Stats NZ has a new website.

For new releases go to

www.stats.govt.nz

As we transition to our new site, you'll still find some Stats NZ information here on this archive site.

  • Share this page to Facebook
  • Share this page to Twitter
  • Share this page to Google+
Appendix 1: Methodology

Data is from the New Zealand General Social Survey

The data for this report is sourced from the 2010 New Zealand General Social Survey (NZGSS). This survey collected well-being information from 8,500 New Zealanders aged 15 years and over between April 2010 and March 2011.

New Zealand General Social Survey has more information.

The NZGSS asked people, in the last four weeks, how often have you felt isolated from others? The response options were:

  • all of the time
  • most of the time
  • some of the time
  • a little of the time
  • none of the time

Given this is a subjective question, the resulting measure is one of loneliness rather than an objective measure of social isolation.

Logistic regression is used to model loneliness

As the five possible response options for respondents to this question have a logical order to them from ‘none’ to ‘all the time’, a cumulative multinomial logistic regression was used. The advantage of using regression analysis is that it holds other factors constant, while looking at the association between the likelihood of feeling lonely and the factor of interest.

A cumulative multinomial logistic regression describes the relationship between the lowest versus all higher categories of the lonely variable and the relationship between the next lowest category and all higher categories, etc. Because the relationship between all pairs of groups is the same, there is only one set of coefficients. Therefore, results from the model refer to the likelihood of feeling lonelier.

Interpretation of odds ratios

The results of the logistic regression analysis are presented in the form of odds ratios. An odds ratio is the odds of an event happening divided by the odds of the opposite event happening. For example, suppose that 400 females felt lonely and 200 did not. The odds of a person feeling lonely are 400/200 = 2, or 2 to 1. In other words, the chances of a female feeling lonely are reasonably good. To give another example, suppose that 500 males felt lonely and 1,000 did not. The odds of a male feeling lonely would be 500/1,000 = 0.5 or 1 to 2. The chances of their feeling lonely are therefore significantly lower than for females.

For continuous explanatory variables (age, mental health, physical health, and ELSI), an odds ratio of greater than 1 indicates a higher likelihood of feeling lonely as the value of the explanatory variable increases and an odds ratio less than 1 indicates a lower likelihood.

For categorical explanatory variables, the odds ratio compares the likelihood of feeling lonely compared with the reference category. An odds ratio greater than 1 indicates a higher likelihood of feeling lonely compared with the reference group, while an odds ratio of less than 1 indicates a lower likelihood.

Variables included in the logistic regression model

Age

As the main variable of interest, a person’s age is included as a continuous explanatory variable in the model.

Discrimination

The NZGSS asks respondents whether they have been treated unfairly or had something nasty done to them because of the group they belong to or seem to belong to in the last 12 months. This is a measure of discrimination. For more information on this measure refer to Statistics New Zealand (2012). Responses to this question are included in the model to test for an association with feeling lonely. The reference category is those who have not been discriminated against.

Economic standard of living

Economic standard of living is included as a continuous variable in the model. Economic standard of living is measured by the Economic Standard of Living Index Short Form (ELSISF). Economic standard of living refers to the material aspect of well-being that is reflected in a person’s consumption and personal possessions – their household durables, clothing, recreations, access to medical services, and so on (Jensen, Spittal & Krishnan, 2005).

Economic standard of living is a more robust measure of material well-being than household income for older people. This is because they often have lower income because of retirement, but reduced expenses, such as no home loan repayments. Given the focus of this analysis is to look at loneliness across the life stage and specifically look at loneliness for older people, it made sense to include ELSISF in the model over household income.

Ethnicity

Whether a person identifies as Asian or not is included in the model as a categorical explanatory variable. The reference category is those who do not identify as Asian. Variables for other ethnic groups were tried in the models but none of them showed a significant association with feeling lonely.

Household size

How many people live in a person’s household is included as an explanatory variable in the model. The reference category is four or more people.

Mental and physical health

Mental and physical health are measured by the short-form health questionnaire (SF-12), which is designed to self-assess symptoms and limitations in everyday activity due to mental and physical health over the previous four weeks. Responses to the SF-12 are summarised in two weighted summary scales – the physical component score (PCS) and the mental component score (MCS). These scores range from 0 to 100, where a 0 score indicates the lowest level of health and 100 indicates the highest level of health.

Migrant status

A person’s migrant status is included as an explanatory variable with multiple categories in the regression model. A person is categorised as a recent migrant when they have arrived in New Zealand within the last four years. The reference category is people born in New Zealand.

Seen family and friends

This variable is derived from responses to questions in NZGSS. It measures whether a person has had face-to-face contact with family or friends who live outside their household in the last week. The reference category is those who have had contact.

Sex

Also included in the regression model is the sex of people. The reference category is females.

  • Share this page to Facebook
  • Share this page to Twitter
  • Share this page to Google+
Top
  • Share this page to Facebook
  • Share this page to Twitter
  • Share this page to Google+