Executive Summary

The relationship between a person's level of education and the level of income this commands has been well explored in the literature. Less well examined are the links between family structure, the educational attainment of the parents in the family and levels of family wellbeing. This study addresses this gap using a series of family wellbeing indicators derived from data available in the five-yearly census as part of the Family and Whānau Wellbeing Project.

The associations between parents' educational attainment and wellbeing (as measured by median equivalised income, the incidence of low income, unemployment, hours worked and level of home ownership) are described for different family types using information derived from census data for the period 1981–2006. The use of census data allows the impacts of parents' educational attainment to be assessed across time and within family types by each of the wellbeing indicators.
The results confirm the findings of other studies, which show that attainment of secondary and post-secondary educational qualifications commands a 'premium' in the labour market. This study indicates how this 'educational premium' is distributed differently over different family types, and suggests that this premium has increased over the period 1981–2006 for most measures of family wellbeing.

Introduction

Purpose of the report

The links between educational attainment and personal income are well documented, as is the rising level of educational participation in New Zealand. Less well understood are the associations between the levels of parental educational attainment in a family and levels of family wellbeing for different family structures. This report aims to explore these links using data from the five-yearly Census of Population and Dwellings conducted by Statistics New Zealand.

The report contains an example of the type of analysis that can be conducted with the data available from this project, and it is hoped that its publication will serve to inspire further analyses of these data by others with specialist knowledge and more detailed primary information in the appropriate areas.

The Family and Whānau Wellbeing Project (FWWP)

FWWP is a five-year research programme supported by the Social Science funding pool of the Foundation for Research, Science and Technology (FRST). The principal goal of this programme is to develop ways to examine and monitor the social and economic determinants of family and whānau wellbeing and how these have changed over the 1981–2006 period.

Background

The 25-year period between 1981 and 2006 covered by this study saw extensive reforms of the education sector in New Zealand, along with − of course − other economic and social reforms. These educational reforms began with a series of reviews in the late 1980s, the outcomes of which served to shape the present structure of education. The reviews included reports on polytechnics, universities, non-formal education and educational administration (Abbott 2006).

With specific regard to universities, the Hawke Report (1988) lay the groundwork for many of the future policies, recommending that universities become more commercial and that universities, rather than government, should set student fee prices (Olssen 2002, 66). Among the recommendations from this series of reports were increased competition between institutions in order to improve efficiencies in education provision, the devolution of control over schools to local communities, and a general intent to expand the number of participants in tertiary education.

Arguably the most influential of the reforms, in terms of its impact on participation rates in tertiary education, was the substantial rise in student fees, beginning with the introduction of a $1,250 flat fee in 1990, followed by a Student Loans scheme in 1992 to enhance students' ability to borrow money for course fees (McTaggart 2006, 84). The year after the flat fee was introduced the Government legislated to allow institutions to begin setting their own fees.

The fees students paid to attend tertiary educational institutions then rose markedly throughout the 1990s, and the New Zealand Union of Students' Associations (2007) has estimated that the increases were on average 13 percent per annum over this period. The higher fees have resulted in the level of student debt increasing considerably over the last 10 years of this study, with the total level of debt held in 2004 being an estimated $5.96 billion (Ministry of Education 2005, 9). According to the Ministry of Education, in June 2004 more than 418,000 people in New Zealand, or 13 percent of those aged 15 years or over, had a student loan debt. This compares with 7.6 percent in 1998 and 1.6 percent in 1993. The median debt in 2004 was just under $10,000 (Ministry of Education 2005, 9).

Overall, in recent decades New Zealand's tertiary education system has changed from one in which the cost of education was low and where allowances were offered to students, to a competitive, market-based system in which universities themselves set fees (McLaughlin 2003, 6).

In addition to these reforms, there has been an emphasis by government on the need to up-skill the population in order to better compete in a changing world economy. As a consequence of the increased emphasis on raising the skill level of the population, the total number of formally enrolled students in tertiary education grew from around 175,000 in 1991 to approximately 368,000 in 2004 (Abbott 2006, 372). Along with rising participation has come a growth in the number of people holding educational qualifications. Newell and Perry note that "From 1981 to 2001, the proportion of New Zealand residents aged 15 years or over with no educational qualification halved from 55.2 to 27.6 percent while the proportion with a university degree tripled from 3.8 to 11.8 percent" (Newell and Perry 2006, 6).

