Running head: FATHERS IN FAMILY THERAPY Father Participation in Family Therapy: A Single-Case (N=1) Study

The present study demonstrates the use of a single-case (N=1) design in exploring the process of decreased father participation in two dual-parent family therapy sessions. Specifically, the quantity of verbal and sub-verbal communications taking place between the therapists and families, in the form of initiations and responses, were examined. The results indicate that the therapists were not more likely to initiate an interaction with mothers than fathers, nor were fathers less likely to respond to therapists than were mothers. Rather, the results indicate that fathers took a less vocal role in the therapeutic process and that the therapists were less likely to reinforce fathers’ participation by responding to them. Fathers in Family Therapy 3 Father Participation in Family Therapy: A Single-Case (N=1) Study As part of a previous study examining the effectiveness of Solution Focused Family Therapy (SFFT) in the treatment of children with disruptive behavior, a subjective observation regarding the participation of fathers was noted (Graham et al., 2001). It appeared to the researchers that fathers of two-parent families appeared to be less involved in the family therapy process than did the mothers. The researchers made the observation that fathers were not as vocal as mothers, and did not take as active a role in the therapeutic process as did mothers. As discussed by van Walsum (2001), there is a growing body of literature looking at the effect of father involvement in family therapy. However, little empirical research has been conducted on why fathers participate less than mothers in family therapy. The literature in this regards is largely theoretical, pointing towards the role of gender stereotypes (c.f., Fox & Bruce, 2001, Goldner, 1985) The present study was conducted to determine the process of therapist-client interactions which contributed to the apparently decreased participation of fathers as compared to mothers in the aforementioned family therapy study (Graham et al., 2001). Additionally, this study will demonstrate the use of a single-case (n=1) design. As stated by Lundervold & Belwood (2000), “single-case (N=1) designs offer a scientifically credible means to objectively evaluate practice and conduct clinically relevant research in practice settings” (p. 92). While large-n studies are often inaccessible to the practicing clinician, n=1 research allows clinicians working outside of a research or academic environment the ability to objectively evaluate their work. While N=1 designs allow for a detailed examination of a single case, they are not generalizable to the general population. Only by conducting large-n studies can generalizability be claimed. Fathers in Family Therapy 4 In light of van Walsum’s (2001) discussion, the following reasons for the decreased participation of fathers are considered. First, the therapist may initiate verbal interactions with the fathers less than with mothers. As the fathers are invited to participate less (by asking fewer questions, inviting their opinions less, etc.) than mothers, they may be less likely to take a vocal role in family therapy. Second, fathers may participate less because their verbal participation is reinforced by the therapist to a lessor extent when compared to mothers. Thus, if therapists don’t respond to fathers as much as to mothers, the fathers may be less inclined to participate. Third, fathers may, of their own volition, be less inclined to respond to therapists when compared to mothers. Finally, the decreased participation of the fathers may be due to a combination of these factors. Method Participants Videotapes of the family therapy sessions of two families from a previous study on family therapy (Graham et al., 2001) were available for examination. The first family consisted of a mother and a father, both age 51, and a 9 year-old son. The child was receiving medication for a diagnosis of attention deficit hyperactivity disorder (ADHD), and his disruptive behaviors were the purpose of their seeking therapy. The therapist was a male advanced psychology doctoral student, following a manualized form of SFFT with a live supervision team watching through a closed-circuit television system. All family members and therapist were Caucasian. Therapy lasted a total of 5 sessions over 7 weeks. The second family was made up of an 8 year-old boy, his 29 year-old biological mother, and his 30 year-old step-father. The child in this family had received a diagnosis of ADHD, and was taking medication for it daily. His disruptive behaviors were the presenting problem for therapy, which lasted a total of 4 sessions over 7 weeks. The therapist was a male advanced doctoral Fathers in Family Therapy 5 student following a manualized form of SFFT with a live supervision team. Both the therapist and the family were Caucasian. Procedure The three study authors coded interactions in the videotapes by consensus. While this procedure does not allow for the calculation of interrater reliability, it is congruent with the