Send Orders for Reprints to Reprints@benthamscience.net Determinants of Parenting Styles of Japanese Fathers and Mothers with Children Aged 0 to 10: Perceived Parenting during Childhood or Dysphoric Mood?

Purpose: To study the determinants of current parenting styles in Japan. Methods: Fathers (n = 312) and mothers (n = 333) of children aged between 0 and 10 attending a paediatric clinic were evaluated with regards to the following: parental attitudes rated by spouses using the Parental Bonding Instrument, current dysphoric mood assessed with the Hospital Anxiety and Depression Scale, life time history of Major Depressive Episode (MDE), and perceived parenting styles when the parents themselves were children. Results: Path models showed that current parenting styles were predicted by parents perception of maternal rearing during childhood. Moreover, in mothers only, they were predicted by the manner in which the women were raised by their fathers. This prediction was indirect, occurring via lifetime history of MDE and current dysphoric mood. Conclusion: This study suggests that the way in which parents themselves were raised was no less important than how they currently felt in determining how they were raising their child. Current dysphoric mood or lifetime history of MDE mediated such effects only among mothers.


INTRODUCTION
Parenting styles are of both clinical and research importance.Children's development is influenced by their interaction with the outer world, one of the most crucial components of which is parenting style.Among the many attributes of parenting styles, many investigators have referred to responsiveness (acceptance/care) and demandingness (control/overprotection) as two of the most significant dimensions.Based on factor analyses of parents' behaviour [1], Baumrind [2] confirmed that parenting styles consisted of these two domains.Responsiveness refers to the extent to which parents foster individuality and self-assertion and this includes parents' warmth towards the child.Demandingness refers to the requirements parents place on children, via behavioural regulation and direct confrontation, regarding integration into society.A similar two-dimensional theory-acceptance/rejection vs control-was proposed by Rohner and Pettengill [3].In a factor analysis of a children's report about their parents' behaviours, Schaeffer [4] identified the following three factors: Acceptance vs Depression has often been associated with parents' attitudes towards their children (Fig. 1A).Parents with depression were found to be more likely to be irritable [9].Mothers who were depressed were more likely to show destructive responses (such as use of criticism, scolding, threatening, withdrawing privileges, and aversive physical control with the intent of stopping the child's behaviour) in a free-play task with a 16-to 18-month-old child [10].Other investigations have shown that the parenting styles of depressed parents differed from those of non-depressed parents [11][12][13].Anxiety may also be associated with poor parenting styles.In this study, we hypothesised that current parenting styles would be predicted by parents' current dysphoric mood state including depression and anxiety.
(A) Current parenting style is predicted by current dysphoric mood.
(B) Current parenting style is predicted by past history of depression.
(C) Current parenting style is predicted by perceived parenting of grandparents (i.e., the enrolled children's grandparents).
(D) Current parenting dysphoric mood and history of Major Depressive Episode are predicted by perceived parenting of grandparents.

Fig. (1).
Major hypotheses of the determinants of current parenting styles.
A second possible determinant of parenting styles is the lifetime history of depressive disorder (Fig. 1B).For example, Kendler, Sham, and MacLean [14] reported that lower levels of parental warmth were predicted by a lifetime parental history of Major Depression and Generalised Anxiety Disorder.Parents with a past history of depressive disorder may be more likely to experience greater current depressive mood and adopt poor parenting styles.The negative effect of past depression on current parenting styles may be direct or indirect, mediated in the latter case by current depressive mood.This issue has yet to be studied empirically.
Parenting styles including abusive behaviours may be transmitted from one generation to another [15][16][17][18][19][20] (Fig. 1C).Adolescent daughters of mothers who exerted affectionless control over them treated their own child in the same way 10 times more frequently than did daughters of mothers who did not exert this type of control [21].Kitamura, Shikai, Uji, Hiramura, Tanaka, and Shono [22] reported that parenting styles in a Japanese population were transmitted directly from grandparents to parents.Hence in the present survey, we hypothesised that current parenting styles would be predicted by the parenting styles of the children's grandparents.
We also hypothesised that the effects of a person's own childrearing experiences on their current parenting styles would be mediated by both the lifetime history of depressive disorder as well as current dysphoric mood.Ample evidence shows that past episodes of depressive disorder and current depressive states are associated with inadequate parenting experienced by individuals during childhood [15,[23][24][25][26][27][28][29][30][31][32][33][34][35] (Fig. 1D).We also hypothesised that effects of the lifetime history of depressive disorder on current parenting styles would be mediated by current mood state.
A methodological drawback of past studies examinig the links between personal variables such as dysphoric mood (depression and anxiety) and childrearing styles is their reliance on the participant as the sole source of information.These individuals who are currently depressed or anxious and those who perceive their own parents' childrearing less affectionate may be more likely to view their own rearing styles as less affectionate.In order to avoid such halo effects, we asked spouses to report on the childrearing styles of their partner.
Our research hypothesis (Fig. 2) is that current parenting styles are predicted by current dysphoric mood (Fig. 1A) and lifetime depressive episodes (Fig. 1B) that mediate (Fig. 1D) the effects of parenting styles perceived when the parents themselves were children (Fig. 1C).This mediation may be partial.In such a case, perceived parenting styles during childhood may also exert a direct effect on current parenting styles.

