Correlational finding on Happiness and Month of the year
Subject code: I06aa01b

StudySakawa et al. (2015): study JP Kinki 2006
TitleActivity, Time, and Subjective Happiness: An Analysis Based on an Hourly Web Survey.
SourceThe Institute of Social and Economic Research, Osaka University, Discussion Paper, 2015, No. 926, 1- 38
URLhttp://papers.ssrn.com/sol3/papers.cfm?abstract_id=2564415
PublicStudents, Osaka, Japan, followed 14 months, 2006-2008
SampleNon-probability sample (unspecified)
Non-Response13%
Respondents N =70

Correlate
Author's labelMonth of the year
Page in Source 1,15,17,25
Our classificationMonth of the year, code I06aa01b
Operationalization
Average of 24 hourly happiness ratings at randomly 
determined moments during one day of each month.

Observed Relation with Happiness
Happiness
Measure
StatisticsElaboration/Remarks
A-ARE-m-sqr-n-11-bOLRC=+/-
               OLRC    p<.
January        +.07    ns
February       +.07    ns
March          +.38    001
April          +.20    05
May            +.28    01
June           +.12    ns
July           -.17    10 
August         +.19    05
September      +.12    ns
October        +.10    ns
November (ref.) 
December       +.21    01

Summarized: in march, april, may, august, and 
december respondents were significantly happier. 
In 5 months of the year happiness is higher than 
in november.

OLRC controlled for:
-other months in the equation
-daily hours 1-24

Extra control for 'daily activities' did not 
effect OLRC significantly.


Appendix 1: Happiness measures used
CodeFull Text
A-ARE-m-sqr-n-11-bselfreport on repeated question

How happy do you feel now?
10 very happy
9
8
7
6
5
4
3
2
1
0 very unhappy


Appendix 2: Statistics used
SymbolExplanation
OLRCORDERED LOGIT REGRESSION COEFFICIENT
Statistic: Ordered logit regression coefficient
Measurement level: Correlate: metric, Happiness: ordered responses
Theoretical range: unlimited

OLRC > 0 A one unit increase in the independent variable corresponds to a higher probability of responding in the highest category of the dependent variable and to a lower probability of responding in the lowest category of the dependent variable.

OLRC< 0 A one unit increase in the independent variable corresponds to a lower probability of responding in the highest category of the dependent variable and to a higher probability of responding in the lowest category of the dependent variable.

OLRC = 0 No relationship between the independent and dependent variable..

Remarks:
The interpretation for the intermediate categories of the dependent variable are ambiguous. It is advised to use marginal effects.
Source:
Ruut Veenhoven, World Database of Happiness, Collection of Correlational Findings, Erasmus University Rotterdam.
https://worlddatabaseofhappiness.eur.nl