Comparing Behavior i Running head: ASSESSING AGGRESSION LEVELS OF FIRST-TIME FATHERS Comparing Behavior Assessment Measures with Behavior Specific Responses to Assess Aggression in First-Time Expectant Fathers

Thirty-four randomly recruited first-time expectant fathers, ages 20 years to 40 years, were administered a battery of survey questions each trimester of their wives’ pregnancy. This study compared Behavior Assessment Measures (Clinical Anxiety Scale, Selfism-Scale, and the Index of Self-Esteem) to Behavior Specific Measures (Non-Physical Abuse of Partner Scale and the Aggression Inventory) to see if any correlations exist among the test data. The study is trying to identify specific test or survey questions that measure attitudinal or behavioral changes in first-time expectant fathers over the nine months of their spouse’s pregnancy. The results indicated a statistically significant correlation between self-esteem and narcissism, non-physical abuse and aggression, and non-physical abuse and anxiety. The Self-Esteem measures appear to have the highest correlations to determine attitudinal changes of first-time expectant fathers. Comparing Behavior 1 Comparing Behavior Assessment Measures with Behavior Specific Responses to Assess Aggression in First-Time Expectant Fathers Until recently, pregnancy was viewed almost exclusively in terms of the psychological, social, and physical changes of the female during pregnancy. Little empirical research has been conducted on expectant fathers to determine if there are attitudinal or behavioral changes during their spouse’s pregnancy. Only a few investigators have reported research that specifically addresses the nature of pregnancy-related symptoms in the prospective fathers. This study will compare aggressive behaviors to non-aggressive behaviors in first-time expectant fathers. Sigmund Freud hypothesized that all humans possessed an aggressive drive from birth, which, together with the sexual drive, contributed to personality development, and ultimately found expression in behavior (Archer & Brown, 2000). Austrian ethologist Konrad Lorenze suggested that aggression was innate, an inherited fighting instinct, as significant in humans as it was in other animals. He contended that the suppression of aggressive instincts, common among human societies, allows these instincts the chance to build up, occasionally to the point where they are released during instances of explosive violence. Many psychoanalysts see aggression as a primary drive, offering the possibility that aggression may be a reaction to frustration of primary needs. Other factors, including learning difficulties, minimal brain damage, brain abnormalities, such as temporal lobe epilepsy, and social factors such as crowding and poverty, have been suggested as contributing factors in certain cases of exaggeratedly aggressive behavior. Psychological investigation into aggressive behavior continues, with significant corrolary studies being performed in endocrinology and in primate research to determine whether hormonal imbalances have an impact on behavior. Each theory may be accurate in part, since aggression Comparing Behavior 2 is believed to have a number of determining factors. Despite the commonality of determinants, there are significant differences between partner violence and other types of violent behavior. Most violence outside the family involves individuals with limited personal contact; partner aggression involves individuals who have an intense, continuing interpersonal relationship. The interpersonal relationship of individuals involved in partner violence usually includes an emotional relationship of attachment, emotional and sexual intimacy, or dependency between partners such that the physical and sexual violence occurs within an intimate relationship context. The relationship context includes a history of prior relationship behavior and expectations and goals for the relationship (West, 2000). Aggression toward a partner includes both physical and psychological aggression. Psychological aggression refers to behavior that is offensive or degrading to the partner usually involving verbal behaviors, such as threats or insults, as well as actions, such as damage to personal property. This behavior, sometimes termed emotional abuse, is usually present in violent couples and has been reported by many women, as having a more severe impact on them than physical aggression (Follingstad, Rutledge, Berg, Hause, & Polck, 1990). Most men and women who exhibit physical aggression toward partners are located at the low end of the severity/frequency continuum of partner aggression and engage in infrequent and minor acts of physical aggression such as pushing, slapping, shoving and hitting on less injurable body parts (i.e. shoulder rather than face). However, a large majority of seriously violent perpetrators are males and victims are females, demonstrated by studies completed by Reiss & Roth, (1993-1994) of individuals who engage in or are the victims of more serious partner violence. Typically, 80 to 90 percent of perpetrators are males. This finding mirrors results of Comparing Behavior 3 serious criminal behavior in which 90 percent of arrestees for violent crimes, including