goal of demonstrating a study which could be conducted by a practicing clinician working alone. Sample. Therapy sessions were coded in 15 second intervals, as this was the longest unit of time that could be consistently analyzed by the raters. In the first session of therapy, the first 10 minutes of the session were coded, as this is the time where therapy norms involving participation might be set. Following the initial 10 coded minutes, 10 minutes of therapy were skipped, followed by 2 minutes of coding. This process (skip 10 minutes, code 2 minutes, etc.) was repeated until the “feedback” segment of therapy. The SFFT treatment manual used by the clinicians included a feedback segment at the end of each session, which was coded in its entirety. The first 2 minutes of subsequent sessions were coded, followed by skipping 10 minutes, coding 2 minutes, etc. The feedback portion of each session was coded in it’s entirety. This resulted in a total of 572 15-second segments being coded. Coding procedure. During each 15-second segment, the person (therapist, father, mother, or child) who spoke the most was noted. During segments in which it was difficult to determine who spent the most time speaking (due to people talking over one another, etc.), this category was left blank. Whether one person initiated contact with another during a given segment was also noted. A person was considered to initiate contact with another by either asking a question, directing a behavior, or beginning a new topic for discussion. In coding, the initiation was coded along with the information about to whom the initiator was directing the contact, as Fathers in Family Therapy 6 evidenced by saying a name, non-verbal indicators, to whom the person orients, and the context of the initiation. Thus, if the therapist asked the mother a question, this would be coded as a therapist-mother initiation. Additionally, when a person did not initiate to any one person, this was coded as an “open-ended” initiation. For example, a therapist who looks around the room at all family members and asks how things have been going over the past week, would be coded as a therapist open-ended initiation. In addition to coding who initiates contact with another, who responded to whom was also noted. Responding was defined as making a verbal statement or a sub-verbal utterance in reply to another’s communication (be it an initiation or response). Utterances were considered the use of a clear nonverbal or one or two-sound response, such as “uh-huh”, while verbal responses were considered any response longer than one or two sounds, or any response that completely answered the other’s inquiry (for example, if the therapist asked the child if he enjoyed playing soccer, and the child said “yes”, a child-therapist response would be coded). A separate examination of verbal responses and utterances revealed highly similar results, and for the sake of brevity, both verbal responses and utterances have been pooled together as “responses” for the remainder of this manuscript. Any one category can only be recorded once for a given time segment. For example, if the therapist responds to the child twice during a segment, only one therapist-child response was coded for that segment. Thus, codings signify only whether or not a given communication occurred within the 15second segment. If a communication began in one segment and continued in the next, both segments were coded as having the communication occur in them. Analysis As separate analyses of the two families yielded identical results, both families were pooled together for the following analyses. This was Fathers in Family Therapy 7 pooling further justified by the fact that each of the therapists used the same treatment manual and the same live-supervision team. Frequencies of the number of segments containing each possible category of interaction were computed. The number of segments containing a therapist initiation to the father was compared to the number of segments containing a therapist initiation to the mother. Likewise, the number of segments containing therapist responses to fathers was compared to therapist responses to mothers, and father responses to therapists was compared to mother responses to therapists. Each of these comparisons was subjected to a chisquare test of statistical significance, using one set of the pair as the expected values for the other. For example, therapists initiated contact with fathers during a total of 29 segments out of 572. The expected values of yes=29 and no=514 would then be used as the expected values for examining the number of segments in which the therapist initiated contact with the mother. To test the hypothesis that therapists might be more likely to respond to mothers than to fathers, the number of therapist responses to each parent was divided by the number of that parent’s responses to the therapist. As parents did not initiate to therapists, the parent responses to the therapists represented the only opportunities for the therapist to subsequently respond to the parent. The division of therapist-parent responses by the “opportunities” therefore