Parent-child contact and post-separation parenting arrangements

Research Report No. 9 – July 2004

You are in an archived section of the AIFS website. Archived publications may be of interest for historical reasons. Because of their age, they do not reflect current research data or AIFS' current research methodologies.

9. The demography of parent-child contact

In an effort to embed the qualitative data in the wider national picture, this chapter examines the extent to which different patterns of care are linked to certain demographic elements in the general population of separated/ divorced parents. Essentially this is a profiling exercise, drawing on data extracted from Wave 1 of the Household, Income and Labour Dynamics in Australia (HILDA) survey.

Analysing the HILDA survey

The HILDA survey collects information related to three broad domains: economic and subjective wellbeing; labour market dynamics; and family dynamics. The first wave of the survey, conducted in the latter half of 2001, examined a range of issues within these three domains, including employment status, work hours and flexibility, personal and household income and their sources, marital history, family (re)formation, and patterns of parental care for children under 18 years whose parents live apart.

The power of the HILDA survey for research into parent-child contact is that it enables national estimates to be obtained across the spectrum of the separated and divorced parent population. The Australian Institute of Family Studies was involved in formulating a set of questions which provide information on separated parents' parenting arrangements and personal circumstances. The chance to draw on these data to explore different patterns of parenting was thus theoretically intriguing and practically inviting.

Sample

The sample on which this analysis is based comprises 1039 parents. These parents had at least one natural or adopted child under 18 years at the time of interview, but were not living with that child's other parent. This sample was derived from a large representative sample of 7,682 households involving 13,969 household member interviews across Australia, yielding a household response rate of 66 per cent (see Watson and Wooden 2002).

Of the 13,969 household members interviewed, 1243 were separated parents with at least one child under 18 years old. Of these, we focused on 'co-parents' (63 parents who indicated that their child spent at least 30 per cent of nights per year with each parent) and the two largest groups in the sample: resident mothers (n=600) and non-resident fathers (n=376).

Both the resident mothers and non-resident fathers were subdivided into three groups according to their reports on father-child contact patterns. The first group had 'standard' contact - whereby fathers have children in their care up to 29 per cent of nights per year, typically every-other-weekend or every weekend. The second group had daytime-only contact - that is, children have no overnight stays at their father's house. The third group had little or no contact - whereby fathers saw their children less than once per year. (HILDA did not collect data on contact during holiday periods. Thus no profile is offered for the 'holiday-only' contact group). In total then, four groups of mothers and four groups of fathers are compared.40

Figure 5. Sampling stratergy

Fig 5. Sampling strategy, described in text

Caveats

Some caveats warrant mention. The samples of separated men and women are independent. That is, the men and women had not been in a relationship with each other (married or cohabiting). This analysis thus focuses on the characteristics and perceptions of one parent - the parent who was interviewed - in examining post-separating parenting arrangements. The extent to which any systematic biases are present within each independent sample is unclear.41

In addition, while the longitudinal design may eventually shed light on any causal directions between parent-child contact and other factors, Wave 1 data are cross-sectional. Thus no claim can be made that certain factors lead to certain patterns of contact or vice versa.

Moreover, cross-sectional snapshots may create the impression that patterns observed are static. The reality, of course, is that patterns of contact are complex, and families can move between different patterns of care. Such complexity is not captured in this single wave of data.

Finally, for reasons of economy and to reduce respondent burden, where respondents had more than one child under 18 potentially in their care, the HILDA survey required respondents to focus on the youngest natural or adopted child.

Results

Tables 2 and 3 present patterns of socio-demographic characteristics of mothers and fathers with the four different types of father-child contact outlined above. A central issue examined below is the extent to which these profiles are consistent with the insights from the focus group material. To this end, simple chi-square tests or analysis of variance tests are computed for the HILDA data to assess whether any overall significant relationship or differences exist on key measures across the groups of interest. But where we say that one group is the 'most (or least) likely' to have certain characteristics, this should not be interpreted to mean that this group is significantly more likely than all other groups to possess such attributes.

