Characteristics Associated with Happiness

This is the draft of a paper for 2009 Conference of the New Zealand Statistical Association, 3 September 2009. The presentation was a PowerPoint based on it. It is part of a study of the impact of drinking by associates undertaken by the  Centre for Social and Health Outcomes, Research and Evaluation (SHORE), Massey University.
 

Authors:  Ru Quan (Ryan) You and Brian Easton (with Sally Casswell and Taisia Huckle)
 

Keywords: Health; Statistics;
 

Introduction
 

This paper reports results from a survey whose primary concern was the impact of drinking on personal welfare. In order to investigate the impact of drinking, we have to control for demographics variables since they influence both drinking behaviour and general behaviour. In this paper we report on how those characteristics influence one particular indicator of welfare – satisfaction with life which, following economics practice, we sometimes abbreviate to ‘happiness’ (although psychologists warn they are not quite the same thing). In particular we control for other characteristics (and drinking) so that the interaction between variables is allowed for (for instance widows tend to be older than married, so the happiness of widows may reflect their age rather than their marital status).
 

The Survey
 

The general population survey was conducted using the SHORE/Whariki in-house Computer Assisted Telephone Interviewing System during 2008/9.  The primary survey population included people aged 12 to 80 years living in private residential dwellings with a connected landline telephone in New Zealand, although a feasibility survey of a secondary random sample of residents with cell phone only access is currently being conducted.
 

The survey instrument, which was developed in collaboration with our Australian colleagues who are doing a parallel survey  included measures of quality of life using the Personal Wellbeing Index and the EQ5-D. Respondents were asked about their drinking behaviour and the heavy drinkers in their lives. Questions on demographic variables were also included
 

Altogether there were 3068 responses with a 64 percent response rate.
 

The Focus of this Paper
 

While the primary purpose of the research project is to evaluate the impact of drinking of associates, it was necessary to control for demographic variables. The drinking behaviour is reported in a subsequent paper. This paper focuses on the relationship between the various personal characteristics and their responses to the satisfaction with life question. Time precludes reporting similar results for the other dimensions of personal well-being.
 

The satisfaction with life question was:
 

I am now going to ask you how satisfied you feel on a scale of Zero to 10.

Zero means you feel completely dissatisfied. 10 means you feel completely satisfied and the middle of the scale is 5 which means you feel neutral (i.e neither satisfied nor dissatisfied).
 

Thinking about your own life and personal circumstances, how satisfied  are you with your life as a whole?
 

 0 = Completely dissatisfied , 1, 2, 3, 4, 5= Neither satisfied nor dissatisfied,  6, 7, 8, 9, 10 = Completely satisfied.
(also scored  Refused  Don’t Read &  Don’t know or Don’t Read)
 

Respondents generally scored a 7, 8 or 9.
 

Notice that this scale is over a finite interval. Consequently the reported variable was scaled to unity and transformed into a normal logit (say Y = LN(x/10/(1-x/10)) where x is the reported score.
 

The impact of the demographic variables (Z) was estimated with the equation
 

                        Y = a + b*Z + u,
 

Where u was a random variable with the usual properties of being normally distributed.
 

Results
 

All the results reported here are relative to a person who is a 50 year old female of Pakeha ethnicity, who is employed as a professional with a university degree, on an income of $60,000 p.a. and who does not drink. On average she would report a satisfaction with life score of 8.2. In the tables bold marks the base category.
 

An asterisk (*) indicates that the other status is significantly different from this at a 5 percent significance level. Note that two estimates may both be significantly different from the base, but not from each other.
 

Gender
 

If the base female is 8.2, the male with otherwise similar characteristics is 8.4. This is not statistically significantly different. It is common to find on average males are less happy than females, but this is because they have more unhappiness prone characteristics rather just their gender.
 

            Life Satisfaction by Gender

<>Gender
<>Female
<>Male
<>Life Satisfaction Score
<>8.2
<>8.4

 

Age
 

A quadratic fitted age against life dimension was statistically significantly with a minimum at 48 years old. Thus a person is happier when they are young, their happiness declines progressively as they get older, until they are about 50, and then increases again. Note that sample consists of those in their own homes (which may be important in interpreting older people’s life satisfaction, and who are not dead (which almost certainly is).
Life Satisfaction by Age

<>Age (years)
<>20
<>30
<>40
<>50
<>60
<>70
<>80
<>Life Satisfaction Score
<>8.8
<>8.5
<>8.3
<>8.2
<>8.3
<>8.5
<>8.9

 

Ethnicity
 

The only ethnic group which is statistically different from the base (is Pakeha) was Asian. This is a cultural response; Asians always tend to score in the middle of survey scales, whereas in the case of happiness non-Asians tend to score more at the top. This is a reminder that there are cultural elements to the responses.
 

