High Recalled Investment in Children: The Development of a New Measure and Associations with Depression in Primary Care
B. H. Green and J. HillFrom the Department of Psychiatry, University of Liverpool.
Address for correspondence:
Dr. B. Green Senior Lecturer,
University Department of Psychiatry,
Royal Liverpool University Hospital,
Liverpool, L69 5GA.
It was hypothesised that mothers who give priority to relationships with their children at the expense of other relationships (high 'investment') will be vulnerable to depression when their children grow up. A new questionnaire, the Relationship with Children Scale (RCS) was devised in order to assess recalled relationships with young children. The RCS was shown to have satisfactory psychometric properties. Samples of women aged 50-70 with and without histories of RDC Major Depression were drawn from primary care, and compared on the PBI Care and Overprotection scales and the RCS Care, Control and Investment scales. A history of Major Depression was associated with low Care, high Overprotection and high Investment. The association between Investment and Major Depression held after subjects with current episodes had been removed, and for those subjects with first episodes after the age of forty. There was some evidence that a link between recalled maternal Overprotection and Major Depression was mediated via high recalled Investment in young children. Replication of these findings is necessary, together with prospective studies in order to test a number of possible explanatory models.
There is substantial evidence that adverse experiences in childhood are associated with an increased risk of adult psychiatric disorder (Bifulco et al 1991, Mullen et al 1993, Kendler et al 1996), and with impaired adult relationships especially cohabiting or other intimate relationships (Quinton et al 1990, Rutter et al 1990, Rodgers 1996). Experiences and choices in the transition from adolescence to adulthood often appear to be crucial to outcome. However this may not be the only developmental task that carries risks for psychopathology. In this paper we consider the possibility that vulnerability to depression may be related to the changing demands made on individuals in the transition from parenthood to later adult life without young children.Social and psychological theories of depression have focused on loss (Bowlby 1980), threat (Brown and Harris 1978), helplessness (Seligman 1974 ), distorted cognitions and attributions (Beck et al 1983), deficits in autobiographical memories (Williams 1996), interacting cognitive subsystems (Teasdale 1993), and negative self-concept (Brown and Harris 1978, 1989). It is implicit in each of these theories that the vulnerability to depression arises significantly within more or less intimate social relationships. They differ however in the extent to which they see the subject as the recipient of adverse experiences such as loss or threat, or as the author of his/her own experiences particularly through the choice and negotiation of relationships. It is probable that the difference is one of emphasis in that often where adversity has been implicated so also have adult relationships (Brown 1992). In looking to relationships for vulnerability and protective factors for depression we are concerned both with the active contribution of the individual and what is experienced in his/her relationships as risk or resource. These considerations imply that a further step may be fruitful, whereby not only the relative contributions of the individual and the environment are examined, but relationships themselves become the object of study. This is an approach that is well established in ethology (Hinde 1979), and family and developmental studies ( Hinde and Stevenson-Hinde 1988, Dunn 1988) but has received relatively little attention in relation to adult depression. Relationships, especially those in the family, change form and function over time so that the possibility arises that relationships that appear at one point in development to be supportive or protective may at another time render the individual vulnerable (Cicchetti and Cohen 1995). If this were the case it would have major implications for what are considered outcomes in studies of adult functioning in relation to childhood psychopathology and experiences.
In this study we considered the possibility that a form of parenting relationship that appears stable and adaptive when children are young might later be a vulnerability factor for depression. We hypothesised that the loss of role of mother to young children might be a vulnerability factor for depression in middle aged women where there had been high emotional investment in relationships with children when they were young, at the expense of other adult relationships. In order to test this hypothesis adequately it would be necessary to identify women with young children who showed such highly exclusive relationships, and in a follow up study, examine the prediction to depression once the children had reached early adult life. A less satisfactory but nevertheless potentially informative and more readily achievable design would assess recalled investment and compare depressed and non-depressed groups, and this was the approach taken in this study.
