-u h! «MW- z :r-UA . .31 1h“ " 1 m .Iv‘ '1 ' . ‘p‘ 'r~' -- 4‘, r , . “it .-“-I'~\:-ru - n . . ' . k _! _ _ 'J' . .' I' 2 ' ‘ ' .3.” «fl 1 ,‘7 .J y' ' a; V11 '5; ‘ ”A? - . v nv’m’ .., W." J '1 =1" '0 v “J. ,.:. w w} . p ‘L 1; 35m“ ' i'it'fit'viJ .X‘ v-‘ ,u‘ \.A '14. I u f VA '1': I. a «on! '4 y ”a ‘..‘ * ' .4 . lla“(. ’ ‘ MN. M "o {J . m-wt‘ 4 1w ..... ., .Iv- J. Evan‘brgegr ‘1‘ ‘ l‘ "3,4.“ I; . l' . #3. 5“ 51%;? ,. ,J ‘v'wg gr I! :fiu I 5 dc u ,. 9:23? , ‘53? W . q$~, . ‘. j ;‘*::\:‘3{ . *0. . I an} ‘ . 3y 6:“ :1 vi!» 1%? ,A {4269; 5; ma. - “H”. H, ‘. . L, ‘mwv‘. ,4 .4":- 45,61; J . l «V! 4”: fix" {1‘ A". IW I" “ ‘é . F , . .I .3“ FLFW' ' V- uamlES 4 «w “"‘l El‘llllElll M \ \\\\\\\\\\\\\\\ \\\\\\\\\\\\\\\\\ E 3 1293 010 5 444 This is to certify that the thesis entitled Comparing Suicidal and Non-Suicidal Adolescents' Beliefs Concerning Family Structure and Parent-Adolescent Communication presented by Kathleen Zawacki has been accepted towards fulfillment of the requirements for MA degree in Family Studies \ «£an QQL Major professor Date February 28, 1992 0-7639 MS U i: an Affirmative Action/Equal Opportunity Institution LIBRARY Michigan State University PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before due due. DATE DUE DATE DUE DATE DUE l MSU I. An Affirmative ActlorVEquel Opportunity Institution l °W@l COMPARING SUICIDAL AND NON-SUICIDAL ADOLESCENTS’ BELIEFS CONCERNING FAMILY STRUCTURE AND PARENT-ADOLESCENT COMMUNICATION Bv Kathleen Zawacki A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Family and Child Ecology 1992 (9,. )2 /':. relatl comr mem. such to ad. Studie comrr herein fEmily Hilario family iWo gr. adoleSc sUiCidal ’eceivi 0/?7- 425'}? ABSTRACT COMPARING SUICIDAL AND NON-SUICIDAL ADOLESCENTS’ BELIEFS CONCERNING FAMILY STRUCTURE AND PARENT-ADOLESCENT COMMUNICATION By Kathleen Zawacki The literature available concerning family structure characteristics related to adolescent suicide is primarily clinical that indicates poor communication and distant or enmeshed emotional bonding among family members. There are relatively few empirical studies that have investigated such family structure characteristics as cohesion and adaptability in relation to adolescent suicide. In addition, there appear to be no available empirical studies which assess cohesion, adaptability, and parent-adolescent communication together within the same study. The research described herein addresses this omission in the research literature by examining the family structure variables of cohesion, adaptability, and communication in relation to adolescent suicide. The purpose of this study was to examine differences in beliefs about family structure and the relationship of these beliefs to suicidal behavior for two groups of adolescents, suicidal and non-suicidal. Forty suicidal adolescents and 101 non-suicidal adolescents participated in this study. The suicidal adolescents were between the ages of 13 and 18 and were receiving inpatient therapy for deliberately inflicting self-injury and/or obsessed with suicidal ideations. The non-suicidal adolescents were also U SI hr YE COI FAA tha ado the reSe adOII between the ages of 13 and 18 and were selected from three high schools geographically near the institutions where the suicidal adolescents were receiving therapy. The instruments used to measure family structure were two Likert- type scales: (1) the Family Adaptability and Cohesion Evaluation Scale Ill (FACES Ill) and (2) the Parent-Adolescent Communication Inventory (PACI). In addition to these scales, demographic information was obtained about the subjects including age, gender, race, who the adolescents lived with at home, and whether the adolescents seriously considered suicide in the past yeah Results of the study indicated significant differences between the suicidal and non-suicidal groups using one-way ANOVA for both the PACI (Ell, 139) = 12.59, p<.0005) and the FACES Ill (El1, 139) = 3.89, Q<.0509) scales. The suicidal adolescents perceived (1) their family communication, using PACI, as lower and (2) their family structure, using FACES III, as more unbalanced in reference to cohesion and adaptability, than the non-suicidal adolescents. These results suggest that the suicidal adolescents, in general, perceive their families as more dysfunctional than the non-suicidal adolescents. The implications are particularly useful for researchers, counselors, and health care professionals involved with adolescents. Copyflghtby KATHLEEN GLADYS ZAWACKI 1992 Idl UNI STE I'll-IE I dedicate this project to my parents, John and Ruth Zawacki. Without their unconditional love and support, I would not have been able to attend graduate school. coll (1). PSYI Donl Wils. Davir Carls t0 pr: Dr0jel SChIa; encou thank: 8“000. Dr. Sc QUailtie rnYlde; Undersn thrOthi generOs. ACKNOWLEDGMENTS I would like to acknowledge the people who have supported and contributed to the completion of this project. First, I would like to thank the adolescents that participated in this project for sharing a part of their personal lives with me. Second, I would like to thank the staff at the five institutions I collected data from. More specifically, I am grateful to the following people: (1) Jeanne A. Ewing, A.C.S.W. and Kathy Hogg, A.C.S.W. at Rivendell Psychiatric Center; (2) Dr. Kang Kwon and Gail Sadler at McLaren Regional Medical Center's Adolescent Psychiatric Unit; (3) Edward Sampson and Donna Six at Bath High School; (4) Timothy Young, Mary Tucker, Rich Wilson, Don Dexter, and Larry Shoemaker at Mason High School; and (5) Dr. David Oegema, Lynn Chapman, Bob Sheehan, Henry Griffin, and Deborah Carlson at Perry High School. These people volunteered their time and effort to provide for a supportive milieu conducive to completing my research project. Third, I am grateful to the professors on my committee, Dr. Lawrence Schiamberg, Dr. Thomas Luster, and Dr. Lillian Phenice for providing me the encouragement and guidance needed to complete this project. Special thanks to Dr. Schiamberg, my academic advisor, for being patient and supportive throughout my academic pursuits at Michigan State University. Dr. Schiamberg’s gentle nature, diplomacy, and expertise are admirable qualities I respect. Fourth, I wish to thank Suzanne Gainforth for taking the time listen to my ideas and read my thesis. Finally, I thank my best friend, Carmine T. lacono, Esq., for being understanding and providing me with the support to build my confidence throughout my academic pursuits. Mr. lacono’s computer expertise and generosity were instrumental for the completion of this project. vi TABLE OF CONTENTS Page LIST OF TABLES ..................................... ix LIST OF FIGURES ..................................... x CHAPTER ONE -- INTRODUCTION ......................... 1 Introduction/Statement of the Problem .................... 1 Purpose of Research ................................. 4 Research Objectives ................................. 5 CHAPTER TWO -- REVIEW OF LITERATURE ................... 6 Overview of Literature ............................... 6 Clinical Research on Adolescent Suicide .................. 10 Empirical Research on Adolescent Suicide ................. 13 Conceptual Model .................................. 14 Research Questions ................................ 19 CHAPTER THREE -- METHODOLOGY AND PROCEDURES ........ 21 Sample ......................................... 21 Sampling Procedures ................................ 21 Research Design ................................... 23 Conceptual and Operational Definitions ................... 24 Research Assumptions .............................. 28 Research Hypotheses ............................... 29 Instrumentation ................................... 29 Data Collection ................................... 31 Scoring Procedures ................................. 33 CHAPTER FOUR -- RESULTS ............................ 36 Organization of Data Analysis ......................... 36 Demographic Factors ............................... 37 Test of Hypotheses ................................ 44 Additional Analyses ................................ 49 vii LIE AF CHAPTER FIVE -- DISCUSSION OF RESULTS ................. 54 Brief Description of the Study ......................... 54 Demographic Findings ............................... 55 Discussion of Hypotheses ............................ 58 Discussion of Results on Cohesion and Adaptability .......... 65 CHAPTER SIX —- CONCLUSION .......................... 68 Limitations of the Study ............................. 68 Summary ....................................... 69 Implications for Further Research ....................... 71 Conclusions ...................................... 72 LIST OF REFERENCES ................................. 74 APPENDICES ....................................... 79 Appendix A - Copies of Instrumentation and Instructions ....... 79 Appendix B - Parent and Adolescent Consent Forms .......... 92 Appendix C - Letters of Permission ...................... 94 Appendix D - UCRIHS Permission Letter ................. 100 viii LIST OF TABLES Page Table 1: Cutting Points and Norms for Individuals on the FACES Ill Instrument ................................. 34 Table 2: Frequencies for People the Adolescents Lived With at Home for the Total Sample .......................... 37 Table 3: Frequencies for the Ages of the Adolescents for the Total Sample ................................... 38 Table 4: Frequencies for Gender, Race and Suicidality by Group Comparison ................................ 39 Table 5: Frequencies of Who the Adolescents Live With at Home by Group Comparison ................................ 41 Table 6: Frequencies for Age by Group Comparison ........... 42 Table 7: One-way ANOVA Results for FACES Ill .............. 45 Table 8: One-way ANOVA Results for PACI ................. 46 Table 9: Mean Scores on FACES III (DFCS) and PACI for Gender by Group Comparison ............................ 47 Table 10: Mean Scores on FACES Ill (DFCS) and PACI for Race by Group Comparison ............................ 48 Table 11: Mean Scores on FACES Ill (DFCS) and PACI for Age by Group Comparison ............................ 48 Table 12: Alpha Reliabilities for PACI, FACES Ill, and FACES lll Dimensions: Cohesion and Adaptability ............. 50 Table 13: T-test Results for the Cohesion and Adaptability Dimensions on FACES III ................................ 51 Table 14: Correlational Matrix for PACI, FACES lll (DFCS), and the Ages of the Adolescents ....................... 52 LIST OF FIGURES Page Figure 1: The Circumplex Model ......................... 17 Figure 2: Diagram of the Four Extreme Family Types from the Circumplex Model ............................ 59 Figure 3: Norms and Cutting Points for Four Levels of Cohesion and Adaptability ............................. 66 CHAPTER ONE -- INTRODUCTION In r i n m n f h Pr I m TAMPA -- If ninth-grade boys had résumés, Justin "J.D." Daniel and Smith Rice would have compiled some impressive credentials. J.D. was an Eagle Scout who worked with the elderly. Smith, a writer who did a mean imitation of rock star Jim Morrison, was considered brilliant by his classmates. The boys, both in their 15th year, had some things in common: Both grew up in affluent, educated families. They attended the same private school. In November, J.D. committed suicide with a gun taken from a family gun cabinet. On Wednesday, so did Smith (Rosen, 1992, p. 18). Adolescent suicide in the United States has increased significantly. Between 1960 and 1980, the suicide rate for the age group between 15 and 24 rose from 5.2 to 12.3 per 100,000, an increase of 136 percent (U. S. Congress, 1984a). As alarming as this statistic may sound, it does not completely reflect the total problem in suicidal behavior. Many researchers believe that there are an estimated 50 to 150 attempts for every successful suicide (McIntire et al., 1977), and even more suicidal ideations. Adolescent suicide is currently the second leading cause of death for the age group between 15 and 24, second only to accidental deaths (Neiger & Hopkins, 1988). Due to these disturbing statistics, researchers and professionals in the helping professions have focused their attention on the problem of adolescent suicide. A wealth of information is available on factors associated with suicide. Studies concerning adolescent suicide have utilized hospitals (Garfinkel, Froese & Hood, 1982). psychiatric institutions 1 2 (Cohen-Sandler, Berman & King, 1982), private practices of counselors and therapists (Pfeffer, 1981), and schools (Nelson, Farberow & Litman, 1988) to analyze why adolescence is such a difficult period in life, given the prosperity of this country. The literature in this area has demonstrated that there are differences in suicidal behavior based on age (Leigh, 1986, p. 67), sex (Schiamberg, 1988, p. 736), and race (Schiamberg, 1988, p. 737). The suicide rate is higher for adolescents in the 15-24 year range (2,286) than it is for adolescents in the 5-14 year range (120) (Vital Statistics of the U.S., 1987). More females than males attempt suicide; however, more males than females succeed in committing suicide (Schiamberg, 1988, p. 736). Also, more white males (141) and females (32) in Michigan committed suicide than black males (24) and females (3) for the year 1988 (Vital Statistics of the U.S., 1988). A recent nationwide study by the Centers For Disease Control (CDC) confirms these earlier findings on racial and gender differences in relation to suicidal behavior (1991). In this CDC study conducted with 11,631 students between the 9th and 12th grade, female students were significantly more likely than male students to report suicidal ideations and/or suicide attempts. Racial differences in response to suicidal ideations and/or suicide attempts for the students in this CDC study were as follows: Ranging from highest to lowest percentages of students reporting suicidal ideations and/or suicide attempts by race were (1) Hispanics (30.4% for suicide ideation and 12.0% for suicide attempts), (2) Whites (28.1% for 3 suicide ideation and 7.9% for suicide attempts), and (3) Blacks (20.4% for suicide ideation and 6.5% for suicide attempts). This study and others reaffirm the wide variety of factors and demographic variables related to the onset of suicide and suicidal behavior. Two familial factors associated with adolescent suicidal behavior are (1) having too low or too high cohesive parent-child relationships (Hepworth, Farley & Griffiths, 1988) and (2) having inflexible or unyielding adaptability to changes in the family (Pfeffer, 1981). Hepworth, Farley, and Griffiths (1988) stated the following about families with suicidal adolescents: Cohesiveness between parents and children is typically low in families of suicidal adolescents. As adolescents strive for autonomy, however, the positive emotional connections between adolescents and their parents tend to erode even further in these families. Thus communication between suicidal adolescents and their parents is generally poor and conflict is high (p. 198). However, Pfeffer (1981) found that unusually close or symbiotic emotional attachments between parents and adolescents were associated with adolescent suicide. Thus, for the familial factor of cohesiveness, unusually low or high cohesiveness is related to adolescent suicidal behavior. Unusually low cohesiveness is referred to as "disengaged” whereas unusually high cohesiveness is referred to as "enmeshed". For the familial factor of adaptability, only unusually low adaptability is related to adolescent suicidal behavior (Pfeffer, 1981; Richman, 1979). Unusually low adaptability is referred to as ”rigid". Olson, Sprenkle, and Russell (1979) devised the Circumplex Model th ad Pal QFOL Teiat 4 attempting to standardize the definitions for the two familial factors of cohesion and adaptability (see Figure 1 in Chapter 2). Within the model, adaptability and cohesion are the two major dimensions while communication serves as the model’s facilitative dimension. Based on this model, cohesion and adaptability are measured using the Family Adaptability and Cohesion Evaluation Scale Ill (FACES Ill) (Olson, Portner, and Lavee, 1983). For purposes of this study, family structure encompasses the dimensions discussed in the Circumplex Model. Although the Circumplex Model alludes to communication as an important facilitative dimension, the FACES III instrument, which measures the dimensions of cohesion and adaptability, does not measure the adolescents' perceptions of family communication. To measure this important dimension in the family for this study, the investigator used the Parent-Adolescent Communication Inventory (PACI) (Bienvenu, 1969). r r h This study investigated the differences in beliefs that two adolescent groups, suicidal and non-suicidal, had about family structure and the relationship of these beliefs to suicidal behavior. The study investigated how adolescents' perceptions or beliefs about the family structure and parent-adolescent communication, whether they were realistic or not, related to their decision to engage in suicidal behavior. Empirically researching family structural factors related to adolescent suicide is necessary because the family plays such a crucial role in the life of an adolescent who is 5 striving for independence. Additionally, how the adolescent acts and behaves, affects the family. ”By identifying the structural and interactional aspects, it becomes clearer how parents, siblings, and the adolescent also participate in the development of self-destructive behaviors" (Leigh, 1986, p. 68). Thinking in terms of each family member acting and affecting the whole family is a perspective of systems thinking. R r h ' iv As stated previously, this study investigated adolescents' perceptions or beliefs about family structure and parent-adolescent communication and whether these beliefs were related to suicidal behavior. It was assumed that the adolescents’ perceptions about their family life, whether they were realistic or not, were significantly related to their own actions (i.e., engaging in suicidal behavior or not engaging in suicidal behavior). Therefore, the objective of this research study was to empirically differentiate between suicidal and non-suicidal adolescents based on their beliefs about or perceptions of family structure and parent-adolescent communication. Distinguishing between these groups empirically may contribute to the understanding of why adolescents engage in suicidal behavior. Such an understanding increases the likelihood of developing effective preventive measures for decelerating the rising rate of adolescent suicide. Thus, by empirically researching and gaining a better understanding of the family system variables operating within suicidal adolescent families, clinical work with these families will be enhanced. 