Correlational findings | World Database of Happiness

Study Ding (2017): study CN 2011

Public: 45+ aged, rural hukou, China, 2011
Survey name: CN-China Health and Retirement Longitudinal Study (CHARLS)
Sample
Respondents: N = 10378
Non Response
Assessment: Interview: face-to-face

Correlate

Authors's Label

Residence

Our Classification

Related specification variables

Operationalization

1. living in urban area
0. living in other area

Observed Relation with Happiness

Happiness Measure

Statistics

Elaboration / Remarks

O-SLW-c-sq-v-5-z

b = +.07 p < .05

REGRESSION COEFFICIENT (non-standardized) by LEAST SQUARES (OLS)

Type: test statistic
Measurement level: Correlate: metric, Happiness: metric
Theoretical range: unlimited

Meaning:
b > 0   A higher correlate level corresponds with a higher happiness rating on average.
B < 0   A higher correlate level corresponds with a lower happiness rating on average.
B = 0   Not any correlation with the relevant correlate.

Living in urban area (vs other)

b controlled for
- age
- edu
- marriage
- health
- income, land, house
- new rural pension program (NRPP)
- other pension
- descendant
- having at least one son

O-SLW-c-sq-v-5-z

b-iv = +.15 ns

REGRESSION COEFFICIENT with INSTRUMENTAL variables

Type: test statistic.
Correlate level: metric, Happiness level: metric
Theoretical range: unlimited

The instrument must be correlated with the endogenous explanatory variables, conditionally on the other covariates. If this correlation is strong, then the instrument is said to have a strong first stage. A weak correlation may provide misleading inferences about parameter estimates and standard errors.
The instrument cannot be correlated with the error term in the explanatory equation, conditionally on the other covariates. In other words, the instrument cannot suffer from the same problem as the original predicting variable. If this condition is met, then the instrument is said to satisfy the exclusion restriction.

See Mardia Kent & Bibby (1979): Multivariate Analysis

Additionally Instrument Variables used for new rural pension program
- availability of new rural pension program
- expects to rely on pension

O-SLW-c-sq-v-5-z

OLRC = +.16 p < .10

REGRESSION coefficient in ORDERED categorical LOGISTIC analysis.

Only the sign of the computed coefficient is informative.

Happiness is an ordered categorical variable. Higher categories correspond to being happier.

OLRC < 0 indicates that the odds of being beyond a chosen happiness category-to- be-ing at or below that category decreases when

1) the corresponding metric correlate increases
2) the corresponding category of a categorical correlate is compared to the reference category.

OLRC > 0 indicates an increase in the odds for both the above cases.

- All sample

O-SLW-c-sq-v-5-z

OLRC = +.21 p < .10

REGRESSION coefficient in ORDERED categorical LOGISTIC analysis.

- Age 45-49

O-SLW-c-sq-v-5-z

OLRC = +.00 ns

REGRESSION coefficient in ORDERED categorical LOGISTIC analysis.

- Age 60-74

O-SLW-c-sq-v-5-z

OLRC = +.03 ns

REGRESSION coefficient in ORDERED categorical LOGISTIC analysis.

- Age 75+

OLRC controlled for
- gender
- age
- edu
- marriage
- health
- income, land, house
- new rural pension program
- other pension
- descendant
- having at least one son

Similar results when using Instrument Variables for new rural pension program
- availability of new rural pension program
- expects to rely on pension

O-SLW-c-sq-v-5-z

OLRC = +.39 p < .01