murder, rape, and aggravated assaults, are males. In an extensive analysis of this issue, Morse (1995) reports that women engage in as much minor and moderate acts of physical aggression as do males, but males engage in a much higher proportion of serious aggression than do females. Thus, 2 percent to 4 percent of males report beating up their partners on average less than once a year. Moreover, women were two to four times as likely to require medical care for an injury as were males. The Pulido (2001) study on Family Violence, held during the period of July 1, 1998, through June 30, 1999, found that out of 858 patients screened, 46 women admitted to be currently involved in a domestic violence relationship and requested assistance. Twenty of these women disclosed physical abuse such as hitting and beating and hair pulling. Twenty-six women described emotional abuse, in which threatening and degrading remarks were experienced; the women’s feelings of inadequacy often were exacerbated by their lack of control over the allocations of money earned and the inability to leave because of the partner’s control over all financial resources. Fifteen women stated that they knew a close friend or family member who was a victim of domestic violence and accepted information on their behalf. This response also could be interpreted as a hesitancy to identify themselves as victims at the time, as a result of their fear of the batterer. Domestic violence is a major health concern in the United States. The American College of Obstetricians and Gynecologists (1998) defines domestic violence as a pattern of assaultive and coercive behaviors including physical, sexual, and psychological attacks, as well as economic coercion, used against current or former intimate partners. Violence does not stop Comparing Behavior 4 during pregnancy. Norton, Peipert, Ziegler, Lima, and Hume (1995) and Jones and Horan (1997) noted that an estimated 7 percent to 17 percent of those screened for domestic violence admitted to abuse during pregnancy. Many providers of obstetric care do not screen every patient, so many domestic violence victims are undetected. Violence during pregnancy may be even more harmful, since it poses a significant additional threat to the fetus. Muhajarien and Nazeem (1999) report that physical abuse during pregnancy increases the risk of miscarrage, abruptio placentae, preterm labour and delivery, fetal fractures and low birth weight. Other adverse consequences for the woman may include rupture of the uterus, liver or spleen, antepartum hemorrhage and pelvic fractures. They found the associations between physical abuse and perceived stress and negative life events reflect the generally unfavorable conditions in which abused women live. The association between stress and physical abuse, where stress was measured both as a generalized global measure and in relation to specific negative events, indicates the profound psychological implications of abuse on women. Bibring (1959) reported that before becoming a father for the first time, a man cares only for himself, and the closest he gets to an individual is his affection for his mate. However, his relationship to his wife changes with the pregnancy and his new parenthood. His wife is now the mother of his child. Together, they will be responsible for raising this child who will be an individual and this child will represent them both. The father and the child will have a special bond, encompassing identifications, concerns, and struggles that over time will widen. During this time in the father’s life, he will need to resolve persistent conflictual relationships he may have had with his own parents. Not all men may go through the same experiences during this Comparing Behavior 5 time; however, it appears tha t all men, especially first-time fathers, undergo considerable behavioral shifts and internal instability, and not many men are the same as they were before the birth of this child. Most men experience significant changes in their sense of responsibility, their relationships with their spouses, their attitudes toward this new child, and their feelings about who they are. Strickland (1987) stated that anxiety related to the additional emotional stress of parenthood was a significant predictor indicating that physical symptoms are evident in fathersto-be; this anxiety is related to additional emotional stress. She also found anxiety to be a positive predictor of hostility and psychological symptoms or measures of emotional well being. In a 1972 study, Trethowan found that fathers-to-be may repress or unconsciously conceal hostility, therefore increasing the expression of symptoms. Although not recent, several studies looking at the development of somatic symptoms in fathers during their wives’ pregnancies have centered on psychological factors connected to symptom formation. For example, Curtis (1955) studied fifty-five fathers-to-be and found that seventeen exhibited serious emotional problems, fourteen exhibited minor problems, and twentyfour with no obvious problems. Among the seventeen more disturbed fathers, psychosomatic symptoms were the prevalent cause of marital stress in many families. These men did not associate their problems with expectant