Report structure

The report is structured as follows. The remainder of section 1 provides information on the data used and how they were accessed, before explaining the strengths and limitations of using census data. Next it provides detailed information on the classification of families and educational qualifications used in the analysis, and finally it supplies details on the family wellbeing indicators used in the report.

In section 2 the relationship between parental educational qualifications, family structure and family wellbeing are described and commented on for the period 1981–2006, for each wellbeing indicator.

Section 3 first discusses the broader literature examining the relationship between educational status and family wellbeing, with a particular focus on the New Zealand experience. It then discusses the results described in section 2 and relates them, where possible, to the wider literature examining the relationship between families' wellbeing and educational status.

Methods

Data source and data access

All data used in this report were derived from the New Zealand Censuses of Population and Dwellings conducted between 1981 and 2006 by Statistics New Zealand. The research team obtained access to confidentialised unit record data through Statistics New Zealand's secure Data Laboratory facility in Auckland. Personal identification information supplied on the original census forms, such as name and address, is not carried over to the computer records held by Statistics New Zealand, and these details are therefore not available to Data Laboratory users.

Using census data to measure wellbeing

The census contains a wealth of information on a wide range of demographic, social and economic issues covering the entire population − or at least those who completed the population census forms. The primary advantages of using census data to assess wellbeing are as follows.

• It allows for an assessment of continuity and change in societal patterns over a long period of time.
• Information obtained from the census covers (almost) all members of the population. It therefore allows us to examine the wellbeing of all New Zealanders, and can provide information on small population groupings in a way that sample surveys rarely can.
• The census collects information on all family members in the household, enabling us to conduct family-level analysis. Such an analysis acknowledges the fundamental interdependence between family members and enables us to see how the impact of the reforms since the mid-1980s has varied according to family type.
• Although the census collects no information on the subjective elements of wellbeing, many of the core outcomes (good jobs, adequate income, education and health) identified by New Zealanders as promoting wellbeing are based on objective living conditions, which are captured (with the limitations outlined below) in the census.

The limitations associated with using census data to measure changes in family wellbeing are linked to the limited range and depth of information collected, the frequency of collection of some data, and the way in which family types are defined and measured.

• The selection of indicators is constrained by the information available through census data. Family and household wellbeing may be influenced by other factors (e.g., the perceived quality of family/household relationships) for which no census information is available. This lack of suitable information also necessitates some indicators being indirect proxy measures of a particular attribute. For example, from the fuller list of indicators examined in Cotterell et al. (2008), the health indicator describes changes in the number of people receiving health-related benefits, rather than being an actual measure of the physical health of a family.
• A lack of data availability may constrain time-series analysis. Some census questions that may be relevant to family/household wellbeing are no longer asked (e.g., on housing insulation), while other census questions (e.g., on smoking) are included only on an irregular basis. This means that the monitoring of changes in some domains (e.g., health, not specifically examined in this report) is less frequent than ideal.
• A lack of in-depth information may place limits on interpreting change in some indicators. For example, because income data are collected in bands rather than in discrete amounts, indicator construction requires some estimation.
• The census definition of 'family' only incorporates those family members who live within the same household. Wellbeing measures constructed from census data may thus be poor indicators for families whose members do not all reside within the one household. In particular, this relates to parents who usually share custody of their children, and children who live across two households. The ability to monitor the wellbeing of those in extended family situations is also constrained by this household-based definition of family.

With the above issues kept in mind, an extensive process of data investigation preceded the construction of wellbeing indicators. Data collected in each census between 1981 and 2006 were checked for consistency and comparability over time. Once this process was complete, a range of indicators, whose purpose was to capture aspects of family and household wellbeing, was constructed.

Wellbeing indicators

With the caveats identified in the above paragraphs in mind, and following the recommendation of Milligan et al. (2006), wellbeing indicators were constructed at both the family and household levels. However, preliminary analyses suggested modifications to the original indicators outlined by Milligan et al. would improve their utility. Principally, several indicators formerly defined only at the household level (i.e., tenure, rental affordability and crowding) were redefined at the family level. In addition, further modifications were designed to make the set of indicators easier to interpret and more consistent. For example, where feasible, indicator definitions were modified to ensure that increases in particular indicators indicated a worsening of levels of wellbeing.

