment fidelity, and (c) focused dependent variables/outcome measures. A multimeasurement strategy should be employed whenever possible, with several different measurement tactics getting at the same construct from different but meaningful perspectives, and at several relevant intervals. This suggested strategy should not be construed as a scattershot approach, with measurement heading in several different directions simultaneously. These three measurement activities will be both conceptually related to and derived directly from the research question(s).

Current Play Therapy Research

21

Stay Close to the Data One’s obtained data should define the limits of one’s conclusions. Straying too far beyond those boundaries opens the door to skepticism at the reader’s end of the process. For example, PT researchers invite doubt when results and discussions are based solely or largely on reported effect sizes without statistically significant findings (or vice versa); this sort of disparity in results is complicated and begs for even greater care and explanation. Theoretical, conceptual, or clinical overstatement of results, with too little discussion of internal-validity threats, is still characteristic of much of contemporary PTR.

Small Is Not Necessarily Bad It is true that much of PTR has used fairly small samples of participants, making it more difficult to obtain statistically significant results. However, the sacrifice of statistical power with small samples can be offset to some degree by determined, tight control of other aspects of the research design. For example, meaningful and statistically significant interactions can be obtained with as few as 20 subjects in each cell of a 2 ⫻ 2 design (e.g., Barnett, 1984). Small, tightly controlled designs often provide foundational support for designs that can justifiably move to scale; the impressive history of work with PCIT exemplifies this strategy. It is a strategy that builds scientific credibility, too.

META-QUESTIONS IN PTR What Is the Theory Or Model of Change in PT? One reads much of the current PTR and is left wondering about implicit and explicit models or theories of change within PT. Presumably, these elements are contained within a theory that guides the PT, but the precise mechanisms and operations are not always clear. For example, in client-centered/nondirective PT much emphasis is placed on the relationship between child and play therapist as the crucible in which healing and growth occur. However, what precisely is the play therapist doing in this relationship work, and what are the expectable links between those therapist behaviors and child behaviors/outcomes (the latter of which could be within-session “micro” outcomes, intermediate therapeutic achievements, or end-of-treatment outcomes)? Do those therapist and child behaviors and the linkages between them change as a function of child problem or a different contextual factor? If not, then is PT operating more like a nonspecific factor in child psychotherapy and the true mechanisms (or “active ingredients”) are to be found in other events, activities, or processes? If it is truly a relationship-building process, can those relationship constructs be articulated and measured?

22

Phillips

Where Is the Play in PTR? It really is quite astonishing to realize how little play actually figures into current or past PTR even though it is the life-blood of PT. There were few researchers who measured anything about play in these investigations, perhaps because PT researchers have thought about play as a means to a therapeutic end, of which the latter is of paramount importance given the clinical contexts in which they usually operate. However, after all, it is play therapy, and so play should figure into the equation fairly prominently in some form. To some degree this is a process issue as opposed to an outcome issue, and historically there has been even less empirical PT process research than outcome research. However it is, more properly, a therapeutic-mechanism linkage issue. For example, the critical components of the cognitive– behavioral approaches to preparing children for medical procedures often included modeling (via dolls, puppets) of impending procedures and coping, as well as children’s opportunities to repeat and rehearse the new experiences in their play. Thus, whether and how children played with the dolls should have figured prominently in the claims made about the generally positive outcomes in this form of PT. Alternatively, there is a long history in PT of special value accorded to pretend or fantasy play. Within some PT frameworks, a child’s intrapersonal and/or interpersonal issues are revealed in the thematic elements of pretend play; there are other ways to think about pretend play as well (e.g., Russ, 2004). The point is the same no matter the framework: if pretend play is critical to treatment, then one ought to measure it and link it systematically to outcomes. And, what of nonpretend play? How, if at all, do such forms of play contribute to PT? Last, if we assume that children’s play will reveal their difficulties, then would it not also reveal their well-being? In this view, play becomes an outcome as well. These are gaping holes in PT’s knowledge base.

