Introduction

Although it is generally accepted that old age is a time of overwhelming loneliness, research shows that adolescents and young adults, especially, tend to feel lonely (Killeen 1998; Le Roux 2001; Taylor et al. 2003). Statistics reveal that 79% of adolescents younger than 18 report feelings of loneliness. By contrast, 53% of adults between the ages of 45 and 54 report loneliness, while only 37% of people older than 55 report feelings of loneliness (Taylor et al. 2003). Adolescents who manifest feelings of loneliness show a predisposition towards psychological disturbances, anxiety, substance abuse, risky sexual practices, aggressive behaviour (Rokach et al. 2001; Saluja et al. 2004), suicidal tendencies (Killeen 1998; McWhirter 1990; Rokach et al. 2001), as well as a susceptibility to physical ailments (Rokach et al. 2001).

One’s attitude towards a variety of social aspects is pivotal to one’s interaction with others (Bennett and Sani 2003). An attitude may be defined as a lasting orientation containing a cognitive, an affective and a conative component (Taylor et al. 2003). The cognitive component of an attitude encompasses one’s thoughts or convictions regarding the attitudinal object (Louw and Edwards 1998). Children develop cognitive schemata or representations in their interaction with their parents (Furman et al. 2002). A schema is the basic and most elementary cognitive unit consisting of an intellectual activity and a congruent action that is outwardly observable (Burden 1997).

The affective component refers to the emotions that are coupled with a specific attitude (Louw and Edwards 1998). A lack of involvement of fathers in the emotional development of their children can give rise to the following negative effects: depression, a poor self-image, violence, divorce anxiety, rejection, self-disdain, affective disturbances and many other problems that undoubtedly retard normal development (Charity 2003). Positive character traits such as empathy, a good self-image (Green 2002; Smith 1996), self-control, psychological well-being, social competence and certain life skills stem from a child’s relationship with an involved father (Evans and Fogarty 1999). Studies found that the availability of father and mother differently influenced relationships (Ducharme et al. 2002; Lieberman et al. 1999). The fathers’ availability was particularly important for lower conflict in friendship.

The conative component of an attitude involves one’s intention to behave in a certain way (Louw and Edwards 1998). Adolescents may react to stressors, such as the loss of a parent through death or divorce, with aggression or repression, inter alia. Repression entails the suppression of emotions, whereby the adolescent cuts him- or herself off emotionally from others (Schoeman and Van der Walt 2001). The loss of a father figure, in particular—which is becoming a common phenomenon in modern society, not only as a result of death or divorce, but also through the lack of involvement of fathers—is a significant stressor for many adolescents (Popenoe 1996).

The more adolescents interact with their parents, the better they succeed in developing their cognitive representations. These schemata or expectations guide the adolescents’ behaviour and influence their cognition, thoughts and emotions (Furman et al. 2002). Seeing that an attitude consists of thoughts, emotions and intentional acts, it may be deduced that adolescents’ behaviour is guided by their attitudes towards one or both of their parents. Fathers, especially, play a cardinal role in child development; and this has a significant impact on children’s social, emotional and intellectual development (Charity 2003; Smith 1996). The cause of loneliness, according to Sullivan (Kara and Mirici 2004), may be found in unpleasant childhood experiences that lead to unfulfilled needs for intimacy, such as the loss of a parent through divorce or death. As mentioned earlier, research has shown that children of divorced parents are at a greater risk of developing loneliness as adults, compared to the children of intact families (Taylor et al. 2003). However, if the father remains involved in the child’s life, even if he is estranged from the child’s mother, the negative consequences of the divorce are considerably reduced (Evans and Fogarty 1999).

