Quality of Family Life as a Factor or Self

This research tested the hypothesis that quality of family life, namely the relationship between an adolescent and his/her parents, is a factor of that individual�s self-esteem. The participants included 103 high school students from a small, Midwestern school and comprised of 44 reported males and 57 reported females. The method for studying family life and self-esteem consisted of surveys of self-reported data. The surveys were distributed to all students of a required, introductory science class. Results of the analysis showed a .352 correlation between the responses that measured family life and the responses that measured self-esteem. The resulting p-value was .003. The implication of this study was that a better understanding of the impact of family dynamics on other aspects of life developed, specifically in the context of self-concept. Self-concept is a critical factor in our ability to solve problems and to become fully-functioning members of society.

A multitude of research studies examine various aspects of familial relationships, including attachment, affection, and, conversely, abuse. Documentation also can be found of studies on self-esteem and an individual�s image of him/herself. The following research, however, attempted to combine the two by correlating quality of family life and self-concept. The survey used in this study was meant to test the hypothesis that family dynamics, namely the relationship between an adolescent and his/her parents, is a factor of that individual�s self-esteem.

Research regarding the quality of family life and its connection to an individual�s self-esteem is essential, because it aids humans in understanding the effects family has on other aspects of life. This information can lead to the elimination or prevention of self-defeating attitudes and can provide early warning signs of negative or dangerous living environments. Also, individuals may obtain a higher comprehension of themselves if they acknowledge the social factors that facilitate self-views and feelings.

Bowlby (1969) explained in Volume I of Attachment and Loss the criteria for patterns of attachment. Bowlby noted that �one of the most obvious criteria� is whether or not he [a child] protests when his mother leaves him for a brief time and how strongly he does so� (p. 333). Attachment behavior can be measured by a child�s greeting toward his mother, his response to interactions with the mother, and exploration and how it is oriented, among others.

Bowlby also defined phases in the development of attachment. Phase I, �Orientation and Signals without Discrimination of Figure,� states that an infant of zero to twelve weeks does not discriminate one person from another (p. 266). She responds to all faces and voices in a friendly manner. Phase II, �Orientation and Signals Directed towards One (or More) Discriminated Figure(s),� lasts through six months, and it includes behaving in a friendly manner to all people, but showing some favoritism toward the mother-figure (p. 266). In Phase III, �Maintenance of Proximity to a Discriminated Figure by Means of Locomotion as well as Signals,� an infant through the age of three orients mobility around her mother, and the infant�s undiscriminating friendliness decreases (p. 267). Other close family members are selected as attachment-figures during this time, although the mother is still the basis for most behaviors. Finally, in Phase IV, �Formation of a Goal-corrected Partnership,� the young child gains some insight into her mother�s motives and feelings (268).

In Volume II of Attachment and Loss, Bowlby (1973) discussed a comparison study, in which children separated from their families were evaluated with children who remained in their homes. Those placed in a residential nursery behaved more aggressively in play situations. ��Episodes of hostile behaviour occurred four times as frequently in the doll play of the separated children as they did in the play of the children living at home� (p. 245). The most interesting finding was that eight of ten separated children attacked a doll that they had previously identified �as a mother or father doll� (p. 245). During reobservation, which took place six weeks after the children returned from the nursery, all of the young participants played less aggressively.

When dealing with self-reliance in children, Bowlby measured this characteristic in terms of two family dimensions. The first dimension assessed �mutual trust and approval between the child and his parents,� and the second measured �consistency of family life� (p. 335). The mutual trust dimension evaluated if the parent accepted and encouraged the child and if the relationship was congenial. The latter considered if the family routines were predictable and if shared activities between parent and child were frequent. The children in his studies who rated the highest in maturity also rated highly in these two scopes of family life. His findings concluded that there is a connection between family patterns and self-reliance in children.

Bowlby�s final volume of Attachment and Loss (1980) dealt with children�s behaviors when facing loss of a parent. He discussed the findings of Hilgard in regard to adjustment after the death of a family member, stating that �her study strongly supports the view that the effect that a parent�s death has on a child is powerfully influenced by the pattern of family relationships to which the child is exposed after it� (p. 315). Hilgard found that the way humans view and respond to distress situations reflect other familial and environmental factors. Within this and other theories of mourning, Bowlby explained the relevance to attachment. He clarified that children of at least sixteen months old have the capacity to mourn. Prior to this age of sixteen months, the infants show distress, not mourning, in the departure of the mother-figure.