resulted in the proportion of segments in which the therapist responded to the parent, given the number of opportunities for responding. These proportions were then compared using a chi-square analysis. The hypothesis suggesting that fathers may take advantage of fewer opportunities to respond to therapists than do mothers was tested by calculating the percentage of opportunities to respond to the therapist taken by each parent. This percentage was calculated by dividing the number of parent responses by the number of segments which contained at least one of Fathers in Family Therapy 8 the following: a therapist initiation to that parent, an open-ended therapist initiation, or a therapist response to that parent. This resulted in a proportion similar to the one described above. These proportions were then compared using a chi-square analysis. Finally, the percentage of segments containing therapist-parent initiations, therapist-parent responses, parent-therapist responses, the percentage of response opportunities taken by the therapist, and the percentage of response opportunities taken by the parents for both mothers and fathers were graphed and examined visually across session phase and therapy session to determine whether the process changed as a result of within-session phase or between sessions. The number of responses in the first phase of therapy examined the first 10 minutes of the first session and the first 2 minutes of each following session. The second phase of therapy included all other segments, save for the feedback segments, which made up the third session phase. Likewise the initial session of therapy comprised the first therapy session. All other therapy sessions were used for the middle therapy session, save for the final therapy session which comprised the final therapy session. Effect sizes were calculated for all tests of statistical significance to provide a measure of effect unbiased by sample size. For all chi-square analyses, the effect size w was calculated. W is a standardized correlation effect size recommended by Cohen (1988) for use with chi-square statistics. According to Cohen’s general guidelines, a w of .10 is considered “small”, .30 “medium”, and .50 “large”. Additionally, to maintain an experiment-wise error rate of p=.05 a Bonferroni correction for 5 tests of statistical significance was employed. In order to be considered statistically significant, individual tests therefore required a p-value of less than .0102. Fathers in Family Therapy 9 Results Mothers talked the most during a greater number of segments than did fathers, with mothers talking the most a total of 151 segments while fathers only talked the most 38 out of 172 segments. This difference was statistically significant with a medium effect size, as summarized in Table 1. ------------------------INSERT TABLE 1 ABOUT HERE ------------------------The results of the chi-squares comparing therapist-parent initiations, therapist-parent responses, and parent-therapist responses are shown in Table 2. As shown, therapists did not initiate statistically or clinically significantly more to mothers or fathers. Therapists responded more to mothers than to fathers, and mothers responded more to therapists than did fathers, each with medium effect sizes. ------------------------INSERT TABLE 2 ABOUT HERE ------------------------Therapists had a total of 198 “opportunities” (parent-therapist responses) to respond to fathers, as compared to a total of 318 opportunities to respond to mothers. When the number of therapist-parent responses was divided by the number of opportunities for responding to that parent, it revealed that therapists responded to 42.1% of mother opportunities, while only responding to 24.2% of father opportunities. The results, shown in Table 3, show that this difference is statistically significant, though with a effect size that is considered small approaching medium. ------------------------INSERT TABLE 3 ABOUT HERE ------------------------Fathers in Family Therapy 10 The Chi-square statistic comparing the number of parent-therapist responses given the number of opportunities to respond to the therapist could not be computed because one of the frequencies was less than 5 (fathers did not respond to an opportunity only 3 times). The percentage of opportunities taken, however, are listed in Table 3. As shown here, fathers responded 98.5% of the opportunities and mothers responded to 94.1% of the opportunities. ------------------------INSERT TABLE 4 ABOUT HERE ------------------------The graphs depicting the percentage of segments containing therapistparent responses, parent-therapist responses, the percentage of response opportunities taken by the therapist, and the percentage of response opportunities taken by the parents resulted in parallel lines, with a consistent difference between mothers and fathers across within-session phases and between sessions. As these results were consistent with the above findings, the graphs are not presented here. The percentage of segments containing therapist-parent initiations across within-session phases is presented in Figure 