According to Asti et al. (2006), unresolved childhood experiences may be reactivated in times of tension and crisis, leading to intense emotional reaction. Unresolved emotional problems may manifest themselves in the form of depression, anxiety, panic attacks, obsessive-compulsive actions, anger, non-involvement, withdrawal, guilt, sadness and other overwhelming emotions such as loneliness (Van Wijk et al. 2004). The available literature shows that the loss of a father figure, whether through divorce, death or a sheer lack of involvement, can lead to loneliness. Loneliness, in turn, may lead to crime, which merely intensifies the feelings of loneliness. Fathers, especially, play a cardinal role in gender-role identification and, consequently, in the ability to develop and maintain romantic relationships. A boy learns from his father what it means to be a man; and if the father is absent, problems concerning gender-role identification tend to occur (Mandara et al. 2005). Girls also derive benefits from strong fatherly care and develop healthy boundaries in sexual relationships, as well as in intimacy (Charity 2003). The involvement of a father in the life of his adolescent daughter helps to reduce psychological problems such as loneliness (Evans and Fogarty 1999).

Research conducted by Marcoen et al. (1987) among 444 adolescents between the ages of 11 and 17, showed that the older the adolescent became, the more parental involvement decreased, leading to an increase in parent-related loneliness. In the same vein, Danziger and Radin (1990) found that fatherly involvement is influenced by the age of the child. These results are corroborated by Jordaan (2004), who found that the older the learners, the more they are inclined to experience loneliness. Heard and Bratter (2006) found that black fathers had a direct influence on their children’s racial identity, and consequently also on their concept of self. Statistics reveal that in every father–child interaction, the degree to which the child identifies with the race and culture of his or her father increases by 34% within a period of a month. Cultural customs are transferred in father–child interactions. In mixed-race marriages, fathers profoundly influence their children’s racial identification, according to Heard and Bratter (2006). Racial identification can possibly lead to social integration and membership of a certain ethnic group. Social loneliness results from defective social integration (Taylor et al. 2003).

Rokach et al. (2001) investigated the influence of culture on loneliness and found that cultural background does, in fact, have an effect on the experience of loneliness. In various other studies, it was found that a strong inverse relationship exists between loneliness and race, although this does not appear to contribute significantly to the prediction of loneliness (Jordaan 2004). Calderon (2000) points out that parents have a significant influence on their children’s language development, as well as their social–emotional development, and that these two aspects of development are interrelated, and thus affect each other mutually. Tempestuous relationships between parents and children tend to occur in families that experience financial pressure, and this has a negative influence on the emotional welfare as well as the social relationships of disadvantaged children (Ackerman, Brown and Izard 2003; Duncan and Brooks-Gunn 2000). Danziger and Radin (1990) stress that socio-economic status is a crucial factor in fathers’ lack of involvement with their children, in the context of disadvantaged communities.

In view of the above discussion, the two research hypotheses that will be examined in the study are as follows: first: there is a relationship between loneliness and the attitudes of adolescents towards their fathers. Second: gender, age, language, race and socio-economic status play a role in the relationship between loneliness and adolescents’ attitudes towards their fathers.

Methods

Participants

The sample was comprised of all learners attending randomly selected secondary schools in the Mangaung area, South Africa. Respondents were between the ages of 11 and 19 years, and were in grades 7–11. In total, 1,217 learners participated in the investigation, but 149 questionnaires were incorrectly completed and could thus not be taken into consideration. The final sample thus consisted of 1068 learners from the different schools. Seven schools were included in the sample, but in terms of an agreement with the Department of Education, their names may not be revealed.

Data-Collection

Permission was granted by the Director of Education, as well as the Director of Quality Assurance in the Free State, South Africa to pilot the study. All the schools and learners participated voluntarily in the project.

The self-administered questionnaires were available in English and Afrikaans and were completed in groups.

The frequencies and percentages of the different levels of the independent variables are reflected in Table 1.

Table 1 Frequencies and percentages of all the variables
Full size table

Because of the limited number of respondents in the 11–13-year age group, this group will no longer be taken into account in this study.