51. fat est exp funt MCK the f ahera establi: faWMy. Separan goo CHAPTER TWO -- REVIEW OF LITERATURE v rvi f Li r r Although there are many factors attributed to the onset of adolescent suicide, family factors are at the core of this social epidemic. Of all the factors attributed to adolescent suicide, including personal factors (low self- esteem and depression) and societal factors (stress related to societal expectations) reviewed by Wodarski and Harris (1987), the structure and function of the family is the most prevalent theme in the literature. McKenry, Tishler, and Kelley (1982) have stated the following about how the familial problems contribute to the onset of adolescent suicide: Certainly, it is the family milieu that, prior to adolescence, provides the widest range of experience and involvement for the child. In fact, descriptive and clinical accounts of adolescent suicide indicate a family background of marital instability, self-destructive behavior, and long and bitter parent- child conflict (p. 267). The family is normally a ”safe haven" for the adolescent to return to after adventuring beyond the boundaries of home. An adolescent establishes an identity by both separating from and connecting with the family. Cooper, Grotevant, and Condon (1983) stated the following about separateness and connectedness for the adolescent in a family: Our findings suggest that the leaving process is facilitated by individuated family relationships, characterized by separateness, which gives the adolescent permission to develop his or her own point of view, in the context of connectedness, which provides a secure base from which the adolescent can explore worlds outside the family (p. 56). 7 An adolescent exploring ”worlds" outside the boundaries of the family still desires to feel as if he/she is a part of the family. Spector (1988) describes the adolescent’s desire of wanting an invitation to be a part of family gatherings, yet being given the freedom to do activities outside the family as a symbol of a secure psycho-social standing within the family. For the adolescent, this paradoxical interplay of separating and connecting with the family is an important developmental process. Additionally, while parents may look forward to even their own autonomy as their children move toward adulthood, the process is one that is very painful and extremely ungratifying because the adolescent's independence is usually gained through simultaneously vacillating between pushing for freedom and pulling for comfort and caregiving. It is the inability of parents and adolescents both to accept change and transition and to coordinate this dance of separation and connection that places families at risk of more serious problematic behaviors (Oster & Caro, 1990, pp. 104-105). This developmental stress that occurs in families where both the parents and adolescents are continually redefining themselves is a difficult transition. Additional disruptions (e.g., severe family conflict or any other situational stress) in this normal adolescent developmental process can be devastating. Because of these family dynamics, it becomes apparent that the components or structure of the family is a logical area to analyze when performing empirical research on adolescent suicide. This study investigated how the structure and communication within the family related to adolescents' decisions to engage in suicidal behavior. Research on the structure of and communication within the family containing suicidal adolescents is primarily clinical. However, before 8 discussing clinical research on adolescent suicide, a brief explanation describing why the systems perspective is used in analyzing adolescent suicidal behavior will be addressed. The systems perspective integrates studying the individual and studying the social environment that the individual operates in. The family is a logical unit for which the systems perspective can be applied. Early social science philosophers, Durkheim and Freud, sanctioned the importance of linking the family with an individual's self-destruction. Durkheim pointed to the necessity of a cohesive family structure and the enactment of certain family roles in order that these moral demands be communicated effectively to the individual. Freud stressed the importance and presence of family members, particularly parents, in the formation of a healthy superego. Both theorists regarded the healthy family as playing an important preventative role with regard to self- destruction (Heillig, 1983, p. 2). More recently, Simon, Stierlin and Wynne (1985) have defined ”system" as the following: The most general definition of system [Greek systema, a composite thing] is the ordered composition (material or mental) elements into a unified whole. . . . The premises of systems theory are based on the insight that a system as a whole is qualitatively different, and ’behaves’ differently, from the sum of the system's individual elements (p. 353). In other words, the family is seen as a system where (1) the whole family is considered greater than the sum of each family member and (2) each family member's actions, whether they are positive or negative, affects the family as a whole. Therefore, to study adolescent families according to the systems perspective, attempts at understanding adolescent behaviors need 9 to be made by looking at the family environment in terms of family structures, interactions, and relationships (Leigh, 1986, p. 39). A recent gallup poll confirms the notion that family conflict is a major factor related to adolescent suicidal behavior (The Gallup Organization, March 1991). This study randomly selected 1,500 adolescents via the telephone and then mailed questionnaires to these consenting adolescents. The response rate was 77% or 1,152 adolescents. The ages of the respondents were between 13 and 19 with 51% being male and 49% being female. Ninety percent were enrolled in public schools. The purpose of the study was to solicit teenagers' attitudes and opinions about suicide based on their own experiences they have had with suicide. Of the total sample, 15% reported coming close to attempting suicide (ideators) and 6% reported that they had attempted suicide. From these ideators and attempters, the six commonly cited factors reported to have led up to their ideations or attempts included the following: -| . Family problems/problems at home (47% or 7% of total sample). 2. Depression (23% or 4% of total). 3. Problems with friends/peer pressure/social relations (22% or 3% of total). 4. Low self-esteem Heeling worthless (18% or 3% of total). 5. Boy/girl relationships (16% or 3% of total). 6. Felt like no one cared (13% or 2% of total) (I). 7). Therefore from the results of this study, it is easy to see why studying the structure, interactions, and relationships in families with suicidal adolescents 10 is so important. i ' h n l n i i Clinical research on adolescent suicide has largely focused on the structure and communication within families. For example, Pfeffer (1981) attempts to explain the onset of suicidal behavior in latency-age children from a systems' perspective. Based upon her many years of treating and studying suicidal and non-suicidal latency-age children, Pfeffer (1981) ". . . presents a formulation of the organizational characteristics of the family system of psychiatrically hospitalized suicidal latency-age children” (p. 330). There are five hypotheses for which Pfeffer defines the features in a suicidal family. They are (1) lack of generational boundaries, (2) severely conflicted spouse relationships, (3) parental feelings projected onto the child, (4) symbiotic parent-child relationships, and (5) inflexible family systems. When these five family characteristics are present in a family system, Pfeffer (1981) argues that there is risk for suicidal behavior. Thus, Pfeffer views adolescent suicidal behavior as a family problem in terms of its structure, interaction, and relationships. In a similar way, Molin (1986) delineated two forms of family dynamics predominantly associated with families of suicidal adolescents. They are the ”symbiotic” and the ”excluded" family relationships. The symbiotic family relationship is a pathological emotional closeness between parent and adolescent, ". . . where the parent has come to rely on the adolescent for a variety of needs, including that of parenting, the onset of 11 adolescence brings with it a distinct threat" (Molin, 1986, p. 179). This is similar to Pfeffer's (1981) hypothesis of the ”symbiotic parent-child relationship”. The excluded family relationship concerns the family that wishes to exclude the adolescent. In this dynamic, the members in the family are so distant and separate from one another that they seek people outside the family with whom to bond emotionally for support. Excluding or alienating the adolescent is usually a result of the family reconstituting or recombining (e.g., a parent remarries and alienates the adolescent in some way from the new family system) (Molin, 1986). This alienation is a defense mechanism the family uses to keep the ”emotional baggage” of the first marriage separate from the new marriage. Thus, for these two dynamics, the structure of the family is a key factor in why an adolescent opts to engage in suicidal behavior. In a study of 100 families, Richman (1979) observed seven characteristics that were present for families with suicidal adolescents. They were (1) an inability to accept necessary change, (2) a disturbed family structure, (3) a depressed family affect, (4) unbalanced intrafamilial relationships, (5) communication problems, (6) double-bind relationships, and (7) an intolerance for crises. Brown (1985) also found some key characteristics in families with suicidal adolescents. They included a lack of (1) clear and overt communication, (2) pleasureful, affectionate contact among family members, (3) mutual familial validation, (4) clear and strong problem-solving skills, (5) clearly defined intergenerational boundaries, and (6) task organization within the family. Leder (1987) also found that there is 12 an association between adolescent suicidal behavior and (1) parental loss, (2) family disorganization due to blocked communication, and (3) divorce. Additional studies report findings similar to those of Leder (1987) (Cohen- Sandler, Berman and King, 1982; Johnson and Maile 1987; Hendlin 1987; and Asarnow and Carlson, 1988). Research confirms the notion that a dysfunctional family system is paramount to adolescent suicidal behavior. Suicide is the result of, "the end product of progressive isolation from meaningful relationships” (Greenlee, 1987, p. 3). Teicher and Jacobs, two researchers interested in how adolescents attempt suicide, developed a three-stage process explaining how adolescents are lead to attempt suicide. The first stage, predisposing factors, consists of a long- standing history of problems from Childhood to adolescence. The second stage is a period of escalation of problems related to adolescence. The final stage occurs during a period of weeks or days that immediately precede the suicide attempt. Common precipitating events of this phase include a rapid breakdown of the adolescent's social supports, including contacts and associations with peers, friends, and family (Hepworth, Farley & Griffiths, 1988, p. 195). Thus, from this three-stage process, adolescents are pathologically attempting to maintain homeostasis within their lives by attempting suicide because of the history of problems that have not been successfully adapted to. Suicidal behavior becomes a form of adaptation for the adolescent. Since the family has been the major influential resource throughout a child's life until the teenage years, it would only seem logical to investigate empirically the structure and communication patterns within the family. 2.. ”Fa. .F. .u-wb‘abv. . 13 Emiri IR rhnAl n ii In a recent study of 76 adolescents hospitalized for suicide attempts, Spirito, Overholser and Stark (1989) found that suicidal adolescents experienced more family conflict with their parents than non-suicidal adolescents. This is consistent with the notion that the structure and communication patterns of suicidal adolescents are more dysfunctional than non-suicidal adolescents. Likewise, Stivers (1987) found that family discord is a precursor to adolescent suicide. In another study, Stivers (1988) concluded that adolescents' low PACI scores (parent-adolescent communication was measured using the PACI) significantly contributed to adolescent depression associated with suicide. These findings point toward family structure and communication as key factors related to adolescent suicidal behavior. In another empirical study using FACES III, Livingston (1989) found that suicidal adolescents' perceptions of family structure and roles within the family were more dysfunctional than non-suicidal adolescents’ perceptions. In the study, Livingston utilized an integrated score of family structure, the DFCS (Distance From Center Score), which is a combined score from responses on a survey measuring family cohesion and adaptability. Spector (1988), in an earlier study using FACES ll, used the raw scores for cohesion and adaptability. Spector found that suicidal adolescents perceived their family cohesion to be unusually low; thus, the suicidal adolescents' perceptions of the emotional bonding in the family was low and infrequent, if at all. Suicidal adolescents in Spector's study perceived their family 14 adaptability to be in the middle on a continuum from rigid to chaotic. Thus, these two studies from Livingston and Spector have found using the FACES instrument that suicidal adolescents perceive their family structure as more dysfunctional than non-suicidal adolescents. In summary, suicidal adolescents perceive their family communication to be ambivalent and conflictual. They perceive their family life as one ambiguous conflict after another without any structure for organizing their future. While it is true that adolescents seek many types of communication without success, suicidal behavior becomes the ultimate form, and last resort, of communication. Husain (1990) convincingly stated, "A poignant interpretation of suicidal behavior in multiple family members is that suicide is a ’Ianguage' that is understood, whereas other methods of communication fail” (p. 123). Thus, dysfunctional family structures and communication are important dimensions of predicting adolescent suicidal behavior. n e lM I Understanding that the family is a system was crucial in providing a conceptual foundation for this project. Andrews, Bubolz, and Paolucci (1980) stated the following about families: "In the family ecosystem, the environed unit is the group of persons who constitute the family, defined as a bonded unit of interacting and interdependent persons who have some common goals and resources, and for part of their life cycle, at least, share living space" (p. 32). Thus, family systems are assessed by analyzing family structures, interactions, and relationships. 