Referring to a family�s impact on self-reliance, Bowlby named two sets of influences on personality (1979). �The first concerns the presence or absence� of a trustworthy figure�� (p. 104). �The second set concerns the relative ability or inability of an individual, first, to recognize when another person is both trustworthy and willing� and, second� to collaborate with that person�� (p. 104). Bowlby also mentioned that those individuals with well-adapted personalities demonstrate �initiative and self-reliance� and come from �closely knit families� with parents who provide encouragement (p. 107). In a study completed in 1960, findings suggested that self-reliance and the ability �to rely on others� are products of strong, supportive families and secure attachments (p. 108).

In response to Bowlby�s The Making and Breaking of Affectional Bonds, Weiss (1994) argued that Bowlby did not study attachment in adults himself; therefore, the Attachment Theory cannot apply to adult relationships. Weiss�s main point stemmed from Bowlby�s description of attachment as being directed toward a person �conceived as stronger and/or wiser� (p. 65).

Weiss acknowledged three differences between adult attachment and child-parent attachment, the first being relationship symmetry versus asymmetry. These terms are in reference to need and response to need in a relationship. The needs of a child are different than the needs of a parent, whereas the needs of two adults are more closely associated. Secondly, children see their caregivers as �stronger and wiser,� but adults rarely view their partners as greatly superior (p. 66). Finally, a child�s security relies heavily on the attentiveness of his parent. In adult relationships, security does not rely as heavily on partner reassurance.

Ainsworth (1991) also studied the attachment behaviors of infants. In one research experiment, Ainsworth located 15 infant-mother pairs. A home visitor, the experiment�s observer, collected data and found that �differences in the security or insecurity of the infants� attachment to their mothers were related to the mothers� behavior� (p. 338). Her research found that when mothers responded in a prompt manner to their infants� needs, the infants �cried relatively little and were securely attached� (p. 338). Babies who were securely attached demonstrated a primitive person permanence, as well, in that they seemed to understand that, even when their mothers were out of sight, they still were available. �The strange situation procedure,� as labeled by Ainsworth, again �highlighted the distinction between secure and insecure infants, and between the two groups of insecure infants�avoidant and ambivalent-resistant� (p. 338).

Another study in the realm of quality of family life looked into the correlation between attachment and early maltreatment. Egeland and Sroufe (1981) used a sample of 267 lower-class mothers and their infants to compare data on attachment and neglect/abuse. The Child Care Rating Scale was employed to measure the meeting of �the child�s health care needs, protecting the child from possible dangerous situations� and not leaving the baby alone� (p. 46). Results found that mothers classified under �the inadequate care group are younger�less well educated� and more were unmarried�� (p. 46). Using Ainworth�s method of measuring attachment, the study showed that �the difference in the proportion of secure, avoidant, and resistant infants in the excellent and inadequate care groups was significant� (p. 47). Those infants classified under the excellent care group demonstrated a higher percentage of secure attachment, and infants in the inadequate care group generally were more unstable.

Looking at affection, Harlow (1958) explained the basic responses of affection that monkeys and human infants share: �nursing, contact, clinging, and even visual and auditory exploration� (p. 574). For his experiment, Harlow separated over 60 rhesus monkeys from their mothers only six to twelve hours after their births. The first notation he made included the discovery that the mortality of these infants �was only a small fraction of what would have obtained had we let the monkey mothers raise their infants� (p. 575).

Within his experimentation, Harlow found that monkeys raised on wire cage floors survived with difficulty, those with a wire-mesh cone in their cages did better, and those with a terry cloth cone in the cage became �husky, healthy, happy babies� (p. 575). Harlow next introduced surrogate mothers to the monkeys; one made only of rough materials, the other wrapped in rubber and terry cloth with a light bulb to radiate heat. The surrogate mother made of terry cloth showed more appeal to the infant monkeys, in that they often clung to the cloth and laid at its base. Harlow found that contact comfort is a very basic affectional variable. A mother provides security and comfort in times of distress; therefore, the cloth surrogate mother took on this role of safety when the wire surrogate could not.

