Stronger Families in Australia study: the impact of Communities for Children
- Previous: Appendix A: Selective attrition
- Next: Appendix C: Using the Adapted PPVT-III in the Stronger Families in Australia project
Appendix B: Outcome measures and their internal consistency and validity
Outcome measures
Healthy young families
Number of child injuries requiring medical attention in the past 12 months was measured through the administration of a single item: ‘In the past 12 months, how many times did your child need medical attention from a doctor or hospital because the child was hurt or injured?’ A value between 0 and 99 was recorded as the response.
Child physical health was measured through administration of the eight physical functioning items of the Pediatric Quality of Life Inventory (PedsQL) (Varni et al. 2003). The PedsQL Measurement Model is a modular approach to measuring health-related quality of life (HRQOL). The 23-item PedsQL Generic Core Scales were designed to measure the core dimensions of health as delineated by the World Health Organization, as well as role (school) functioning: physical functioning, emotional functioning, social functioning and school functioning. Parents reported how much each item was a problem for their child on a 4-point Likert-type scale, ranging from 0=‘Never a problem’ to 4=‘Almost always a problem’. Sample items included: ‘Taking part in active play or exercise’ and ‘Having hurts or aches’. A summary score was derived as the sum of responses to all eight items.
Child emotional and behavioural problems were assessed using the parent-reported Strengths and Difficulties Questionnaire (SDQ) (Goodman 1997). The SDQ is a very widely used instrument for screening emotional and behavioural adjustment problems in children and included 20 attributes covering four domains of behaviour problems: hyperactivity, emotional problems, conduct problems and peer problems. Sample items included: ‘Restless, overactive, cannot sit still for very long’ (hyperactivity subscale); ‘Often fights with other children or bullies them’ (conduct problems subscale); ‘Picked on or bullied by other children’ (peer problems subscale); and ‘Nervous or clingy in new situations, easily loses confidence’ (emotional problems subscale). Respondents rated each item on a scale from 0 to 2 to indicate how much it applied to the study child (0=‘Not true’, 1=‘Sometimes true’, 2=‘Certainly true’). The SDQ also includes a prosocial scale, which was used in the current study (see below).
One way of assessing the validity of the mental health variable was to test the extent to which interviewer observations of whether the child had a negative mood32 and whether they were fearful33 was associated with parental reports of child mental health problems (as measured by the SDQ). The interviewers at Wave 3 were asked to report on the duration and intensity of the study child’s negative mood and fearfulness. These items were derived from the Home Observation for Measurement of the Environment (HOME) (Caldwell & Bradley 1979) and have been used in the Longitudinal Study of Australian Children (LSAC).
Overall, there was a clear association between the interviewers’ observations of child negative mood and fearfulness and parent reports of child mental health problems on the SDQ. Children who showed low negative mood had a mean total SDQ score of 9.0, while children who showed some or more intense negative mood were rated as having greater mental health problems by their parent (a SDQ total score of 10.5 and 11.5 respectively, F(2)=22.19, p<0.001). Similar associations were also found with interviewer observations of fearfulness (F(2)=13.47, p<0.001).
Child prosocial behaviour was measured by the Prosocial subscale of the SDQ (Goodman 1997). The mean of 5 parent-rated items was used to assess the child’s propensity to behave in a way that is considerate and helpful to others, with items scored as for the mental health variable (0=‘Not true’, 1=‘Sometimes true’, 2=‘Certainly true’). There was a positive relationship between interviewer-observed child positive mood and prosocial behaviour as assessed by the SDQ.
Child overweight was determined by using the child’s length (or height), weight and girth to calculate body mass index (BMI). These measures were taken using measuring boards and weighing scales during administration of the Wave 3 survey.
Parent general health was measured using a single item: ‘In general, how would you say your health is?’, rated on a 5-point Likert-type scale, where a score of 1=‘Excellent’ and a score of 5=‘Poor’.
Parent mental health was measured using the Kessler-6 (K-6) (Kessler et al. 2002), which comprises six items and has been widely used and validated in many epidemiological studies (for example, Furukawa et al. 2003). Items were rated on a frequency scale where 1=‘All of the time’ and 5=‘None of the time’, and included ‘Hopeless’ and ‘So sad that nothing could cheer you up’. Responses to each item were summed to produce a total parent mental health score.
Supporting families and parents
Hostile parenting was assessed through parent self-ratings on five items measuring the frequency of irritable and hostile behaviour directed at the study child. Sample items included: ‘You have been angry with the child’ and ‘You have lost your temper with [child]’. The items were sourced from LSAC and the National Longitudinal Survey of Children and Youth, a 20-year longitudinal survey of a nationally representative sample of Canadian children and youth (Statistics Canada & Human Resources Development Canada 1995). Items were rated on a 10-point Likert-type scale, where a score of 1=‘Never’ and 10=‘All the time’.
Interviewer observations about the behaviour of the respondent parent were used to validate the measure of parenting hostility. After the interviewers left the home, they completed a single item according to the extent to which they observed the parent scolding, shouting at, belittling or hitting the child. Parents who were observed scolding, belittling or hitting their child had significantly higher levels of self-rated hostile parenting than those who did not (3.71 compared to 3.23; t(1,791)=–5.38, p<0.0001).