15 This study described how suicidal and non-suicidal adolescents perceived their family structure and communication using FACES Ill (and PACI) which is derived from the Circumplex Model. The Circumplex Model is defined using cohesion, adaptability, and communication as key dimensions within the model. ”In this model, families are assessed on the dimensions of -> cohesion -> adaptability (change), and communication" (Simon, Stierlin, & Wynne, 1985). Olson, Sprenkle, and Russell (1979) developed the Circumplex Model ”as a tool for clinical diagnosis and for specifying treatment goals [in family therapy] with couples and families" (p. 3). Cohesion, adaptability, and communication were key concepts consistently found in the literature on family theory and therapy. Thus, their purpose for this model was to integrate diverse theoretical and therapeutic concepts of the family. The Circumplex Model utilizes the two central dimensions of family cohesion and family adaptability. Family cohesion is defined as "the emotional bonding members have with one another and the degree of individual autonomy a person experiences in the family system” (Olson, Sprenkle, & Russell, 1979, p. 5). Family cohesion is viewed on a continuum from high bonding and low individual autonomy, referred to as enmeshment, to extreme separateness, with low bonding and high individual autonomy, referred to as disengagement. These extreme forms of cohesion are considered dysfunctional for the family. Family adaptability is defined as, "the ability of a marital/family system to change its power structure, role relationships, and relationship 16 rules in response to situational and developmental stress" (Olson, Sprenkle, and Russell, 1979, p. 12). The family adaptability dimension assesses the family systems' ability to reorganize in times of stress produced by certain situations (i.e., a death of a family member) or stress produced by developmental changes in the family (i.e., a child beginning the adolescent years while the parents face a midlife transition). Family adaptability is also viewed on a continuum. At one end of the continuum is extreme rigidity, where the family ”sees" any change as a threat to survival. At the other end of the continuum is extreme chaos, where the family reorganizes too much in response to stress and does not appear to have any stability. These two dimensions, cohesion and adaptability, are drawn as two intersecting lines displaying the continuum from low to high for both dimensions. Families who are moderate in response to cohesion and adaptability fall in the center of the two intersecting lines. More dysfunctional families fall more toward the outer boundaries of the intersecting lines. Figure 1 shows a graphical representation of the Circumplex Model. From this figure, cohesion and adaptability are broken down into four separate levels, giving the model 16 different types of family systems. From the 16 family types, families are then categorized into one of three global types of families. The first type is the Balanced families. These families fall into the two central cells of cohesion and adaptability. The balanced families are considered the most functional. The second type is the Mid- range families. These families fall into an extreme cell on one dimension and 17 Figure 1: The Circumplex Model (from OS. Russell, 1979). CIRCUMFLEX MODEL or mums mum! 3mm e—Inw coneslou HIgI'I —e _ DISENGABED SEPARATED CONNECTED ENIESHED FACES III ,0 2, 35 mucmneo ‘. p... w sw'mm" .Wmdu...’ «— Lou —— ADAPTABILITY —IIIgII——e Clinical Rating Scale Em [mm-(mama: -axnm hWMMthhmhmmMmgflcb—dflumm mam-mum Fanny Sock! Science . M an an. "l0 [5“ 290 "cue-I Hell. Unlnnfly of Minnesota St. Paul. MN 55108 a central cell for the other dimension. These families may or may not function adequately and are "on the edge" of functioning capability. The third type is the Extreme families. The extreme families fall into the outer L... 18 cells on both dimensions (Olson, McCubbin, Barnes, Larsen, Muxen, & Wilson, 1989). These extreme families are considered dysfunctional and it is hypothesized that most families with suicidal adolescents will be categorized in this group. The third dimension mentioned in the Circumplex Model is communication. This is considered the facilitative dimension because it facilitates movement along the cohesion and adaptability dimensions. Communication is crucial to the model becauSe most changes occur through communication within the family. In general, positive communication skills are seen as helping marital and family systems facilitate and maintain a balance on the two dimensions. Conversely, negative communication skills prevent and minimize marital and family systems from moving into the central areas and thereby, increase the probability that extreme systems will remain extreme. Positive communication skills include sending clear and congruent messages, empathy, supportive statements, and effective problem-solving skills. Conversely, negative communication skills include nonsupportive (negative) statements, poor problem-solving skills, and paradoxical and double-binding messages (Olson et al., 1989, p. 68). Although the Circumplex Model mentions communication as a facilitative dimension and gives it importance within the model, an instrument for measuring communication within this model is lacking. FACES lIl measures the two dimensions of adaptability and cohesion within the Circumplex Model. Thus, this study utilized the concepts and the instrument, FACES lll, drawn from the Circumplex Model to analyze suicidal and non-suicidal adolescents’ beliefs on family structure. However, for the facilitative dimension of communication within the 19 Circumplex Model, there does not exist a way of measuring communication. For purposes of this study, PACI was used to measure beliefs on family communication for suicidal and non-suicidal adolescents. Suicidal and non- suicidal adolescents’ beliefs about family communication were analyzed. It was hypothesized that (1) negative communication patterns contained more dysfunction in the family, while (2) positive communication patterns contained less dysfunction in the family. Using the Circumplex Model, this study proposed that suicidal adolescent beliefs about their family structure would fall in the extreme family type, while the non-suicidal adolescent beliefs about their family structure would fall more toward balanced to mid-range family types. This study also proposed to show that suicidal adolescent beliefs about their family communication would be more negative, while the non-suicidal adolescent beliefs about their family communication would be more positive. R r h i II To accomplish the primary research objective of differentiating between suicidal and non-suicidal adolescents' beliefs on the structure and communication within the family, specific research questions were addressed. 1) Are suicidal and non-suicidal adolescents' beliefs about family structure different? 2) Are suicidal and non-suicidal adolescents’ beliefs about family communication different? 20 3) If suicidal and non-suicidal adolescents’ beliefs about family structure are different, then are the beliefs of suicidal adolescents more dysfunctional than the beliefs of non-suicidal adolescents? 4) If suicidal and non-suicidal adolescents’ beliefs about family communication are different, then are the beliefs of suicidal adolescents more negative than the beliefs of non-suicidal adolescents? CHAPTER THREE -- METHODOLOGY AND PROCEDURE 53mm The two adolescent groups that this study utilized in measuring family structure and parent-adolescent communication were as follows: (1) Experimental Group (Suicidal Adolescent Group)--consisting of 40 adolescent clients affiliated with Rivendell Psychiatric Center (located in St. Johns, Michigan) and McLaren Regional Medical Center (located in Flint, Michigan). These adolescents were inpatient clients receiving treatment for engaging in suicidal behavior. The ages of these adolescents were between 13 and 17; and (2) Comparison or Control Group (Non-suicidal Adolescent Group)-- consisting of 101 selected high school students from Bath High School, Mason High School, and Perry High School. The high schools were in towns near the Lansing area. The students were between the 9th and 12th grades. Sampling Prggeggres Due to the sensitive nature of the research, choices of where to collect data from were limited. Prior to the selection of adolescents to participate in this study, approval from all the institutions the investigator was interested in collecting data from had to be obtained. The following is a list of area high schools the investigator was declined permission to collect data from for the non-suicidal comparison group in this project: (1) Lansing 21 22 School District, (2) DeWitt Public Schools, (3) Grand Ledge Public Schools, (4) Holt Public Schools, (5) Haslett Public Schools, (6) Leslie School District, (7) Owosso School District, (8) Webberville School District, (9) St. John's School District, (10) Portland School District, and (11) Jackson School District. The three high schools that did grant permission for the investigator to collect data from for the non-suicidal comparison group were (1) Bath High School, (2) Mason High School, and (3) Perry High School (see Appendix C for letters of permission). The following is a list of psychiatric institutions the investigator was declined permission to collect data from for the suicidal experimental group in this project: (1 ) Pine Rest Christian Hospital, (2) St. Lawrence/Dimondale (closed the adolescent unit), (3) Mercywood Neuropsychiatric Hospital, (4) Foote Hospital, and (5) Forest View Psychiatric Hospital. The two psychiatric institutions that did grant permission for the investigator to collect data from for the suicidal experimental group were (1) Rivendell Psychiatric Center and (2) McLaren Regional Medical Center (see Appendix C for letters of permission). Along with receiving permission from the institutions the investigator collected data from, permission was required through the University Committee on Research Involving Human Subjects (UCRIHS) at Michigan State University. Approval was granted on August 5, 1991 (see Appendix D for UCRIHS permission letter). Thus, the locations of where the data was collected was a direct result of the institutions that granted permission for the investigator to perform the research. There were a total of 40 suicidal adolescents and 101 non-suicidal 23 adolescents who participated in the study. The only criterion for selection of suicidal and non-suicidal adolescents was that parental and adolescent consent forms had to be signed prior to administering the questionnaires (see Appendix B for copies of Parental and Adolescent Consent Forms). Thus, any form of random or stratified sampling did not occur as a result of the sensitive nature of this research. Demographic questions inquiring as to the adolescents age, sex, and race were completed by the adolescents. Additional questions the investigator asked on this questionnaire were (1) who the adolescents were living with at home and (2) if the adolescents had seriously considered committing suicide in the past year (see Appendix A). These questions were the only questions the researcher asked the adolescents to fill out other than the instruments mentioned below. R§§g§r§h Dg§ign The research design of this study was descriptive and of a component type. This design was chosen to get a better understanding of both suicidal and non-suicidal adolescents’ perceptions of their family. Descriptive (nonhistorical, nonexperimental) research has as its goal the investigation of the characteristics of a given ”as is" universe or sample of interest. The researcher seeks such measures as parameter values, distributions of attributes, differences between groups, associations between variables, and communality or clustering of variables. Descriptive studies are rarely definitive, but they often exhibit important patterns or relationships (Grosof & Sardy, 1985, p. 109). Being descriptive and cross-sectional, this study compared suicidal and non- suicidal adolescents' beliefs about the family through an analysis of family 24 structure. This research is classified as a component type because it studied only one family member’s beliefs on family structure, the adolescent's beliefs. The adolescent is the member or component of the family system. ”Component analyses seek to solve problems of how components of a system work individually and interactively” (Grosof & Sardy, 1985, p. 118). Thus, the adolescent’s perceptions of family structure, realistic or not, will delineate how the adolescent sees herself/himself operating in the family, both individually and interacting with othertfamily members. n In rinlDfiniin Below is a list of the major variables and definitions that were used in this research project. For this study the independent variables were considered the two separate groups of adolescents, suicidal and non-suicidal adolescents, because they were the given variables. The dependent variables were the scores obtained on the instruments because these scores were a direct result of which group the adolescents were in. Theoretically, the researcher assumes that family dysfunction and poor parent-adolescent communication leads to adolescent suicidal behaviors. Independent Variables i i l A I an Conceptual: Persons between the ages of 13-18 who deliberately inflicted self-injury and/or were obsessed with suicidal ideations (thoughts) with a suicidal intent as reported by the subject. Operational: Persons between the ages of 13-18 who were 25 hospitalized and receiving inpatient therapy for deliberately inflicting self- injury and/or obsessed with suicidal ideations. The adolescents chosen for this experimental group had either (1) inpatient records that the investigator could access in order to select adolescents according to the above requirements or (2) the supervisor for the adolescent unit gave adolescent referrals to the investigator according to the above requirements Non-suicidal Adglesgents Conceptual: Persons between the ages of 13-18 who were not presently receiving inpatient therapy or hospitalized for engaging in suicidal behavior. Operational: Persons between the ages of 13-18 who were not presently receiving inpatient therapy or hospitalized for engaging in suicidal behavior. The adolescents chosen for this comparison group were high school students between 9th and 12th grade. Dependent Variables Family Strugtgre (Scores) Conceptual: The conceptual definition of Family Structure contains three important components: cohesion, adaptability, and communication. The degree to which individuals within a family are separated from or connected to his or her family system is cohesion, the first component. The extent to which the family system is flexible and able to change in response to stress is adaptability, the second component. The third component, 26 family communication is defined below. Operational: Family structure was measured using Olson, Portner, and Lavee's (1983) Family Adaptability and Cohesion Evaluation Scale lll (FACES Ill). The two components of Family Structure, Cohesion and Adaptability, are measured using FACES III, which is a 20-item Likert scale with a 5-point response format. Cohesion is measured by assessing beliefs on family boundaries, coalitions, emotional bonding, time, space, friends, decision- making, interests and recreation. Adaptability is measured by assessing beliefs on family power, negotiation style, role relationships and relationship rules. Odd items are summed to arrive at a Cohesion Score and even items are summed to arrive at an Adaptability Score. These scores are then substituted into a formula called the Distance From Center Score (DFCS) = ((Cohesion score - 39.8)2 + (Adaptability score - 24.1 )2)"2 (see Scoring Procedures in this chapter for further details). The larger the DFCS score, the more dysfunctional the family type is perceived by adolescents (Livingston, 1989). Based on the answers of FACES lIl, families are classified into three general types: balanced, mid-range and extreme (Olson etaL,1989L Family Qammgniaatian (Scores) Conceptual: Family Communication is the third component of family structure. It is a facilitative dimension, meaning that communication within the family facilitates movement along the two central dimensions of cohesion and adaptability. Positive communication skills (i.e., empathy, reflective listening, 27 supportive comments) enable couples and families to share with each other their changing needs and preferences as they relate to cohesion and adaptability. Negative communication skills (i.e., double messages, double binds, criticism) minimize the ability of a couple or family members to share their feelings and thereby, restrict their movement on these dimensions (Olson et aL,1989L Operational: Family Communication is a facilitative component of family structure and is not operationally defined using FACES Ill. Along with measuring the structure within the family using FACES I", this study measured adolescents’ beliefs on communication with their parents. Communication in this respect was measured using Bienvenu's (1979) Parent-Adolescent Communication Inventory (PACI). Thus, PACI is used to operationalize communication. The PACI is a three-point Likert scale with 40 items designed to assess levels of communication between parents and their adolescent children. Responses are scored from one to three, with higher scores corresponding to a more favorable or positive parent- adolescent communication. (Touliatos, Perlmutter, & Straus, 1990). “Communication is defined as the process of transmitting feelings, attitudes, facts, beliefs, and ideas” (Touliatos, Perlmutter, & Straus, 1990). Other Definitions Eam_ily - An organized household of persons who interact, interrelate and are interdependent of one another to function as a system. For purposes of this study, each household of persons must have an adolescent and at least one parent living in the same household. Eunatianal Family - Being classified, according to FACES III, as either 28 Balanced or Mid-range in family type and scoring moderate to high on family communication as measured by the PACI. anfgnatianal Family - Being classified, according to FACES III, as Extreme in family type and scoring low to moderate on family communication as measured by the PACI. r A m i n Below is a list of assumptions for this thesis. 