Deviating from attachment and affection, Stockhammer, Salzinger, Felman, and Mojica (2001) examined efficient ways of measuring child maltreatment for intervention and prevention. One hundred families with children ranging in age from nine to eleven participated in this research. Comparing parent interview data with Child Protective Service records, significant differences arose within the information regarding abuse. The �CPS records provide more reliable information about risk factors more proximal to the child� while parent interviews provide better information about� family risk factors� (p. 334).

Using an ecological framework, the �abuse and the context in which it occurs predicts the functioning of physically abused children more strongly than information about child abuse alone� (p. 337). The implication of this research is the learned necessity to locate multiple sources for information regarding �child abuse, family context, and child functioning to obtain a reasonably comprehensive assessment of the effect of maltreatment on children�s functioning� (p. 340). A child�s outcome should be evaluated based on family factors as well as �specific abuse incidents� (p. 340).

Alpay (2000) defined self-esteem by first explaining self-image and self-concept. �Self-image is the individual�s awareness of person attributes. This is developed at an early stage through the influences of the parents or guardians� (p. 1). Self-concept is �the sum total of a person�s perceived and desired mental and physical characteristics�� and �the person�s perceived worthiness from these�� (p. 1). Self concept, therefore, encompasses self-image and self-esteem.

Alpay (2000) constructed self-esteem from three stages of development. In early childhood, the child focuses on specific characteristics and behaviors. His/her self-esteem may reflect the parents� perception of that individual at this time. In middle childhood, the second stage, self-image acknowledges emotions, relationships, and other traits. It is at this stage that issues may arise of self-esteem. Avoidance, compensation, low motivation, and resistance are all behaviors of self-esteem issues. During the last stage, which is adolescence, images of self become more stable, but �depression, moodiness, and sensitivity� may emerge (p. 2).

Burns (1982) suggested that academics and social factors facilitate self-esteem in adolescence. Although success may be a factor of high self-concept, self-esteem also can influence performance, in that motivation and expectations develop from positive self thoughts. High achievement, therefore, may result from high expectations and elevated self-esteem.

Miyamoto, Hishinuma, Nishimura, et al. (2001) examined the cross-cultural equivalencies and constructs of self-esteem and major life events. In this research, the Rosenberg (1965) Self-Esteem Scale was employed to demonstrate that experiences can affect self-esteem, but it also established that self-esteem can influence events. The results showed that the similarities measured �did not provide support for the cultural specificity of self-esteem, the experiencing of major life events, and their interrelations� (p. 161). The researchers acknowledged that psychological processes may vary cross-culturally, in that, for some Asian American cultures, individual efforts and concepts are overshadowed by collectivism. This variance can alter thoughts of self-esteem.

In other research, Colwell and O�Connor (2003) compared teacher communication in normal and nurturing classrooms in an attempt to predict characteristics that create a more conducive climate for self-esteem. The researchers looked at teacher responses, teacher praise, behavior management, non-verbal cues, as well as other descriptions. Nurture group teachers appeared to encourage problem-solving skills as a result of this experiment. This classification of teachers also demonstrated more �positive non-verbal behaviour� and less deprecatory remarks (p. 123). In addition, the nurture group teachers dealt with inappropriate classroom behavior in calm, positive ways. While the climate seemed to rejuvenate the self-esteem of the students, the study cannot be certain of this claim. A self-report survey may be useful in further studies.

Parker, Low, Walker, and Gamm (2005) explored the relationship between an individual�s level of self-worth and his/her measure of jealousy when reacting to a close friend�s relationship with another peer. The researchers used the Friendship Jealousy Questionnaire (Parker, 2005), the Children�s Social Desirability Scale (Crandall, Crandall, & Katkovsky, 1965), and the Global Self-Worth subscale of the Self Perception Profile for Adolescents (Harter, 1998). The study found that adolescents with reported lower self-worth displayed greater jealousy, and that those with �chronically poor self-esteem� place less trust in their friends� commitment to them�� (p. 239). Conversely, the individuals with higher self-regard reported a lower feeling of being in competition with their friends and others.

Campbell and Lavallee�s (1993) research stated that people with lower self-esteem are more reactive, in general, to social environment feedback and are �more conservative and cautious in their orientation of this environment� (p. 10). Feedback included measures of acceptance by their peers in the form of accuracy and personal attributions. The accuracy and attribution characteristics were defined for the purposes of this study as the level of congruence between the participants� and their partners� ratings of their behavior in a social situation.