Participants
This questionnaire survey was conducted in collaboration with the Kumamoto Paediatric Association.We solocited the participation of Association members and 20 clinics agreed to take part.Paediatricians handed the questionnaire to the parent(s) of each child aged less than 11 years who visited the clinic.Thus, the present sample was convenient.If only one parent was present, this parent was given another questionnaire so that each partner would have a copy.Fathers and mothers were asked to fill in the questionnaire independently, and then return it in a stamp-added envelope.All the participating clinics provided generalised child care as opposed to specialised treatment of particular paediatric conditions such as developmental disorders.In Japan, children with serious diseases are referred to specialised institutions such as university hospitals and therefore, most of the children in this study were unlikely to be suffering from serious medical conditions.
In 759 families, at least one parent responded to the questionnaire survey.Overall, 395 fathers and 731 mothers responded.In 367 families both parents did so.There were 176 boys (51.4%) and 184 girls (45.6%).The mean age (SD) of the children was 3.4 (2.7) years with a range between 0 and 10 years.The gender of the remaining seven children was not reported.In order to avoid shared observer bias, we used the spouse's assessment of each individual's parental style.Thus, we used the data of 312 (79%) fathers whose wives reported their (fathers') parenting styles and 333 (46%) mothers whose husbands reported their (mothers') parenting styles.The mean (SD) ages of the fathers and mothers were 35.2 (5.8) and 33.6 (5.2) years, respectively.The fathers were significantly older than the mothers (p < 0.001).
This research project was approved by the Ethical Committee of Kumamoto University Graduate School of Medical Sciences.

Current Parental Styles
The Parental Bonding Instrument (PBI) [24,36] is a measure used to retrospectively assess how an individual perceives their father's and mother's parenting (separately).There are two subscales: Care and Overprotection.Care consists of 12 items (rated on a 4-point scale) related to a parenting style that ranges from coldness, indifference, and neglect, to affection, emotional warmth, empathy, and reciprocity.Overprotection consists of 13 items (rated on a 4-point scale) ranging from parental control and overprotection, intrusion, and infantilization, to parental allowance, independence, and the development of autonomy.The Japanese version of this scale was used [7,37].
To assess the parents' current parenting styles, we used the spouse-rated PBI in order to avoid shared observer bias.Thus, we rephrased sentence into the present tense and then asked each spouse to rate their spouse's parenting styles towards the child who consulted the clinic.Mothers, therefore, were asked to rate fathers' parenting styles, while fathers were asked to rate mothers' parenting styles.Modification of the PBI from the past to present tense has been reported in previous investigations [14].Use of a spouse as an informant on parental styles has been adopted in previous investigations [e.g., 22,38].As an example of the process, an original item, "My father spoke to me with a warm and friendly voice" would be transformed into "My husband speaks to the child with a warm and friendly voice".Fathers and mothers agreed with each other to a moderate extent about their parenting style assessments: maternal Care r = 0.55; maternal Overprotection, r = 0.57; paternal Care r = 0.38; paternal Overprotection, r = 0.38 [37].Missing PBI values were substituted with the mean of the item only for those cases in which fewer than six items (20% of the total) were missing.The mean was substituted for 5.1% cases of fathers' ratings of mothers' parenting styles and 8.1% cases of mothers' ratings of fathers' parenting styles.

Parental Dysphoric Mood
The Japanese version [39,40] of the Hospital Anxiety and Depression Scale (HADS) [41] is a measure of current affective and cognitive symptoms of depression and anxiety with an established factor structure [42].Parents used the HADS to assess their current mood state.This instrument has two subscales: Depression and Anxiety, each with seven items.The scale is rated from 0 (low depression or anxiety) to 3 (high depression or anxiety) and the total scores of the HADS Depression and HADS Anxiety can therefore range from 0 to 21. Validity of the HADS was found to be correlated with other commonly used questionnaires and rating scales such as the Beck Depression Inventory, Symptom Checklist 90 Scale, State-Trait Anxiety Inventory, Montgomery-Asberg Depression Rating Scale, and Structured Clinical Interview for DSM-III [43][44][45].
Missing HADS values were substituted with the mean of the item only in cases where fewer than three items (25% of the total) were missing.The mean was substituted for 1.9% of assessments of fathers' dysphoric mood and 1.8% of those of mothers' dysphoric mood.