The Relationship with Children Scale – A New Measure
As a measure of recalled relationships with children was not available BG and JH designed a questionnaire, the Relationship with Children Scale (RCS). The questionnaire (which is shown in the Appendix) has 30 statements referring to the person's recalled relationship with his/her children when they were between the ages of 5 and 10, and the respondent is asked to tick one of four boxes ranging from 'Very Like' to 'Very Unlike'. The 'Investment' statements reflect the intensity of the relationships, and their importance to the person relative to other adults especially spouse or friends. Examples of items to which the response 'Very Like' would contribute to high Investment include 'They were my main source of emotional satisfaction' and to low Investment 'I was closer to my husband than my children'.Items which it was anticipated would reflect 'Care' and 'Control' were included, for four reasons. Firstly we anticipated that the validity of the Investment scores would be increased if the statements were embedded in a questionnaire that was experienced as addressing a range of parenting issues and not just one relatively narrow area. Secondly it was important to establish whether a dimension of Investment could be identified that was relatively independent of other dimensions of parenting that have previously been investigated (Baumrind 1971, Parker et al 1979). Thirdly the inclusion of Care and Control Scales provided a direct parallel with the Parental Bonding Instrument (Parker et al 1979), and hence possibilities of studies of links between recalled childhood experiences and recalled parenting relationships using both instruments. Fourth, in studies of depression if there were an effect of low mood on reporting of parenting it would be most likely to be reflected in the perception of having been a bad parent, that is to say of having provided low care.
In this paper we report the extent of association among the RCS scales and their internal reliability, a comparison of the RCS scores of depressed and non-depressed women on each of the scales, and the relative roles of PBI and RCS scores in the prediction of depression.
Twenty five women aged 50-70 who were known to have had children and who had been diagnosed as depressed during the previous year were randomly selected from the computer database of a busy inner city general practice in central Liverpool. These twenty five patients were invited by letter to take part in the study and of these twenty agreed. A further thirty women in the same age range who had not been recorded as depressed over the previous year were randomly selected from the same database, and of these twenty five agreed to participate. Subjects were included in the depressed group if they had had at least one episode of RDC Major Depression since age 18 (derived from the depression section of the SADS-L, see below), and all of those identified by the GP's met this criterion. A further four patients from those originally identified as controls had had previous episodes of Major Depression and were reassigned to the depressed group, so that the 'ever depressed' group was comprised of twenty four women (mean age 62.5 years )and the non-depressed control group of twenty one (mean age 61.0 years).
MeasuresAdult lifetime RDC Major Depression was identified using the depression section of a modified version of the SADS-L (Endicott and Spitzer 1978, Harrington et al 1989) which has established inter-rater reliability and validity. It was administered by BG who had been trained by an experienced interviewer.
At the same interview subjects were given the Parental Bonding Instrument (Parker et al 1979a) and the Relationship with Children Scale. The Parental Bonding Instrument is psychometrically robust (Parker 1989), and it does not appear to be strongly affected by current mood (Parker 1981, Gotlib et al 1988, Duggan et al 1998). Associations of the scales with adult psychopathology, particularly depression, have been demonstrated in referred and general population samples (Mackinnon et al 1989, Kendler et al 1993, Rodgers 1996). The PBI has 25 items referring to recalled relationships with each parent and in most studies the dimensions of ‘Care’ and ‘Overprotection’, which were originally generated by factor analysis, have been used. Some recent studies have indicated that a three factor structure is more appropriate (Kendler 1996, Murphy et al 1997) but in this study, for comparability with the majority of reported findings, we have used the two factors. In some studies the combination of low care and high control ‘affectionless control’ has been associated with depression (e.g. Gottlib et al 1988), however in others (e.g. Mackinnon et al 1989) only low care has been an independent predictor.