1. It was assumed that the experimental group of adolescents who were receiving therapy for suicidal behavior were genuinely suicidal (However, 40% of the suicidal adolescents denied being seriously suicidal in the last yead. 2. It was assumed that the comparison group of adolescents who were selected from the high schools were not suicidal (However, 30.7% of the non-suicidal adolescents reported having seriously considered suicide in the last year). 3. It was assumed that engaging in suicidal behavior was not "normal” for the general population of adolescents in the United States. 4. It was assumed that adolescents’ reports on the FACES Ill and PACI instruments were genuine and that fabricating did not confound the results of the study. 29 Reaaarah Hypathaaaa The following are the list of research hypotheses for this study. A p < .05 was required to reject all null hypotheses. Hal: Adolescents from the suicidal experimental group will report the same scores as the non-suicidal comparison group on family structure, using the DFCS from FACES lll. Holl: Adolescents from the suicidal experimental group will report the same scores on family communication as the non-suicidal comparison group, using the PACI. H,l: Adolescents from the suicidal experimental group will report more dysfunction than the adolescents from the non-suicidal comparison group in relation to beliefs on family structure, using the DFCS from FACES lll. H1Il: Adolescents from the suicidal experimental group will report lower scores on family communication than the non-suicidal comparison group, using the PACI. n r m n i n Two instruments were used for this study: Family Adaptability and Cohesion Evaluation Scale Ill and the Parent-Adolescent Communication Inventory (see Appendix A for copies of the instruments). Eamily Agamaaility ang Qahaaian Saala Ill (FACES lll) - FACES III is used as an instrument to assess family cohesion and adaptability. Olson, Portner, and Lavee (1983) developed this instrument as a way of measuring the two major dimensions within the Circumplex Model. This instrument was 30 originally designed to assess family functioning levels in therapy; however, it has also been used for empirical research (Livingston, 1989). Individuals, couples, or families may be administered the test. FACES III is a third version of FACES I and II and was developed to improve reliability, validity, and clinical utility. The correlations of the two dimensions of cohesion and adaptability with each other is .03. Cronbach’s alpha is reported to be .77 for cohesion and .62 for adaptability, and .68 for the total scale (Touliatos, Perlmutter, & Straus, 1990). ”Test-retest reliability is reported at .83 for cohesion and .80 for adaptability. Content validity is rated by Olson as very good" (Livingston, 1989, p.47). Results on the scores from FACES lll allow families to be classified into one of three groups: balanced, mid-range and extreme. As stated previously, the most healthy functioning families are the balanced type, while the most dysfunctional families are the extreme type. These classifications are derived from the Distance From Center Score (DFCS) formula (see scoring procedures for further explanation). Earant-Adalaaaant Qammgniaatign lnvantary (PACI) - The measurement of the third component in the Circumplex Model of family functioning, communication, is missing in FACES llI. Thus, Bienvenu's (1979) PACI is used to measure family communication in this study. PACI was originally developed in 1968 and then revised in 1979. Validity information was obtained from various populations: normal, delinquent, high-achieving, and low-achieving adolescents. Test-retest reliability for three weeks is .78 and two-weeks is .88. Cronbach’s alpha is reported to be .93. PACI can be 31 used for both groups and individuals by counselors, educators, and researchers. The items on the 40-item test "refer to various styles and characteristics of parent-adolescent communication, such as listening habits, self-expression, understanding, acceptance, criticism, sarcasm, and trust” (Touliatos, Perlmutter, & Straus, 1990, p. 66). D lle i n Pr r All the data were collected personally by the investigator of the study. The dates for data collection for the non-suicidal comparison group were as follows: (1) Bath High School on September 18, 1991: (2) Mason High School on September 20, 1991; and (3) Perry High School on September 30, 1991. The adolescents chosen from Bath High School were students taking a psychology course. The students chosen from Mason High School and Perry High School were from 1 required course occurring at each grade level, 9th through 12th grade. The parental consent forms were handed out at least 4 days prior to administering the questionnaires. The students were given a brief description of the research, handed out parental consent forms, and asked to return them to their teacher prior to the date the questions were administered. At the scheduled times and dates of administering the questionnaires, the adolescents were then handed out consent forms to be signed themselves. After signing the consents, each student was given a "Questionnaire Packet” containing all the instruments and instructions (see Appendix A), a pencil, and a computer scan-tron sheet to mark the answers to the questions on. After each class completed their packet, questions 32 were answered and the data collection was completed. All measures of confidentiality were followed. Initial data collection for the suicidal adolescents began on August 22, 1991 and terminated on January 9, 1992. The procedures for data collection at Rivendell Psychiatric Center and McLaren Regional Medical Center occurred differently. At Rivendell Psychiatric Center, the person who performed the intake interviews with the adolescents entering Rivendell Psychiatric Center presented the consent forms to the adolescents' parents. The unit supervisor then collected the signed parental consent forms and gave them to the investigator upon arrival to the unit. The investigator then selected the subjects for the study according to the signed parental consent forms. The adolescents were given an explanation of the research, asked to participate voluntarily, and then handed the adolescent consent forms to be signed. After the adolescent agreed to participate in the study by signing the consent form, the adolescent was then handed the ”Questionnaire Packet”, a pencil, and the computer scan-tron sheet to answer the questions on. After completing the packet, the adolescents’ questions were answered and the investigator left the unit. The number of suicidal adolescents filling out questionnaires at one time ranged from 1 to 3 adolescents per investigator visit. All measures of confidentiality were followed. At McLaren Regional Medical Center, the investigator was given access to the adolescents records on the unit. The investigator then _ selected adolescents for the study that fit the definition of Suicidal Adolescents discussed in the "Conceptual and Operational Definitions” 33 section of this chapter. The investigator arrived on the unit at times the parents were visiting the adolescents in the unit. At that time, the investigator approached both the parentlsl and the adolescent. The investigator explained the research project and gave the parent a consent form to sign. Upon this signature, the investigator arranged with the adolescent an appropriate time to meet to administer the questionnaires. At the time of this meeting, the adolescent was given the adolescent consent form to sign. After the adolescent agreed to participate in the study by signing the consent form, the adolescent was then handed the "Questionnaire Packet", 8 pencil, and the computer scan-tron sheet to answer the questions on. After completing the packet, the adolescent’s questions were answered and the investigator left the unit. W FACES Ill - Data for this instrument were scored in the following manner: From the 20 questions in this instrument, the individual has a choice of selecting one of five likert-type answers, ranging between 1 and 5. The five answer choices are "1 - Almost Never", "2 - Once in Awhile", "3 - Sometimes”, "4 - Frequently", and "5 - Almost Always". Odd items are summed to arrive at a Cohesion Score and even items are summed to arrive at an Adaptability Score. These scores are then substituted into a formula called the Distance From Center Score (DFCS) = ((Cohesion score - 39.8)2 + (Adaptability score - 24.1)2)“2 This formula was derived from established norms for the scale. ”The Distance from Center (DFC) is a linear score 34 indicating the distance of an individual's cohesion and adaptability score from the center of the Circumplex Model. . . .' The main limitation of the score is that it indicates the distance, but not the direction of the score from the center of the Circumplex Model" (Olson, et al., 1985, p. 31). Table 1 gives a representation of the cutting points and mean scores for individuals on the FACES lll instrument. From this table, the number "39.8" in the DFCS formula is derived from the mean scores on the cohesion dimension. Also from this table, the number "24.1" in the DFCS formula is derived from the mean scores on the adaptability dimension. These mean scores are subtracted from the individual scores, then squared and added to each other. The square root is then calculated from this resulting number to arrive at the distance from center score. Table 1 in in nNrmfrlnivi lnhFAEllllnrmn INDIVIDUAL SCORES Circumplex Cutting __Iyp_e_a Points Percentages Parents Adolescents Parents Adolescents Balanced < 6 < 7 47 49 Mid-Range > 6< 11 > 7 <13 42 41 Extreme >11 >13 11 10 35 Table 1 Continued INDIVIDUAL NORMS FOR FACES lIl Mean SD (Standard DeviationL Cohesion 39.8 5.4 Adaptability 24.1 4.7 *This table is taken from Olson, et al., 1985, p. 32. PACI - Data for this instrument were scored in the following manner: From the 40 questions in this instrument, the individual has a choice of selecting one of three likert-type answers, ranging between 1 and 3. The five answer choices are "1 - Yes, Usually", "2 - No, Seldom", and "3 - Sometimes”. If an individual responds with the answer "1", a score of 3 is given; if an individual responds with "2", a score of 0 is given; and if an individual responds with ”3", a score of 2 is given for questions 1, 2, 5, 10, 11, 14, 15, 17, 18, 20, 22, 25, 26, 27, 30, 31, 34, 35, 36, and 39. If an individual responds with the answer ”1", a score of 0 is given; if 'an individual responds with "2", a score of 3 is given; and if an individual responds with "3", a score of 1 is given for questions 3, 4, 6, 7, 8, 9, 12, 13, 16, 19, 21, 23, 24, 28, 29, 32, 33, 37, 38, and 40. The scores for each item is summed to arrive at a total sum score. The possible range for the total sum scores on this inventory is between 0 and 120. The higher the total score, the higher the level of parent-adolescent communication (Bienvenu, 1969). CHAPTER FOUR -- RESULTS r niz in fD Anl i The results of the data are organized into three main parts: (1) demographic analyses, (2) testing of the hypotheses, and (3) additional or post-hoc analyses. For the demographic section, the total sample size composition and the differences between the suicidal and non-suicidal ' groups were explored for gender, race, age, suicidal intent, and for whom the adolescents lived with at home. For the analyses of the hypotheses section, one-way ANOVAs were employed to test for group mean differences on FACES Ill and PACI. Also included in this section are results of testing for any possible two-way interactions for the demographic variables of gender, race, and age. For the additional or post-hoc analyses section, alpha reliabilities, t tests, and correlations were conducted to test the reliability and validity of the instruments for the sample used in this study. The t tests were conducted on the differences between group means for the separate dimensions of the FACES Ill instrument: cohesion and adaptability. The total sum scores on the instruments were compared with the ages of the respondents in the study for the correlational analyses. The following analyses are based on a total sample size of 141 adolescents. There were 40 suicidal adolescents and 101 non-suicidal adolescents. 36 37 DmrhiF r Total Sample For the total sample of 141 adolescents, 18.4% were from Rivendell Psychiatric Center (3 =26), 9.9% of the adolescents were from McLaren Regional Medical Center (a = 14), 22% of the adolescents were from Bath High School (_r1=31), 40.4% of the adolescents were from Mason High School (n=57), and 9.2% were from Perry High School ([1:13). Also, for the total sample, 45.4% were male (n=64) and 54.6% were female (a: 77). For the race of the adolescents in the total sample, 92.9% were White (3:131), 2.1% were Black (3:3), 2.8% were Hispanic (g=4), and 2.1% were Native American (3 =3). None of the adolescents reported being of Asian descent. Thirty-nine percent of the total sample of subjects reported having seriously considered committing suicide in the past year, while 61% of the total sample did not seriously consider committing suicide in the past year (This includes both the suicidal and non-suicidal adolescent groups). Tables 2 and 3 give the frequencies of who the adolescents lived with at home and the ages of the adolescents, respectively. Table 2 Fr niefrP lhAdl entsLiv Wih HmfrhT l Sample Living with at Home jreouencv' EercentagL Mother 119 84.4 Father 94 66.7 Table 2 continued 38 Ljngg with at Home Frequencv* Percentaga_ Brother(s) 76 53.9 Sisterls) 58 41.1 Halfbrotherls) 14 9.9 Stepfather 13 9.2 Halfsisterls) 10 7.1 Grandmother 10 7.1 *The only people listed are those with frequencies of 10 or more. The people excluded were grandfather, auntls), unclels), stepmother, stepbrotherls), stepsisterls), female cousinls). male cousinls), mother's boyfriend, father's girlfriend, mother's boyfriend's children, father’s girlfriend’s children, female friends, male friends, and other people not listed. Table 3 r n i f r h A f h A n f r h T m l Age FrequenCIL Percentage 13 10 7.1 14 25 17.7 15 28 19.9 16 45 31.9 17 33 23.4 The data from Table 2 indicate that the majority of the total sample lives with either their mother or father. The mean age for the total sample was 15.47 (a: 1.23). 39 Group Comparisons There were a total of 28.4% suicidal adolescents (3:40) and 71.6% of non-suicidal adolescents (Q=101). Gender,race, and whether the adolescents seriously considered committing suicide in the past year (Suicidality) frequencies are shown in Table 4. Table 4 Fr nifr nchen iili r mrin Suicidal Group Gender Fr en Non-suicidal Group Gender _e_qu__c_vFr en Male 18 Male 46 Female . 