Notes

For example, recommendation number 10 of the Report by the Commission on the Measurement of Economic Performance and Social Progress (Stiglitz et al. 2009) states the following: "Measures of both objective and subjective well-being provide key information about people’s quality of life. Statistical offices should incorporate questions to capture people’s life evaluations, hedonic experiences and priorities in their own survey." Similarly, the World Happiness Report (Helliwell et al. 2012) presents data on subjective well-being for more than 150 countries and recommends more general collection by national statistical agencies.

The 2009 GSS was also fielded in the territories; however, those respondents are not included in the analysis for this study so as to keep the sample consistent across surveys.

Some years are exceptions. The 2007 GSS collected information from respondents aged 45 or older, and 618 of the 23,404 interviews were conducted by proxy.

Because the 2007 GSS sample was restricted to respondents aged 45 or older, data from that year are not included in most of the analysis in this study.

For example, in 2002, respondents were asked: "Are you very satisfied, satisfied, dissatisfied or very dissatisfied with your life in general?"

Except for the fact that the wording "right now" was omitted from the question in the 2011 GSS.

Prior to 2009, life satisfaction in CCHS was measured on a five-point scale.

The Ordered Probit model of life satisfaction estimated using the pooled sample yielded qualitatively similar results to those obtained from the OLS model.

As mentioned earlier, the 2007 GSS interviewed exclusively individuals aged 45 or older.

In the GSS data it is the number of children aged 17 or younger. In the CCHS data it is the number of children aged 18 or younger.

Labour force activity is measured as main activity during a reference week in all surveys except the 2003 and 2006 GSS where it is determined based on respondent’s main activity in the previous 12 months.

In the GSS for the years from 2003 to 2008, household income information is available only in broad categories, precluding the option of calculating the real annual income of households. Broad categories of nominal household income are included in regressions nonetheless as control variables. The results are not qualitatively different when this set of variables is excluded from the models.

Measures of the quality of one’s relationships with friends and family as well as levels of trust and feelings of belonging are well documented covariates of life satisfaction. These correlations are not studied in this paper since the required variables are not available in all of the surveys considered. Rather, it is a priority to study the comparability of estimated relationships between life satisfaction and a smaller set of covariates available and measured consistently across all ten surveys.

Barrington-Leigh (2010, 2013) reported on the consistency of coefficients estimated using the 2003 and 2008 GSS samples.

That is, the coefficient is negative in cases where married individuals became separated or widowed, and is positive in cases where unmarried individuals became married.