Table 2. Characteristics of separated/divorced resident and co-parent mothers by care arrangements of their children (n=632) a
  Little or no father–child contact
(n=246)
Daytime-only father–child contact (n=114 ) "Standard” contact (n=240) Shared care a (n=32) p
Employment status
Full-time 24.80% 14.60% 27.80% 46.60% *
Part-time 22.20% 16.40% 33.50% 28.20%  
Not employed 53.00% 69.00% 38.70% 25.10%  
Able to work at home (employed)
Yes 18.80% 19.50% 26.20% 34.70%  
No 81.20% 80.50% 73.80% 65.30%  
Educational attainment
Degree or higher 12.80% 11.90% 18.30% 45.00% **
Other qualification 37.60% 40.40% 38.20% 21.20%  
No qualification 49.60% 47.70% 43.50% 33.80%  
Age of youngest child (years)
0-4 26.60% 42.80% 22.00% 16.90% **
5-11 38.30% 31.70% 46.40% 60.10%  
12-14 14.60% 10.90% 15.20% 12.10%  
15-17 20.50% 14.70% 16.30% 11.00%  
Distance between parents
<10 km 12.40% 38.30% 26.90% 62.30% **
10-49 km 21.10% 36.30% 34.20% 23.80%  
> 49 km 66.50% 25.40% 38.90% 13.90%  
Living with a partner
Yes 41.70% 19.50% 36.70% 24.30% **
No 58.30% 80.50% 63.30% 75.70%  
Age (mean years) 34.8 35.7 36.7 37.4  
 Housing tenure
Fully own/purchasing 39.20% 36.40% 51.00% 67.40% **
Rent 60.80% 63.60% 49.00% 32.60%  
Number of bedrooms in HH (mean) 3.1 3.2 3.3 3.1  
 Personal income ($)
≤15,000 54.00% 63.90% 45.40% 29.10% **
15,001- 35,000 32.80% 27.10% 38.20% 36.10%  
>35,000 13.20% 9.00% 16.40% 34.80%  
Relationship  with former partner b
Satisfied (0-3) 29.70% 29.10% 36.50% 47.40%  
Mixed feelings (4-6) 20.30% 27.40% 25.20% 20.10%  
Dissatisfied (7-10) 50.00% 43.50% 38.30% 32.50%  
Satisfied (0-3) 18.00% 23.00% 32.40% 41.60%  
Mixed feelings (4-6) 12.30% 21.60% 22.40% 17.70%  
Dissatisfied (7-10) 30.30% 34.30% 34.00% 28.60%  
Not stated 39.40% 21.10% 11.30% 12.00%  
Child support (received)
Yes 29.40% 53.70% 65.70% 50.30% **
No 70.60% 46.30% 34.30% 48.80%  
Any financial support (received) c
Yes 33.90% 62.00% 73.00% 66.60% **
No 66.10% 38.00% 27.00% 33.40%  
Annual child support received per childd 2708 2543 3535 2980  
Views about amount of contact
Nowhere near enough 35.10% 20.30% 20.40% 11.90% **
Not enough 8.90% 14.50% 22.30% 4.50%  
About right 53.80% 56.90% 52.00% 74.50%  
A little/way too much 2.30% 8.30% 5.30% 9.10%  

Notes:
weighted data; analysis adjusted for the design effect;
a Shared care = 30 per cent overnight threshold used by ABS (“co-parents”;
b Ratings of satisfaction are on 0-10 point scale (0=completed dissatisfied; 10=completely satisfied);
c Includes (regular) child support and any other financial support;
d where child support is received;
** p < .01 level (X2 test);
* p < .05 level (X2 test);
# Different from shared care group at .05 significance level
Source: HILDA Wave 1 (2001)