1. As shown here, during the initial phase of each therapy session, therapists initiated contact with fathers more than mothers. Subsequent phases show the therapist responding more to the mother than the father. -------------------------INSERT FIGURE 1 ABOUT HERE -------------------------The percentage of segments containing therapist-parent initiations across therapy sessions are presented in Figure 2. As shown here, therapists initiate to mothers and fathers nearly the same during the first session, initiate to mothers more than fathers during the “body” of therapy, and Fathers in Family Therapy 11 initiate to fathers slightly more than mothers during the final “wrap-up” session. -------------------------INSERT FIGURE 2 ABOUT HERE -------------------------Discussion As expected by the authors, mothers participated more actively than did fathers in these family therapy sessions. Contrary to what the authors anticipated, however, this difference in participation was not due to the number of times in which the therapist initiated contact with the parents. The results shown in Table 2 indicate that the therapist initiated contact an equal number of segments with the father and the mother. It was found, however, that therapists did respond more to mothers than to fathers, and that mothers did respond more to therapists than did fathers. This information alone, however, is insufficient to determine the source of the increased participation by mothers. While the question of therapist-parent initiations can be tentatively ruled out, it is unknown if the difference in participation was due primarily to mothers responding more to therapists or therapists responding more to mothers. The results presented in Tables 3 and 4 answer this question. Therapists respond to mothers a higher proportion of the time than they did to fathers, given the respective number of parent responses. Although a chisquare statistic for the number of parent-therapist responses given the number of opportunities (segments which contained a therapist-parent initiation, therapist-parent response, or therapist open-ended initiation) could not be computed due to a violation of the chi-square assumption that all frequencies must be greater than 5, the resulting percentages suggest that mothers and fathers responded equally given the number of opportunities. This suggests that the decreased participation of fathers when compared to Fathers in Family Therapy 12 mothers could have been the result of fewer therapist-father responses, not fewer father-therapist responses or therapist-father initiations. The visual interpretation of the differences between fathers and mothers revealed that the above results were stable across within-session phases and across sessions, with one exception. Therapists initiated more to fathers during the initial phase of each therapy session. This phase is marked by information gathering, asking the family to report their progress over the previous weeks. That fathers were engaged more than mothers during this information gathering phase, suggests that therapists believed that fathers are more responsible for “business” activities such as reporting progress. Additionally, therapists initiated more to fathers during the final therapy session, a session marked by reviewing the therapy process and making plans for continued improved family functioning. While this could be due to fathers being seen by therapists as more instrumental or responsible for “making plans”, it may also be an artifact of the fact that, by the end of therapy, the therapists became aware of the decreased participation of fathers and made an attempt to get them more involved in the therapeutic process. Surprisingly, therapists did not initiate more to fathers than mothers in the first session of therapy, a session marked by information gathering and reporting. In these two family therapy cases, the “culprit” for reduced father participation appeared to be a low number of therapist responses to the father. Responding to another’s comments in therapy validates their participation, and serves to encourage that person’s active participation in the therapeutic process. The fathers in the present study were not positively reinforced for participating in therapy as much as were the mothers, and participated less as a result. Fathers in Family Therapy 13 The reduced amount of reinforcement may be due to a number of factors. Research on gender roles and socialization suggests that women are socialized to be more relationship-oriented, while men are socialized to be more instrumental, or “problem-solving”. As family therapy often focuses more on relationships than direct problem-solving, it may be that mothers’ relational-oriented comments were more in line with the therapists’ ideas of therapy than fathers’ instrumental comments, and were consequently rewarded to a greater extent. Alternatively, it might be the case that the gender stereotypes were not true, and that the participation of both fathers and mothers was qualitatively equivalent. In this case, the therapist might be less likely to reward the father than the mother