Single mothers by choice: Mother-child relationships and children's psychological adjustment.

Procedure

The families were assessed at home. Written informed consent to participate in the investigation was obtained from each parent and verbal assent was obtained from the child. Ethical approval was granted by the University of Cambridge Psychology Research Ethics Committee. Each parent was administered an audiorecorded standardized interview that lasted approximately 1.5 hr and standardized questionnaires and participated in a video-recorded observational task with the child that lasted 5–10 min. Teachers completed a questionnaire designed to assess the children’s psychological adjustment. Written informed consent was obtained from teachers. To provide interrater reliability ratings for the interview and observational measures, data from 30 randomly selected families were coded by a second interviewer who was blind to family type.

Mothers’ psychological wellbeing

The Trait Anxiety Inventory ( Spielberger, 1983 ), the Edinburgh Depression Scale ( Thorpe, 1993 ), and the short form of the Parenting Stress Index ( Abidin, 1990 ) were completed by the mothers to assess anxiety, depression and stress associated with parenting, respectively. Each of these instruments, for which higher scores represent greater difficulties, has been shown to have good reliability and to discriminate well between clinical and nonclinical groups.

Parent–child observations

The Etch-A-Sketch task ( Stevenson-Hinde & Shouldice, 1995 ) was used to obtain an observational assessment of interaction between the mother and the child. The Etch-A-Sketch is a drawing tool with two dials that allow one person to draw vertically and the other to draw horizontally. The mother and child were asked to copy a picture of a house, each using one dial only, with clear instructions not to use the other dial. The sessions were video-recorded and coded using the Parent–Child Interaction System ( Deater-Deckard & Petrill, 2004 ) to assess the construct of mutuality; that is, the extent to which the parent and child engaged in positive dyadic interaction characterized by warmth, mutual responsiveness, and cooperation. The following variables were rated on a 7-point scale ranging from 1 ( no instances ) to 7 ( constant, throughout interaction ): (a) child’s responsiveness to parent assessed the extent to which the child responded immediately and contingently to the mother’s comments, questions or behaviors; (b) mother’s responsiveness to child assessed the extent to which the mother responded immediately and contingently to the child’s comments, questions or behaviors; (c) dyadic reciprocity assessed the degree to which the dyad showed shared positive affect, eye contact and a “turn-taking” (conversation like) quality of interaction; and (d) dyadic cooperation assessed the degree of agreement about whether and how to proceed with the task. It was not possible to calculate Cohen’s Kappas for these variables due to restriction of the range of the scores as most families obtained scores at the top end of the scales. However, this did not reflect low interrater reliability as agreement within one point was 100% for dyadic reciprocity, above 95% for child’s responsiveness, above 90% for dyadic cooperation and above 85% for mother’s responsiveness.

Psychiatric ratings

The child’s psychological adjustment was also assessed during the interview with the mother using a standardized procedure ( Rutter, Cox, Tupling, Berger, & Yule, 1975 ). Detailed descriptions were obtained of any emotional or behavioral problems shown by the child. These descriptions of actual behavior, which included information about where the behavior was shown, severity of the behavior, frequency, precipitants, and course of the behavior over the past year were transcribed and rated by a child psychiatrist who was unaware of the nature of the study. A high level of reliability ( r = .85) between ratings made by social scientists and those made “blindly” by a child psychiatrist has been demonstrated for this procedure and validity has been established through a high level of agreement between interview ratings of children’s psychological problems and mothers’ assessments of whether or not their children had emotional or behavioral difficulties ( Rutter et al., 1975 ). Psychological problems, when identified, were rated according to severity on a 3-point scale ranging from 0 ( no disorder ) through 1 ( slight disorder ) to 2 ( marked disorder ) and type (anxiety, conduct/oppositional disorder, mixed disorder, Autism Spectrum Disorder, ADHD and speech delay).

Parenting

As shown in Table 2 , the positive parenting variables (expressed warmth, overt mother–child warmth, overt child-mother warmth, mother’s enjoyment of play, amount of interaction, and quality of interaction) were entered into a MANCOVA with family type (one-parent vs. two-parent) as the between-subjects factor and mother’s age, child’s age and mother’s working status as covariates. Wilks’ λ was not significant, F (6, 92) = 0.61, p = .72, showing that there was no difference in the level of positive parenting between the solo mother and two-parent families. When the negative parenting variables (frequency of battles, level of battles, conflict, and criticism) were entered into a MANCOVA with family type (one-parent vs. two-parent) as the between-subjects factor and child’s age, child’s gender, mother’s age, mother’s educational level, and perceived financial difficulties as covariates, Wilks’ λ was significant, F (4, 93) = 3.38, p = .01. One-way ANCOVAs identified a significant difference between groups for frequency of battles, F (1, 96) = 12.91, p = .001, d ′ = 0.51, reflecting less frequent battles between mothers and children in solo mother than in two-parent families. When this analysis was repeated without covariates, there remained a significant difference in negative parenting between the two family types, F (4, 98) = 2.64, p = .04, again reflecting less frequent battles in solo mother than in two-parent families F (1, 101) = 6.67, p = .01. There were no significant differences between family types for level of battles, criticism, or conflict. When this analysis was repeated without covariates, there remained a significant difference in negative parenting between family types, again reflecting less frequent battles in single than in two-parent families.