Table 3. Characteristics of separated/divorced non-resident fathers and co-parent fathers by care arrangements of their children (n=407) a
  Little or no father–child contact
(n=117)
Daytime-only father–child contact (n=57 ) "Standard” contact (n=202) Shared carea (n=31) p
Employment status
Full-time 60.50% 62.60% 79.10% 67.50% **
Part-time 4.10% 11.00% 6.20% 10.40%  
Not employed 35.40% 26.50% 14.70% 22.10%  
Able to work at home (employed)
Yes 24.30% 20.50% 26.70% 44.90%  
No 75.70% 79.50% 71.30% 55.10%  
Educational attainment
Degree or higher 7.50% 14.20% 13.90% 20.30% **
Other qualification 32.80% 50.40% 52.80% 56.10%  
No qualification 59.70% 35.40% 33.30% 23.60%  
Age of youngest child (years)
0-4 22.00% 33.20% 15.10% 24.00%  
5-11 49.50% 36.00% 44.90% 59.10%  
12-14 14.00% 15.80% 22.20% 10.3  
15-17 14.50% 15.00% 17.80% 6.50%  
Distance between parents
<10 km 12.70% 41.70% 26.50% 69.00% **
10-49 km 20.60% 27.70% 31.40% 29.60%  
> 49 km 66.70% 30.60% 42.10% 1.40%  
Living with a partner
Yes 60.80% 24.30% 44.00% 17.50% **
No 39.20% 72.70% 56.00% 82.50%  
Age (years) 36.1 37.7 40.2 38.6 **
 Housing tenure
Fully own/purchasing 35.10% 41.50% 54.40% 74.10% **
Rent 64.90% 58.50% 45.60% 25.90%  
Number of bedrooms in HH (mean) 2.9 2.5 2.9 3.3 **
 Personal income ($)
≤15,000 43.90% 38.30% 21.70% 39.70% **
15,001-35,000 24.80% 36.60% 25.90% 28.70%  
> 35,000 31.30% 25.10% 52.40% 31.70%  
Relationship  with former partner b
Satisfied (0-3) 24.30% 26.70% 31.70% 22.50% **
Mixed feelings (4-6) 13.50% 24.60% 33.10% 52.90%  
Dissatisfied (7-10) 62.30% 48.70% 35.20% 24.60%  
Satisfied (0-3) 16.70% 21.50% 28.50% 20.10%  
Mixed feelings (4-6) 9.30% 19.80% 29.80% 47.40%  
Dissatisfied (7-10) 42.90% 39.10% 31.60% 22.00%  
Not stated 31.10% 19.60% 10.10% 10.60%  
Child support (paid)
Yes 53.40% 71.30% 86.20% 53.80%  
No 46.60% 28.70% 13.80% 46.20%  
Any financial support (paid) c
Yes 59.10% 84.90% 97.70% 58.90%  
No 40.90% 15.20% 2.30% 41.10%  
Annual child support paid per child d $2,250 $3,189 $3,631 $2,378  
Satisfaction  with contact 
Nowhere near enough 73.70% 59.20% 49.10% 11.80% **
Not enough 6.20% 14.20% 25.70% 18.20%  
About right 20.20% 24.20% 25.20% 67.10%  
A little/way too much 0% 2.50% 0% 2.90%  

Notes: weighted data; analysis adjusted for the design effect;
a Shared care = 30 per cent overnight threshold used by ABS (“co-parents”);
b Ratings of satisfaction are on 0-10 point scale (0=completed dissatisfied; 10=completely satisfied);
c Includes (regular) child support and any other financial support;
d where child support is received;
** p < .01 level (X2 test); * p < .05 level (X2 test);
# Different from shared care group at .05 significance level
Source: HILDA Wave 1 (2001).

Shared care

The focus group data presented in Chapter 3 suggest that 50/50 residential care arrangements are often logistically complex, and that those who opt for shared care appear to be a relatively distinct subgroup of separated/divorced parents. Virtually all of the 12 parents in the 50/50 care groups ('co-parents') adopted a shared care arrangement from the time of separation, and set up this arrangement without any involvement with the legal system.

The qualitative data point to a number of factors - relational and structural = that appear to be conducive to making shared care a viable option for separated/divorced parents. These factors include:

  • geographical proximity;
  • the ability of parents to get along sufficiently well to develop a business-like working relationship;
  • child-focused arrangements (with children kept 'out of the middle', and with children's activities forming an integral part of the way in which the parenting schedule is developed);
  • a commitment by everyone to make shared care work;
  • family-friendly work practices for both mothers and fathers;
  • reasonably well off financially (particularly for women); and
  • shared confidence that the father is a competent parent.

This profile was largely replicated by data from Wave 1 of the HILDA survey. These data show that parents with 'shared care' do indeed appear to be a relatively small but select group of parents.

As noted earlier, an important caveat is that parents with joint physical care of their children are a relatively rare group in Australia. This means that obtaining a large, representative sample of co-parents is extremely difficult, and that most surveys, even those of substantial rigour such as HILDA, typically yield co-parent sample estimates that are not statistically reliable. At the same time, there is so little empirical data on shared parenting that to disregard small pockets of data runs the risk of leaving our knowledge of joint parenting poorly informed. The HILDA coparent data are thus presented, but with the caveat that these data should be interpreted with caution - especially in relation to between-group comparisons.