22 Female 55 Bass Fr U n Bass accuses! White 32 White 99 Black 2 Black 1 Hispanic 3 Hispanic 1 Native Native American 3 American 0 Suiudflitx W 0 Yes 24 No 16 Suicidalltx W n Yes 31 No 70 There was not a significant difference between the suicidal and non-suicidal groups for gender (x’ (1, N=141l = .003, p< .95). There was a significant 40 difference between the suicidal and non-suicidal groups for race If (1, bl=141) = 15.02, g<.001). Twenty percent of the suicidal group of adolescents were Black, Hispanic, and Native American. Only 2% of the non-suicidal group of adolescents were either Black or Hispanic, and zero reported being Native American. As to be expected, there was a significant difference between suicidal and non-suicidal adolescents in reference to the question, ”Have you seriously considered committing suicide in the past year?" 06 (1, N=141l = 10.34, g<.001). Sixty percent of the suicidal subjects answered "yes” to this question and 30.7% of the non-suicidal subjects answered ”yes” to this question. Based on these demographic statistics there was a difference between the two groups in relation to race and suicidality. To analyze the differences between the suicidal and non-suicidal groups in relation to whom the adolescents lived with at home, five classifications of family groupings were employed. The first group is termed the ”Nuclear” group. This group contains the mother, father, brother, and sister. The second group is termed the "Other Relatives" group. This group contains the grandmother, grandfather, auntls), unclels), female cousinls), and male cousinls). The third group is termed the ”Remarried" group. This group contains the stepfather, stepmother, stepbrother, stepsister, halfbrother, and halfsister. The fourth group is termed the "Nonlegal" group. This group contains the mother's boyfriend, father’s girlfriend, mother's boyfriend’s children, and father's girlfriend’s children. The fifth group is termed the ”Unrelated" group. This group contains female friendls), male 41 friendls), and others not listed. Table 5 represents the classification of the people the adolescents lived with at home. The table frequencies are the sum of the adolescents indicating that they lived with 1 or more people within each group classification. Table 5 n i fWh h A l n Liv Wi h m r m ri n Suicidal Group Non-Suicidal Group Live with at Home Freouencv’ Freouencv’ Nuclear 36 98 Other Relatives 10 11 Remarried 11 22 Nonlegal 6 3 Unrelated 6 4 *The frequency of the adolescents indicating living with one or more person in the group. There is a significant difference between the suicidal and non-suicidal adolescents in living with their nuclear family ()r2 (4, N=141l = 12.62, Q< .013). Ten percent of the suicidal adolescents reported not living with anyone from their nuclear family as compared to 2.97% of the non-suicidal adolescents. There is also a significant difference between the suicidal and non-suicidal adolescents in living with nonlegal family members (x’ (2, _l_\l_=141) = 7.55, n< .02). Fifteen percent of the suicidal adolescents 42 reported living with nonlegal family members as compared to 2.97% for the non-suicidal adolescents. Also, the suicidal adolescents reported living with unrelated people in their homes significantly more than the non-suicidal adolescents (x’ (1, [51:141) =5.30, p<.021). Fifteen percent of the suicidal adolescents reported living with unrelated people in their homes as compared to 3.96% for the non-suicidal adolescents (All the differences described above were significant at the p< .05 level). Thus, in comparing the two groups in relation to who the adolescents live with at home, the suicidal adolescents were more likely than the non-suicidal adolescents to be: (1) not living with anyone from their nuclear family, (2) living with nonlegal family members, and (3) living with unrelated people not affiliated with their family. Table 6 gives the frequency in ages for both the suicidal and non- suicidal adolescents. Table 6 Fr n i f r A r m ri n Suicidal Non—Suicidal Aoe Group Group 13 10 ' 0 14 _ 4 21 15 I 1 1 17 16 10 35 17 5 28 43 The mean age for the suicidal adolescent‘group was 14.9 (a: 1.37) and the mean age for the non-suicidal group was 15.69 (a: 1.09). A 2-tailed t test revealed significant differences between the means (t(139) =-3.60, n< .000). One plausible reason for the difference in mean ages is related to the non-suicidal adolescent group not having any 13 year-olds to compare with the suicidal adolescent group. Having 13 years-olds in the non-suicidal group would decrease the mean and probably eliminate the significant difference. Given the small sample size of the comparison group and the significant differences just described, being able to match the adolescents on age, sex, and race would be impossible. For age, there were no 13 year- olds to compare with the suicidal group. For sex, the comparison group was already comparable: 45% were suicidal males and 45.5% were non-suicidal males. For race, there were hot enough Black, Hispanic, and Native American adolescents in the comparison group to match with the suicidal group. Thus, if there had been a comparable group of non-suicidal adolescents (in respect to age, sex, and race), the subsequent statistical results, testing the hypotheses on mean differences with FACES Ill and PACI, would have used matched groups. Tag; 91‘ Hyggthesaa Hypothesis H1l H,I: Adolescents from the suicidal experimental group will report more dysfunction than the adolescents from the non—suicidal comparison group in relation to beliefs on family structure by using the DFCS from FACES III. Results A one-way ANOVA for the FACES Ill instrument was used to test the difference between the two means for the suicidal and non-suicidal adolescent groups using the DFCS . The results in Table 7 support the rejection of the null hypothesis at the p< .05 level (rounding to the nearest hundredth). Thus, there was a significant difference between the group means for the FACES lll instrument. The mean score on FACES III for the suicidal group was 15.03 (a=7.57) with minimum and maximum scores of 5.04 and 32.07. The mean score on FACES III for the non-suicidal group was 12.34 (§=7.19) with minimum and maximum scores of 1.20 and 28.64. Thus, FACES lll did discriminate between the suicidal and non- suicidal group in using the DFCS means. As to be expected, according to the mean DFCS for the suicidal and non-suicidal adolescent groups, the suicidal group of adolescents reported more dysfunction in their families. This result supports hypothesis l-I1I. 45 Table 7 n-w AN VAR ltfrFAE lll Mean Degrees of Significance MES Freedom F Value of F Value Between 206.7025 1 3.8785 *.0509 Within 53.2943 139 *Significant at the g< .05 level (when rounded to the nearest hundredth). Hypothesis H1ll H,ll: Adolescents from the suicidal experimental group will report lower scores on family communication than the non-suicidal comparison group by using the PACI. Resuhs A one-way ANOVA for the PACI instrument was used to test the differences between the two means for the suicidal and non-suicidal adolescent groups using the total sum score. The results in Table 8 support the rejection of the null hypothesis at the g< .05 level. Thus, there was a significant difference between the group means for the PACI instrument. The mean score on PACI for the suicidal adolescent group was 63.47 (§=24.65) with the minimum and maximum scores of 15 and 107. The mean score on PACI for the non-suicidal group was 79.18 la=23.31) with minimum and maximum scores of 20 and 117. Thus, PACI did discriminate 46 between the suicidal and non-suicidal group in using the total sum score means. As to be expected, the suicidal group scored significantly lower on the PACI than did the non-suicidal group. This result supports hypothesis H,ll. Table 8 n-w AN VARe l forPAI Mean Degrees of Significance Gwyn; Squares Freedom F Value of F Value Between 7065.4457 1 12.5892 *.0005 Within 506.2285 139 *Significant at the Q<.O5 level. Two-way Interactions EAQESll - ANOVA statistical procedures were performed to test if their were significant interactions for the DFCS obtained by the suicidal and non- suicidal groups in reference to gender, race, and age. There were not significant two-way interactions between the two adolescent groups with reference to gender (5(1): .005, g<.943), race (5(2) = .243, g<.792), or age (El2)1.08, Q<.362). Tables 9, 10, and 11 give the mean scores for each of these variables. Thus, for the FACES lll instrument, the adolescents' gender, race, or age did not affect the outcome of the scores in either the suicidal or non-suicidal group. PAQ - ANOVA statistical procedures were performed to test if their were 47 significant interactions for the total PACI score obtained by the suicidal and non-suicidal groups in reference to gender, race, and age. There were not significant two-way interactions between the two adolescent groups with reference to gender (E(1)=.246, p<.621), race (El2) = .263, a<.769), or age (El3)1.24, n<.297). Tables 9, 10, and 11 give the mean scores for each of these variables. Thus, for the PACI instrument, the adolescents’ gender, race, or age did not affect the outcome of the scores in either the suicidal or non-suicidal group. Table 9 Maa Saaraa an FACES lll (DFCS) ang PACI far Gander ay Gragg m ri n Suicidal Group Non-suicidal Group Mean Mean Bender ore Seeder Scores Male (n =18) Male (n =46) DFCS 14.25 DFCS 11.68 PACI 65.67 PACI 83.74 Female (n = 22) Female (n = 55) DFCS 15.67 DFCS 12.90 PACI 61.68 PACI 75.36 48 Table 10 nun or- on ll OF ._no A Ifr R - . r _. oumri on Suicidal Grouo Non-suicidal Group Mean Mean Base Segre: Base Seems White In =32) White In =99) DFCS 15.50 DFCS 12.41 PACI 62.13 PACI 79.23 Black (n = 2) Black (n =1) DFCS 12.70 DFCS 13.57 PACI 71.50 PACI 67.00 Hispanic (n = 3) Hispanic (n =1) DFCS 12.54 DFCS 4.96 PACI 71.00 PACI 86.00 Native American (n =3) Native American (n =0) DFCS 14.07 DFCS 0.00 PACI 65.00 PACI 0.00 Asian (n =0) . Asian (n =0) DFCS 0.00 DFCS 0.00 PACI 0.00 - PACI 0.00 Table 11 11°41 of" n FA III DF .n- PA lfoer: o r0 o oo-mri o 5"IIG II-"IIG Mean Mean Ass Secures A99. Scores 13 years (n =10) 13 years (n =0) DFCS 14.15 DFCS 0.00 PACI 69.90 PACI 0.00 14 years (n =4) _14 years (n = 21) DFCS 22.34 DFCS 12.94 PACI 40.75 PACI 75.62 Table 11 Continued Suicidal Group 49 Non-suicidal Group A93 15 years (n=11) DFCS PACI 16 years (n =10) DFCS PACI 17 years (n = 5) DFCS PACI Mean Scares 13.33 77.18 15.87 50.90 13.00 63.80 Mean Ase Scores 15 years (n =17) DFCS 11.28 PACI 85.59 16 years (n = 35) DFCS 12.25 PACI 77.23 17 years (n = 28) DFCS 12.66 PACI 80.39 Alpha Reliabilities A iinlAnl To test for the reliability of the instruments used in this study, Cronbach's a was employed. Cronbach's 0 (alpha) is a measure used to test the internal consistency of items in a test, scale, or instrument. It can be viewed as the correlation between this test or scale and all other possible tests or scales containing the same number of items, which could be constructed from a hypothetical universe of items that measure the characteristic of interest . . . . Cronbach's a can be computed using the following formula: (klcov/var a II 1 +(k-1lcov/var where k is the number of items in the scale, cov is the average covariance between items, and var is the average variance of the items (Norusis, 1990, p. B-190-191). 50 Cronbach's alpha was used to test the internal consistency of the items in PACI, FACES Ill, and the two dimensions within FACES lll: Cohesion and Adaptability. Table 12 shows these reliabilities. Table 12 :lo . R-.lioili ' or A I : II .no F A, ll I-im in: . - on and Agamagility W Alpha Standardized Item Alpha PACI .9319 .9333 FACES lll (total scale) .8497 .8474 FACES Ill: Cohesion .8824 .8825 FACES lll: Adaptability .6743 .6785 The alpha reliability for the PACI instrument is very high. The alpha reliability for the total FACES lIl scale is high, but not as high as the PACI instrument. When the FACES Ill scale is tested for internal consistency in relation to each dimension, the Adaptability dimension appears to decrease the total scale’s reliability level because of its moderate to alpha reliability. Ttests for Cohesion and Adaptability Ttests were conducted for each dimension on the FACES Ill scale, Cohesion and Adaptability, to investigate the differences between the mean scores obtained by the suicidal and non-suicidal groups. The rationale 51 behind performing these separate t tests was to see if there were significant differences for each dimension on the FACES llI scale. Since the DFCS utilizes both dimensions for its calculations, it neglects to show the strength of either dimension in distinguishing between the suicidal or non-suicidal adolescents. Table 13 displays the results of these calculations. Table 13 T- R sult f r h hesi nan Ad ilit Dim n i n n FA E III Cohesion Results Standard Degrees of Significance r n vi in Fr m Suicidal (n=40) 26.45 8.86 139 -2.51 *.013 Non-suicidal (n =101) 30.62 8.94 Adaptability Results Standard Degrees of Significance r vi i n Fr m T V I Suicidal (n =40) 24.60 5.25 139 1.36 .175 Non-suicidal (n =101) 23.07 6.29 *Significant at the p<.05 level. Table 13 shows that there is a significant difference between the suicidal and non-suicidal adolescents' group means for Cohesion but not for Adaptability. Thus, with significance, the non-suicidal adolescents reported 52 higher scores on the Cohesion dimension of FACES Ill: and without significance, the non-suicidal adolescents reported lower scores on the Adaptability dimension of FACES lll. Correlational Analyses The total scores on PACI, FACES lIl (using the DFCS), and the ages of the adolescents were correlated to test for (1) linear relationships between the instruments and (2) linear relationships between the adolescents ages and the instruments. The resulting correlational matrix is presented in Table 14. Table 14 rrlinerifrAlFAElllF nhA fh A l n Scale PACI DFCS AGE PACI 1.00 DFCS -.6760* 1.00 lp=.001) AGE .0640 -.0652 1.00 *Significant at the n< .05 level. A moderately strong significant negative correlation exists between the total scores obtained on PACI and the DFCS. Thus, for this sample, as the perceptions of adolescents' parent-adolescent communication increases 53 (indicated by increased PACI scores), the adolescents perception of family functioning improves (indicated by a decreasing DFCS) (A causal relationship for this negative correlation must not be interpreted as correlational analyses show only the relationships between variables and not whether one variable causes a change in another variable). CHAPTER FIVE -- DISCUSSION OF THE RESULTS ri f ri i n f h This study investigated adolescents’ beliefs on family structure and parent-adolescent communication. In particular, the study explored the differing beliefs of family structure and parent-adolescent communication for two adolescent groups: Suicidal and Non-suicidal. The objective of the study was to differentiate between suicidal and non-suicidal adolescents based on their beliefs concerning family structure and parent-adolescent communication. Only the adolescents' beliefs were obtained in this study based on the presumption that the adolescents' beliefs, whether they were realistic or not, would be sufficient enough to discriminate between the two groups. The following hypotheses were tested to accomplish the objective of this research: Hypothesis 1: Adolescents receiving inpatient treatment for suicidal behavior will report more dysfunction in their family structure than the non- suicidal adolescents not receiving treatment for suicidal behavior. Hypothesis 2: Adolescents receiving inpatient treatment for suicidal behavior will report less positive parent-adolescent communication than non- suicidal adolescents not receiving treatment for suicidal behavior. In addition to testing these hypotheses, the researcher also obtained selected demographic information from the adolescents to explore any distinctions in the two groups. The demographic information obtained from these adolescents were (1) gender, (2) race, (3) suicidal intent, (4) who the 54 55 adolescents lived with at home, and (5) age. The post-hoe analyses related to the t tests performed on the separate dimensions on the FACES lll instrument will also be discussed. The instruments used to measure family structure and parent- adolescent communication were the Family Adaptability Cohesion and Evaluation Scale Ill and the Parent Adolescent Communication Inventory. These instruments were administered to a total of 141 adolescents between the ages of 13 and 17. The suicidal group consisted of 40 adolescents receiving inpatient therapy from Rivendell Psychiatric Center (located in St. Johns, Michigan) and McLaren Regional Medical Center (located in Flint, Michigan). The non-suicidal group consisted of 101 adolescents attending either Mason High School, Bath High School, or Perry High School (located in Mason, Bath, and Perry, Michiganuareas served by the two psychiatric institutions just mentioned). The organization of this chapter will (1) discuss the demographic results, (2) discuss of the results related to the hypotheses, and (3) discuss the t test post-hoe analyses related to the FACES lll instrument used in the study. D m r hi Fin in The first demographic discussion is related to the suicidal and non- suicidal adolescent groups’ gender. The results revealed the 45% of the suicidal group were male and 55% were female. This proportion similarly matched the non-suicidal group at 45.5% male and 54.5% female adolescents. Thus, there was not a significant difference between the 56 groups. The ratio of females to males (1.2:1) in the suicidal group of adolescents is rather low compared to the literature on gender issues. Ratios of female to male suicidal attempters is cited in the literature at 4:1 with females ingesting pills as their primary method of attempt (Garfinkel, Froese, and Hood, 1982). The second demographic discussion is related to the group comparisons on race. The results did reveal a significant difference between the groups in relation to race revealing that 20% of the suicidal group was either Black, Hispanic, or Native American, while only 2% of the non-suicidal group was either Black or Hispanic. None of the non-suicidal adolescents reported being Native American. These results are consistent with the literature in that Whites consistently have a higher incidence of suicide than Blacks; and Hispanics and Native Americans are considered the "at-risk” groups for suicide and other violent deaths (Berman and Jobes, 1991). The third demographic discussion relates to the comparison of the suicidal and non-suicidal group in relation to their ”suicidal intent”. Adolescents were asked, "Have you seriously considered committing suicide in the past year?". The suicidal group did answer "yes” significantly more than the non-suicidal group; however, 30.7% of the non-suicidal group reported "yes" to this question. This high percentage is alarming but consistent with a recent study conducted by the Centers For Disease Control (1991) where 27.3% of the 11,631 high school students surveyed reported