fatherhood. For some time, it has been suggested that expectant fathers are afflicted by physical symptoms similar to those experienced by women during pregnancy. This phenomenon has been dubbed Couvades Syndrome. Couvade consists of the male simulation of his mate’s childbirth labor and delivery and the observance of certain proscribed dietary restrictions by the father Comparing Behavior 6 during the postnatal period (Strickland, 1986). The social background of an expectant father may also influence his proclivity toward symptom manifestation. Less well-educated expectant fathers seem to report more symptoms and seek care for couvadelike symptoms more often than the more highly educated (Trethowan, 1972). Strickland (1986) found that planning a pregnancy, social class, and racial background are associated with symptoms manifestation in expectant fathers. Men faced with unplanned pregnancies had significantly more symptoms than those with planned pregnancies. Working class men reported significantly more somatic and psychological symptoms than middle-class men during the entire course of pregnancy and in each stage of pregnancy, except in late pregnancy where the higher mean number of psychological symptoms reported by working-class men was not statistically significant. African-American respondents consistently reported more symptoms than Asian respondents, differences were more marked in early pregnancy and decreased significantly over time in relation to Asian expectant fathers who reported increases in symptoms over time as delivery became more imminent. The differences between AfricanAmerican and Asian expectant fathers’ reported symptoms support the idea of cultural differentiation in expectant fathers’ responses to pregnancy only if the assumption is correct that the African-American subculture in the United States has maintained cultural traits of ancestors. If symptom manifestation by expectant fathers is a response to anxiety and concern about pregnancy, the differences in trends between African-American and Asian respondents may reflect the points during pregnancy when these groups of expectant fathers are more likely to translate their anxieties into symptoms. A study conducted by Morse (1999) at the University of Melbourne paid particular attention to how pregnancy impacted the fathers. She found that up to 15% of the men Comparing Behavior 7 in her study suffered depression and anxiety from the middle stage of pregnancy upwards, but, four month after birth, depression and anxiety in men had reduced to 6%. In comparison, up to 20% of women studies also suffered depression during pregnancy, which reduced to about 12% after birth. Morse reported the cause of depression and anxiety was due mostly to the age, emotional state and family background of the fathers, with younger fathers and those with poor family relationships more likely to suffer. Within the context of a court psychiatric clinic, Hartman and Nicolay (1966), found that fathers-to-be committed crimes of a sexual nature such as exhibitionism, voyeurism, and rape more frequently than any other types of crime. In their clinic, they found that expectant fathers had a common hyper masculine facade. From the research, it appears that stress factors in expectant fathers can have a tremendous impact on their ability to adapt to fatherhood. Early identification of an expectant father’s behavioral profile might help determine whether the father will easily adapt as a parent and a husband or whether his attitudinal or behavioral changes will affect his ability to adapt. If it can be determined during the pregnancy what might trigger these aggressive and violent behaviors in expectant fathers, it may be possible to address issues that cause stress on the family’s future. Hence, there is good reason to study behavior of first-time expectant fathers. Hypotheses This study compared Behavioral Assessment Measures with reported Behavior Specific responses to see if aggression in first time fathers-to-be can be assessed with these instruments. Therefore, it is hypothesized that the Clinical Anxiety Scale (CAS), Selfism (NS), and Index of Self-Exteem (ISE) measures will show a statistically significant correlation when compared with the Non-Physical Abuse of Partner Scale (NPAPS) and Aggression Inventory (AI) measures. Comparing Behavior 8 Method Participants The subjects in this study were comprised of 34 married, first-time fathers-to-be, between the ages of 20 and 40. The participants in this research were recruited from the Upper Ohio Valley (West Virginia, Ohio, and Western Pennsylvania). Each participant was interviewed at the initial session and tested each trimester during the course of the pregnancy. Participation was voluntary and the subjects were not financially or otherwise rewarded for their participation. The participants were unknown to the examiner prior to the study. Instrumentation The Behavioral Assessment Measures that each participant completed were the California Psychological Inventory (CPI), the Clinical Anxiety Scale (CAS), Selfism (NS), and the Index of Self-Esteem (ISE). The Specific Behavior Measures included the Non-Physical Abuse of Partner Scale (NPAPS) and the Aggression Inventory (AI). The Attitude Measure