Lifetime History of Major Depressive Episode
To identify the lifetime history of Diagnostic and Statistical Manual of Mental Disorders 3 rd version Revised (DSM-III-R) Major Depressive Episode (MDE), we used an ad hoc questionnaire consisting of 16 items that tapped criteria for current or past MDE [46].Each item was rated on a 2-point scale (true/false) for the same period and we diagnosed MDE using an algorithm according to the DSM-III-R MDE criteria.This method is very similar to the lifetime version of the Inventory to Diagnose Depression (IDDL) [47], the validity of which was reported to be high using the interview-based DSM diagnosis of MDE.The IDDL's current version was translated into Japanese and its validity was confirmed using the Structured Clinical Interview for DSM-III-R as an external validation criterion [48,49].It is of note that the questionnaire provides three items for criterion A2 of MDE (loss of interest): (a) lost interest in things, (b) lost pleasure in things, and (c) lost interest in opposite-sex partners (loss of interest in sex).We defined this criterion as being met if all three A2 items were fulfilled.

Grandparents' Parenting Style
To measure the parenting styles that fathers and mothers perceived as children, we used the Parental Bonding Instrument (PBI; [24]) in its original form.The participating parents were asked to report how they had perceived their own parents' (i.e., the enrolled children's grandparents') attitudes towards them when they were aged 16 or less.
Missing PBI values were substituted with the mean of the item only for those cases with fewer than six items (20% of the total) missing.The mean was substituted for 1.3% and 2.9% cases of paternal grandfathers' and paternal grandmothers' parenting styles, respectively and for 2.4% and 3.6% cases of maternal grandfathers' and maternal grandmothers' parenting styles, respectively.

Statistical Analyses
After calculating means and SDs of all the variables used in this study, the subscales of the current parenting styles rated by the spouse -Care and Overprotectionwere correlated with predictor variables for the fathers and mothers separately.The predictors included (1) the parents' and children's ages, (2) current dysphoric mood including depression and anxiety, (3) lifetime history of MDE, and (4) perceived rearing experienced by the parents during childhood.
Because these predictor variables were correlated with each other and occurred sequentially during the life course, we performed path model analyses in the context of structural equation modelling (SEMs) (Fig. 2) following our theoretical consideration.We posited that the children's grandparents' parenting styles (i.e., the parenting styles experienced by the parents during their childhood) would appear first in the life course of individuals.This would then be followed by development of MDE before current Depression and Anxiety.Finally, current parenting styles would appear.We also posited four latent variables: Grandfathers' and Grandmothers' Parenting Styles (composed of Care and Overprotection), Paternal and Maternal Parenting Styles (composed of Care and Overprotection), and current Dysphoric Mood (composed of Depression and Anxiety).It was shown later in this study as well as in previous investigations that the indicator variables of each latent variables were moderately correlated with each other (e.g., [7]).We also surmised that what would matter most were not the different effects of Care and Overprotection on the child but the overall picture of the nurturing environment.We hypothesised that in the path model, variables appearing earlier in the life course would predict all the variables that appeared later.In addition, children's and parents' ages were expected to predict the current parenting styles.Finally we set covariances between children's and parents' ages as well as Grandfathers' and Grandmothers' Parenting Styles (Fig. 2).
All the statistical analyses were conducted using the Statistical Package for the Social Sciences (SPSS) version 20.0 and Amos 20.0.

RESULTS
Means and SDs of all the variables used in this study are shown in Table 1.In fathers, spouse-rated Care was correlated with child's younger age, higher grandmothers' Care, and lower grandparents' Overprotection.In fathers, spouse-rated Overprotection was correlated with child's younger age and higher grandfather's Overprotection.In mothers, spouse-rated Care was correlated with child's younger age and higher grandmother's Care.
Because many of the predictor variables were intercorrelated (data not shown) and we theoretically considered that some predictors preceded others chronologically, we next performed path analyses using AMOS models for fathers and mothers separately (Figs. 3, 4).