An RCT in PTR? Like all types of research, RCTs have their weaknesses (e.g., not ruling out Hawthorne effects), and some research questions are better addressed by other designs. RCTs are, however, one of the best strategies in the methodological repertoire for approximating direct causality. What would it take to bring this best and most convincing experimental method to PTR? Focused research questions, thorough planning, painstaking organization, funding, careful oversight, to be sure—indeed, lots of each of these elements, and probably others as well (e.g., maybe multisite collaboration). However, they can be accomplished. It is surprising and certainly unfortunate that so few RCTs involving PT have been undertaken. RCTs can be accomplished on small and large scales, and PTR would do well to keep its “eyes” on this experimental “prize” no matter what scale is planned. What would a multisite, large scale PT RCT look like? What empirical work would need to precede such an endeavor? What PT research questions warrant or could propel RCTs? The benefits to the field and to children and families would be enormous.

Current Play Therapy Research

23

REFERENCES Baggerly, J. (2004). The effects of child-centered group play therapy on self-concept, depression, and anxiety of children who are homeless. International Journal of Play Therapy, 13, 31–51. Barnett, L. A. (1984). Young children’s resolution of distress through play. Journal of Child Psychology and Psychiatry, 25, 477– 483. Bonner, B. L., Walker, C. E., & Berliner, L. (1993). Children with sexual behavior problems: Assessment and treatment. Washington, DC: Administration of Children, Youth, and Families, Department of Health & Human Services. Bonner, B. L., Walker, C. E., & Berliner, L. (1999). Children with sexual behavior problems: Assessment and treatment (Final Report, Grant No. 90 –CA–1469). Washington, DC: Administration for Children, Youth, and Families, DHHS. Bratton, S., & Ray, D. (2000). What the research shows about play therapy. International Journal of Play Therapy, 9, 47– 88. Bratton, S. C., Ray, D., Rhine, T., & Jones, L. (2005). The efficacy of play therapy with children: A meta-analytic review of treatment outcomes. Professional Psychology: Research and Practice, 36, 376 –390. Burstein, S., & Meichenbaum, D. (1979). The work of worrying in children undergoing surgery. Journal of Abnormal Child Psychology, 7, 121–132. Carpentier, M. Y., Silovsky, J. F., & Chaffin, M. (2006). Randomized trial of treatment for children with sexual behavior problems: Ten-year follow-up. Journal of Consulting and Clinical Psychology, 74, 482– 488. Casey, R., & Berman, J. (1985). The outcome of psychotherapy with children. Psychological Bulletin, 98, 388 – 400. Cassell, S. (1965). Effect of brief puppet therapy upon the emotional responses of children undergoing cardiac catheterization. Journal of Consulting Psychology, 29, 1– 8. Constantino, G., Malgady, R. G., & Rogler, L. H. (1986). Cuento therapy: A culturally sensitive modality for Puerto Rican children. Journal of Consulting and Clinical Psychology, 54, 639 – 645. Danger, S., & Landreth, G. (2005). Child-centered group play therapy with children with speech difficulties. International Journal of Play Therapy, 14, 81–102. Dougherty, J., & Ray, D. (2007). Differential impact of play therapy on developmental levels of children. International Journal of Play Therapy, 16, 2–19. Eisenstadt, T. H., Eyberg, S. M., McNeil, S. B., Newcomb, K., & Funderbuck, B. (1993). Parent-child interaction therapy with behavior problem children: Relative effectiveness of two stages and overall treatment outcome. Journal of Clinical Child Psychology, 22, 42–51. Eyberg, S. M. (1988). Parent-child interaction therapy: Integration of traditional and behavioral concerns. Child & Family Behavior Therapy, 10, 33– 46. Fall, M., Balvanz, J., Johnson, L., & Nelson, L. (1999). A play therapy intervention and its relationship to self-efficacy and learning behaviors. Professional School Counseling, 2, 194 –204. Fall, M., Navelski, L. F., & Welch, K. K. (2002). Outcomes of a play intervention for children identified for special education services. International Journal of Play Therapy, 11, 91–106. Flahive, M. W., & Ray, D. (2007). Effect of group sandtray therapy with preadolescents. Journal for Specialists in Group Work, 32, 362–382. Gariepy, N., & Howe, N. (2003). The therapeutic power of play: Examining the play of young children with leukaemia. Child: Care, Health and Development, 29, 523–537. Garza, Y., & Bratton, S. C. (2005). School-based child-centered play therapy with Hispanic children: Outcomes and cultural considerations. International Journal of Play Therapy, 14, 51–79. Johnson, P. A., & Stockdale, D. F. (1975). Effects of puppet therapy on palmar sweating of hospitalized children. Johns Hopkins Medical Journal, 137, 1–5. Kaduson, H. G., & Finnerty, K. (1995). Self-control game interventions for attention-deficit hyperactivity disorder. International Journal of Play Therapy, 4, 15–29. Karcher, M. J., & Lewis, S. S. (2002). Pair counseling: The effects of a dyadic developmental play therapy on interpersonal understanding and externalizing behaviors. International Journal of Play Therapy, 11, 19 – 41. Kazdin, A. E., Bass, D., Ayers, W. A., & Rodgers, A. (1990). Empirical and clinical focus of child and adolescent psychotherapy research. Journal of Consulting and Clinical Psychology, 58, 729 –740. Kazdin, A. E., & Weisz, J. R. (2003). Evidence based psychotherapies for children and adolescents. New York, NY: Guilford Press. Knell, S. M. (1998). Cognitive-behavioral play therapy. Journal of Clinical Child Psychology, 27, 28 –33. LeBlanc, M., & Ritchie, M. (2001). A meta-analysis of play therapy outcomes. Counseling Psychology Quarterly, 14, 149 –163. Lebo, D. (1953). The present status of research on nondirective play therapy. Journal of Consulting Psychology, 17, 177–183.