having thought seriously about attempting suicide. The fourth demographic variable studied related to who the 57 adolescents lived with at home. The results showed that suicidal adolescents reported significantly more than the non-suicidal adolescents of not living with their nuclear family and living more with nonlegal family members (mother’s boyfriend, father’s girlfriend, or the children of these nonlegal family members) or other unrelated people. This result is interesting yet it does not seem surprising since one factor cited extensively in the literature related to adolescent suicide in the family is recent losses (i.e., parental separation/divorce)(Farberow, 1987; Richman, 1986; Leder, 1987). The dynamics of having to live with someone new and having lost a mother or father (at least living in the same home together) through divorce or separation can be devastating for many adolescents. This study did not inquire as to any recent divorces or separations in the adolescents' families, therefore, making any definitive conclusions as to why suicidal adolescents reported living more with nonlegal and unrelated people in their homes would be premature. However, this finding is a measure of instability within the home since 15% of the suicidal group reported living with nonlegal or unrelated people compared to 2.97% and 3.96% for the non-suicidal group. The last demographic variable to be discussed is the age of the adolescents in the study. There were significant differences noted in age, where the suicidal adolescents were younger Igg=14.9) than the non-suicidal adolescents (x_=15.69) in this study. A valid explanation for this difference can be attributed to the fact that the non-suicidal adolescent sample was collected only from high school students, where the beginning age to start 9th grade is 14. If an 8th grade class had been sampled, differences in the 58 ages of the adolescents for the two groups may not have been significant. Qiaggaaign 9f Hypglhaaaa Hypothesis 1 Hypothesis 1 stated that suicidal adolescents will report more dysfunction in their families than the non-suicidal adolescents. To measure family functioning, the researcher employed the concepts of the Circumplex Model and the instrument FACES III to operationalize these concepts. To be able to use linear analyses in testing for significant difference between the suicidal and non-suicidal adolescents’ responses on FACES III, the Distance From Center Score (DFCS) was used. This score combines the scores on both Cohesion and Adaptability to form a linear measurement beginning at the center of the Circumplex Model and ending at the furthest point from the center of the Circumplex Model. Using the DFCS, families can fall into one of three family types: Balanced, Midrange, or Extreme. Balanced types are the most functional and have the lowest DFCS, while the Extreme families are considered the most dysfunctional and have the highest DFCS. In this study there was a significant difference between the suicidal and non-suicidal adolescent groups’ mean distance from center scores (p<.05). The non-suicidal adolescent group's mean DFCS was 12.34. In using the norms from Table 1 for adolescents, this would place the non- suicidal adolescent group's overall rating to be in the Mid-Range Type. The suicidal adolescent group’s mean DFCS was 15.03. This would place the suicidal adolescent group's overall rating to be in the Extreme Type. These Figure 2: Diagram of the Four Extreme Family Types from the Circumplex 59 Model (All text and the concept taken from Olson, 1989, p. 15). Chaotlcally Disengaged mmMmmmmm -Erratic leadership, unsuccessful parental control, Ineffective discipline, Inconsistent consequences, Impulsive decisions andess negotiation, lack of role clarity, role reversals, frequent rule changes mummmmmm -Extrame emotional separateness, lack of family loyalty, very little involvement with each other, very little sharing of feelings, lack of parent-child closeness, separateness preferred, independent decision making, very little time together, lack of family loyalty. Rigidly Disengaged i ' n ili -Authoritarian leadership, highly controlling parents, strlct consequences and limited negotlations. decisions imposed by parents Strictly defined roles, generally traditional male - female roles and unchanging rules. 9‘ I C | . -Extreme emotional separateness, lack of family loyalty, very little involvement with each other, very little sharing of feelings. lack of parent-child closeness, separateness preferred, independent decision making, very little time together, lack of family loyalty. +<4-P-W>HV)U>¢ o-COHESION—O Erratic leadership, unsuccessful parental entrol, Ineffective discipline, Inconsistent onsequences, Impulsive decisions less negotiation. lack of role clarity, role reversals, frequent rule changes h h i ~Extreme emotional closeness, loyalty to famlly demanded, very dependent of one ranother, little private space permitted, lack of generational boundaries, energy n'lainly focused inside the family, with few individual friends permitted, very reactive emotionally, decisions are subject to the wishes of the whole group. Rigidly Enmashad i l n A ill -Authoritarian leadership, highly controlling parents, strict consequences and limited negatlations, decisions imposed by parents Strictly defined roles, generally traditional male - female roles and unchanging rules. Winn -Extreme emotional closeness, loyalty to family demanded, very dependent of one another, little private space permitted, lack of generational boundaries. energy mainly focused inside the family, with few individual friends permitted, very reactive emotionally, decisions are subject to the wishes of the whole group. so results support the hypothesis that the suicidal adolescents will report more family dysfunction than the non-suicidal adolescents. These results do not indicate, however, what cell within the Circumplex Model-thegsuicidal and non-suicidal adolescents generally fall in. Since the suicidal adolescents are classified in the Extreme Type on the Circumplex Model, their perceptions can be considered one of four integrated types in the model: (1) Chaotically Disengaged, (2) Chaotically Enmeshed, (3) Rigidly Disengaged, or (4) Rigidly Enmeshed. These four types are diagrammed above in Figure 2. Since the non-suicidal adolescents are classified in the Mid-Range Type on the Circumplex Model, their perceptions can be considered one of eight types in the model: (1) Chaotically Separated, (2) Chaotically Connected, l3) Flexibly Enmeshed, (4) Structurally Enmeshed, (5) Rigidly Connected, (6) Rigidly Separated, (7) Structurally Disengaged, and (8) Flexibly Disengaged. A description of the "WW” type is as followsz There is a good balance between too close and too separate. There is a balance of time together with time apart with some involvement between members. There is some independent decision making and some joint decisions. There is a balance between energy focussed inside and outside the family. Loyalty to the family is expected but not demanded (Olson, 1989, p. 15). A description of the "Structured ta Flaxiala an Adaptaaility" is as follows: Leadership is sometimes shared and democratic. The roles and household responsibilities are stable but may be shared. The rules are predictable and fair, but can be flexible when needed. The children’s feelings are sometimes taken into account when making decisions (Olson, 1989, p. 15). . The four extreme family types that the suicidal adolescents would be classified in this study are consistent with the literature available on 61 adolescent suicide and the family. Pfeffer (1981) discussed "symbiotic parent-child" relationships, which is similar to enmeshed on cohesion; and "inflexible family systems", which is similar to rigid on adaptability. Molin (1986) discussed the ”excluded family relationship" which is similar to disengaged on cohesion. Kerfoot (1980) discussed the role reversals of parents and adolescents, which is similar to chaotic on adaptability. Thus, the results of this research support the clinical research just mentioned; however, this study does not indicate which of the four extreme types the suicidal adolescents would be generally characterized in. There has been empirical research on problem families that found significant results using versions of the FACES instrument. Reinherz, Stewart-Berghauer, Pakiz, Frost, and Moeykens (1989), in their longitudinal study of 185 boys and 193 girls, investigated adolescent depressive symptomatology from age 5 to 15. They incorporated a battery of tests in the design of their study to identify risk factors and mediators of depression in adolescents. One of the instruments they used was the Cohesion dimension from the FACES lll scale. Using a hierarchical regression analysis, family cohesiveness accounted for 8.5% of the total variance for depressive symptoms, with higher levels of family cohesion associated with fewer depressive symptoms in the adolescents. In another study assessing schizophrenic, neurotic, and no therapy families, the FACES ll instrument was used (Clark, 1984; cited in Olson, 1989). Both dimensions in the Circumplex Model were used and the results significantly revealed that a higher proportion of neurotic and schizophrenic 62 families fell in the extreme family type than the no therapy families. The no therapy families were found significantly more in the balanced family type. In yet another study, Garbarino, Sebes, and Schellenbach (1985) compared 27 high risk families to 35 low risk families for destructive parent- child relations in adolescents using the FACES instrument. The results of the study showed that the majority of the high risk families were Classified into the extreme family type, while low risk families were classified into the balanced family type. These studies confirm the construct validity of using FACES lll instruments for differentiating between problem and non-problem families. Hypothesis 2 Hypothesis 2 stated that the suicidal adolescents will report less positive parent-adolescent communication in their families than the non- suicidal adolescents. To measure parent-adolescent communication, the researcher utilized the PACI as a measure of parent-adolescent communication because it assessed on one scale, the adolescents' perceptions about communication between the adolescent and (1) the mother, (2) the father, and (3) both parents. Barnes and Olson (1982) also developed a parent-adolescent communication instrument. Since Olson is one of the developers of the Circumplex Model, it would seem logical to use that parent-adolescent communication instrument. However, the Barnes and Olson (1982) parent-adolescent communication instrument's major shortcomings are that adolescents' perceptions about their mother and father are asked on two separate scales, and both of these scales lack 63 questions related to how the adolescents perceive communication among both parents. The PACI instrument does reflect all three perceptions of parent-adolescent communication (see Appendix A). Therefore, because of the shortcomings just described, it was the researcher's belief that the PACI instrument better delineates systemic framework components of parent- adolescent communication than the Barnes and Olson (1982) instrument. Peterson (1986) stated, "those who advocate 'systems thinking' focus their attention on the organization and relationships between components (i.e., the individuals) of family systems' (pp. 26—27). The PACI instrument, by addressing questions about the mother, father, and both parents, seems to address the systemic view. In this study, using PACI as the parent-adolescent communication instrument, there was a significant difference between the suicidal and non- suicidal adolescents groups' mean scores (n< .05). The non-suicidal adolescent group’s mean PACI score was 79.18 and the suicidal adolescent group’s mean PACI score was 63.47. Since the total PACI score ranges between zero and 120, the midpoint on the scale would fall at a PACI score of 60. Thus, a low to moderate PACI score could be interpreted as any PACI score below or equal to 60, while a moderate to high score could be interpreted as any PACI score greater than 60. Using these cut-off points for this study, both the suicidal and non-suicidal adolescent groups scored in the moderate to high PACI score range. However, there is a 15.71 difference in mean PACI scores between the two groups. This difference was found to be significant at the .0005 64 level. Also, the standard error of the mean (SEM) for the suicidal group was 3.8969 and the SEM for the non-suicidal group was 2.3191. This gives 95% confidence intervals for the suicidal group to be at 55.5928 and 77.3572. The 95% confidence intervals for the non-suicidal group is at 74.5772 and 83.7793. From these confidence intervals, there is 2.8 value overlap concluding that with 95% confidence, the suicidal adolescents in this sample would score lower on PACI than the non-suicidal adolescents. Therefore, although both the suicidal and non-suicidal adolescent groups scored in the moderate to high range on PACI, the suicidal adolescents scored significantly lower than the non-suicidal adolescents on PACI. Clinical literature supporting the notion of poor parent-adolescent communication with families of suicidal adolescents are Richman (1986) and Yusin (1972; cited in Heillig, 1983). An empirical study conducted by Stivers (1988) used PACI scores correlated with the Suicide-Depression Inventory (SDI) on 43 adolescents. The SDI is a 50-item questionnaire "which serves as a screening instrument for depression and suicide proneness" (p. 293). This study revealed that the adolescent's perceptions of the parent-adolescent communication was significantly correlated with the scores on the SDI (r= .2512, p= .001). Thus, further empirical research is needed to be conducted with PACI to support the relationship between poor adolescent-communication and adolescent suicidal behavior. 65 Dic ssi n fR sul n hei n n A tablli To test the scores for the individual dimensions of the FACES lll instrument, t tests were performed between the mean scores on Cohesion and Adaptability. The results showed significance for the Cohesion dimension but not for the Adaptability dimension (see Table 13). Norms for the different levels on the cohesion and adaptability dimensiOn are found in Figure 3. In looking at this figure for the Cohesion dimension, the mean score of 26.45 would place the suicidal adolescent group clearly within the disengaged level. This level is considered an extreme or dysfunctional family level within the Circumplex Model. A mean score of 30.62 for the non- sbicidal adolescent group would place this group somewhere in between the disengaged or separated levels. The separated level is considered as one of the center levels, indicating more balance within the Circumplex Model. Thus, for the cohesion dimension, the hypothesis that the suicidal adolescents will report more dysfunction in their families than the non- suicidal adolescents is supported. Spector (1988) also found that the suicidal adolescents mean scores fell within the disengaged level. The empirical literature discussed in chapter 2 of this report also supports these findings. However, in looking at this figure for the Adaptability dimension, this same hypothesis is not supported. The mean score of 24.60 for the suicidal group places this group somewhere in between the structured and flexible levels. These two levels are within the center of the Circumplex Model, indicating more balance in the family on adaptability. The mean score of 66 Figure 3: Norms and Cutting Points for Four Levels of Cohesion and Adaptability (From Olson et. al, 1985, p. 26). lie-Ila Cabesfaa (rev scares aad aercaaffles) to1214161820222426281332‘M3638wanna!) 1-‘M3 'l AllSteces 16.31. 2.9mm“ Adolescents 18.61. 3‘ 33;, I721. managed Family Maptablllta (rev scares aad percaatlles) 1012M1618w2224262830323436364042444648 LAMB :}:}:§:§:}:§:;':3:§"""ii‘i""“""' -r -~ -' - '° ' 5 , , - “15W“ 363’- 5532: A , jeii'::ff;':;-'-. 2.?erentsa Adolescents! 13.97. 3.10m: Couples 13.2% Rigid - 23.07 for the non-suicidal group places this group within the structured level, also indicating balance in the family on adaptability. Both the suicidal and non-suicidal adolescents obtained scores within the two more balanced levels on the Circumplex Model. The result obtained on the Adaptability dimension differs from the results obtained using the DFCS method. Spector (1988) and Livingston (1989) also did not find significant differences in their suicidal and non-suicidal adolescents mean scores for the 67 Adaptability dimension. The scores they obtained also had their suicidal adolescents falling within structured levels within the Adaptability dimension. Therefore, the mean score obtained in this sample, along with the empirical studies just mentioned, does not support the notion that the Adaptability dimension is useful in discriminating between suicidal and non- suicidal adolescents. CHAPTER SIX -- CONCLUSION imi i n f h The following are a list of limitations for this study: 1) The sampling procedure for this study was not random, which limited being able to generalize to a population of suicidal adolescents. However, due to the nature of this study, it was impossible to conduct research on random samples of suicidal adolescents. 