Cambell and Lavallee (1993) also contended that individuals with high self-esteem hold more consistent beliefs than those with low self-images. When rating themselves on groups of adjectives uniformly positive or negative, the participants with high self-esteem consistently endorsed or rejected the words, whereas the participants with low self-esteem demonstrated greater variation. Similarly, low self-esteem participants exhibited less internal consistency when responding to self descriptions with the reactions of �me� or �not me� (p. 9).

Rosenberg (1965) used the results of a self-attitude study to look at self-esteem in adolescents. The research was completed for the purpose of gauging how social experiences affect self-values and self-concept. Rosenberg explained that self-attitudes, as measured in the experiment, can be examined within eight dimensions: content, direction, intensity, importance, salience, consistency, stability, and clarity. The researcher observed, within these dimensions, that people with low self-esteem often appeared more depressed and expressed more feelings of discouragement or unhappiness. In this study, �80 percent of those with the lowest self-esteem scores were �highly depressed� according to the measure� (p. 22). Through his work, Rosenberg (1965) defined self-esteem in terms of high and low. Individuals with high self-esteem respected themselves and consider themselves worthy; individuals with low self-esteem felt rejected, dissatisfied, and disagreeable.

Combining the theories of attachment and self-concept, more recent research shows possible correlation between family factors and personal qualities, such one�s image of him/herself. Love and Murdock (2004) hypothesized that parental attachment is one factor in the psychological well-being of an individual. A group of college students completed four separate surveys, including the Parental Bonding Inventory, the Comprehensive Affective Personality Scale, the Brief Life Satisfaction Scale, and the Conflict of the Family Environment Scale. Data analysis supported the hypothesis that �individuals from step-families have less secure attachments than individuals from intact families,� and the analysis also demonstrated that �family type was found to be a significant predictor of well-being� membership in an intact, biological family is associated with higher levels of well-being� (p. 604). While overprotection did not act as a significant predictor of psychological well-being, �perceived attachment to parents� did act as a predictor within this research (605).

Brody and Flor (1997) predicted that supportive mother-child relationships and predictable family routines would correlate with children�s self-regulation and adjustment. In this study, the researchers conducted two home visits, in which they observed mothers and children in three activities: playing a board game together, the mother reading a story to the child, and constructing Lego models. The Rosenberg Self-Esteem Scale (1965) and Children�s Self-Control Scale (Humphrey, 1982) both were used. The research supported the hypothesis that family routines and mother-child relationships are connected with children�s self-regulation and adjustment. In addition, family income associated with the self-esteem and depressive symptoms of the mother.

Other research, completed by Flouri (2004), measured the connection between a child�s characteristics, including self-esteem, emotional well-being, and behavioral well-being, and the mother�s and father�s involvement in that child�s life. �Children with high self-esteem tended to report high father�s involvement,� (p. 155) according to this study. Flouri also found that the interaction between the structure of a family and the child�s emotional and behavioral well-being predicted a father�s involvement. In addition, female child participants reported that the mother�s involvement was higher than the father�s.

In a cross-cultural study, Al-Simadi and Atoum (2000) looked at the impact of family and environment on self-concept. Specifically, the researchers measured, �What type of relationship exists between youth living in Palestinian refugee camps and their families,� and �What is the level of self-concept of youth living in Palestinian refugee camps?� (p. 380). The results of two self-report scales showed that the relationships between parents and adolescents appeared to be authoritarian in nature and that lower levels of self-concept emerged. The study proposed that the self-esteem of Palestinian youths may be related to parenting style, in that authoritarian ways often reflect a more inflexible and disconnected family life.

Based on this literature, the hypothesis was generated, which stated that the quality of family life, specifically the relationship between parent and child, is a factor of an individual�s self-esteem. Several of the previously mentioned authors suggested that attachment, affection, and abuse contribute to later psychological well-being or dysfunction; so the current research continues the study that poor family relationships influences the lowering of self-concept.

Measurement of the participants� quality of family life was based on Bowlby�s (1973) two family dimensions. The research looked at the adolescents� ratings of parental disciplinary actions, honesty between parent and adolescent, and self-reported overall relationship. The study also used birth order, parents� occupation and work hours, divorce within the family, and make up of primary caretakers. When measuring self-esteem, the research employed self-reported feelings of respect, love, and worthiness.