assessed was the Love Attitudes Scale (LAS). The Relationship Assessment Scale (RAS) and the Index of Marital Satisfaction (IMS) were administered to measure Relationships. The Multidimensional Scale of Perceived Social Support (MSPSS) was given to measure Social Support Perception. The current study looked for relationships between the Behavior Assessment Measures (CAS, NS, ISE) and the Behavior Specific Measures (NPAPS and AI). Clinical Anxiety Scale (CAS) The Clinical Anxiety Scale (CAS) is a 25 item scale that focuses on measuring the amount, degree, or severity of clinical anxiety reported by the subject, with higher scores indicating higher amounts of anxiety. The CAS is simply worded and easy to administer, score, Comparing Behavior 9 and interpret. The items for the CAS were psychometrically derived from a larger number of items based on the criteria for anxiety disorders in DSM III. The CAS has a clinical cutting score of 30 (+/-5). This instrument is particularly useful for measuring general anxiety in clinical practice. The CAS has excellent internal consistency, with a coefficient alpha of .94. The standard error of measurement (SEM) of 4.2 is relatively low, suggesting a minimal amount of measurement error. The CAS had good stability, with two-week test-retest correlations that range from .64 to .74. The CAS has good group validity, discriminating significantly between groups known to be suffering from anxiety and lower-anxiety control groups. Using the clinical cutting score of 30, the CAS had a very low error rate of 6.9% in distinguishing between anxiety and control groups. Analysis of the CAS in relation to demographic variables such as age, sex, and education reveals that scores on the CAS are not affected by those factors. Selfism (NS) The Selfism (NS) is a 28 item scale designed to measure narcissism, referred to by developers of this instrument as selfism. Selfism is viewed as an orientation, belief, or set affecting how one construes a whole range of situations that deal with the satisfaction of needs. A person who scores high on the NS views a large number of situations in a selfish or egocentric fashion. At the opposite end of the continuum are individuals who submerge their own satisfaction in favor of others. The NS samples beliefs across a broad range of situations and is not targeted toward a specific need area. Based on a review of the literature, impressionistic sources, and the work of cultural observers, the original 100 items were narrowed down to 28 Comparing Behavior 10 based on low correlations with the Marlow-Crowne Social Desirability Scale, high correlations with NS total scores, and a reasonable spread over the five response categories. The NS has very good internal consistency, with split-half reliabilities of .84 for males and .83 for females. The NS also has excellent stability, with a four-week test-retest correlation of .91. The NS has fair concurrent validity, correlating significantly with the Narcissistic Personality Inventory and the Religious Attitude Scale. Also, the NS demonstrated a form of known-groups validity by correlating positively with observers’ judgements of their close friends’ narcissistic characteristics. The NS also distinguished between respondents who were high and low on cynicism regarding the motive of individuals in need of help. Index of Self Esteem (ISE) The ISE is a 25 item scale designed to measure the degree, seve rity, or magnitude of problem the subject has with self-esteem. Self-esteem is considered as the evaluative component of self-concept. The ISE is written in very basic language, is easily administered, and easily scored. Because problems with self-esteem are often central to social and psychological difficulties, this instrument has a wide range of utility for a number of clinical problems. The ISE had two cutting scores. The first score is 30 (+/-5): Scores below this point indicating absence of a clinically significant problem in this area. Scores above 30 suggest the presence of a clinically significant problem. The second cutting score is 70. Scores above this point nearly always indicate that clients are experiencing severe stress with a clear possibility that some type of violence could be considered or used to deal with problems. The ISE has a mean alpha of .93, indicating excellent internal consistency, and an excellent (low) standard error of measure (SEM) of 3.70. The ISE also has short-term Comparing Behavior 11 stability with a two-hour test-retest correlation of .92. The ISE has a good known-groups validity, significantly distinguishing between clients judged by clinicians to and not to have problems in the area of self-esteem. Further, the ISE has very good construct validity, correlating poorly with measures with which it should not and correlating well with a range of other measures with which it should correlate highly: depression, happiness, sense of identity, and scores on the Generalized Contentment Scale. Non-Physical Abuse of Partner Scale (NPAPS) The Non-Physical Abuse of Partner Scale (NPAPS) is a 25 item instrument that is designed to measure the degree or magnitude of perceived non-physical abuse that clients report they have inflected on a spouse or partner. The NPAPS was