4  Data issues and descriptive statistics

4.1 Defining who is Maori

Chapple and Rea (1998), in their study on differences in labour market outcomes between Maori and non-Maori, explore the critical issue of how ‘ethnicity’ is defined. They note that the question related to ethnicity in the Household Labour Force Survey is only asked of ‘household heads’ on behalf of other household members. They claim that the question appears to be misleading, because it implies that the respondent should choose only a single ethnicity even though up to three ethnic groups may eventually be chosen. Despite this perhaps unintentional emphasis on single ethnicities, the authors find considerable evidence on the importance of multiple ethnicities for defining who is Maori.

The conventional procedure for defining ethnicity in the HLFS, and in many other statistical surveys in New Zealand, is to place Maori ethnicity in a clear hierarchical position. Individuals who identify themselves as Maori are classified as Maori, and not as one of the other ethnic groups with which they might identify. In other words, those defined as Maori include those with ‘mixed’ ethnicities. Those defined as non-Maori exclude anyone with Maori ethnicity. In this way, the broadest possible interpretation is used in defining Maori identity.

Chapple and Rea found that this broad Maori classification increased in the HLFS from around 10% to 10.5% of the working-age population between 1985 and 1998. Over this same period, the share of Maori reporting other ethnic affiliations increased from 22% to 27%. The authors claim that this is an important definitional issue, because Maori with mixed ethnicities may be quite different from Maori without other ethnic affiliations. Chapple and Rea (p.130) reach the conclusion that “… On average, the sole Maori group faces substantially greater labour market disadvantage than the mixed Maori group.” The authors also note that those with multiple ethnicities comprise a much larger proportion of all Maori in the South Island. This implies that the issue of defining who is Maori is even more important with the CHDS, where the vast majority of respondents reside in the Canterbury region.

Alexander et al. (2001) picked up this issue of an arbitrarily broad classification of Maori ethnicity when they asked Statistics New Zealand to separately classify sole and mixed Maori. They included indicator variables for both Maori groups, Pacific Islanders and other non-European groups in their hourly earnings regressions. Without controlling for sample selection bias, the authors were unable to find consistent statistical evidence of the greater disadvantage for sole Maori relative to mixed Maori hypothesised by Chapple and Rea.[7]

Unlike Alexander et al., we test the sensitivity of our empirical results to this issue of ethnic identification by using two different ethnic classification schemes for Maori. First, we use the conventional approach of defining ‘Maori’ to include anyone who claims that they are Maori at the time of the survey at age 21.[8] We call this our ‘two-way’ classification scheme. Second, we separate this broad ethnic group into ‘sole Maori’ and ‘mixed Maori’. We call this our ‘three-way’ classification scheme. Note that the residual group of ‘non-Maori’ is consistent between these two ethnic classification schemes.

These panel data provide an important advantage in considering ethnic identity. In most cross-sectional data sets, ethnic affiliation is self-reported and there is little that can be done to ‘verify’ this information. In the CHDS we have both the self-reported ethnicity of the youth and their parents from earlier surveys. In fact, we have information on the ethnic identities from both birth parents, and from ‘parental figures’ at the ages of 7 and 14 of the CHDS children. These parental figures could be the same as the birth parents, or they could include the relevant custodial adults in the family in these later years (eg, subsequent partners of birth mothers). This allows us to compare the self-reported ethnicities of children with that of both their birth parents and subsequent parental figures.

Table 1 displays the information on ethnicity for the CHDS youth who provided valid information on all of the variables that will be needed for subsequent analyses in this study. A two-way classification scheme is used for the ethnicities of these 21 year-olds. Of these 973 individuals, 109 (11.2%) reported that they had at least some Maori identity. However, 61 youth (56.0% of those identified as Maori) also reported European, Pacific Island or Other Ethnicity. This means that Sole Maori comprise a minority of those falling under this broad, conventional classification of Maori ethnicity. By far the most common additional ethnicity among Maori is European (55.0%). The vast majority of non-Maori also identified with European ancestry (97.5%). These proportions are significantly different from one another at better than a 1% level (indicated by the two ‘asterisks’ in the Maori column). Although Maori are less likely to have some Pacific Island ethnicity than non-Maori, these proportions are not statistically significant.

