2.2 MARCO REFERENCIAL
2.2.1 INTERNACIONAL
2.2.1.1 Centro Deportivo de Alto Rendimiento ¨La Loma¨
Ethical approval was sought from Camden and Islington Community NHS Trust, and the Joint University College London/University College London Hospitals Trust ethics committees. Copies of the letters of approval can be found in Appendix 1. Approval was also sought and granted by Islington Education Authority.
The two-phase recruitment process was designed to allow easy opt-out for unwilling participants. Written consent was required before commencement of the interview procedure, and this could only be obtained once the participant had
been given a full opportunity to ask questions regarding the research, and felt they had received satisfactory answers to all their questions (see appendix 3. for a copy of the form). The investigator was aware that in a sample the size of the one used m this study, there was a likelihood of one or more of the participants having experienced some form of abuse or other severe negative experience as a child, and that being asked to describe their parents may potentially cause some degree of distress. Furthermore, that one or more of the parents may have abused or been involved in a severe negative experience with their own child, and so describing the child and particularly a recent conflict with the child may also lead to potential distress. However, it was felt that the details given in the information sheet would provide an opportunity for individuals to opt out if they were unwilling to describe their parents, self, or chÜd. Also, the nature of the interview meant that there was no probing of descriptions, allowing participants to reveal as httle or as much information as they felt comfortable with.
Design
A naturalistic (non-experimental), correlational design was adopted for this phot study. All participants completed the full interview schedule which measured level of maternal depression, and used two methods, to infer representations of self, child, and own mother: a structured rating scale (repertory grid), and verbal free-response descriptions. Representations were measured on a number of different dimensions. The number of dimensions used to measure representations reflects the pilot nature of the research, where the aim is to explore the possibility
of some aspect of representations being associated with discrepancy in child behaviour rating. The participants were asked to rate their child’s behaviour, from which a criterion rating was subtracted to produce a discrepancy score.
The associations between the level of maternal depression, and the dimensions of representations and ratings of chüd behaviour were assessed using correlation and regression analysis.
Measures
Measures of Child Behaviour
The Pre-school Behaviour Checklist (PBCL), (McGuire and Richman 1986), was
used as the primary measure of chüd behaviour (see Appendix 4.). It comprises 22 items, scored either 0, 1, or 2, assessing a range of potential emotional and behavioural problems in pre-school chüdren. For most items the rater is given a choice of three descriptions of chÜd behaviour, in order of increasing frequency and/or severity of the problem, leading to a total possible score of 44. The exceptions are the ‘activity’ item where four choices are given to include a description of extreme inactivity which, along with the description of extreme activity, would also receive a score of 2. The ‘habits’ item is scored by summing the number of habits rated as occasional or frequent. In addition, cases where the first language is not English are taken into account when scoring the ‘speech’ items.
McGuire and Richman (1986), have demonstrated good inter-rater agreement between nursery staff (r = .68, p > .001) using the PBCL, and found no significant difference in ratings between teacher and nursery nurses/assistants when both members of staff knew the child well. Internal consistency using the Spearman-Brown split half formula was high (r = .83). Test-retest reliability was not examined. However, the Behaviour Checklist (BCL), (Richman, 1977), firom which the PBCL was developed, has shown reasonable reliability on this concept (r = .81, over a 4 week interval, Richman, 1977), and (r = .54, over a 12 month interval, Spiker, 1992).
Further examination has demonstrated the PBCL to have comparable
discriminant validity when compared to clinical interview ratings of behaviour problems. It has also been shown to discriminate well between clinical and non- clinical groups of children. Concurrent validity as checked against the Pre-school Behaviour Questionnaire (PBQ), (Behar and Springfield 1974) was high (r = .89), (McGuire and Richman, 1986),
There are a number of other specific reasons for en^loying the PBCL in this research. First, it has been designed specifically for use with the subject group examined in this study, and is one of only a few measures of child behaviour that can be rated adequately by mothers and nursery staff. Third, other widely used measures of child behaviour have been validated on American populations (e.g. the CBCL, Achenbach et al. 1987a). The PBCL, has been validated not only on
an English population, but more specifically on an inner London population largely similar to the sample in this study.
Accessing Representations
In keeping with previous research this study adopts both a language based measure of representations, and the more structured repertory grid procedure.
Inferring Representations Through Language: Free-Response Descriptions
This study adopts a free-response method for inferring representations from verbal descriptions, adapted from the method developed by Blatt and colleagues (Blatt, Wein, Chevron and Quinlan, 1979, Quinlan, Blatt, Chevron and Wein, 1992, Bers, Blatt, Sayward, and Johnston, 1993, and Blatt, Bers and Schaffer, 1993).