The participants for this research were 103 high school students from the Southern Illinois area. The participants were chosen based on convenience/proximity sampling. They ranged in age from 14 to 16 and consisted of 44 males and 57 females, not including those individuals who chose not to respond to the question of gender. Approximately 77.5% of the participants categorized themselves as Caucasian, 13.7% marked African American, 2.0% answered Asian, 2.0% stated Native American, and the remaining 4.9% categorized themselves as Other or a non-listed ethnicity. The participants remained anonymous for the purpose of analysis.

The materials for this study included the Family Inventory, specifically created for high school age participants. This survey incorporated the Burns Self-Esteem Inventory (Burns, 1995). The participants were instructed to respond to the 4 demographic questions, 22 questions pertaining to their family lives, and 5 questions pertaining to self-esteem. The survey was completely based on self-report, and if at any time the participants felt uncomfortable answering a question, they were free to leave it unanswered.

The surveys were dispersed by the participating teacher to the high school participants, all students in a required, introductory science class, over the course of one day. The individuals were asked to respond to 31 items, including demographic questions, for the Family Inventory. A brief statement explaining confidentiality and withdrawal procedures and the directions for each of the survey�s four sections were included with the questionnaire. A copy of this survey is included in the appendices section. Once completed, the answer forms were returned to the surveyor, who entered all data into SPSS for analysis. Analysis was run to determine frequencies, descriptives, internal reliability, comparisons of means, and correlations between qualities of family life and self-esteem.

The researcher first noted significant frequencies before analyzing the results for the Family Inventory. Beginning with ethnicity categorization of the sample, 77.5% (valid percent after missing responses) of participants categorized themselves as Caucasian. Less than one fourth of the participants claimed to be one of several minority ethnicities, including African American, Asian, or Native American.

In regards to parent-adolescent relationships, when asked, �Do you have a father or other primary caretaker, besides your mother/mother-figure, with whom you currently live,� 27.2% of the participants responded with, �No.� For the same question, but regarding the mother, only 4.9% responded with, �No.� Approximately 31.4% of participants claimed that their parents are divorced, and 22.5% claimed that their parents never married, leaving only 46.1% of participants who stated that their parents have not gotten a divorce.

Also significant, 9.2% of participants responded negatively to the question, �What is your favorite thing to do with your parent(s)?� In rating their relationships with their mothers, participants gave a mean response of 5.57 on a 7.00 Likert scale, 1 being �Very Poor� and 7 being �Excellent. In rating their relationships with their fathers, participants gave a mean response of 5.18 on the same 7.00 Likert scale. When asked, �Do you feel that your mother/primary caretaker disciplines you fairly,� participants gave a mean answer of 4.90 on a 7.00 scale. For a similar question regarding the father, the mean response was 5.00.

In reference to the self-esteem section, the mean total of responses equaled 19.21 out of 25.00. The lowest mean rating in this portion of the survey resulted from the statement, �I feel worthwhile,� while the highest rating resulted from, �I respect myself.� There were no significant differences among the self-esteem statements.

Another interesting finding pertained to the question, which asked, �What is your favorite thing to do with your parent(s)?� The greatest response was �go out to eat,� with 19 participants giving the open-ended answer. Also, 9 of the 103 participants gave a negative response, by stating, �I hate my step-dad,� �stay away from them as long as humanly possible,� or �never any family time.� The table below shows a breakdown of responses.

In running analysis on internal reliabilities for the Family Inventory, the items pertaining to subject-parent relationships were significantly correlated, with p-values at the .0001 level. Correlating the rating of parental disciplinary actions to the adolescents� relationships with their mothers resulted in an alpha coefficient of .616 and a p-value of .0001. Correlating the same qualities, but for the father, resulted in a coefficient of .618 and a p-value of .0001. The overall reliability alpha coefficient was .849.