Parents were asked about their ethnicities at the birth of their children in 1977. This information is recorded for fathers even if they weren’t present in the household at the time of this birth.[9] For youth who claimed Maori ethnicity as age 21, more than one-third of their mothers (39.4%) and one-half of their fathers (54.1%) claimed Maori ethnicity at the birth of these children. Unlike the ethnicity question asked of youth at age 21, the ethnic categories for parents were ‘mutually exclusive’ (ie, they could not tick multiple categories). This means that a parent who ‘primarily identified’ with another ethnicity was unable to list Maori as ‘secondary choice’. This could result in some measurement error in defining the ethnicities of parents.

** Maori mean significantly different from non-Maori mean at 1% level.

* Maori mean significantly different from non-Maori mean at 10% level.

Notes: These data are taken from the 973 respondents in the CHDS who provided valid information for the purposes of this study. Youth are defined as ‘Maori’ in this table if they identify Maori as at least one of their ethnicities at age 21. All other youth are considered to be ‘non-Maori’. Ethnicity information is available for the birth parents, and the maternal and paternal figures at ages 7 and 14 of the child. Paternal figures may differ from birth parents. Youth can choose multiple ethnicities, but the ethnic categories for adults are mutually exclusive (i.e.. Maori/Part Maori, Pacific Island and Other).

More than four-fifths of Maori youth (80.7%) had at least one birth parent claiming some Maori ethnicity. This proportion is substantially greater than the 1.7% of non-Maori youth with at least one birth parent who claimed some Maori ethnicity. The difference between the means is statistically significant.

One of the reasons for any ‘discrepancies’ between the ethnicities of children and their parents is that children may base their ethnic identification later in life on the ethnicities of ‘parental figures’ other than their birth parents. To test this proposition, we use data on the recorded ethnicities of the custodial adults or parental figures in the families in which the CHDS children resided at ages 7 and 14. This is the only information available on the ethnicities of these adults after the birth of the child. Variables on the ethnicities of the parental figures are constructed using data from the three observation points, and are shown in Table 1. A higher proportion of all children report some Maori ethnicity associated with their paternal and maternal figures, when compared to that reported from their parents at the time of their birth.[10] For youth identifying themselves as Maori at age 21, 80.7% had either a mother or father who was self-identified as Maori at the time of their birth. A similar statistic is 90.8% for either maternal or paternal figures up to age for 14 for these children. The correlation between the reported Maori ethnicities of youth and their paternal figures (0.82) is slightly higher than the correlation between the reported Maori ethnicities of youth and their birth parents (0.81).

Table 2 displays similar statistics for the three-way split in youth ethnicity. The group identified as ‘Maori’ in the first column of Table 1 is now split into ‘sole Maori’ and ‘mixed Maori’ in the first two columns of Table 2. For youth who identified Maori as their only ethnicity at age 21, 93.8% had at least one parent who identified him or herself as Maori at birth. All of the youth in this sole Maori category had at least one paternal figure identify him or herself as Maori by age 14. The intergenerational relationship in Maori identity is much weaker for mixed relative to sole Maori. For youth who identified Maori as only one of their ethnicities at age 21, 70.5% had at least one parent identify him or herself as Maori at birth, and 83.6% had at least one paternal figure who identified him or herself as Maori by age 14.

How do we use this information on the self-reported ethnicities of both youth and their parents in this study on the labour market outcomes of these youth? Table 3 summarises the available information on the ethnic identities of youth and their parents, and provides a set of possible alternatives for defining youth as ‘Maori’. At one extreme, if we insisted that both birth parents claim at least some Maori ancestry and youth claim only Maori ancestry, then the number of observations would fall to 13 (1.3% of the entire sample). At the other extreme, if we consider youth who claim some Maori ancestry, along with those whose parents or parental figures claim some Maori ancestry, the number of observations would climb to 137 (14.1% of the sample).