DISCUSSION
The present study showed that in bivariate correlations current parenting styles-Care and Overprotection-were not predicted by parents' current level of depression or anxiety.However, SEM indicated that in mothers, Maternal Dysphoric Mood (composed of Depression and Anxiety) significantly predicted Current Maternal Parenting Style (composed of Care and Overprotection).This was not the case in fathers.This suggests the importance of dysphoric mood rather than depression per se.Past investigations have paid more attention to the effects of depression on parenting styles and less to the effects of anxiety.It may be beneficial to investigate the effects of a broader range of mood states on parenting styles, including fear and anger.
As expected, current Maternal Dysphoric Mood was predicted by lifetime history of MDE in both fathers and mothers.The link between MDE and lower levels of parental warmth reported by Kendler, Sham, and MacLean [14] may be explained by the possibility that current dysphoric mood mediates the effects of lifetime history of MDE on poorer parenting styles.
The lack of a link between current mood state (depression and anxiety) and parenting styles among fathers in this study may be explained by the fact that we removed the shared source variance by having spouses report on parenting style.This was a unique method that has been used infrequently.Alternatively it may be because we did not aim to recruit parents with high depression or anxiety scores.A third possibility is that fathers who were depressed or anxious may have been less likely to participate in this study so that the resulting participants may have been healthier.Using a path model, this study showed that parenting styles perceived when parents were children predicted current parenting styles directly and, in mothers, indirectly as well.In mothers, poorer parenting styles of grandfathers and grandmothers predicted a lifetime history of MDE: this in turn predicted current dysphoric mood and led to poorer current parenting styles.A lifetime history of MDE also had a direct path towards poorer current parenting styles.Thus, the lifetime history of MDE itself mediates the effects of perceived parenting during childhood on current parenting styles.
This study also showed a direct link between grandmothers' parenting styles perceived when parents were children and the current parenting styles of fathers and mothers.Parenting styles of grandmothers perceived when parents were children may have influenced current parenting styles through learning.Alternatively, the link between the two may be mediated by variables that were not examined in the present study including parental personality [22], and marital adjustment [54].Investigation into these potential mediators is of importance not only for research purposes but also clinical ones.This is because such mediators may be used as a target for effective parenting programmes.
Parental gender specificity was observed in this study.Thus, the fathers' Current Parenting Style was predicted only by how they remembered being raised by their own mothers whereas the mothers' Current Parenting Style was predicted directly by how they remembered being raised by their own mothers and indirectly by how they remember being raised by their own fathers.It has been suggested that father-son, father-daughter, mother-son, and mother-daughter are distinct dyads [55].
The strengths and limitations of this study should be noted.First, the present sample used spousal reports as the measure of the current parenting styles.The PBI scores are associated, though modestly, with the tendency toward social desirability [37].Yet direct observation of parenting by a researcher is time-consuming and may result in distortion of behaviours in parents.However, spouses have rarely been used as informants on the other parent's parenting styles [37,38].Spouses may observe parenting behaviours in more naturalistic settings and shared observer bias may be reduced by the use of a spouse as a reporter.
A drawback of this study is that it used a convenient sample.The representativeness of the parents is in question.The attrition rate of parents was substantial because we selected families in which both parents returned the questionnaire.Another drawback is this study's crosssectional research design.Parents who are depressed or anxious may be more likely to report past MDEs and to describe the way they were raised from a more negative perspective.However, there are reports showing that memories of childhood rearing are minimally influenced by current mood state [56][57][58][59].We used questionnaires whose reliability and validity have been reported.However, the questions regarding lifetime history of MDE were constructed based on IDDL [47].A methodological limitation of this study is the fact that we did not diagnose the history of MDE using a structural diagnostic interview.The questionnaire assessing the history of clinical depression was developed ad hoc and has not fully tested in terms of its psychometric properties.Despite the validity of an instrument of similar design, the questionnaire used in this study should not be used to form definitive conclusions.Hence, its reliability and validity should be evaluated in future studies.
Since this study used a non-clinical population, reported parenting difficulties may have been fairly mild in nature.The HADS Depression and Anxiety scores in this study were slightly lower than those reported in a Japanese undergraduate population [53].Further studies should be carried out in a clinical population where severe psychopathology result in a different picture.
Combinations of fathers and mothers and sons and daughters should also be investigated in the future.This is because gender difference between parents and children have been less extensively studied [55].
Caution should also be exercised before forming conclusions based on this study because of the age range of the children examined.The PBI was used for parents of children aged 0-10.Many PBI items, however, may not be valid for parents of very young children.At least five of the items are about autonomy (e.g., the child being free to make his or her own decisions, choose his or her own clothing, go out alone, etc.).These items would clearly be irrelevant for children under two and would be questionable for preschoolers.Perhaps restricting the age range of the children to 2 -10 years would be more appropriate.Therefore, we repeated our initial statistical analyses after selecting only parents of children aged two or older.SEM results for fathers and mothers individually showed virtually the same pattern as the SEMs outcomes in this study (data not shown).A few estimates (e.Taking into consideration these methodological shortcomings, we believe this study suggests that the way in which parents themselves were raised was no less important  than how they felt currently in determining how they were raising their child.Current dysphoric mood or lifetime history of MDE mediated such effects only among mothers.Among fathers, perceived parenting in the past was found to be the sole determinant of current parenting style.
g., Paternal Grandmothers' Parenting Style on Current Paternal Parenting Style, estimate = 0.80; Maternal Grandmothers' Parenting Style on History of MDE, estimate = -.12)lost their significance possibly due to the smaller number of cases (n = 222 and 237 for fathers and mothers, respectively).

g. 4 .
Path model of mothers' current parenting styles with their potential predictor variables