developed for use with heterosexual or homosexual couples who are dating or living together as married or unmarried couples. This scale is one of the few to examine perceptions of the abuser as to the amount of abuse he or she perceives as inflicting. As such, it can be very useful as a device for tracking the abuser’s perception over time during an intervention program. The NPAPS has excellent internal consistency, with an alpha in excess of .90. The NPAPS is reported as having good content and factorial validity, as well as beginning evidence of construct validity. Aggression Inventory (AI) The Aggression Inventory (AI) is a 30 item instrument designed to measure different aggressive traits. Subjects rate the items on a five-point scale, ranging from “does not apply to me” to “applies exactly to me.” The AI consists of four subscales: physical aggression (PA), verbal aggression (VA), impulsive/impatient (II), and avoidance (A). Because of possible Comparing Behavior 12 gender differences in many aspects of aggression, scores on the AI must be considered separately for women and men. The AI has fair to good internal consistency. For men the alpha coefficients were PA = .82; VA= .81; II = .76, and A = .70. The validity of the AI subscale has been supported by factor analysis and differences between men and women. The latter serves to suggest the AI has fair known group validity where men and women significantly differed on each subscale and on all but six of the individual items. Procedure This project is part of a large-scale study designed by Dr. Robert Rodriquez, Ph. D. It was divided into sections and conducted by a cohort group of 13 Marshall University graduate students. A longitudinal study of 34 expectant fathers was conducted during early, middle, and late pregnancy. Fathers who initially agreed to participate in this research were recruited from the Upper Ohio Valley. Potential participants were contacted at home, given an explanation of the study; a quick screening was conducted to see if they were eligible for the study, and then were asked to participate. Each man in the study was given a demographic data questionnaire that gathered information and a general behavioral history (See Appendix C). During session one (first trimester) and session three (third trimester) the following set of tests were administered: CPI, CAS, NS, and ISE. Within a three day span the next set of tests were administered: NPAPS, and AI. The subjects were given a 15 minute break and the last set of tests was administered: LAS, RAS, IMS, and MSPSS. During the second trimester, each participant was administered: CAS, NS, ISE, and LAS. The subjects were given a 15 minute break and the following tests were administered: Comparing Behavior 13 RAS, IMS, and MSPSS. All subjects were assigned double codes to assure confidentiality and anonymity. All data and information was centrally controlled and analyzed. The subjects agreed that information gathered could be shared by other cohort members and the university committee. The examiner was to avoid counseling at any time during the interview and testing protocol. This study compared the Behavior Assessment Measures: CAS, NS, and ISE to the Behavior Specific Measure; NPAPS and AI to see if any correlation coefficients between the test data occur. Both parametric and non parametric test were utilized to ensure validity across the testing spectrum. The Non-parametric test that was administered was the Wilcoxon, Paired Sample T-Test. This test was used to look at the individual measures. The parametric test of Pearsons r was also used to see if a correlation existed between the individual test measure. Together, these test were applied to each question and attribute area to discern any significant level of differences. The compilation of this data will enhance the summation of results to support the proposed theory. Results All analyses were done using SPSS software (SPSS Incorporated, 1999). Paired t tests were used to see if attitudinal or behavioral changes occurred in the fathers over the three trimesters of the pregnancy. The total scores of each individual test (CAS, NS, ISE, NPAPS, and AI) for each trimester were entered into the computer; no significant changes were found over time on any of the individual test. If a significant statistical correlation had been found, a factorial Analysis of Variance over time would have been run. A t test for related means found that there was no significant difference in husband’s narcissism across the three trimesters of pregnancy. The first related means t test compared the Comparing Behavior 15 difference in the first and second trimesters and found no significant difference in narcissism t(29)=.15, p=.88 (two-tailed). There was also no significant difference between trimesters two and three t(29)=1.63, p=.12 (two-tailed); and between trimesters one and three t(29)=1.36, p=.19 (two-tailed) (See Appendix A). A t test for related means found that there was no significant difference in husbands’ anxiety across the three trimesters of pregnancy. The first related means t test compared the difference in the first and second trimesters and found no significant difference in anxiety t(33)=1.35, p=.19 (two-tailed). There was also