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 3: Results on model building

This appendix discusses the two ways of building the determinants of life satisfaction models.

As shown in table 1, we included 21 independent variables (ie including age squared) in the model of drivers of overall life satisfaction.

Two ways of building the model on the drivers of well-being are:

  • ordinary least square (OLS) method
  • logistic regression.

The most logical method to use for this study is the ordered logistic regression, since the dependent variable (overall life satisfaction) is ordinal. This means that 0, which means completely dissatisfied, is lower than 10, which means completely satisfied. Discrete steps are located between 0 and 10. The constraint in using this method is its complexity to interpret the results, especially if the ordered categorical responses have more than five levels that are in a clear order (eg levels of well-being with 0 as the lowest category and 10 the highest).

The main advantage of OLS is that interpreting the regression results is more simple and straightforward than the ordered logistic regression.

In this study we used both ordered logistic regression and OLS as a sensitivity check for the robustness of the OLS results, since some of the assumptions for OLS may not hold for the ordered well-being data.

This report adopted the methodology used in What makes for a better life? (Boarini, Comola, Smith, Manchin, & de Keulenaer, 2012) and Measuring national well-being – What matters most to personal well-being? (Oguz, Merad, & Snape, 2013).

Results of the ordered logistic regression were similar to those in OLS. The statistical significance, signs, and relative sizes of the regression coefficients were very similar. Choosing the regression method made little difference to the overall results.

Figures 2 to 5 used the results of the regression model found in table 4.

Table 4
4 Estimated regression coefficients

 Parameter

Estimate

 Standard error

 t Value

 Pr > |t|

 Intercept

 8.842629

 0.546084

 16.19

 <.0001

 sex 0

 0.187181

 0.04828

 3.88

 0.0001

 sex 1

 0

 0

 *

 *

 DUM_children 0

 -0.11199

 0.052781

 -2.12

 0.0341

 DUM_children 1

 0

 0

 *

 *

 urban_area 1

 -0.21953

 0.08914

 -2.46

 0.014

 urban_area 2

 -0.1102

 0.121656

 -0.91

 0.3653

 urban_area 3

 -0.07387

 0.103528

 -0.71

 0.4757

 urban_area 4

 0

 0

 *

 *

 SOLEnoughIncome_r 11

 -0.81782

 0.1092

 -7.49

 <.0001

 SOLEnoughIncome_r 12

 -0.47014

 0.082871

 -5.67

 <.0001

 SOLEnoughIncome_r 13

 -0.14718

 0.071824

 -2.05

 0.0407

 SOLEnoughIncome_r 14

 0

 0

 *

 *

 DUM_unemployment 0

 0.175988

 0.100414

 1.75

 0.08

 DUM_unemployment 1

 0

 0

 *

 *

 HEAHealthStatus_r1 11

 -1.1284

 0.178825

 -6.31

 <.0001

 HEAHealthStatus_r1 12

 -0.79339

 0.107737

 -7.36

 <.0001

 HEAHealthStatus_r1 13

 -0.43776

 0.075463

 -5.8

 <.0001

 HEAHealthStatus_r1 14

 -0.19844

 0.063727

 -3.11

 0.0019

 HEAHealthStatus_r1 15

 0

 0

 *

 *

 DV_highest_qualr 1 No qualification

 0.294426

 0.08086

 3.64

 0.0003

 DV_highest_qualr 2 Level 1 - 4 cert

 0.182567

 0.065641

 2.78

 0.0055

 DV_highest_qualr 3 Level 5 - 6 dipl

 0.168811

 0.119238

 1.42

 0.1572

 DV_highest_qualr 4 Level 7 / bachel

 0

 0

 *

 *

 DUM_partnered 0

 -0.27409

 0.070639

 -3.88

 0.0001

 DUM_partnered 1

 0

 0

 *

 *

 WHALoneliness_r 11

 -0.90106

 0.393102

 -2.29

 0.0221

 WHALoneliness_r 12

 -1.19607

 0.179309

 -6.67

 <.0001

 WHALoneliness_r 13

 -0.8713

 0.098859

 -8.81

 <.0001

 WHALoneliness_r 14

 -0.46253

 0.0683

 -6.77

 <.0001 

 WHALoneliness_r 15

 0

 0

 *

 *

 DUM_Contact 0

 -0.06081

 0.072392

 -0.84

 0.4011

 DUM_Contact 1

 0

 0

 *

 *

 DUM_Volunteering 0

 -0.07144

 0.04888

 -1.46

 0.1442

 DUM_Volunteering 1

 0

 0

 *

 *

 DUM_Crime 0

 0.115977

 0.074703

 1.55

 0.1209

 DUM_Crime 1

 0

 0

 *

 *

 DVImpOfCulture_r 11

 -0.37197

 0.114291

 -3.25

 0.0012

 DVImpOfCulture_r 12

 -0.47981

 0.094258

 -5.09

 <.0001

 DVImpOfCulture_r 13

 -0.37409

 0.080062

 -4.67

 <.0001

 DVImpOfCulture_r 14

 -0.33216

 0.080427

 -4.13

 <.0001

 DVImpOfCulture_r 15

 0

 0

 *

 *

 DUM_AncestralMarae 0

 -0.02043

 0.055183

 -0.37

 0.7113

 DUM_AncestralMarae 1

 0

 0

 *

 *

 DVSpeakTeReo_4grps_r 11_12

 -0.11736

 0.100578

 -1.17

 0.2436

 DVSpeakTeReo_4grps_r 13

 -0.11862

 0.096239

 -1.23

 0.2181

 DVSpeakTeReo_4grps_r 14

 -0.1578

 0.107853

 -1.46

 0.1438

 DVSpeakTeReo_4grps_r 15

 0

 0

 *

 *

 DV_agenbr

 -0.05065

 0.009063

 -5.59

 <.0001

 age_sqrd

 0.000586

 0.000103

 5.69

 <.0001

 log2_hhd_income_r

 0.024467

 0.027481

 0.89

 0.3735

 HOUProblems

 -0.0891

 0.019055

 -4.68

 <.0001

 qCDTTrust

 0.080095

 0.016971

 4.72

 <.0001

 qCDTInstTrust_Courts

 0.053979

 0.012419

 4.35

 <.0001

 Symbol: * No value for the reference group.
Source: Statistics New Zealand
  • 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+