The items within the self-esteem portion of the Family Inventory also showed internal reliability. Correlating the results of the statements, �I like myself,� and �I feel good about myself,� gave an alpha coefficient of .830 and a p-value of .0001. Also, comparing the results of, �I feel worthwhile,� to the overall self-esteem totals gave an alpha coefficient of .815 with a p-value of .0001. The overall reliability alpha coefficient was .896.
Correlations

Analysis of the Family Inventory data did not disprove the hypothesis that quality of family life is a factor of self-esteem in adolescents. Correlating the total from all parent-related questions to the total from all self-esteem related questions produced a p-value of .003 and a correlation coefficient of .352. The positive coefficient represents the result that as subject-parent relationships are rated higher, self-esteem criteria also are rated higher. Correlating the totals from the mother-focused section to the self-esteem totals resulted in a correlation coefficient of .323 and a p-value of .001. Running correlations between the father-focused totals and the self-esteem totals produced a coefficient of .247 and a p-value of .033. These findings are all significant at the .05 level.

Analyzing the data with the use of ANOVAs and independent sample t-tests also produced significant results. When running an ANOVA to compare the means of the self-esteem totals within responses to, �Overall, how comfortable do you feel when you are with your family,� a p-value of .001 resulted. Using the data regarding birth order, significant differences were found when comparing means of the self-esteem totals. For example, t-tests showed that first born participants had a higher self-esteem total than last born/youngest participants with a p-value of .048. Conducting an ANOVA resulted in an overall p-value of .042 when comparing self-esteem means among birth orders. There were no significant differences between first born and middle child participants or middle child and only child participants.

According to the results of the Family Inventory, this research supported the hypothesis that an individual�s quality of family life, specifically when looking at the relationship between parent and adolescent, acts as a factor of self-esteem. By measuring an adolescent�s view of his/her parents� discipline patterns, his/her ability to express honesty at home, and that individual�s self-reported rating of the overall parent-adolescent relationship, researchers can make predictions about that individual�s level of self-esteem. The research suggested that as the relationship between a parent and an adolescent positively increases, the adolescent�s self-esteem may increase, as well.

These findings are relevant, in that teachers, coaches, and other role models can help students with self-esteem issues more effectively if they understand the root of the problem. Making the connection between an adolescent�s family dynamics and his/her level of self-esteem also can be early warning signs of a negative home environment. At-risk programs are good examples of this theory; they employ the notion that the environment in which an individual is raised can influence academic achievement. As studied in past research, such as the work of Burns (1982), self-esteem and scholastic ability may be factors of each other.

The results of the research completed with the use of the Family Inventory are supported by additional previous studies. The inquiries of Al-Simadi and Atoum (2000), Love and Murdock (2004), and Flouri (2004) all agree that quality of family life is a factor of an individual�s level of self-esteem and overall well-being. While this research and others support the overlying hypothesis, there were limitations of this study. First and foremost, the data only reflected one population. The researcher is uncertain if significant results will develop if the Family Inventory is distributed to additional populations. Also, as stated, the participants did not reflect an ethnically diverse sample. Over three-fourths of participants fell into one category of ethnicity. A more random sample will need to be used for further analysis.

In sum, the use of the Family Inventory supported the hypothesis that family dynamics, namely the relationship between parent and adolescent, affect self-esteem. As stated, p-values showed significance at the .05 level through the use of several bivariate correlations and independent sample t-tests. Self-reported comfort at home, birth order, and overall parent-child relationship ratings, all may act as factors of an individual�s self-esteem.

1. Do you have a mother or other primary caretaker, besides your father/father-figure,

**If you answered �yes,� please proceed to question 2. If you answered �no,� please

4. Overall, how would you rate your relationship with your mother/primary caretaker?

7. Do you have a father or other primary caretaker, besides your mother/mother-figure,

**If you answered �yes,� please proceed to question 8. If you answered �no,� please

10. Overall, how would you rate your relationship with your father/primary caretaker?

2. If yes, your parents did divorce, how old were you at the time? (If this question does not apply to you, please skip to question 3.)

3. During your childhood (ages 0-10), who was/were your primary caretaker(s)? Mark

4. During your adolescence (ages 11+), who has/have been your primary caretaker(s)?

5. How many siblings/half-siblings (but not including step-siblings) do you have? _____

8. What is your favorite thing to do with your parent(s)? ________________________

9. Approximately, how often do you do this activity together? (Please choose only one response).

Part IV-Write a check mark or X to indicate how much each statement describes how you have felt in the past week, including today. Please choose only one response for each statement.

	1	2	3	4	5
I feel worthwhile.
I like myself.
I respect myself.
I feel lovable.
I feel good about myself.