In addition, a MANCOVA was carried out for the variables relating to the mothers’ psychological wellbeing (Trait Anxiety Inventory, Edinburgh Depression Scale, and Parenting Stress Index total scores) with family type (one-parent vs. two-parent) as the between-subjects factor and perceived financial difficulties as a covariate. Wilks’ λ was not significant, F (3, 91) = 0.56, p = .64, showing that there was no difference in parental wellbeing between the solo mother and two-parent families.

The variables relating to the construct of mutuality from the observational assessment of the quality of mother–child interaction (mother responsiveness, child responsiveness, dyadic reciprocity, and dyadic cooperation) were also entered into a MANCOVA with family type (one-parent vs. two-parent) as the between-subjects factor and mother’s age, and child’s age as covariates. Wilks’ λ was not significant, F (4, 81) = 1.05, p = .38, showing that there was no difference in mother–child interaction between the solo mother and two-parent families.

Child Adjustment

A two-way ANCOVA with family type (one-parent vs. two-parent) and child’s gender (male vs. female) as between-subjects factors, and perceived financial difficulties entered as a covariate, was carried out for the total score of the Strengths and Difficulties Questionnaire (SDQ) rated by mothers. There was a significant main effect for child’s gender, F (1, 96) = 8.99, p = .003, showing higher levels of child adjustment problems among boys than girls. However, there was no significant difference between family types for the total score of the SDQ, F (1, 96) = 0.29, p = .59, and the interaction between family type and child’s gender was not significant, F (1, 96) = 0.61, p = .43.

Teachers’ total SDQ scores were also entered into a two-way ANCOVA with family type (one-parent vs. two-parent) and child’s gender (male vs. female) as between-subjects factors. The main effect for child’s gender was significant, F (1, 51) = 5.63, p < .05, indicating higher levels of child adjustment problems among boys than girls. Neither the main effect for family type, F (1, 54) = 1.23, p = .27, nor the interaction between family type and child’s gender, F (1, 54) = 1.70, p = .19, was significant. Although only 57% of the teachers completed the SDQ, the correlation between the mothers’ and teachers’ SDQ scores for those children for whom both questionnaires were available was significant (Pearson’s r = .53, p < 0.001). There was no significant difference in mothers’ total SDQ scores between those children for whom teachers’ SDQ scores were available and those for whom it was not, and no significant difference between family types in the proportion of teachers who did not complete this questionnaire.

The ratings of psychiatric disorder by a child psychiatrist found five children of solo mothers to show a slight disorder (one with speech delay, one with Autism Spectrum Disorder (ASD), one with conduct/oppositional defiant disorder (ODD), one with a mixed disorder including ADHD with ASD and ODD traits, and one with mixed disorder including ASD and speech delay) and two to show a marked disorder (one with speech delay and the other with a mixed disorder including speech delay, ADHD and anxiety). For the two-parent families, two children showed a slight disorder (one with ADHD and the other with speech delay) and two showed a marked disorder (one with speech delay and the other with a mixed disorder including ASD and speech delay). There was no difference in the proportion of children in solo mother and two-parent families rated as having a slight or marked psychiatric disorder, χ 2 (2) = 0.95, p = .62.

Parenting and Child Adjustment

Hierarchical regression analysis was used to examine factors associated with child adjustment problems in both family types (see Table 3 ). The outcome variable was children’s SDQ scores as completed by mothers. The demographic variables that were significantly correlated with SDQ scores, perceived financial difficulties and child’s gender, were entered in the first step. The variables added in the second step were a dummy-coded variable representing family type as well as the parenting and mothers’ wellbeing variables from the interview (positive parenting, negative parenting, maternal depression [Edinburgh Depression Scale], maternal anxiety [Trait Anxiety Inventory], and maternal parenting stress [Parenting Stress Index]). In the third step, the interaction terms between family type and each of the five independent variables representing parenting and mothers’ wellbeing were included. Perceived financial difficulties and child’s gender jointly explained 13% of the variance in children’s adjustment and the omnibus test for the first step was significant: F (2, 92) = 6.83, p = .002, showing greater adjustment difficulties for children whose mothers were experiencing financial hardship and also for boys. The parenting and mothers’ psychological wellbeing variables explained an additional 15% of the variance in children’s adjustment and the omnibus test for the second step was also significant: F (8, 86) = 4.14, p < .001. After accounting for the effects of perceived financial difficulties and child’s gender, maternal parenting stress was significantly related to children’s adjustment problems. The interaction terms explained an additional 4% of variance in child adjustment difficulties and the omnibus test for the third step was significant: F (13, 81) = 2.87, p = .002. None of the interaction terms was a significant predictor of child adjustment difficulties showing that there were no differences between family types in the relations between the parenting and maternal wellbeing variables and child adjustment.