Parenting self-efficacy was measured through administration of a single item: ‘Overall as a parent, do you feel you are …’, with five response categories: 1=‘a very good parent’, 2=‘a better than average parent’, 3=‘an average parent’, 4=‘a person who has some trouble at being a parent’, and 5=‘not a good parent’. Self-reports of parenting efficacy were positively correlated with interviewer observations of parental warmth, and were negatively correlated with parental harshness and with child fearfulness and distress. For example, parents who were observed praising their child had significantly higher levels of parenting self-efficacy than those who were not observed praising their child (3.88 versus 3.61; t(1,791)=–6.55, p<0.0001).
Parental relationship conflict was assessed as the frequency of verbal and physical arguments, anger and disagreements between mothers and fathers, using five items rated on a 5-point Likert-type scale, ranging from 1=‘Never’ to 5=‘Always’. Sample items included: ‘Do you or your partner argue?’ and ‘Is there anger and hostility between you and your partner?’ Responses to all items were summed to produce a total relationship conflict score. The measure has been used in LSAC.
Living in a jobless household was derived from responses to items relating to the parent(s) participation in paid work. A jobless household was one where no parent living in the household was employed in the last week.
Early learning and care
Receptive vocabulary achievement and verbal ability were assessed using the LSAC Short-Form of the Peabody Picture Vocabulary Test (PPVT).34 This test measures a child’s knowledge of the meaning of spoken words, and his or her receptive vocabulary for Standard American English. Items include a PPVT Test Kit with up to 71 plates of displayed pictures. The child is not required to define words but to show what they mean by saying or pointing to a picture that best represents the meaning of the words. The PPVT takes approximately 10 minutes to administer. A standardised Rasch-modelled score was derived based on interviewer administration of the PPVT (Short-Form) (Appendix C).35
Quality of the home learning environment was assessed through a series of questions about participation in four activities over the past week, developed for LSAC. Sample items included: ‘Read to [study child] from a book’ and ‘Drew pictures or worked on art or craft activities’. Items were rated on a 4-point Likert-type scale, where 1=‘None’, 2=‘1–2 days’, 3=‘3–5 days’, and 4=‘Every day (6–7 days)’. Responses to each item were summed to derive an overall home learning environment measure.
Child-friendly communities
Support in raising children was measured as the frequency with which the respondent parent received support from: partner/spouse, spouse/partner’s parents, own parents, other family, friends, neighbours, and community organisations. Responses were recorded on a 5-point Likert-type scale, ranging from 1=‘Never’ to 5=‘Always’. Responses to individual items were summed to produce a total support in raising children score. This measure was also used in LSAC.
Involvement in community service activity was measured using a single item specifically developed for the LSAC study: ‘Do you or [partner] participate in any ongoing community service activity (for example, volunteering at school, coaching a sports team, or working with a church or neighbourhood association)?’ A yes/no response format was used.
Neighbourhood as a place to bring up children was measured by a single item: ‘How do you feel about your neighbourhood as a place to bring up children?’ Responses were recorded on a 5-point Likert-type scale, ranging from 1=‘Very good’ to 5=‘Very poor’. This question has been used in other large-scale Australian studies such as LSAC, Sure Start and the Millenium Study (for example, Belsky et al. 2006).
Community social cohesion was measured in terms of parent perceptions of trust, reciprocity and shared values existing between individuals in the local community. Respondents were asked to indicate the degree to which they agreed with five statements on a 5-point Likert-type scale, ranging from 1=‘Strongly agree’ to 5=‘Strongly disagree’. Sample items included: ‘People around here are willing to help neighbours’ and ‘People in this neighbourhood can be trusted’. Responses to all items were summed to produce an overall community social cohesion score. The scale of community social cohesion was developed by Sampson, Raudenbush and Earls (1997). It is widely used and has good reliability and validity.
Community facilities were measured in terms of access to basic services, shops, public transport and parks. Respondents were asked to indicate the extent to which they agreed with four items on a 5-point Likert-type scale, ranging from 1=‘Strongly agree’ to 5=‘Strongly disagree’. Sample items included: ‘There are good parks, playgrounds and play spaces in this neighbourhood’ and ‘There is access to close, affordable, regular public transport in the neighbourhood’. Responses to all items were summed to produce an overall community facilities score. The reliability and validity of this measure, using data from Wave 1 of LSAC, has been established (Edwards 2006).
Unmet service needs was measured along four dimensions: medical services; services for a developmental issue, such as a learning, speech or behaviour problem; family support services; and adult mental health services. Respondents were asked to indicate whether in the past 12 months the service type was needed or not. If the service was required, respondents then indicated whether there was any difficulty accessing the service, using a yes/no format where 0=‘No’ and 1=‘Yes’. Responses relating to all five service types were aggregated to produce an overall unmet service needs score, where scores of 1 and above were defined as ‘unmet service need’ and a score of 0 was defined as ‘no unmet service need’.
Internal consistency reliabilities
Internal consistency reliabilities (Cronbach’s α) for all scaled outcome measures (that is, derived from multiple variables) are presented in Table B1.
Note: Information about children’s emotional and behavioural problems (SDQ) and prosocial behaviour were not collected in Waves 1 and 2, hence the empty boxes.
- Previous: Appendix A: Selective attrition
- Next: Appendix C: Using the Adapted PPVT-III in the Stronger Families in Australia project
[ top ]