Participants were asked to give descriptions of; Self, Child, and Own Mother. As the study focuses primarily on the relationship between mother and child, it was important to elaborate on this relationshÿ, rather than rely solely on specific descriptions. Therefore, in addition to the free-response descrçtions, partic^ants were asked to give a description of the last conflict they remembered between themselves and their child.
The analysis of the descriptions is based on that used by Blatt et al. (1993) and adapted to compare inter-generational representations, by incorporating scales from the AAI, and a Reflective-Self Scale (Fonagy et al. 1991). Finally, a global rating for each description which combined both the content and structural features of the
representation is included. The 14 scales used in this study rated the following dimensions of each of the three descriptions:
From Blatt and Colleagues
1. Predominant Mode of Representation - A four point scale which indicates which mode of representation was given most enphasis in the description, 1 = physical, 2 = behavioural, 3 = personality, 4 = internal qualities.
2. Substantiality - A four point scale which measures how substantial the description is. A low score indicates a flat unintegrated description using only one mode of representation, a high score represents a multidimensional description, where many modes are used and integrated.
3. Conceptual Level - A nine point scale which assesses the highest cognitive developmental level of the description. The scale ranges from a score of one which is given to a description at a sensorimotor-preoperational level, through a score of three where the description reaches an external iconic level, up to a score of nine where the description is judged to have reached a conceptual representation, where the individual is able to be understood on a variety of dimensions.
4. Tolerance of Contradictory Aspects - A five point scale measuring the eprtent to which contradictions are referred to and tolerated in the description. A low score is given if no contradictions are mentioned, a high score is given to descriptions where a number of diverse, and apparently contradictory aspects are mentioned, but tolerated as part of the individual’s personality.
5. Negative/ Positive Regard - A seven point scale measuring the extent to which the individual is described in positive terms. A higher score represents a more positive description.
6. Critical - A five point scale which measures the extent to which the description is critical of the individual A high score represents a severely critical description.
7. Striving/Ambitious - A five point scale measuring the extent to which the
individual is described as striving or ambitious. A low score indicates that the individual is described as non-striving, a high a score indicates a strongly striving individual
From the AAI
8. Loving/Unloving - A nine point scale measuring the extent to which the individual is described as loving. A low score represents a person very lacking in love, a high score represents a very loving individual.
9. Rejecting - A nine point scale measuring the extent to which the individual is described as rejecting. A low score represents an individual described as not at all rejecting, a high score is given to a description of a person who is extremely rejecting.
10. Role Reversal - A nine point scale measuring the extent to which an individual seeks to reverse generational roles. The greater the extent of the role reversing, the higher the score.
11. Coherence - A nine point scale measuring the extent to which the description of the individual is a coherent one. A low score represents a highly incoherent description, a high score a highly coherent description.
12. Idealisation - A nine point scale which measures the extent to which the individual is idealised in the descriptiom A low score represents no idealisation, a high score very strong idealisation.
FromFonagye^fl/.
13. Reflective-Self Scale - (used for rating descrÿtions of self only) A nine point scale depicting the extent to which the participant is able to reflect on the
intentions and motivations of themselves and others. A low score is given if a person gives no evidence of this ability, a high score is given to a person who demonstrates a consistent understanding of the motivations of self and others.
From WMCI
14. Global Categorisation - This is a scale which integrates a number of dimensions of the descrÿtions and places them into one of three categories:
1. Balanced which represents a rich and ençathic descrÿtion, 2. Disengaged, which represents a descr^o n suggesting detachment and emotional distance, characterised by an extreme absence of interest in the individual, 3. Distorted, characterised by internal inconsistency in the representation, the description may show over-concem for the individual, or feelings of being anxiously overwhelmed.
These scales are summarised below in table 1.
Table 1.
Scale Size of Scale
Scales from Blatt and Colleagues I. Modes of Description
Predominant Mode of Description 4-point
Substantiality 4-point
Conceptual Level 9-point
n.
Cognitive VariablesTolerance of Contradictory Aspects 5-point
in.
General ViewNegative/Positive Regard 7-point
Critical 5-point
Striving/Ambitious 5-point
Scales from AAI
Loving/Unloving 9-point
Rejecting 9-point
Involving/Role Reversal 9-point
Coherence 9-point
Idealisation 9-point
From WMCI
Global Categorisation 3-categories
FromFonagye^iz/.
The rating procedure for rating descriptions of others (Blatt et al. 1979), has been demonstrated to possess reasonable inter-rater reliability for each of the ratings (ranging fromr = .69, to r = .95). Furthermore, the analysis of the descriptions was shown to discriminate between types and levels of depression in individuals, using both the Depressive Experiences Questionnaire (Blatt et a l 1979), and the Zung Depression Inventory (Zung, 1972). The rating scale developed for descriptions of self, showed similar psychometric properties. Inter-rater reliability for the scales was acceptable ranging fromr = .74, to r = .99, (Blatt 1993). Furthermore, the scales were tested on both clinical and non-clinical populations and found to discriminate well between the two populations.