due to gender role expectations. If the therapist holds the belief that mothers are more relationship-oriented, and are more the family caretakers than are the fathers, the therapist might support the mother more during therapy. Other explanations are also possible, though the fact remains that father involvement and active participation in family therapy is important. Fathers constitute half of the parental dyad, and are no less important than are mothers. The possibility that fathers may come to family therapy already feeling “inadequate” in the relational domain only exasperates the problem. Increased father participation in the cases in question could likely be facilitated by an increase in the number of supportive responses made by the therapist. By being rewarded with therapist responses, fathers may be more likely to feel validated as a valuable part of family therapy. Thus, the fathers’ self-esteem could be increased, and he could be more likely to continue being an active part of the therapeutic process. Being a single-case study, the above results are not generalizable beyond these two family-therapist couplings. Additionally, the fact that both therapists were male, and that a manualized form of SFFT was used in these Fathers in Family Therapy 14 therapy sessions might have contributed to the results. The N=1 design, however, is easily accessible to the independently practicing clinician. It does not require a large investment of resources, and can be invaluable in tracking client progress, responses to interventions, and, as demonstrated here, the therapeutic process. N=1 designs are particularly suited to suggesting practice, as they provide a personalized objective analysis of an individual or family. Though the results of this study are not generalizable beyond these two families, they do raise the question of whether these results might hold true for the population at large, over a variety of different family therapy orientations, or over different presenting problems. Certainly, more family therapy process research (both single-case and large n) is desperately needed on the roles of fathers in family therapy. Despite the prevailing public belief that fathers do not hold as much relational power as do mothers, fathers are an equally important part of the family, and their equal participation in family therapy is a worthy goal. Fathers in Family Therapy 15 ReferencesCohen, J. (1988). Statistical power analysis for the behavioral sciences (2ed.). New York: Academic Press.Fox, G., & Bruce, C. (2001). Conditional fatherhood: Identity theory andparental investment theory as alternative sources of explanation offathering. Journal of Marriage and The Family, 63, 394-403.Goldner, V. (1985). Feminism and family therapy. Family Process, 24, 31-47.Graham, J.M., Conoley, C.W., Craig, M.C., Neu, T., O’Pry, A., Cardin, S.A,,Brossart, D.F., & Parker,R.I. (2001). The Effectiveness of SolutionFocused Family Therapy with Four Aggressive and Oppositional ActingChildren. Paper presented at the annual meeting of the AmericanPsychological Association, San Francisco, CA.Lundervold, D.A., & Belwood, M.F. (2000). The best kept secret in counseling:Single-case (N=1) experimental designs. Journal of Counseling andDevelopment, 78, 92-102.Van Walsum, K. (2001). Where have all the fathers gone? Father involvement in family therapy. Paper presented at the annual meeting of the AmericanPsychological Association, San Francisco, CA. Fathers in Family Therapy 16 Table 1Number of Segments in Which One Parent Talked the Most ____________________________________________________________ Freq.Chi-Asymp.No Yes Square df Sig. wFather 534 38Mother 421 151 114.893 1 .000 .448 Note: Results in bold are statistically significant at the .0102 level. Fathers in Family Therapy 17 Table 2.Coding Category Frequencies by Father and Mother ____________________________________________________________ Freq.Chi-Asymp.No Yes Square df Sig. wTherapist InitiationsTo Father 543 29To Mother 541 31 .136 1 .712 .015Therapist ResponsesTo Father 524 48To Mother 438 134 72.080 1 .000 .355 Parent Responses toTherapistBy Father 374 198By Mother 254 318 101.976 1 .000 .422____________________________________________________________ Note: Results in bold are statistically significant at the .0102 level. Fathers in Family Therapy 18 Table 3Proportion of Therapist-Parent Responses, Given the Number of Parent-Therapist Responses ____________________________________________________________ Therapist Parent % Opp. Chi-Asymp.Responses Responses Taken Square df Sig. wFather 4819824.2%Mother 13431842.1% 25.630 1 .000 .284____________________________________________________________ Note: Results in bold are statistically significant at the .0102 level. Fathers in Family Therapy 19 Table 4Proportion of parent-therapist responses given the number of therapist-parent initiations,responses, or therapist open-ended initiations. __________________________________________ Parent Parent % Opp. Chi-Responses Opp.Taken SquareFather 19820198.5%Mother 25427094.1% * _______________________________________ *Unable to calculate because some cells are <5.