Note that the estimate does not tell us that, in actuality, Maori are as happy as Pakeha, only Maori with the same demographic characteristics are the same.
 

                        Life Satisfaction and Ethnicity

<>Ethnicity
<>Pakeha
<>Maori
<>Pasifika
<>Asian
<>Life Satisfaction Score
<>8.2
<>8.1
<>8.3
<>7.8*

 

Marital Status
 

Marital status matters. The married are more satisfied with life on average. and this is statistically different from other the states. Women who are separated (or widowed or divorced) are statistically less happy than married women but much the same level of happiness applies to single women. Separated men are even less happy than women, but single men are about half way between married men and separated men in life satisfaction terms.
 

Life Satisfaction and Marital Status

<>Marital Status
<>Female
<>Male
<>Married
<>Separated
<>Single
<>Married
<>Separated
<>Single
<>Life Satisfaction Score
<>8.2
<>7.6*
<>7.6*
<>8.4
<>7.2*
<>7.8*

 

 

Employment Status
 

It turns out that the only two employment situations which are statistically different from being employed full time are being sick or being a full time male parent.
 


Life Satisfaction and Employment Status

<>Employment Status
<>Life Satisfaction Score
<>Employed Full Time
<>8.2
<>Employed Part time
<>8.3
<>Student
<>7.9
<>Unemployed
<>7.8
<>Sick
<>7.0*
<>Retired
<>8.4
<>Female Parenting
<>.8.4
<>Male Parenting
<>7.3*

 

Education and Training
 

Other than men with a trade certificate, educational and vocational training has little direct effect on one’s satisfaction with life. It may have an indirect effect since marital status, employment status, occupation and income are all affected by qualifications.
 

Life Satisfaction and Education and Training

<>Education and Training
<>Life Satisfaction Score
<>Post Graduate Degree
<>8.4
<>Degree
<>8.2
<>Professional Certificate
<>8.4
<>Diploma
<>8.2
<>Male Trade Certificate
<>8.8*
<>Female Trade Certificate
<>7.9
<>Secondary Education Certificate
<>8.4
<>No Secondary Education Certificate
<>8.4

 

Occupation
 

Occupations divide into two categories. Occupations higher in the hierarchy – professions and  their allies and skilled workers – are more satisfied with life than those lower in the occupational hierarchy —  trades workers and labourers
 

Life Satisfaction and Occupation

<>Occupation
<>Life Satisfaction Score
<>Professional (degree)
<>8.2
<>Professional (other)
<>8.3
<>Director
<>8.1
<>Managerial
<>8.3
<>Technical
<>8.3
<>Skilled Worker
<>8.1
<>Trade  occupations
<>7.8*
<>Labourers
<>7.8*
<>Others
<>7.7*

 

Incomes
 

A statistically significant quadratic on log income gives the patters shown in the following table. Certainly one’s happiness rises with increased income, but the increase is surprisingly small. For instance quintupling income form $20,000 p.a. to $100,000 per year, increases happied by .2 units, whereas being sick relative to being employed reduces it by six times that and being single relative to married three times that. One might conclude it is happier to be married healthy and poor than living alone sick and ric – on average of course. .
 

Life Satisfaction and Income

<>Annual Income
<>$20,000
<>$40,000
<>$60,000
<>$80,000
<>$100,000
<>$120,000
<>Life Satisfaction Score
<>8.1
<>8.1
<>8.2
<>8.3
<>8.3
<>8.4

 

Conclusions
 

1. The results are not markedly different from similar overseas studies.
 

2. There are New Zealand studies but they tend to ignore the interaction effects.
 

3. For small changes the impacts are additive so that other combinations of characteristics can be calculated providing the total change is not to large (where the log-normal effects become non-linear).
 

4. A similar analysis can be done for other dimensions of personal welfare.
 

5. This is preparation for the main purpose of the survey which is the impact of drinking on personal welfare.
 

Go to top