24

Phillips

Li, H. C. W. (2007). Evaluating the effectiveness of preoperative interventions: The appropriateness of using the Children’s Emotional Manifestation Scale. Journal of Clinical Nursing, 16, 1919 –1926. Li, H. C. W., & Lopez, V. (2008). Effectiveness and appropriateness of therapeutic play intervention in preparing children for surgery: A randomized controlled trial study. Journal for Specialists in Pediatric Nursing, 13, 63–73. Li, H. C. W., Lopez, V., & Lee, T. L. I. (2007a). Effects of preoperative therapeutic play on outcomes of school-age children undergoing day surgery. Research in Nursing & Health, 30, 320 –332. Li, H. C. W., Lopez, V., & Lee, T. L. I. (2007b). Psychoeducational preparation of children for surgery: The importance of parental involvement. Patient Education and Counseling, 65, 34 – 41. Lonigan, C. J., Elbert, J. C., & Johnson, S. B. (1998). Empirically supported psychosocial interventions for children. [Special Issue]. Journal of Clinical Child Psychology, 27(2). Milos, M. E., & Reiss, S. (1982). Effects of three play conditions on separation anxiety in young children. Journal of Consulting and Clinical Psychology, 50, 389 –395. Nathan, P., & Gorman, J. M. (2002). A guide to treatments that work (2e). New York, NY: Oxford University Press. Nicol, A. R., Smith, J., Kay, B., Hall, D., Barlow, J., & Williams, B. (1988). A focused casework approach to the treatment of child abuse: A controlled comparison. Journal of Child Psychology and Psychiatry, 29, 703–711. Pelham, W. E., Jr., & Fabiano, G. A. (2008). Evidence-based psychosocial treatments for attentiondeficit/hyperactivity disorder. Journal of Clinical Child and Adolescent Psychology, 28, 184 – 214. Pelham, W. E., Jr., Wheeler, T., & Chronis, A. (1998). Empirically supported psychosocial treatments for attention deficit hyperactivity disorder. Journal of Clinical Child Psychology, 27, 190 –205. Phillips, R. D. (1985). Whistling in the dark?: A review of play therapy research. Psychotherapy: Theory, Research, & Practice, 22, 752–760. Phillips, R. D. (1988). Play therapy in health care settings: Promises never kept? Children’s Health Care, 16, 182–187. Post, P. (1999). Impact of client-centered play therapy on the self-esteem, locus of control, and anxiety of at-risk 4th-, 5th-, and 6th-grade students. International Journal of Play Therapy, 8, 1–18. Pumfrey, P. D., & Elliott, C. D. (1970). Play therapy, social adjustment and reading attainment. Educational Research, 12, 183–193. Rae, W. A., Worchel, F. F., Upchurch, J., Sanner, J. H., & Daniel, C. A. (1989). The psychosocial impact of play on hospitalized children. Journal of Pediatric Psychology, 14, 617– 627. Ray, D. C. (2006). Evidence-based play therapy. In C. E. Schaefer & H. G. Kaduson (Eds.), Contemporary play therapy: Theory, research and practice (pp. 136 –157). New York, NY: Guilford Press. Ray, D. C., Bratton, S., Rhine, T., & Jones, L. (2001). The effectiveness of play therapy: Responding to the critics. International Journal of Play Therapy, 10, 85–108. Ray, D. C., Schottelkorb, A., & Tsai, M.