The possibility that the differences between the depressed and non-depressed groups arose from an effect of current mental state on RCS scores was examined by comparing the six subjects with current RDC Major Depression and the eighteen with lifetime depression but no current episode. The means for Investment, Overprotection and Care of the currently depressed and previously depressed were 20.8 vs 22.8, 13.2 vs 13.0 and 25.3 and 27.6. None of these was significant which is not surprising given the small number of currently depressed subjects. Where there were trends they were in the opposite of the predicted direction if mental state were to explain high Investment and in the predicted direction if it were to explain low Care. The comparison between the ever-depressed and non-depressed groups, omitting those currently depressed, was repeated and the difference on the Investment scale was slightly higher and highly significant (Table 2) whilst the differences on the Overprotection and Care scales remained non-significant.
Table 2 Comparison of mean PBI Care and Control Scales and RCI Investment Scale for a) the whole sample, and after removal of b) subjects with current RDC Major Depression and c) subjects with first episode before age 40.
No Adult Lifetime Depression
Adult Life Depression
|(a) Whole Sample N=24||
(b) Current Depression Omitted (N=18)
|(c) First Episode <40 Omitted (N=17)|
Means were compared using two tailed between groups t-tests, and significance of differences between ever depressed and never depressed groups are indicated by p<0.05=*, p<0.01 =**, and p<0.001 = ***.
Our hypothesis had been that vulnerability to depression would arise from a combination of high investment in children when they were young, coupled with the development of their subsequent independence. Therefore we tested to see if the results applied for those whose first episode had occurred after the age of 40 when in most cases children will at least have reached adolescence. Seven out of the 24 ever-depressed women had had their first episode before the age of 40, and in a comparison of those with onset before and after 40 there were no significant differences on any of the RCS scales. Women with first episode of depression at over age 40 and never-depressed groups were compared and the difference on the Investment scale remained substantial and significant (Table 2) whilst those for the other two scales were small and non-significant.
Parental Bonding InstrumentIn a comparison of the ever-depressed and never-depressed groups the means for maternal Care were 23.4 and 29.6, and for maternal Overprotection 14.5 and 8.6, and both of these differences were significant at p<0.01 using two tailed t-tests (Table 2). These differences were smaller and non-significant once the currently depressed subjects had been removed but comparable and significant when only those with first episodes over age 40 were included (Table 2). Nine of the 24 depressed subjects reported Affectionless Control and one of the 21 controls (c2 = 6.95, df = 1, p = 0.012 (Fisher’s Exact Test)
The relationship between the RCS and the PBI scores was examined. The RCS Care scores were correlated with the PBI maternal Care scores (r=0.56, p<0.001) and the RCS Investment scores were correlated with the PBI Control (r= 0.39, p<0.01). Next we examined whether there was evidence that links between RDC Major Depression in middle aged women and recalled childhood experiences was mediated via high Investment as parents. This was done using dichotomised variables so that possible interactions could be examined. Median splits were used to maximise power and the cut-offs were 27-28 for maternal Care, 11-12 for maternal Overprotection, and 18-19 for Investment in children. Of the 24 women with histories of depression 16 (67%) reported low care compared with 6/21 (29%) of those without depression (c2 = 6.51, df = 1, p = 0.011), and 15/24 from the depressed group had high Overprotection compared with 3/21(14%) of the non-depressed group (c2 = 10.85, df = 1, p = 0.001). In a forward stepwise logistic regression with RDC Major Depression as the dependent variable and low Care and high Overprotection as predictors only high Overprotection was entered (Wald = 9.36, p = 0.002), and there was no interaction between Care and Overprotection. Only Overprotection was included in further analyses. Table 3 shows the numbers of subjects with high Overprotection and Investment in relation to lifetime Major Depression. Fifteen out of the 26 with high Investment also recalled high maternal Overprotection compared with 3/19 (16%) of those with low Investment (c2 = 8.03, df = 1, p = 0.005). A forward stepwise logistic regression with high Investment and high maternal Overprotection as predictors and Major Depression as the dependent variable was carried out. High Investment was entered first (Wald = 10.60, p = 0.001) and then Overprotection (Wald = 3.90, p = 0.048), and there was not interaction between the predictor variables. This suggested that high Investment was a mediator between high Overprotection and Major Depression.