Of all types of arrangements, co-parenting appeared the most likely to occur when the youngest child was aged between five and 11 years (that is, of primary school age). This situation applied to the majority of co-parent mothers and fathers (59-60 per cent) and to less than half the parents in all other groups. This pattern makes sense: infants and pre-school age children are likely to have a stronger psychological attachment to one parent (typically the mother), and moving infants between two households on a regular basis is also logistically more difficult for parents than moving older children between households. In addition, children older than primary school age are likely to have particularly strong needs for close involvement with their friends, a situation that may work against dual-residence living.42

Consistent with the focus group data, socio-economic resources appear to be a critical facilitator of shared parenting arrangements. Parents with shared care were the most likely of all parents to have a university degree (45 per cent of coparent mothers versus 13-18 per cent of mothers in the other groups; 20 per cent of co-parent fathers versus 8-14 per cent of fathers in the other groups). They were also most likely to be home owners or purchasers (67 per cent of co-parent mothers vs other mothers 39-51 per cent; 74 per cent of co-parent fathers vs other fathers 35-54 per cent); and to live within ten kilometres of their former partner, which itself may be related to financial resources (62 per cent co-parent mothers vs other mothers 12-38 per cent; 69 per cent of co-parent fathers vs other fathers 13-42 per cent).

In addition, co-parent mothers were the most likely of the female groups to be in full-time employment (47 per cent of co-parent mothers vs other mothers 15-28 per cent), and to have incomes of $35,000 or more (35 per cent vs 9-16 per cent), while co-parent fathers were the most likely of the male groups to have a larger home in terms of the number of bedrooms (co-parent fathers mean=3.3 bedrooms vs other fathers mean=2.5-2.9 bedrooms).

Furthermore, although most parents in the sample were unable to work from home, co-parents appeared to be more likely to be able to do this than parents with little or no father-child contact (co-parent mothers 35 per cent vs other mothers 19-26 per cent; co-parent fathers 45 per cent vs other fathers 24-27 per cent).43

While satisfaction with one's relationship with another person should be viewed as a crude proxy for quality of the relationship, trends in satisfaction with relationships with former partners were consistent with the notion emerging from the focus group data that co-parents are able to adopt a cooperative or business-like relationship. Of the men who participated in the HILDA survey who indicated how they felt about their relationship with their former partner, co-parent fathers were the most likely to express mixed feelings or indifference (neither good nor bad: co-parent fathers 53 per cent vs other fathers 14-33 per cent) and were the least likely to express dissatisfaction (co-parent fathers 25 per cent vs other fathers 35-62 per cent). While the pattern of results was not significant for mothers, it was consistent with that for fathers. Close to half the co-parent mothers viewed their relationship with their former partner positively (co-parent mothers 47 per cent vs other mothers 30-37 per cent) and, like coparent fathers, only a minority of co-parent mothers expressed dissatisfaction (33 per cent vs 38-50 per cent).

Along with 'resident parenting', co-parenting may well reduce the desire and opportunities for repartnering. Both types of parenting typically involve far greater investment of time in the children's lives than the different forms of 'non-resident parenting', and it may well be that resident parents and parents with shared care prefer to invest themselves in their children rather than in a new relationship. Such a preference would certainly be in keeping with a childfocused orientation.44 In addition, finding a partner who is willing to take on the role of full-time or near half-time resident step-parenting may not be easy. It is thus not surprising that co-parent fathers were the least likely of all the fathers (all of whom were non-resident parents) to have repartnered (18 per cent vs 24-61 per cent). The repartnering rate for co-parent mothers was somewhere in between that for the other female groups (24 per cent vs 20-42 per cent), a trend that is not surprising given that all other groups were resident parents.

Consistent with focus group data suggesting a strong motivation by both parents to make this shared care work well, parents with this arrangement - albeit the least common pattern of care - were the most likely of all groups to believe that the amount of parent-child contact that that was occurring was 'about right' (co-parent mothers 75 per cent vs other mothers 52-57 per cent; co-parent fathers 67 per cent vs other fathers 20-25 per cent).

To sum up, well-educated dual career former couples who lived near each other and who had primary school aged children were the most likely to have shared care arrangements. The co-parent mothers, in particular, tended to be the best off in terms of personal income of all female groups, while the co-parent fathers were the most likely to be single and to have some work flexibility, and the least likely to express dissatisfaction with their relationship with their former partner. This might suggest that they were able to adopt at least a working businesslike relationship with their former partner.