The Measuring Instrument

The Loneliness Scale

The Loneliness Scale, developed by this researcher, was used to measure the occurrence of loneliness in the respondents. Research on different dimensions of loneliness and other aspects thereof was taken into account in compiling the research instrument. This instrument has already been used in various investigations (Jordaan 2004; Le Roux 1998). The alpha coefficients were calculated in each of these studies and ranged from 0.88 to 0.91. It thus compares very favourably with other measuring instruments of the same nature. In conjunction therewith, the internal validity rendered a correlation of rpn = −0.6116 (Jordaan 2004). The scale is a one-dimensional measuring instrument consisting of 30 items, and is balanced, with 15 positively-worded and 15 negatively-worded items.

The Cronbach alpha coefficient of the Loneliness Scale was once again investigated in the current study and was calculated at 0.88.

The Fathers Questionnaire

A one-dimensional questionnaire consisting of 30 items was used to measure the adolescents’ attitudes towards their fathers. Seeing that no existing questionnaire to measure adolescents’ attitudes towards their fathers was available, this researcher compiled such a questionnaire, after a thorough examination of the relevant literature. The items were chosen in such a manner that cognitive, affective and conative components were reflected therein (see also the section “Discussion” on attitudes toward fathers). The fact that the measuring instrument was balanced fulfilled a further requirement—the questionnaire had 15 positive and 15 negative items, in order to take the acquiescence effect into account.

The Cronbach alpha for reliability for the different racial groups amounted to 0.942 for coloured, 0.921 for black, and 0.953 for white respondents, with a collective average of 0.941. The validity for this instrument is currently under investigation.

A biographical questionnaire was used to obtain information on the gender, age, race, language and socio-economic status of each respondent.

Statistical Techniques

The hypotheses, as formulated above, were examined by means of a stepwise multiple-regression analysis, as well as a multiple-variance analysis. In cases where significant differences were encountered in the variables with more than two levels, the Scheffé test for post hoc comparisons was applied.

Stepwise Multiple-Regression Analysis

Stepwise multiple-regression analysis is used when various potentially independent variables are present, and an investigation is carried out in order to determine which variables exert a statistically significant effect on the overall correlation. In this study, the criterion variable is loneliness, and the predictor variables are attitudes towards fathers, gender, age, race, language and socio-economic status respectively.

Multiple-Variance Analysis

A multiple-variance analysis was carried out on the data obtained. This statistical technique measures the interactive and independent effects of two or more independent variables on several dependent variables, with a view to determining whether there are differences in the dependent variables. In this case, the dependent variables are loneliness, as well as attitudes towards fathers, while the independent variables are gender, age, race, language and socio-economic status.

Results

Intercorrelation Matrix

The results reflecting the links between loneliness and attitudes towards fathers, as well as the independent variables, are provided in Table 2 in the form of an intercorrelation matrix.

Table 2 Inter-correlation matrix of all the independent variables
Full size table

According to Table 2, all the independent variables show a connection with loneliness. However, the strongest link occurred in the case of attitudes towards fathers, namely −0.372, which is a highly significant inverse correlation. From this, it may be deduced that adolescents who have a negative attitude towards their fathers are more inclined to experience loneliness, and vice versa (adolescents suffering from loneliness also tend to have a negative attitude towards their fathers).

The independent variable with the second highest correlation with loneliness is that of socio-economic class. This inverse correlation of −0.217 indicates that the lower the adolescent’s socio-economic class, the greater the adolescent’s tendency towards loneliness. The converse is also true. Those in the higher socio-economic classes are less susceptible to loneliness. Thirdly, race, as an independent variable, also correlates with loneliness. The inverse correlation of −0.135 between loneliness and race is also significant, indicating that race is possibly a strong predictor of loneliness. Gender also shows an inverse correlation with loneliness (−0.106). This suggests that male adolescents possibly experience loneliness to a higher degree than female adolescents.

Loneliness and age show a positive correlation of 0.100. The older the adolescents, the greater the tendency to be lonely. Lastly, it can be seen that language also shows a positive correlation with loneliness. Language may therefore also be a good predictor of loneliness.