2) The sample of suicidal adolescents chosen for the study may not have reflected the entire population of suicidal adolescents because they were chosen from a group of adolescents who were receiving therapy from psychiatric institutions for engaging in suicidal behavior. 3) This study was limited to addressing only those adolescents who had survived in engaging in suicidal behavior. 4) This study was limited in that it did not distinguish between high and low lethal suicidal behaviors for the suicidal group of adolescents. 5) This study was descriptive in nature, therefore, hypothesis seeking. Being able to establish causal relationships was not possible. 6) Scores obtained by the FACES Ill and PACI instruments were only obtained by the adolescents in the families under study. This was only one family member's perception of family structure and communication. 7) Only a small sample of suicidal and non-suicidal adolescents were used for the research making it impossible to match the groups on demographic variables of age, sex, and race. 68 69 8) The two groups of adolescents, suicidal and non-suicidal (30.7% of the non-suicidal adolescents reported having seriously considered suicide in the last year), and race (only 2% of the non-suicidal adolescents reported being non—white, while 20% of the suicidal adolescents reported being non-white) were confounding variables within this study. Further studies should address this issue. mm r From this small sample of adolescents, significant differences between suicidal and non-suicidal adolescents’ perceptions of their families were found. The results of this study clearly provide evidence that the sample of suicidal adolescents perceived their families as more dysfunctional and less communicative. The correlational analyses (see Table 14) did show a negative relationship between overall PACI scores and DFCS. As the adolescents' perceptions of parent-adolescent communication increases, the adolescents' perception of their family functioning also increases (indicated by lower DFCS). Research into this type of relationship needs to be conducted on a larger scale to confirm this relationship. The suicidal adolescents’ composition in the other variables measured in this sample were also different from the non-suicidal adolescents. More suicidal adolescents reported being Black, Hispanic, and Native American than the non-suicidal adolescents even with differing group sizes (see Table 4). There were 40 total suicidal adolescents and 101 total non-suicidal adolescents. Suicidal adolescents also reported living with nonlegal family 70 members and unrelated people in the family more than non-suicidal adolescents (see Table 5). At the same time, suicidal adolescents indicated that they lived less with their nuclear families than non-suicidal adolescents. This reflects a difference in family stability, at least as far as family membership. Suicidal adolescents represented more of the ”unstable" family compositions than the non-suicidal adolescents in this sample. The last, not surprising, difference noted between suicidal and non-suicidal adolescents in this sample was that suicidal adolescents reported "having seriously considered committing suicide in the past year" (see Appendix A) more than the non-suicidal adolescents. However, these results did indicate that 30.7% of the non-suicidal adolescents sampled from high schools in cities surrounding the Lansing, Michigan area reported having seriously considered committing suicide in the past year. Taking this sample’s results and the results from a larger scale study by the CDC (1991) using high school students, where 27.3% reported having seriously thought about suicide, emits a powerful signal that the youth in society need to be taught better ways of coping with their problems than resorting to the ultimate escape: suicide. Adolescence, indeed, is a difficult transitional period from childhood to adulthood and more research needs to be conducted on these "normal" adolescent populations to further explain this phenomenon. 71 Impliaatians far Farther Rasearah This research did indicate that there are relationships between family problems and adolescent suicide. However, additional research beyond the scope of this thesis project on families of suicidal adolescents needs to be conducted to get a better explanation of this social tragedy. Using larger sample sizes, using all the family members in the research to getter a better understanding of the entire family system dynamics, and controlling for the lethality in suicidal attempts when conducting this research are only a few methods by which a better understanding of how families contribute to adolescent suicide can be accomplished. More research is being conducted in the area of family members who have mental health problems and its relationship to adolescent suicide. Stiffman (1989) conducted a study of 291 adolescents utilizing runaway shelters. Of these 291 adolescents, 87 or 30% reported having attempted suicide. When analyzing differences between these two groups of runaways, suicidal and non-suicidal, the suicidal group of runaways significantly reported more of their family members as depressed, drug- abusing, and/or having antisocial personality traits. Also, one out of three of the suicidal runaways reported having had a family member attempt suicide. Pfeffer (1990), in an article addressing when to hospitalize suicidal or adolescent youth, mentions as one of the factors predicting psychiatric hospitalization to be parental psychopathology. A case study in this article describes a 9-year olds' precipitating events that lead to her hospitalization. This case study states, "Nine-year-old Anna attempted suicide a year after 72 her parents separated. Her mother was seriously depressed and had difficulty concentrating at work; she tended to withdraw from involvements with Anna and Anna's 15-year-old brother" (p. 144). These studies address the problem of suicidality in youth not only as mental health problems for the suicidal youth, but as a family mental health problem. Therefore to address the problem of suicidal behavior in adolescents, the family system problems need to be analyzed. Further empirical research in the area of linking family dysfunction (using FACES Ill), poor parent-adolescent communication (using PACI) with family mental health problems is warranted. Possibly utilizing personality inventories with the FACES Ill and P'ACI instrument may help to distinguish further the "at risk" group of adolescents for suicide. Qancluaian The results of this research indicate that family dysfunction was perceived to be more prevalent in the suicidal group of adolescents than the non-suicidal group of adolescents. This indicates that adolescent suicidal behavior is a family system problem. To address this problem, treatment facilities (i.e., runaway shelters and psychiatric institutions) need to realize the importance of the family in therapy. Since adolescent suicidal behavior is, in part, a response to family dysfunction, then it should be a requirement for all treatment facilities to have family therapy as one form of indicated treatment. Effective primary prevention programs aimed at educating youth and families about positive communication and relational skills may facilitate 73 prevention of suicidal behavior in adolescents altogether. However, without the support of other larger social systems (i.e., school systems and the government) in encouraging prevention programs for adolescent suicide, the families of these adolescents will suffer. Andrews, Bubolz, and Paolucci (1980) poignantly stated: The family system includes both personal attributes of its members such as health and skills, and structural attributes of the family, i.e., authority patterns, roles, goals, aspirations, affectual relationships, and patterns of decision making. The family system produces human resources, and with the help of other social systems transforms these resources to provide the human capital reserves of society. If the required sources are not available or the family is not able to utilize resources, family members do not develop in a positive direction and instead of becoming human capital reserves of a society, become costs to society (p. 39). LIST OF REFERENCES EFR Andrews, M. P., Bubolz, M. M., & Paolucci, B. (1980). An ecological approach to the study of the family. M rria F mil R v' w, 3, 29-49. Asarnow, J. & Carlson, G. (1988). Suicide attempts in preadolescent child psychiatry inpatients. i i n Lif Thr nin B h vi r, 18, 129-136. Barnes, H. & Olson, D. H. (1982). Parent-Adolescent Communication. In Olson, D. H., McCubbin, H. l., Barnes, H., Larsen, A., Muxen M., & Wilson, M. Family lnvantgriaa (pp. 51-66). St. Paul, MN: Family Social Science. Berman, A. L. & Jobes, D. A. (1991). Agalaacant agigiaa: Aaaaaamant and intarvantian. Washington, DC: American Psychological Association. Bienvenu, M. J. (1969). Measurement of parent-adolescent communication. Family Cagrginatar, 1_8_, 117-121. Brown, S. (1985). Adolescents and family systems. In M. Peck, N. Farberow, & R. Litman (Eds.), Yggth aaicige (pp. 71-80). New York: Spflngen Centers For Disease Control (1991). Attempted suicide among high school students - United States, 1990. M r i i n M r Raggrt, _4_Q(37), 633-635. Cohen-Sandler, R., Berman, A. L., & King, R. A. (1982). Life stress and symptomatology: Determinants of suicidal behavior in children. Jggrnal Qf Iha Amarigan Agademy Qf thlg Payghiatry, .2_‘_I_(2), 178- 186. Cooper, C. R., Grotevant, H. D., & Condon, S. M. (1983). Individuality and connectedness in the family as a context for adolescent identity formation and role-taking skill. In H. D. Grotevant & C. R. Cooper (Eds), Agalaagant gavalagmant in tha family (pp. 43-59). San Francisco: Jossey-Bass. 74 75 Farberow, N. L. (1987). The role of the family in suicide. In R. F. Diekstra & K. Havvton (Eds.). Sgiaiga in adaleacanca (pp. 139-152). Hingham, MA: Kluwer Academic Publishers. Garbarino, J., Sebes, J., & Schellenbach, C. (1985). Families at risk for destructive parent-child relations in adolescents. thlg Davalagma 1, 5.5.. 174-183. Garfinkel, 8., Froese, A., & Hood, J. (1982). Suicide attempts in children and adolescents. American Jagrnal af Payahiatry, 142, 643-644. Greenlee, L. F. (1987). Tgward an gnglaratanging af agalaagant garaagigiaa. (Report No. CG 019 911). California: Biola University. (ERIC Document Reproduction Service No. ED 182 465). Grosof, M. S. & Sardy, H. (1985). A resaarch grimar far the agaial ang aahavigral aaianaaa. Orlando, FL: Academic Press. Heillig, R. J. (1983). Adglaaaant agicidal behaviar: A family ayatama mam-Ll. Ann Arbor, MI: UMI Research Press. Hendlin, H. (1987). Youth Suicide: A psychosocial perspective. Suicide ang Lifa Threataning Bahaviar, 12, 151-165. Hepworth, D. H., Farley O. W., & Griffiths, J. K. (1988, April). Clinical work with suicidal adolescents and their families. MM; Tha alggrnal gf Qantamggrary Sggial Wgrk, 195-203. Husain, S. A. (1990). Current perspective'on the role of psychosocial factors in adolescent suicide. Psychiatric Annals. 20(3), 122-127. Johnson, S. & Maile, L. (1987). icide n h h l . Springfield, IL: Charles C. Thomas. Kerfoot, M. J. (1980). The family context of adolescent suicidal behavior. Jggrnal af Aaalaaaanca, _2_, 365-370. Leder, J. (1987). Dead aarigga: A Izaak far taanagara aaggt taanaga agiaide. New York: Atheneum. Leigh, G. K. (1986). Adolescent involvement in family systems. In G. K. Leigh & G. W. Peterson, Agglaacanta in familiaa (pp. 38-72). Cincinnati: South-Western Publishing. 76 Livingston, S. E. (1989). Family structure and adolescent family roles: A comparison of suicidal and non-suicidal adolescents. Diaaertatien Anagraeta Internatienal, _5_1_(3), 1015A. (University Microfilms Order No. DA9012924). McIntire, M. et al. (1977). Recurrent adolescent suicidal behavior. Bedlam: .69. 605-608- McKenry, P. C., Tishler, C. L., & Kelley, C. (1982). Adolescent suicide. Clinical Pegiatries, 21(5), 266-270. Molin, R. S. (1986). Covert suicide and families of adolescents. Adeleaeenge, 21(81), 177-185. Neiger, B. L. & Hopkins, R. W. (1988). Adolescent suicide: Character traits of high-risk teenagers. Ageleaeence, 23(90), 467-475. Nelson, F. L., Farberow, N. L., & Litman, R. E. (1988). Youth suicide in California: Comparative study of perceived causes and interventions. Qemmgnily Mental Health Jegrnal, 23(1), 31-42. NoruSis, M. J. (1990). P P + tai i 4. fr h IBMP XT AT n P§[2. Chicago, IL: SPSS Inc. ' Olson, D. (1989). m l x m d I ff mil m Vlll- mil m n n in rv ni n. (Available from Family Social Science, University of Minnesota, 290 McNeal Hall, St. Paul, MN 55108 - Haworth Press) Olson, D. H., McCubbin, H. l., Barnes, H. L., Larsen, A. S., Muxen, M. J., & Wilson, M. A. (1989). Familiea: what makea them werk. Newbury Park, CA: Sage Publications. Olson, D. H., McCubbin, H. I., Barnes, H. L. Larsen, A. S. ., Muxen, M. J., & Wilson, M. A. (1985). F mil inv n lnv n ri in natienal survey at families acresa the family life eyele. St. Paul, MN: Family Social Science. Olson, D., Portnor, J., & LaVee, Y. (1985). EAQES lll. St. Paul, MN: University of Minnesota. Olson, D., Sprenkle, D., & Russell, C. (1979). Circumplex model of marital and family systems: I. Cohesion and adaptability dimensions, family type and clinical applications. Family Preeeaa, _1_8_, 3-28. Oster, G. D. & Caro, J. E. (1990). n r n in n r in r ageleaeenta and their familiea. New York: John Wiley & Sons. 77 Peterson, G. W. (1986). Family conceptual frameworks and adolescent development. In Leigh, G. K. & Peterson, G. W. Ageleaeenta in familiea. (pp. 13-35). Cincinnati, OH: South-Western. Pfeffer, C. R. (1990). Clinical perspectives on treatment of suicidal behavior among children and adolescents. Payehiatrie Annala, 20(3), 143-150. Pfeffer, C. R. (1981). The family system of suicidal children. Ameriean Jegrnal ef Payehetherapy, 15(3), 330-341. Reinherz, H. 2., Stewart-Berghauer, G., Pakiz, 8., Frost, A. K., Moeykens, B. A., & Holmes, W. M. (1989). The relationship of early risk and current mediators to depressive symptomatology in adOlescence. rnlfth Amri nA dem fhil Adl nPs hir, 25(6), 942-947. Richman, J. (1986). Family therapy fer aeicigal people. New York: Springer Publishing. Richman, J. (1979). Family therapy of attempted suicide. Family Preeeaa, 1.8. 131-142. Rosen, M. (1992, January 11). School reeling after 2 suicides. The 51, Peteraegrg Timefi. PD. 18, 68. Schiamberg, L. B. (1988). Chile ana adeleaeent develegment. New York: MacMillan Publishing. Simon, F. B., Stierlin, H., & Wynne, L. C. (1985). The langgage at family therapy: A ayatemie veeaeglary ang spurgebgek. New York: Family Process Press. Spector, R. D. (1988). The relationship between family dynamics and adolescents who attempt suicide. Diaaertaxien Agatraeta Internatignal, 19(8), 34588. (University Microfilms Order No. 8823037). Spirito, A., Overholser, J. & Stark, L. (1989). Common problems and coping strategies ll: Findings with adolescent suicide attempters. rn l fA n rmal hil P h l ,Agril, 213-221. Stiffman, A. R. (1989). Suicide attempts in runaway youths. MM Life-Threatening Behavier, 13(2), 147-159. Stivers, C. (1987). Adolescent suicide: An overview. Marriage ang Family Review, fl, 135-142. 78 Stivers, C. (1988). Parent-adolescent communication and its relationship to adolescent depression and suicide proneness. Adeleeeeece, 23(90), 291-295. The Gallup Organization, Inc. (1991, March). W Exeeetive summary. Princeton, NJ: The Gallup Organization. Touliatos, J., Perlmutter, B. F., & Straus, M. A. (1990). Hangeoek ef Eamily Meaaerement Techniguea. Newbury Park, CA: Sage Publications. U. S. Congress, Senate, Committee on Judiciary, & Subcommittee on Juvenile Justice. (1984a, October). Hearings on teenage suicide. Testimony by A. L. Berman on behalf of the American Psychological Association. Vital Statistics of the United States. (1987). Volume Il-Mortality, Part A. U.S. Department of Health and Human Services. Vital Statistics of the United States. (1988). Volume ll-Mortality, Part B. U.S. Department of Health and Human Services. Wodarski, J. S. & Harris, P. (1987, November - December). Adolescent suicide: A review of influences and the means for prevention. Mal Wgrk, 477-484. APPENDICES APPENDIX A COPIES OF INSTRUMENTATION AND INSTRUCTION 79 QUESTIONNAIRE PACKET 80 When reading the following instructions and questions please take your time and read them carefully. Answer each question as honestly and carefully as you can. There is no "right" or "wrong" answer. If you have a problem with answering any of the questions, please put your hand up and wait quietly until the researcher comes to assist you. 81 INSTRUCTIONS FOR THE FIRST QUESTIONNAIRE This is the first part of the questionnaire. The first part of this questionnaire consists of 40 questions. It is numbered 1 through 40. Please make certain that the question number is the same as the number on the computer card. You will note that the 3 answers you have to choose from are "1 - yes, usually", "2 - no seldom", and "3 - sometimes". When answering questions 1 through 40 use bubbles marked 1, 2, or 3 for your answer. Please mark only one answer per question. For example: The question states, "Do you eat breakfast?" If you normally eat breakfast, you would fill in "1 - yes, usually". If you rarely eat breakfast, you would fill in "2 - no seldom". If you only eat breakfast sometimes, you would fill in "3 - Sometimes". Remember, only fill in 1 choice per question. It is up to you to decide which one best answer describes you. L“:(D ., 21* PLEASE NOTE Copyrighted materials in this document have not been filmed at the request of the author. They are available for consultation, however, in the author’s university library. Appendix B, Parent-Adolescent Communication Inventory, 82—85 Appendix C, Faces III Instrument, 87—88 University Microfilms International 82 Parent-Adolescent Communication Inventory (PACI) by Millard J. Bienvenu, copyright 1969. Questions 1 to 40 ask questions about your parents, mother and father. If you do not live with or have contact with either your mother or father, think of someone in your family who would take either your mother’s or father’s place and answer the questions in that manner (eg., mother’s boyfriend or your grandfather who would take your father’s place in the family). YES NO SOME- usually seldom TIMES (1) (2) (3) 1. Is family conversation easy and pleasant at mealtimes? 