Discussion

The children of single mothers by choice in the present study were found to experience similar levels of parenting quality to the comparison group of children in traditional two-parent families. For positive aspects of parenting as assessed by interview, there was no difference between the two family types, with mothers and their children in both the solo mother and two-parent families showing high levels warmth and interaction. Neither was there a difference in the quality of mother–child interaction as assessed through observation. There was a difference, however, for negative aspects of parenting as assessed by interview, with a lower frequency of conflict between mothers and their children in solo mother than in two-parent families. This finding remained when the analysis was repeated without covariates, suggesting that this represents a genuine difference between the two family types. Whereas this may seem to be an anomalous finding, in a previous study of a different sample, solo mothers with 2-year-old donor-conceived children were rated as feeling less anger toward their children than were a comparison group of partnered mothers and the children showed fewer emotional and behavioral problems ( Murray & Golombok, 2005b ). Although divorced and unmarried single mothers have been shown to experience raised levels of psychological problems, this was not found to be the case in the present study of single mothers by choice. The solo mothers did not differ from the partnered mothers in terms of anxiety, depression, or stress associated with parenting.

With respect to the psychological wellbeing of the children, no differences were found between the children in solo mother and two-parent families for emotional and behavioral problems as assessed by the Strengths and Difficulties Questionnaire completed either by mothers or the children’s teachers. In addition, the assessment of the presence of psychiatric disorder by a child psychiatrist showed no difference between family types in the proportion of children rated as having a psychiatric disorder. As the child psychiatrist was unaware of the child’s family type, these findings provide important validation for the mother’s reports. It is important to note that almost two thirds (64%) of the children with a slight or marked psychiatric disorder had a developmental disorder which is unlikely to be related to family type. When the four children with a marked psychiatric disorder were removed from the sample, the findings did not change for any of the parenting or child adjustment variables.

Although the children of solo mothers were no more likely to experience psychological problems than were those with a mother and father, higher levels of financial difficulties, and higher levels of parenting stress were each associated with higher levels of children’s emotional and behavioral problems within the solo mother families. As discussed above, financial hardship is one of the key predictors of psychological problems in the children of divorced or unmarried single mothers ( Hetherington & Stanley-Hagan, 2002 ; McLanahan & Sandefur, 1994 ; Pryor & Rodgers, 2001 ; Amato, 2000 , 2005 ), and it seems from the present study that financial hardship also has a negative impact on children’s psychological wellbeing in families where the mother has made an active decision to parent alone. The finding that stress associated with parenting was associated with child adjustment problems in the solo mother families is consistent not only with previous research on single mothers by choice ( Chan et al., 1998 ) but also with the broader research literature on divorced and unmarried single-mother families ( Amato, 2000 , 2005 ; Dunn et al., 1998 ; Hetherington & Stanley-Hagan, 1999 ; Pryor & Rodgers, 2001 ). It is important to emphasize, however, that the solo mother families did not differ from the two-parent families in the association between financial hardship, parenting stress, and child adjustment problems, indicating that these risk factors were operating in a similar fashion in both family types.

A potential risk factor for the children of single mothers by choice that does not exist for children from other types of single-mother family is their donor conception. As a result, the children grow up unaware of the identity of their biological father. The low level of psychological problems among the children of single mothers by choice in the present study suggests that lack of knowledge of the identity of their biological father does not have a negative impact on their psychological wellbeing. However, the mean age of the children was only 5-1/2 years at the time of study. Research on adoption has found that adopted children show an increased interest in their biological parents at adolescence, the time at which issues relating to identity become salient ( Brodzinsky, 2011 ; Grotevant & Von Korff, 2011 ). This suggests that the children of single mothers by choice may similarly become more interested in their biological father at adolescence and the absence of information about his identity may produce challenges at that time. In a qualitative study of solo mothers with 3–11-year-old donor-conceived children in Israel, Weissenberg and Landau (2012) reported that all of the children expressed a wish for a father.

A potential difficulty with the study is that differences between the solo mother and two-parent families may not have been identified due to the modest sample sizes. However, it would have been possible to detect a d (standardized difference between means) as small as 0.27 as statistically significant, for a power of 0.80. Thus, to the extent that significant differences between family types may not have been detected due to insufficient power, these differences would have been small. For the regression analysis which involved 103 families, 2 predictors in step one and 13 predictors in Step 3, it was possible to detect effects sizes as small a