no significant difference between trimesters two and three t(33)=-1.406, p=.17 (two-tailed); and between trimesters one and three t(33)=-.64, p=.53 (two-tailed) (See Appendix A). A t test for related means found that there was no significant difference in husbands’ self—esteem across the three trimesters of pregnancy. The first related means t test compared the difference in the first and second trimesters and found no significant difference in self-esteem t(31)=1.17, p=.25 (two-tailed). There was also no significant difference between trimesters two and three t(30)=-.928, p=.36 (two-tailed); and between trimesters one and three t(30)=.61, p=.55 (two-tailed) (See Appendix A). A t test for related means found that there was no significant difference in husbands’ aggression across the three trimesters of pregnancy. The first related means t test compared the difference in the first and second trimesters and found no significant difference in aggression t(31)=1.52, p=.14 (two-tailed). There was no significant difference between trimesters two and three t(31)=-.18, p=.86 (two-tailed). There was no significant difference between trimesters one and three t(31)=-1.08, p=.29 (two-tailed); and between trimesters one and four t(30)=.29, p=.78 (two-tailed) (See Appendix A). Comparing Behavior 15 A t test for related means found that there was no significant difference in husbands’ physical abuse of partner across the three trimesters of pregnancy. The related means t test compared the difference in the first and third trimesters and found no significant difference in physical abuse t(32)=.63, p=.53 (two-tailed) (See Appendix A). Correlation is basically a measure of relationship between two variables. Relationships among the variables behavior assessment measures, and behavior specific measures were correlated using Pearson’s r. This study was designed to determine if a relationship exists between these two measures, also each measure was correlated to see how they all compare against each measure to look for a significant change. A basic condition necessary for the computation of the Pearson r is that there be a linear relationship between the two variables. In everyday usage an r of .8 and above is considered a high coefficient, and r around .5 is considered moderate, and an r of .3 and below is considered a low coefficient. Comparing Behavior 16 Table 1 Intercorrelations Between The Dimensions of the Behavioral Construct ____________________________________________________________________________ AX1 AX2 AX3 NS1 NS2 NS3 SE1 SE2 SE3____________ AX 1 ---AX 2 .721** ---AX 3 .870** .570** ---NS 1 .183 .273 .408 ---NS 2 .265 .240 .509* .850** ---NS 3 .234 .320 .302 .772** .821** ---SE 1 .432* .304 .608** .570** .530** .462* ---SE 2 .110 .229 .280 .590** .463* .444* .771** ---SE 3 .350 .295 .507* .654** .487* .474* .812** .817** ---_____________________________________________________________________________ Note. The * symbol represents significance at the .05 level, the ** at the .01 level. AX = Anxiety, NS = Narcissism, SE = Selfesteem. n = 23 Only one area of significant correlation was observed between Anxiety and Narcissism. This occurred between Anxiety in the third trimester and Narcissism in the second trimester. Anxiety shows some significant correlation with self-esteem in three areas: anxiety in the first trimester and self-esteem in the first trimester; anxiety in the third trimester and self-esteem in the first trimester; and anxiety in the third trimester and self-esteem in the third trimester. All measures of narcissism showed significant correlations with all measures of self-esteem in all three of the trimesters. Comparing Behavior 17 Table 2 Intercorrelations Between The Dimensions of the Behavior Specific Measures ____________________________________________________________________________ NPAPS1 NPAPS3 AIPA1 AIPA3 AIVA1 AIVA3 AIII1 AIII3 AIA1 AIA3 NPAPS 1 -----NPAPS 3 .849** ------AIPA 1 .254 .362 -----AIPA 3 .367 .553** .668** -----AIVA 1 .431* .334 .691** .520* -----AIVA 3 .417* .347 .663** .451* .892** -----AIII 1 .547** .447* .151 .142 .161 .053 -----AIII 3 .575** .591** .196 .513* .176 .160 .466* ----AIA 1 -.341 -.496* -.132 -.217 -.197 -.268 .172 -.260 ----AIA 3 -.246 -.254 -.085 -.143 -.253 -.344 .008 -.272 .559** --Note. The * symbol represents significance at the .05 level, the ** at the .01 level. n = 23 Non-Physical Abuse of Partners is abbreviated as NPAPS, Physical Aggression is abbreviated as AIPA, Verbal Aggression is abbreviated as AIVA, Impulsive/Impatient Aggression is abbreviated as AIII, and Avoidance Aggression is abbreviated as AIA. Significant correlations occurred between Non-physical Abuse and Physical Aggression in the third trimester of both measures, between Verbal Aggression and Non-Physical Abuse during the first trimester of both measures and Verbal Aggression the third trimester and NonPhysical Abuse the first trimester. Significant correlations also occurred between Impulsive/Impatient Aggression first & third and the Non-Physical Abuse first and third trimester, and the Avoidance and Non-physical abuse third trimester. Table 2 indicates a relationship between the behavior specific measures of Non-Physical Abuse and Aggression. Comparing Behavior 18 