Inferring Representations Using Rating Scales: Repertory Grid Procedure
A repertory grid procedure was adopted as the method for inferring
representations through rating scales. Elements were chosen for participants to rate that would allow for construing of ‘real’ individuals to be examined, but also enable the relationship between ‘real’ and imagined ‘ideal’ elements to be
explored. Table 2. presents the elements rated by participants, elements are given in the order of presentation in the research procedure.
Table 2.
Elements Used in the Study, Presented in Order o f Presentation to Participants.
Elements My Child
My Child as I would ideally wish them to be Myself as I am now
Myself as I would ideally wish to be Myself as 1 imagine 1 was as a child My Mother
My Father
Elements were rated against eight constructs that were suppUed by the investigator, these are presented in table 3. below. The constructs were
presented as pairs of dichotomous adjectives, except the final construct relating to independence which was presented as ‘Very independent' vs. ‘Needs a lot of help’. In ratings of child elements the phrase ‘for their age’ was added to the negative pole. Each element was rated against each construct on a 7-point, bi polar scale.
Table 3.
Construct Supplied fo r Rating o f Each Element, Presented in Order o f Rating,
by the Participants.____________________________ Constructs Good - Bad Strong - Weak Active - Passive Kind - Unkind Intelligent - Unintelligent Forgiving - Unforgiving Warm - Cold
Very independent - Needs a lot of help
The constructs were chosen based on a literature review of studies that had used similar methods. Constructs were chosen on the basis of which demonstrated the most discriminatory properties, between elements and between individuals. Furthermore, based on previous literature (Osgood et al. 1957, Quinlan, Blatt, Chevron and Wein, 1992, Bers, Blatt, Sayward, and Johnston, 1993), the chosen constructs were assumed to combine into two groupings one labelled ‘agency’
which included the constructs: strong, active, and independent, and a second labelled ‘benevolence’ which included: forgiving, warm, kind, and good. The construct ‘intelligent’ was assumed to be independent of these groupings. A factor analysis employing VARIMAX rotation was performed to test these associations.
Reasonable test re-test rehabihty has been demonstrated on a number of different dimensions of the grid (FranseUa and Bannister, 1977). Rating of elements using supphed constructs have demonstrated re-test correlation coefficients as high as r = .86, over a period of six weeks, (Bannister and Mair, 1968). Similar reliability has been demonstrated for self ratings. Sperlinger (1976) demonstrated a re-test reliabihty coefficient of .95, for perceived similarity between self and others, over a period of seven months. Furthermore, Bannister and Mair (1968), have
demonstrated that much greater reliability is obtained by grids enq)loying fewer elements and constructs (a 10 x 6 matrix, as opposed to a 15 x 6 matrix). In the hght of this, this study employs an 8 x 8 matrix.
In terms of validity of the repertory grid method, in an extensive hterature review Winter (1992), concludes that there is,
“considerable evidence of the validity of repertory grid measure in relation to a wide range of characteristics and aspects of the behaviour of ‘normal’ subjects” (p.65).
Measure of the Level of M aternal Depression
The Beck Depression Inventory (EDI), (Beck and Steer, 1993), was used to
measure the level of maternal depression in the mothers taking part in the research. The EDI is a 21 item self-completion questionnaire designed to measure the severity of depression in adults (Beck, Rush, Shaw and Emery,
1979). Each item consists of a group of three statements describing a synq)tom of depression in increasing order of severity. The respondent is asked to indicate which statement best describes the way they have been feeling over the past week. Each item is rated on a four point scale leading to a possible maximum score of 63. The higher the score, the higher the indicated level of depression.
The inventory is one of the most widely used measures of severity of depression (Steer, Beck and Garrison, 1985). It is particularly apt for this study since it remains one of the few measures of depression whose psychometric properties have been tested on ‘normal’ as well as psychiatric populations (Beck, Steer, and Garbin, 1988). Lightfoot and Ohver (1985) have demonstrated a test-retest rehability coefficient of .90 over a period of two weeks in ‘normal’ young adults. Furthermore, it has been shown to correlate well with other measures of
depression including clinical ratings (mean r = .60, for non-clinical samples), and to show reasonable properties of construct, content and discriminant vafidity (Beck, Steer, and Garbin, 1988).
Demographic Data
In order to examine whether a satisfactory cross-sectional sample had been gathered in the research, the following demographic data was collected: age and sex of the child, age, social grade, employment status, marital status, and ethnic group of the mother, as well as number and ages of any other children hving at home. (See Appendix 5.)