A new instrument, the Relationship with Children Scale, was developed in order to assess recalled care and control of, and emotional investment in, young children. Two of the three scales showed substantial variability of scores and all three scales demonstrated moderately good internal reliability. In this primary care sample of women aged 50-70, Major Depression in adult life was associated with a markedly increased level of recalled emotional Investment in children. This difference persisted after account had been taken of current depression, and where only women with first episodes in middle age were included. High PBI Overprotection was associated with RCS high Investment and with Major Depression. Logistic regression models with binary predictor variables suggested that the link between high maternal Overprotection and Major Depression was mediated via high Investment in children.The discussion will focus on the sample, the measures and the interpretation of the findings. In many respects the sample is likely to have been representative of patients in the age range 50-70 presenting with depression in primary care. The practice from which the subjects were drawn is a large group practice serving a relatively stable population, spanning a very deprived community (Toxteth) and a more affluent area (Aigburth). Patients who had very recently registered with the practice would have been excluded from the sampling process through lack of records, which therefore may have introduced a slight bias towards more 'settled' individuals. There is no reason to expect that the practice population was unrepresentative of the general population of the city, except in under-representing the highest socioeconomic groups. It is likely that some women with depression had not been identified by their G.P., however it is unlikely that this would have introduced systematic bias with respect to PBI or RCS scales. Any that had been missed and included initially in the control group would have been moved into the depressed group on the basis of the SADS-L assessment. The comparability of this sample with those of previously reported studies is indicated by the findings of associations between Major Depression and PBI Care and Overprotection scales and 'Affectionless Control' (Mackinnon et al 1983). However it was unusual in that there was a stronger association of depression with the binary Overprotection variable than with the Care variable. In general people are not good at recalling accurately previous episodes of depression. Green et al (1994) found that only about 10% of people over the age of 65 correctly reported an episode of depression three years before. Whilst recall of any adult lifetime depressive episodes is likely to be better than that, it is possible that despite the efficacy of the SADS-L (Endicott and Spitzer 1978) some subjects who had experienced lifetime depression were included in the control group. The internal reliability of the RCS was moderately good although further studies using larger samples are required in order to establish whether it is satisfactory. Face validity appeared to be good and construct validity was not attempted because, as far as we are aware there are not any equivalent measures. The RCS appeared acceptable to the subjects and was generally as easy to administer as the PBI. For some people at the outset some explanation of the possible responses was necessary. This might suggest that if the RCS were administered postally a comparison of postal replies and those given in a setting where queries could be addressed might be needed. In interpreting the findings it is crucial to emphasise that replication and clarification are needed. Replication will also entail the investigation of possible sources of variation, for instance across social class or economic circumstances. The meaning of the results will be clarified where studies also assess actual changes in relationship with children and spouse, and where they examine the prospective predictive power of high investment in children. Some of the RCS items refer to priority given to relationships with children over the marital relationship. Prospective studies will be able to determine whether the prediction from Investment holds after controlling for quality of marriage. It will be important to devise studies to disentangle several possible causal links. An aspect of personality functioning might be manifested as investment and also render the individual vulnerable to depression, or vulnerability might stem from the loss of the relationships with children, or from lack of intimacy with spouse, or a combination of these.
Our findings raise questions concerning the experiences of the children of the mothers in our study. Did they experience high investment as high care or overprotection? Have they shown evidence of psychopathology or relationship difficulties in childhood or adulthood, or is high Investment more of a personal attribute with no particular implications for parenting? Studies of intergenerational patterns of relationships could address these issues and hence illustrate further mechanisms in the intergenerational transmission of psychological health and disorder.