'Standard' contact

Of the other three forms of contact examined in this analysis of HILDA data, 'standard' contact was the most commonly reported form for fathers, while for mothers, much the same numbers reported 'standard' contact and little or no contact. A number of parents in the focus group with 'standard' contact noted how this pattern of care was practical given parents' work and new family commitments.

Indeed, the HILDA data suggest that fathers reporting standard contact were the most likely to be in full-time employment ('standard' contact fathers 79 per cent vs other fathers 61-68 per cent) and to have incomes in excess of $35,000 (52 per cent vs 25-32 per cent), and were the second most likely to be home owners/purchasers (54 per cent vs 35-42 per cent).45 According to mothers' reports, fathers with 'standard' contact were also the most likely group to support their children financially (73 per cent vs 34-67 per cent), a trend that is consistent with the (non-significant) pattern of reports by fathers on this issue.

Of those who reported how they felt about their relationship with their former partner, fathers with 'standard' contact were the most likely of all fathers to view this relationship positively (32 per cent vs 23-27 per cent), while mothers with 'standard' contact were the second most likely of all mothers to provide favourable views (37 per cent compared with 29-47 per cent). Nevertheless, most fathers with 'standard' care were not particularly happy about their relationship with their former partner: one third indicated mixed feelings or indifference, and around another one third expressed dissatisfaction. Mothers with 'standard' care arrangements were more likely to express either satisfaction or dissatisfaction (37-38 per cent) than mixed feelings or indifference (25 per cent).

In summary, fathers with 'standard' contact appeared to be best off financially, a trend that may partly stem from their being a little older and that would certainly help in their ability to have the children stay overnight. While only a minority of parents in all groups seemed happy with their relationship with their former partner, mothers in this group provided the most favourable evaluations of all female groups.

Little or no father-child contact

The focus group data presented in Chapter 4 suggest that high levels of interparental conflict and physical distance between parents often underpin little or no father-child contact. Resident mothers also reported disinterest on the part of fathers as a significant factor.

Again the qualitative profile was largely replicated by the HILDA data. These data suggest that parents who report little or no father-child contact tend to have a distinct demographic profile, as characterised by the three Rs - repartnered, relocated, and residual bad feelings towards each other. To this list could also be added a fourth 'r' - relative economic disadvantage.

Specifically, non-resident fathers who reported little or no contact with their children were the most likely of all groups of men to have no post-secondary school qualifications (60 per cent vs 24-35 per cent of men in other groups), to be without paid work (35 per cent vs 15-27 per cent), to be earning less than $15,000 per year, and to be renting their home (65 per cent vs 26-59 per cent). Resident mothers whose children rarely or never saw their father also tended to have low socio-economic status, although on some socio-economic indicators, they seemed better off, or no worse off, than those whose children had daytime only contact with their father. They were the most likely of all female groups to have no post-secondary school qualifications (50 per cent vs 34-48 per cent); more than half were not in paid work (53 per cent vs 25-69 per cent), and had incomes of less than $15,000 per year (54 per cent vs 29-64 per cent). While lack of a job and low personal income may not necessarily indicate financial deprivation (for example, a new partner may be a high income earner), it is noteworthy that 61 per cent were renting their home (compared with 33-64 per cent of other mothers).

Both mothers and fathers in the little or no father-child contact group were the most likely of all groups to be living with a new partner (mothers: 42 per cent vs 24-37 per cent; fathers: 61 per cent vs 18-44 per cent), and to be living 50 kilometres or more from their former partner (mothers: 67 per cent vs 14-39 per cent; fathers: 67 per cent vs 1-42 per cent). In fact, 17 per cent of mothers and 8 per cent of fathers who reported little or no father-child contact were not able to provide information on the whereabouts of their former partner, suggesting that they had lost contact with them.

Of those providing views about their relationship with their former partner, non-resident fathers with little or no contact were more likely than all other fathers to report being dissatisfied (62 per cent vs 25-49 per cent ). Although mothers' views about relationships with their former partner did not vary significantly with contact arrangements, the pattern of overall results was consistent with that for the fathers: 50 per cent of mothers who reported that their children rarely or never saw their father expressed dissatisfaction, compared with 33-44 per cent of mothers in the other groups.