Moreover, race, language and income all show a correlation with attitudes towards fathers. Language, as an independent variable, shows an inverse correlation with attitudes towards fathers. In the statistical analysis of the data, Afrikaans and English were coded as 1 and 2, respectively, while the indigenous languages were coded as 3, 4 and 5. From this, it may be assumed that those who speak the indigenous languages will tend to have more negative —albeit still positive—attitudes towards their fathers. As far as race is concerned, there is a positive correlation of 0.175. In terms of the coding, coloured and black respondents are indicated by the numbers 1 and 2 respectively, while number 3 refers to white respondents. According to the data, white adolescents tend to have positive attitudes towards their fathers. Lastly, there is a positive correlation of 0.215 between attitudes towards fathers and income. Those in the high socio-economic classes tend to have positive attitudes towards their fathers.

Stepwise Multiple-Regression Analysis

In Table 3 the results of a stepwise multiple-regression analysis are shown. These results contain particulars in respect of the multiple correlations (R), the proportions of variance explained (R 2), the standard assessment errors and the F-values for determining the significances.

Table 3 Stepwise multiple-regression analysis with loneliness as criterion variable
Full size table

According to Table 3, attitudes towards fathers could significantly explain 0.135 of the variance of loneliness. Income, gender and language, in turn, could significantly explain 0.015, 0.006 and 0.003 of the variance of loneliness, respectively. The variable, attitudes towards fathers, thus appears to be the best predictor for loneliness. The above-mentioned four variables could explain 15.9% of the variance of loneliness. Race as a variable was not taken into account in the comparison.

The co-efficients of the regression analysis, with loneliness as the criterion variable, are given in Table 4.

Table 4 Co-efficients of regression analysis with loneliness as criterion variable
Full size table

According to Table 4, it is clear that each of the four variables is highly significant and thus makes a significant contribution to the variance of the criterion variable, which, in this case, is loneliness.

Multi-Variable Variance Analysis

The results of the multi-variable variance analysis carried out to test the research hypotheses, as postulated in section “Results”, are given in Table 5.

Table 5 Multi-variable variance analysis for testing main effects for loneliness and attitudes towards fathers as dependent variables
Full size table

According to Table 5, statistically significant differences occurred in respect of all the independent variables, i.e. gender (p = 0.001); age (p = 0.003); race (p = 0.000); language (p = 0.000); as well as socio-economic class (p = 0.000). A post hoc Scheffé test was carried out on each individual independent variable in order to determine which groups differed significantly, and which did not differ at all. These results are shown in Tables 6 and 7.

Table 6 Scheffé test results for clarification of differences in respect of race
Full size table
Table 7 Scheffé test results for the clarification of differences in respect of socio-economic status
Full size table

Gender and age consist of only two categories each; and it can thus be assumed that females differ significantly from males, and that adolescents in the 14–16 age group differ from those in the 17–19 age group.

According to Table 6, the statistics for the coloured group differ significantly from the statistics for both the black and white racial groups. Moreover, the black group differs significantly from the white and coloured racial groups. In addition, there are significant statistical differences between the white and the black groups, as well as between the white and the coloured groups.

With regard to socio-economic status, the Scheffé results in Table 7 show that those who fall within the low socio-economic class (with incomes ranging from less than R999 to R4999 per month) differ significantly from those belonging to the average socio-economic class (R5000–R8999 per month). Those who fall within the average socio-economic class (R5000–R8999), in turn, differ from those belonging to both the low and high socio-economic classes (R9000–more than R11000). However, those from the high socio-economic class (with incomes of R9000 to more than R11000) only differ significantly from those in the average socio-economic class (R5000–R8999).