Baumrind, D. (1971). Harmonious parents and their preschool children. Developmental Psychology 41 (1) 92-102.Beck, A.T., Epstein, N. and Harrison, R. (1983). Cognitions, attitudes and personality dimensions in depression. British Journal of Cognitive Psychotherapy 1(1), 1-16. Bifulco, A. , Brown,, G. W. and Adler,Z. (1991). Early sexual abuse and clinical depression in adult life. British Journal of Psychiatry 159, 115-122. Bowlby, J(1980) 'Loss: Sadness and Depression' Vol 3 of Attachment and Loss. New York: Basic Books. Brown,G.W. (1992) 'Social support: an investigator-based approach' In The Meaning and Measurement of Social Support (eds H.O.F. Veiel and U. Baumann) pp235-257. Washington: Hemisphere. Brown G.A and Harris T.O. (1989) Life Events and Illness. New York: Guilford. Duggan, C., Sham, P., Minne, C., et al (1998) Quality of parenting and vulnerability to depression: Results from a family study. Psychological Medicine, 28, 185 - 191. Dunn, J. (1988) The Beginnings of Social Understanding Oxford: Blackwell. Endicott, J. and Spitzer, R.L. (1978). A Diagnostic Interview. The Schedule for Affective Disorders and Schizophrenia. Archives of General Psychiatry 35, 837-844. Gotlib, J. I., Gotlib, I. H., Mount, J. H., Corvy, N. I., and Wiffen, V. E. (1988). Depression and Perceptions of Early Parenting: A longitudinal investigation. British Journal of Psychiatry, 152, 24-27. Green, B.H., Dewey, M.E., copeland, J.R.M. et al. (1994). Factors accociated with recovery and recurrence of depression in older people. International Journal of Geriatric Psychiatry 9, 789-795. Harrington, R., Hill, J., Rutter,M., John,K., Fudge, H., Zoccolillo, M. and Weissman, M. (1988). The assessment of lifetime psychopathology: a comparison of two interviewing styles. Psychological Medicine 18, 487-493. Hinde, R.A. (1979) Towards Understanding Relationships. London: Academic Press. Hinde, R.A. and Stevenson-Hinde, J. (1988). Relationships within Families: Mutual Influences Oxford: Oxford University Press. Kendler, K. S. (1996). Parenting: a genetic-epidemiological perspective. American Journal of Psychiatry, 153, 11-20. Kendler, K. S. (1995). Genetic epidemiology in psychiatry. Archives of General Psychiatry, 52, 895 - 899. Kendler, K. S., Kessler, R. C., Neil, M. C., et al (1993) The prediction of major depression in women: toward an integrated etiological model. American Journal of Psychiatry, 150, 1139-1148. Kendler, K.S., Neale, M.C., Prescott, C.A.., Kessler,R.C., Heath,A.C., Corey, L.A. and Eaves, L.J. (1996) Childhood parental loss and alcoholism in women: a causal analysis using a twin-family design. Psychological Medicine 26,79-95. Mackinnon, A.J., Henderson, A.S. Scott,R. and Duncan-Jones, P. (1989). The Parental Bonnding Instrument (PBI): an epidemiological study in a general population sample. Psychological Medicine 19, 1023-1034. Mackinnon, A.J., Henderson, A.S. and Andrews, G. (1993). Parental affectionless control as an antecedent to adult depression: a risk factor refined. Psychological Medicine 23,135-141. Main, M. and Goldwyn, R. (In Press) Adult Attachment Rating and Classification Systems. In A Typology of Human Attachment Organization Assessed in Discourse, Drawings, and Interview.(Ed M. Main) New York: Cambridge University Press. Minuchin, S. (1974) Families and Family Therapy. Cambridge, Massachussetts: Harvard University Press. Mullen, P., Martin, J.L., Anderson, J.C., Romans, S. E. and Herbison, G.P. (1993). Childhood sexual abuse and mental health in adult life. British Journal of Psychiatry 163, 721-732. Murphy, E., Brewin, C. R., and Silka, L. (1997). The assessment of parenting using the parental bonding instrument: two or three factors? Psychological Medicine, 27, 333-342. Parker, G, Tupling, H. and Brown,L.B (1979) A parental bonding instrument. British Journal of Medical Psychology 52, 1-10. Parker, G. (1979). Parental characteristics in relation to depressive disorders. British Journal of Psychiatry 134, 138-147. Parker, G. (1981). Parental reports of depressives: an investigation of several explanations. Journal of Affective Disorders 3, 131-140. Parker, G. (1989). The Parental Bonding Instrument: psychometric properties reviewed. Psychiatric Developments 4, 317-335. Quinton, D., Rutter, M. and Gulliver, L. (1990). Continuities in Psychiatric Disorders from Childhood to Adulthood in the Children of Psychiatric Patients. In L. Robins and M. Rutter Eds. Straight and Devious Pathways from Childhood to Adulthood. Cambridge: Cambridge University Press. Rodgers, B. (1996). Reported parental behaviour and adult affective symptoms. 2. Mediating factors. Psychological Medicine 26, 63-77.. Rutter, M., Quinton, D. and Hill, J. (1990). Adult outcome of Institution-Reared Children: Males and Females Compared. In L. Robins and M. Rutter Eds. Straight and Devious Pathways from Childhood to Adulthood. Cambridge: Cambridge University Press. Seligman, M. E. P. (1974). Depression and Learned Helplessness. In: The Psychology of Depression: Contemporary Theory and Research. Ed. R. J. Friedman and M. M. Katz. Washington: Winston-Wiley, 83-113. Teasdale, J.D. (1988). Cognitive vulnerability to persistent depression. Cognition and Emotion. 2, 247-274.