-H. (2007). Play therapy with children exhibiting symptoms of attention deficit hyperactivity disorder. International Journal of Play Therapy, 16, 95–111. Reams, R., & Friedrich, W. (1994). The efficacy of time-limited play therapy with maltreated preschoolers. Journal of Clinical Psychology, 50, 889 – 899. Reddy, L. A., Files-Hall, T. M., & Schaefer, C. E. (2005). Empirically based play interventions for children. Washington, DC: American Psychological Association. Reyes, C. J., & Asbrand, J. P. (2005). A longitudinal study assessing trauma symptoms in sexually abused children engaged in play therapy. International Journal of Play Therapy, 14, 25– 47. Russ, S. W. (2004). Play in child development and psychotherapy: Toward empirically supported practice. Mahwah, NJ: Erlbaum. Scott, T. A., Burlingame, G., Starling, M., Porter, C., & Lilly, J. P. (2003). Effects of individual client-centered play therapy on sexually abused children’s mood, self-concept, and social competence. International Journal of Play Therapy, 12, 7–30. Shen, Y.-J. (2002). Short-term group play therapy with Chinese earthquake victims: Effects on anxiety, depression, and adjustment. International Journal of Play Therapy, 11, 43– 63. Shirk, S. R., & Russell, R. L. (1992). A re-evaluation of estimates of child therapy effectiveness. Journal of the American Academy of Child & Adolescent Psychiatry, 31, 703–709. Silverman, W. K., & Hinshaw, S. P. (2008). Evidence-based psychosocial treatments for children and adolescents: A ten-year update. [Special Issue]. Journal of Clinical Child & Adolescent Psychology, 37(1). Trostle, S. L. (1988). The effects of child-centered group play sessions on social-emotional growth in

Current Play Therapy Research

25

three- to six-year-old bilingual Puerto Rican children. Journal of Research in Childhood Education, 3, 93–106. Tyndall-Lind, A., Landreth, G. L., & Giordano, M. A. (2001). Intensive group play therapy with child witnesses of domestic violence. International Journal of Play Therapy, 10, 53– 83. Weisz, J., Weiss, B., Alicke, M., & Klotz, M. (1987). Effectiveness of psychotherapy with children and adolescents: A meta-analysis for clinicians. Journal of Consulting and Clinical Psychology, 55, 542–549. Weisz, J., Weiss, B., Han, S., Granger, D., & Morton, T. (1995). Effects of psychotherapy with children and adolescents revisited: A meta-analysis of treatment outcome studies. Psychological Bulletin, 117, 450 – 468. Young, M. R., & Fu, V. R. (1988). Influence of play and temperament on the young child’s response to pain. Children’s Health Care, 16, 209 –215. Zahr, L. K. (1998). Therapeutic play for hospitalized preschoolers in Lebanon. Pediatric Nursing, 24, 449 – 454.