Discussion

In the present research study, the relationship between loneliness and adolescents’ attitudes towards their fathers was examined, taking several independent variables into account, namely gender, age, race, language and socio-economic status. The most important finding of the study is that one’s attitude towards one’s father is the best predictor of loneliness during adolescence. The variable, attitudes towards fathers, accounts for 13.5% of the variance in respect of loneliness. The results obtained also showed that there is a highly significant inverse correlation between loneliness and attitudes towards fathers. Adolescents who have a negative attitude towards their fathers are inclined to be lonelier. The converse is also true: lonely adolescents tend to be those whose attitudes towards their fathers are negative. According to Lieberman et al. (1999) there is a changing nature of the father–child relationship in adolescence. For example, mothers remain emotionally involved with their children, while the quality of affect towards their fathers were rated lower by both sons and daughters, thus influencing loneliness.

In addition, it was found that socio-economic status is the second best predictor of loneliness. A highly significant inverse correlation can be observed between loneliness and socio-economic status. Danziger and Radin (1990) stress socio-economic status as a critical factor influencing the lack of involvement of fathers with their children in the context of disadvantaged communities. Insufficient income exposes parents to negative life experiences and stressors, which, in turn, inevitably lead to psychological discomfort in parents. This discomfort hinders the parents’ capacity to react sensitively and responsively to their children (Ackerman et al. 2003). Thirdly, race displays a statistically significant inverse correlation with loneliness, and can therefore be regarded as a good predictor for loneliness. The results also show that black adolescents are significantly lonelier than the other two groups, and that the white adolescents are significantly less lonely than the coloured and black adolescents. These findings concur with those of Jordaan (2004), who found a strong inverse correlation between loneliness and racial group. However, Jordaan (2004) also found that the variable, racial group, was not meaningful in the prediction of loneliness. This contradicts the findings of the present study.

Gender, as an independent variable, also appears to have a significant influence on loneliness as a dependent variable. The results indicate that male adolescents experience a significantly higher degree of loneliness than their female counterparts. Gender accounts for 0.06% of the variance of loneliness. According to Killeen (1998), it is generally assumed that the female gender is more susceptible to loneliness than the male gender. Similarly, it was found (cf. Inderbitzen-Pisaruk et al. 1992) that there is no difference between the genders concerning the experience of loneliness, but that gender differences did come to the fore in the variables that predict loneliness during adolescence. This does not tally with the findings of the present study, however. The results also show that loneliness correlates positively with age. The older the adolescents, the more susceptible they are to loneliness. Marcoen et al. (1987) found that as the adolescent became older, a decrease in parental involvement occurred, resulting in an increase in parent-related loneliness. These results are corroborated by Jordaan (2004), who found that the older the learners, the more inclined they are to experience loneliness.

In the present study, adolescents’ attitudes towards their fathers appear to be the best predictor of loneliness during adolescence. The results showed that there is a highly significant inverse correlation between loneliness and adolescents’ attitudes towards their fathers. This implies that adolescents with a negative attitude towards their fathers are more inclined to be lonely, and vice versa. This researcher agrees with Charity (2003) that to bring up a capable and well-adjusted child in present-day society is a highly complicated task that is made even more difficult if a single parent (usually the mother) is obliged to perform this task alone.

The breakdown of the nuclear family, the lack of involvement on the part of fathers, the rising divorce statistics and the increasing mobility of modern society all contribute to loneliness reaching epidemic proportions. These phenomena are especially problematic during adolescence; and because of the vulnerability of adolescents, it is of crucial importance that further research should be conducted in order to investigate the disturbing findings that have been brought to light in this study. Fathers are urged to become more involved in child rearing in order to contribute to their physical as well as their psychological well-being. Professionals who render services to reach this goal, should be informed about the crucially important role fathers have to play in this regard. Healing programmes should also include the father, which will be a critical step towards the prevention and palliation of deeply seated emotional problems, like loneliness.

Shortcomings include that only adolescents in the Mangaung area were included in the study. Therefore, the results cannot be extrapolated to all South African adolescents. Thus, it is recommended that in future studies, the sample should be extended to include adolescents from other cities and even other provinces. In this research, no distinction was drawn between geographical areas. In future studies, country districts, as well as informal settlements, could also be included in the research.