III II Teasdale J. D.(1993). Emotion and two kinds of meaning: cognitive therapy and applied cognitive science. Behavioural Research and Therapy 31 (4) 339-354.
Appendix This questionnaire asks about how things were with your children when they were small, say between the ages of five and ten.
Please read these statements one-by-one. See if they are like the way things were with you and your children when they were small. Please tick the box that fits best.
|Very Like||Moderately Like||Moderately Unlike||Very Unlike|
|1)They needed my help in nearly everything they did.||I|
|2) They needed protection for longer than most children.||III||II||I|
|3)They were my main source of emotional satisfaction.||3||2||1|
|4) I talked to them a lot.||iii||ii||i|
|5) I read to them when they were small.||iii||ii||i|
|6) I made sure that I had my own life.||1||2||3|
|7) I preferred to spend time with my children rather than my friends.||3||2||1|
|8) I was closer to my husband than my children.||1||2||3|
|9) I trusted my children more than my husband.||3||2||1|
|10) They were the most important thing in my life.||3||2||1|
|11) They got as much freedom as they needed.||I||II||III|
|12) They were happy to go off to school.||I||II||III|
|13) I insisted that they did exactly as I told them.||III||II||I|
|14)They needed a lot of guidance.||III||II||I|
|15) I feel that children should be able to make their own mistakes.||I||II||III|
|Very Like||Moderately Like||Moderately Unlike||Very Unlike|
|16) My children confided in me.||iii||ii||i|
|17) They needed a good deal of affection.||iii||ii||i|
|18) I helped them.||iii||ii||i|
|19) I didn't like them growing up.||3||2||1|
|20) I wanted to be the most important person in their lives.||3||2||1|
|21) When they were upset I tried to comfort them.||iii||ii||i|
|22) I praised them.||iii||ii||i|
|23) They kept some things to themselves.||I||II||III|
|24) I needed to control their behaviour most of the time.||III||II||I|
|25) I had to watch very carefully over them||III||II||I|
|26) I always made sure I was there to look after them.||3||2||1|
|27) We laughed together.||iii||ii||i|
|28) I lived for my children||3||2||1|
|29) I made sure I always knew where they were.||iii||ii||i|
|30) I liked to hear how they felt about things.||iii||ii||i|
Scores for the Care scale are recorded in dotted Roman numerals thus:- i,ii,iii.Scores for the Control scale are recorded in Roman numerals thus:- I,II,III.
Scores for the Investment scale are recorded in Arabic numerals thus:- 1,2,3.
Table 3 Adult Lifetime RDC Major Depression in Relation to Maternal Overprotection and Investment in Children.
|Low Maternal Overprotection||High Maternal Overprotection|
|RDC Major Depression||Low Investment||High Investment||Low Investment||High Investment|
Version 1.0 published 28.9.2000
Version 2.0 published 2.11.2000