2. Do your parents wait until you are through talking before "having their say"? 3. Do you pretend you are listening to them when actually you have tuned them out? _ 4. Does your mother tend to lecture and preach too much to you? 5. Does your family have good times? 6. Is it hard for you to respect your parents opinion? 7. Do your parents ever laugh at you or make fun of you? 8. Does your mother wish you were a different kind of person? 9. Do your parents think that you are bad? CONTINUE WITH QUESTIONS ON THE NEXT PAGE 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 83 Does your family talk things over with each other? Do your parents discuss personal problems with you? Does your father wish you were a different kind of person? Do you talk to your parents in a disrespectful manner? Do your parents show an interest in your interests and activities? Do you discuss personal problems with your parents? Is it hard for your mother to praise and say nice things to you? Do your parents ask your opinion in deciding how much spending money you should have? Do your parents try to discuss sex with you? Is it hard for your parents to trust you? Do you help your parents to understand you by saying how you think and feel about things? Is it hard for your father to praise and say nice things to you? Does your mother have confidence in your abilities? CONTINUE WITH QUESTIONS ON THE NEXT PAGE 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 84 Are your parents sarcastic towards you? Is it hard for your father to trust you? Does your father have confidence in your abilities? When a difference arises are you and your parents able to discuss it together (in a calm manner)? Do your parents consider your ideas in making family decisions? Does your father criticize you too much? Is it hard for your mother to see your side of things? Do your parents allow you to get angry and blow off steam? Do your parents consider your opinions in making decisions which concern you? Does your mother criticize you too much? Do you find their tone of voice irritating? Do they try to make you feel better when you are "down in the dumps"? Do you really try to see your parents side of things? Do your parents encourage you to talk about your problems? Is it hard for your father to see your side of things? CONTINUE WITH QUESTIONS ON THE NEXT PAGE 85 38. Does your father tend to lecture and preach too much? 39. Do you accept their reasons for decisions they make concerning you? 40. In talking with your parents, is it hard to say what you really feel? CONTINUE WITH QUESTIONS ON THE NEXT QUESTIONNAIRE .. LI‘IHIr APPENDIX C LETTERS OF PERMISSION 86 INSTRUCTIONS FOR THE SECOND QUESTIONNAIRE This is the second part of the questionnaire. The second part of this questionnaire consists of 20 questions. It is numbered 41 through 60. Please make certain that the question number is the same as the number on the computer card. You will note that the 5 answers you have to choose from are "1 - Almost Never", "2 - Once in Awhile", "3 - Sometimes", "4 - Frequently", and "5 - Almost Always". When answering questions 41 through 60 use bubbles marked 1, 2, 3, 4, and 5 for your answer. Please mark only one answer per question. It is up to you to decide which one best answer describes you. 87 FACES IH Instrument, by David H. Olson, Joyce Portner, and Yoav Lavee. Family Social Science, 290 McNeal Hall, University of Minnesota, St. Paul, MN 55108, copyright 1985. l 2 3 4 5 Almost Never Once in Awhile Sometimes Frequently Almost Always DESCRIBE YOUR FAMILY NOW: 41. Family members ask each other for help. 42. In solving problems, the children’s suggestions are followed. 43. We approve of each other’s friends. 44. Children have a say in their discipline. 45. We like to do things with just our immediate family. 46. Different persons act as leaders in our family. 47. Family members feel closer to other family members than to people outside the family. 48. Our family changes its way of handling tasks. 49. Family members like to spend free time with each other. 50. Parent(s) and children discuss punishment together. 51. Family members feel very close to each other. 52. The children make the decisions in our family. 53. When our family gets together for activities, everybody is present. 54. Rules change in our family. 55. We can easily think of things to do together as a family. 56. We shift household responsibilities from person to person. CONTINUE WITH QUESTIONS ON THE NEXT PAGE 88 57. Family members consult other family members on their decisions. 58. It is hard to identify the leader(s) in our family. 59. Family togetherness is very important. 60. It is hard to tell who does which household chores. CONTINUE WITH QUESTIONS ON THE NEXT QUESTIONNAIRE 89 INSTRUCTIONS FOR THE THIRD QUESTIONNAIRE This is the third part of the questionnaire. The third part of this questionnaire consists of 10 questions. It is numbered 61 through 70. Please make certain that the question number is the same as the number on the computer card. For questions numbered 61, 62, and 63 please fill in one answer per question. For questions 64 through 69 you may mark as many answers as is necessary. For Example: The question reads, "Who is living with you at home?" If you live with your Mother, brothers, sister, and mother’s boyfriend, then for question 64 you would fill in bubbles "1", "3", and "4". You would then leave questions 65, 66, and 67 blank. For question 68, you would fill in bubble "1". Question 69 would be left blank. 90 Please answer the questions below on the computer card as honestly as you can concerning you and your family. 61. What is your gender, male or female? Answer "1" for male and "2" for female on the computer card 62. Please indicate on your computer card the best choice that describes your race. If none of the choices below apply to you, please leave this question blank. 1. White 2. Black 3. Hispanic 4. Native American 5. Asian 63. Have you seriously considered committing suicide in the past year? Answer "1" for YES and "2" for NO on the computer card CONTINUE WITH QUESTIONS ON THE NEXT PAGE 91 FOR QUESTIONS 64 TO 69, BUBBLE IN ON YOUR COMPUTER CARD ALL THOSE THAT APPLY. EEO IS LIVING WITH YOU AT HOME: 64. 1. Mother 2. Father 3. Brother(s) 4. sister(s) 65. 1. Grandmother(s) 2. Grandfather(s) 3. Aunt(s) 4. Uncle(s) 66. 1. Stepfather 2. Stepmother 3. Stepbrother(s) 4. Stepsister(s) 67. 1. Half brother(s) 2. Half sister(s) 3. Female cousin(s) 4. Male cousin(s) 68. 1. Mother’s boyfriend 2. Father's girlfriend 3. Mother's boyfriend’s children 4. Father’s girlfriend's children 69. 1. Female friend(s) of the family 2. Male friend(s) of the family 3. Other people not listed 70. What is your age? To answer this question, turn the computer card over (where questions 1 through 56 were answered). Underneath the area labelled "SECTION", bubble in your age in the first two columns starting from the left side. Please leave the last column blank on the right side. THANK YOU FOR TAKING THE TIME TO ANSWER THESE QUESTIONS IN AN HONEST MANNER. APPENDIX B PARENT AND ADOLESCENT CONSENT FORMS 92 Parental Consent Form I, , agree to allow my son/daughter, , to participate in the research project being conducted on the structure and communication within the family. I understand that the project requires my son/daughter to fill out three questionnaires: (1) one on family structure; (2) one on family communication patterns; and (3) one short demographic questionnaire, with one question pertaining to thoughts on suicide. The project will take my son/daughter 30 to 45 minutes to complete. Participation in this project is voluntary and I understand that my son/daughter may withdraw at any time. I also understand that all the information given to the researcher will be kept confidential. My name or my son/daughter’s name will never be used when this research is being conducted or discussed. This project is being conducted through the Department of Family and Child Ecology at Michigan State University. Kathleen Zawacki is the principal investigator. If I have any questions concerning this project, I will contact her at (517)-882-4064. Parent’s signature Date 93 Adolescent Consent Form I, , agree to participate in the research project on family structure and family communication being conducted through the Department of Family and Child Ecology at Michigan State University with Kathy Zawacki as the principal investigator. I have chosen to participate freely and voluntarily. The purpose of the project and its procedures have been explained to me. I understand that I will be answering questions about my family and the way we interact and communicate. It will take me approximately 30 to 45 minutes to complete these questionnaires. I understand that I may withdraw from participation in this project at any time. I also understand that any information that I provide will be kept confidential. My name will never be used when conducting the research or discussing the research. If I have any questions about this research project, I understand that I can contact Kathy Zawacki to answer my questions. I have read and understood this consent form. This sheet will be kept separate from the questionnaires I will be filling out. Participant’s signature Date 94- RIVENDELL PSYCHIATRIC CENTER .AIfinnhkytknmmujbrAdbkmxnu April 29, 1991 Kathleen C. Zawacki Graduate Student, M.S.U. 2513 Skye Road Lansing, MI. 48911 Dear Kathleen: I an writing to infora you Rivendell of Michigan has agreed to allow you to conduct your research at our psychiatric facility. As we discussed, on April 29th, your first visit will be here on Friday, May 17th at 11:00 a.n. At that time we will begin planning the actual implementation of your testing/research proposal. I look forward to meeting you then. Sincerely, ('12 z/ / M4621“) Karen Gallag er Patient Rights Advocate KG/bb C: Mike Crosby Jeanne Ewing I01 We“ Townsend Road 0 St. Johns. Michigan 48879 0 (517) 224-1177 0 Fax (5”) 2244280 95 MCLAREN REGION MEDICAL CENTER July 2, 1991 Kathleen G. Zawacki 2513 Sky Road Lansing, MI 48911 Dear Ms. Zawacki: The Institutional Review Committee and the Executive Committee at McLaren Regional Medical Center have approved the protocol you submitted entitled 'Comparing Suicidal and Non-Suicidal Adolescents' Beliefs Concerning Family Structure and Parent-Adolescent Communication. It is their understanding that the study will be under the direction of Dr. Kang Kwon, Medical Director of the Adolescent Psychiatric Unit. 1 have also sent a copy of the protocol to Dr. Tai Kang, Chairman of the Department of Psychiatry, at McLaren Regional Medical Center for his information. If you have any questions, please contact me through Sharon Davis, Medical Staff Secretary, Robert 1.. Cross, M.D. Chairman lsd 401 South 3.1km Highway 0 Flint. Michigan 48532-3685 0 (313) 762.2000 MWTdekukM-MUdmCdkpeflwm 96 Bath Community (Schools "Where Children Count" CflbccfmcdquMcmtmt DO. box 139 DuhlfimgmifiBOO (517) 6416721 June 21, 1991 Ms. Kathleen Zawacki 2513 Skye Road Lansing, Michigan 48911 Dear Ms. Zawacki: 1 am writing this letter to inform you that Mrs. Donna Six is willing to allow you to present your graduate survey to her psychology classes at the beginning of the 1991—92 school year. I understand that you need a letter to present to your ucrhis committee to verify approval for presentation of your survey to Bath School's psycho1ogy classes. This letter should serve as your verification. If 1 can be of further help to you or your committee, please do not hesitate to contact me. Sincerely: é§;:::7Sampson, Principal Bath High Sshool ES:Js “WW muscle-n1 MEN manna? maunue hhcumfluunun mmnugnasx bunny-aux noses» anuwm anuwm hasapuem mans-n: 97 RRRY RBLIC SCHOOLS 277$II1TTOS IOAD°PEIIY.NICHIGA.\ 1811‘} WWW PERRY HIGH SCHOOL 517/625-3106 SHAFI'SHI'RG El£\1E\'1’.ARY 517/625-3101 PERRY MIDDLE SCHOOL 517/625~6196 it? ‘h‘S-Sl Ii VAX 517/625-6256 W 6' 1991 Kathy Zawacki 2513 Skye Rd. Lansing. MI 4891l Dear Kathy. This letter is to confirm that Perry High School will assist in yourreseardipsojectsbyadvisingywtouseselectedstudents as participants. This permission is based on approval by the university and permission free the parents of those students involved. his project will occur the end of Septuber of 1991. If you have any quotients. please contact as at 625—3104. 8 y. 1 .. j “gal/h ((ip'l'LL David M. Oegena. PhD. Principal Perry High School IND/ea Sl'PEthTE'A‘DEVT PIMPAIS BOARD Of EDUCATION Jacldyn Hut-d David Oegema, High .fdwl bbert .Ashley. herd-u! _ Gerald Biennium? High Sdaal Ted jagmin. limhw'dml gem: CLI’LUM (1101101le Donald Vernon. Judd: ideal Catherine Hahn. W' ‘ Karen Milton. Run-Elaamkrn' Dan Peabody. Tm oommnmunon Cheryl mwmm' HerbenSmlth.M Mtduel Shaft Gerald loses. hm James Nutter. 11m 98 [Will 111111111111 Willi ”a: my... Timothy Vbung - Principal Assisi-u Principals: Patricia Milbourn 0 Sandra Hargrova June 26, 1991 Ms. Kathleen 6. Zawacki 2513 Shya Road Lansing, MI 48911 Dear Ms. Zawacki: l have reviewed the materials you sent as regarding your research project on suicidal behavior. The aaterials seen appropriate for students in grades 9 through 12. Mason High School will participate in your research next Fall. Plefse contact as in aid-August to finalize the de- tai s. Sincerely. //.: (0» Tie Y Princip Mason High School 1001 South Barnes Street . Mason. Michigan 48854 - 15171 616 - 9055 99 m. UNIVERSITY OF MINNESOT rainy Boast Science memes - 290MeNaall-Ul tsasautordAvama t SLPOULMSSIOU (612) 625-7250 PERMISSION TO USE FACES III I am pleased to give you permission to use FACES 111 in your research project. teaching. or clinical work with couples and families. You can either duplicate the materials directly or have them retyped for use in a new format. If they are retypcd. acknowledgement should be given regarding the name of the instrument. the developer's name. and the University of Minnesota. In exchange for providing this permission. we would appreciate a copy of any papers. thesis. or reports that you compete using these inventories. This will .zcip us in staying abreast of the most recent development and research With these scales. Thank you t‘ or your cooperation. in closing. 1 hope you find FACES 111 of value in your work with couples and families. 1 would appreciate hearing from you as you make use of this inventory. Sincerely? ’ Md H. Olson. Ph.D. Professor DHmew =AMtLY tNVENTORiES PROJECT mm film: Davie H. Olson. PhD. APPENDIX D UCRIHS PERMISSION LETTER 100 MICHIGAN STATE UNIVERSITY OWICEOI‘VICEHESIDENTI’OIIBEAICH EASILANSINGOIICHIGANOdlfle-l“ AND DEAN Of 1'“! GRADUATE SCHOOL August 1h, 1991 Kathleen Zawacki 2513 Skye Road Lansing. MI a89ll RE: COMPARING SUICIDAL AND NON-SUICIDAL.ADOLESCENTS' BELIEFS CONCERNING FAMILY STRUCTURE AND PARENT-ADOLESCENT COMMUNICATION, 1R3 ‘91-363 Dear Ms. Zawacki: UCRIHS' review of the above referenced project has now been completed. I am pleased to advise that the rights and welfare of the human subjects appear to be adequately protected and the Committee, therefore. approved this project at it meeting on August 5, 1991. You are reminded that UCRIHS approval is valid for one calendar year. If you plan to continue this project beyond one year. please make provisions for obtaining appropriate UCRIHS approval one month prior to August 5. 1992. Any changes in procedures involving human subjects must be reviewed by the UCRIHS prior to initiation of the change. UCRIHS must also be notified promptly of any problems (unexpected side effects, complaints, etc.) involving human subjects during the course of the work. Thank you for bringing this project to our attention. If we can be of any future help, please do not hesitate to let us know. ely, ‘**‘£;) é:;-. (J\;) '\| David 3. "right. Ph.D.. Chair University Committee on Resea Involving Human Subjects (UCRIHS) DEV/den cc: Dr. Lawrence Schiamberg MSU 'm . Mfmiar Action/Equal Opportunity Institution UNIV. LI RRRIES iljnnuulmlwfilm HICHIGQN mu Will 3129 10554446