Table 3 Correlations between Non-Physical Abuse, Aggression and Anxiety _______________________________________________________________________ Anxiety 1 Anxiety 2 Anxiety 3___ NPAPS 1 .392 .415* .469* NPAPS 2 .423* .506* .545** AIPA 1 .072 -.060 .304 AIPA 3 .187 .032 .434* AIVA 1 -.167 -.057 .051 AIVA 3 -.030 -.001 .102 AIII 1 .155 .195 .376 AIII 3 .480* .287 .551** AIA 1 -.169 -.376 -.030 AIA 3 -.144 -.416* -.059 Note. * indicates significance at the .05 level. The ** symbol represents significance at the .01 level. NPAPS is Non-Physical Abuse of Partner Scale AIPA is Physical Aggression, AIVA is Verbal.Aggression, AIII is Impulsive/Impatient Aggression, and AIA is Avoidance Aggression. n = 23 Significant correlations occurred between Anxiety compared to the Non-Physical Abuse across the scales, except the first trimester of Anxiety and Non-Physical Abuse. Significant correlations were found in four areas on the Anxiety compared to the Aggression scales, which could have happened by chance. Significant correlations were also found on Anxiety and Physical Aggression third trimester, Anxiety first trimester and Impulsive/Impatient Aggression third trimester, Anxiety and Impulsive/Impatient Aggression third trimester, and Anxiety second trimester and Avoidance Aggression third trimester. Comparing Behavior 19 Table 4 Correlations between Non-Physical Abuse, Aggression and Narcissism _______________________________________________________________________ Narcissism 1 Narcissism 2 Narcissism 3 NPAPS 1 .286 .309 .240 NPAPS 2 .397 .361 .304 AIPA 1 .431* .605** .460* AIPA 3 .275 .443* .242 AIVA 1 .276 .400 .277 AIVA 3 .296 .431* .424* AIII 1 .253 .398 .180 AIII 3 .017 .105 .007 AIA 1 -.058 -.038 -.190 AIA 3 -.109 -.207 -.326 Note. * indicates significance at the .05 level. The ** symbol represents significance at the .01 level. NPAPS is Non-Physical Abuse of Partner Scale, AIPA is Physical Aggression, AIVA is Verbal Aggression, AIII is Impulsive/Impatient Aggression, and AIA is Avoidance Aggression. n = 23 Table 4 shows that there were no significant correlations on any trimester between Narcissism and Non-Physical Abuse. Significant correlations comparing Narcissism and Aggression occurred in five areas: Physical Aggression first trimester and Narcissism all three trimester, Narcissism second trimester and Physical Aggression third trimester, Verbal Aggression third trimester and Narcissism second and third trimesters. Comparing Behavior 20 Table 5 Correlations between Non-Physical Abuse, Aggression, and Self-Esteem _______________________________________________________________________ Self-Esteem 1 Self-Esteem 2 Self-Esteem 3 NPAPS 1 .557** .378 .302 NPAPS 2 .488* .329 .369 AIPA 1 .261 .339 .280 AIPA 3 .269 .120 .249 AIVA 1 .166 .368 .214 AIVA 3 .296 .446* .356 AIII 1 .441* .351 .157 AIII 3 .392 .212 .226 AIA 1 .012 -.069 .071 AIA 3 -.126 -.109 .047 Note. * indicates significance at the .05 level. The ** symbol represents significance at the .01 level. NPAPS is Non-Physical Abuse of Partner Scale. AIPA is Physical Aggression, AIVA is Verbal Aggression, AIII is Impulsive/Impatient Aggression, and AIA is Avoidance Aggression. n = 23 Significant correlations between the Non-Physical Abuse scale compared to the SelfEsteem scale occurred on Self-Esteem in the first trimester and Non-physical abuse first trimester, and Self-Esteem first trimester and Non-Physical Abuse second trimester. On the Aggression scale compared to the Self-Esteem scale, significant correlations occurred in the Self-Esteem second trimester and the Verbal Aggression third trimester, and Self-Esteem first trimester and the Impulsive/Impatient Aggression first trimester. Since there are only a few significant correlations on these scales there it is possible they happened by chance. Comparing Behavior 21 Discussion The hypothesis that behavioral assessment measure would correlate with behavior specific measures held true for some of the test administered. When comparing the correlations between self-esteem and narcissism they are moderately to highly correlated. There is minimal correlation between anxiety and self-esteem and the same is true between anxiety and narcissism. A moderate to high correlation was found between Non-Physical Abuse and Aggression, as was to be expected. A significant correlation was noted on the Non-Physical Abuse and the Impulsive/Impatient subtest on the Aggression Scale. Both of these test were behavior specific measures. This indicates that either test can be used to assess this measure. Both test do not have to be administered. A moderate correlation was also found on the Non-Physical Abuse and the Anxiety scales. There was little correlation found between Anxiety and Aggression, Narcissism and Aggression, and Self-Esteem and Aggression, Narcissism and Non-Physical Abuse. The few correlations that occurred on these scales could have occurred due to chance. Due to the n of 23, small sample size, the chance of making a Type II error was possible. The chance of this happening is equal to the value of alpha, .05. This means that there is a 5% chance of saying there is statistical significance when there is none. Further research in behavioral changes of first-time expectant fathers is indicated on