For the purposes of this analysis, a subset of what were considered analytically relevant indicators was selected. These indicators were: median equivalised income, low income, families without paid work, long hours worked and home ownership (see Table 1.1). The remaining indicators that were used in the previous FWWP results report but are not examined here include low rental affordability, household crowding, and receipt of health-related benefits; the education domain indicators are instead used in this report to break up the population (Cotterell et al. 2008). Appendix C provides detailed information on how the wellbeing indicators presented in this report were constructed.

Classifying educational qualifications

This report presents wellbeing indicators at the family level and relates these to educational qualifications. As such, a method had to adopted for attributing education − very much a personal trait − to families. Education information was collected to varying degrees of detail in the six censuses from 1981 to 2006: in each, secondary school and post-secondary qualifications were asked about separately; overseas qualifications were recorded separately from 1991 onwards; trade certificates and vocational qualifications were coded variously, and only from 1986 onwards; the New Zealand Certificate of Educational Achievement (NCEA) levels were included in the 2001 and 2006 secondary qualification categories, alongside the 'old school' assessments, and so on.

For the purposes of these analyses, to provide the best compatibility across time (and to make them less overwhelming) it was decided to examine three levels of qualification: post-secondary qualification, secondary school qualification and no educational qualification. This three-way categorisation is not the only schema possible and different categorisations of qualifications could be used for future research. For example, a useful categorisation would include a degree/non-degree post-secondary qualification breakdown, in addition to the no educational and secondary school qualification categories.

To relate these to families, we used the education levels of the parent(s) to give: at least one parent with a post-secondary qualification; at least one parent with a secondary school qualification (no parent with post-secondary); and no parent with any educational qualification. This was based on the assumption that the income, educational and employment characteristics of parents were likely to have the most influence on the level of family wellbeing.

Family/household classifications

A family consists of a nucleus of parent(s) and child(ren) who reside in the same dwelling. Statistics New Zealand notes that:

'Family type' is a derived variable that classifies family nuclei according to the presence or absence of couples, parents and children. A 'family nucleus' is a couple, with or without child(ren), or one parent and their child(ren) usually resident in the same dwelling. The children do not have partners or children of their own living in the same household. People who usually live in a particular dwelling, and are members of a family nucleus in that dwelling, but who are absent on census night, are included, as long as they are reported as being absent by the reference person on the dwelling form.<¹

In contrast, a household is defined as any group of families or individuals living in the same dwelling, regardless of their relationships to one another. Therefore, census families are wholly contained within households.

Family and household units in the census are further classified according to their structure, the dependency status of children, and retirement status. Child dependency is defined in terms of age and work and labour-force status. All children in family nuclei aged 14 years and under are classed as 'dependent children'. Those aged 15 to 17 years inclusive are also dependent unless they are in the full-time labour force. Children over the age of 17 years are classed as 'independent children' or 'adult children'.²

In this report the primary focus is the census family unit, and indicators were compiled for these units and cross-classified by family type and child dependency status. This classification consists of five categories: couple without children, couple with dependent children, couple with only independent children, one-parent family with dependent children, and one-parent family with only independent children.

Where particular family types were living in multiple-family households they were classified separately into their constituent family types. This action was taken because the focus in this report is on changes in the level of family wellbeing. Note, however, that multiple-family households  that is, two, three or more families living in a single dwelling  make up a very small proportion of the total household count.

Table 1.2 shows the numbers of the various family/household types that existed in each of the censuses under study, and includes the numbers of those families and households in the categories that are not analysed in this report in the second half of the table.

The family types used in this report are: couples without children; couples with dependent children; couples with only independent children; one-parent families with dependent children; and one-parent families with only independent children. These classifications and their associated definitions are detailed in Table 1.3 below.

Footnotes

1. 2006 Census Information Releases, Statistics NZ website.
2. See Statistics New Zealand, 2001b for definitions of 'child in a family nucleus' and 'labour force status'.
3. See Appendix C.4 for a definition of retirement.