This pattern of response may be indicative of inter-parental conflict or disappointment. It is noteworthy that both mothers and fathers in the little or no contact father-child contact group were the most likely of all groups to be unable to rate the quality of their relationship with their former partner (mothers: 39 per cent vs 12-21 per cent; fathers: 31 per cent vs 10-20 per cent).46 This pattern of response, along with geographic distance and repartnering, suggests a 'clean break' at many levels. Nevertheless, the clean break from a former partner did not necessarily reflect a desire for a clean break between children and their fathers. Mothers and especially fathers who reported little or no father-child contact were the most likely of all groups to see the amount of father-child contact as 'nowhere near enough' (mothers: 35 per cent vs 12-20 per cent; fathers: 74 per cent vs 12-59 per cent).

In summary, the HILDA data suggest that conflict, emotional and physical distance, new partners, and socio-economic disadvantage feature prominently in the profile of parents who report little or no father-child contact.

Daytime-only contact

Parents who reported daytime-only father-child contact closely resembled the little or no contact group on several dimensions - most notably, housing tenure and relationship quality with former partner and, for mothers, personal income. That is, resident mothers and non-resident fathers who reported daytime-only contact were among the most likely to be renting (mothers: 64 per cent vs other mothers 33-61 per cent; fathers 59 per cent vs other fathers 26-65 per cent) and to express dissatisfaction with their relationship with their former partner (of those who reported their views on this issue, mothers: 44 per cent vs other mothers 33-50 per cent; fathers: 49 per cent vs other fathers 25-62 per cent). In addition, the mothers were the least likely of all female groups to be earning more than $35,000 per year (9 per cent vs 13-35 per cent).

It is noteworthy that fathers who reported daytime-only contact with children had fewer bedrooms than other fathers (mean = 2.5 bedrooms vs mean = 2.9-3.3 bedrooms).

But most conspicuous in the profile of parents who report daytime-only father-child contact is the over-representation of children aged 0-4 years (mothers: 43 per cent vs 17-27 per cent; fathers: 33 per cent vs 15-24 per cent). This suggests that daytime-only contact may also be a function of children's young age and level of emotional dependence on a primary carer.

Summary

Material resources and conflict look to be important correlates of particular patterns of care. The maturity of the child also seems to be important, with daytime-only contact being most common when the child is of pre-school age, and 'shared care' applying when the child is of primary school age. All of this highlights the complexity surrounding different patterns of father-child contact, socio-demographic factors and parental relationships after separation.

Overall, these data provide strong empirical support for the idea that practical and relationship factors play an important role in the form that parent-child contact takes. The Parent-Child Contact Study's focus group data are therefore largely supported by data collected from the large representative HILDA sample of separated and divorced parents in Australia.

However, ideally, any discussion of contact should also be grounded in the views of children since it is they who are often mute in the decisions that affect their lives, and most affected by parenting arrangements (Kaltenborn 2001, 2004; Mason 2000; Parkinson et al. forthcoming; Smart et al. 2001; Wallerstein and Blakeslee 2003).

40. It should be noted that the data were weighted using the responding person population weight. This weight - the inverse of the probability of selection - is adjusted for the probability of response to household and person level benchmarks (see Watson and Fry 2002). In addition, to address HILDA's stratified cluster design, estimates of the variance were adjusted for the design-effect using Stata 7.

41. Piecemeal evidence exists in Australia to suggest that involved non-resident parents may be far more likely to participate in divorce research than those less-involved (the Australian Divorce Transitions Project is a case in point). That the HILDA survey covered a broad range old issues (not specifically related to divorce) may help to minimise such biases. The use of the words such as 'divorce', 'child support', or 'contact' in the sample recruitment phase may act as alarm bells to less involved parents.

42. However, it may be that primary-school-aged children today are more likely than similarly aged children in previous years to live in a shared care arrangement. Should such a 'cohort effect' exist, then future teenagers would be more likely than today's teenagers to experience shared care, all other things being equal.

43. The ability to work from home was significantly more likely for fathers in shared care than for fathers in each of the other groups (taken separately). While the results were not significant for mothers, the pattern is the same. (The overall relationship between ability to work at home and caring arrangement was neither significant for mothers nor fathers.)

44. Our focus group data suggest that while their high involvement with their children may be off-putting to some women, co-parent fathers' personality styles and generally positive, sensitive demeanour may not.

45. As noted earlier, those with shared care were the most likely to be home owners (74 per cent).

46. This question was included in the self-completion questionnaire (rather than the face-toface interview). These percentages exclude mothers and fathers who did not return the self completed questionnaire (5-6 per cent of mothers in all groups, and 10-12 per cent of fathers).