both larger and more diverse populations. A larger more diverse sample size for this study would have been ideal. The sample does not represent diversity in education, income, or ethnicity. For this reason, the study results are not generalizable to the population at large. Initially, each graduate student involved in this study was to obtain ten first-time fathersto-be to participate in this study; the intent was to have over 100 participants in the study. Letters were sent to OBGYN doctors and clinics in the Upper Ohio Valley explaining the Comparing Behavior 22 purpose of this study and asking if they would be willing to participate. If the OBGYN doctors agreed to help, letters were provided to the office manager to pass out at initial visits written specifically to first time expectant fathers. Over 200 letters were sent out per graduate student to different OBGYN offices throughout the Ohio Valley. The fathers were asked to respond by calling the number provided. The graduate students met individually with the OBGYN doctors and explained this study; the doctors seemed interested and eager to help. Not one phone call was made from a prospective father through this process. Subjects were obtained through personal contacts with co-workers and friends of family members. The number was far below what the study had initially expected, due to the fact that it was difficult to obtain subjects to agree to participate when there was no compensation provided to the fathers to agree to meet for three different sessions for one to two hours, throughout the pregnancy. Several men who initially agreed to participate dropped from the study because they felt that some of the information obtained was too evasive and there were too many questions to answer. Subject 0106 dropped from this study because during the first trimester an Alpha Fetal Protein Test was conducted, and the fetus tested positive for Down Syndrome. The doctor told them there would be a 1 in 136 chance the baby would be Downs. The father was very stressed and felt he could not deal with the additional stress of this study. Actually, these men would have been ideal subjects for this study. It is suggested if this project is repeated, the use of a control group should be considered. A control group would need be a group of males that plan on some day becoming fathers, they would have to agree to stay in the study when they become expectant fathers and repeat the study. Comparing Behavior 23 Surveys have become an established means of collecting data and have earned the reputation for yielding valuable information about human behavior. Nonetheless, a few major problems can occur. The first has to do with sampling considerations. A biased sample can produce misleading results. The second problem concerns response bias, or social desirability bias. Sometimes people respond to a survey question in a way that reflects not how they truly feel or what they truly believe, but how they think they should respond. A third major problem in survey research concerns the content of the items contained in the survey. Sometimes survey writers try to include too much of an item, resulting in an item that actually asks for two responses at once (Goodwin, 1998). One difficulty confronted in the study on the assessment of partner violence is the private nature of most abusive behavior. This study, like most others, relied on self-reports of physical aggression. A confounding question to ask in this study would be whether the fathers were actually honest when answering questions that addressed them actually being abusive in their relationship. On the abusive questions administered, it would have been an interesting comparison to see if the mothers would have answered in the same way as the fathers. Another confound of this study was incorrect responses given by some of the subjects; due to this, some of the results had to be thrown out because tests were not completed as the directions stated, or some of the test items were not answered at all. When running the correlations for the test comparisons, 23 actual sets of data were completed correctly throughout all subjects. A limitation of this study is the homogeneous nature of the race, socioeconomic status, and sexual orientation sample. The results of this study should be interpreted in light of the fact that the samples for this study were drawn from a population that included only white, healthy Comparing Behavior 24 married expectant fathers who volunteered to participate. It would be important to determine if the findings apply to other groups. Another limitation of this study is that it is very difficult for a cohort group of 13 to do the same study. It was difficult to get a consensus many times. It was difficult to get all 13 together at the same time. Not every member of the group was able to obtain subjects for the study, therefore, some member spent hours giving the battery of test each trimester, and others did not contribute to the data collected. As the project was designed, three members of the group were doing correlations between test clusters, and they were depended on the other members of the group to score and run their data before they could complete their statistics Comparing Behavior 25