Correlational finding on Happiness and Married vs never married
Subject code: M02aa01

StudyBradburn (1969): study US 1963
TitleThe Structure of Psychological Well-Being.
SourceAldine Publishing, 1969, Chicago, USA
Public21-60 aged, urban areas, USA, 1963 - 64
Sample
Non-Response 20%, Attrition 30%
Respondents N =2787

Correlate
Author's labelMarried vs. never married
Page in Source 149
Our classificationMarried vs never married, code M02aa01
Operationalization
0 Never married
1 Married

Observed Relation with Happiness
Happiness
Measure
StatisticsElaboration/Remarks
O-HL-c-sq-v-3-aaG=+.40 p < .01
% very happy:      males    females
- never married    18       18
- married          35       38
A-BB-cm-mq-v-2-aDMr=+ p < .05
Happiness level in average ridits (RT)
                          males     females
- never married           .41       .44
- married                 .48       .42
                          (05)      (ns)
O-HL-c-sq-v-3-aatc=+.08 p < .01
A-BB-cm-mq-v-2-aDMr=+
                          males     females
- never married
  < age 25                .42       .45
  > age 25                .45       .46

- married
  < age 25                .59       .49
  > age 25                .51       .47

Ridit analysis compares distribution in category 
with distribution in total sample. RT above .50 
means relative high level, RT below .50 relative 
low level.
O-HL-c-sq-v-3-aaG=+.38 p < .01
Males
O-HL-c-sq-v-3-aaG=+.42 p < .01
Females
         
O-HL-c-sq-v-3-aaD%=+
% very happy:      males    females
- never married
  < age 25         15       25
  > age 25         20       15

- married
  < age 25         42       47
  > age 25         34       37


Appendix 1: Happiness measures used
CodeFull Text
A-BB-cm-mq-v-2-aSelfreport on 10 questions:

During the past few weeks, did you ever feel ....? (yes/no)
A Particularly exited or interested in something?
B So restless that you couldn't sit long in a chair?
C Proud because someone complimented you on something
you had done?
D Very lonely or remote from other people?
E Pleased about having accomplished something?
F Bored?
G On top of the world?
H Depressed or very unhappy?
I That things were going your way?
J Upset because someone criticized you?

Answer options and scoring:
yes = 1
no = 0
Summation:
-Positive Affect Score (PAS): A+C+E+G+I
-Negative Affect Score (NAS): B+D+F+H+J
-Affect Balance Score (ABS): PAS minus NAS
Possible range: -5 to +5

Name: Bradburn's 'Affect Balance Scale' (standard version)
O-HL-c-sq-v-3-aaSelfreport on single question:

Taken all together, how would you say things are these days? Would you say that you are....?
3 very happy
2 pretty happy
1 not too happy


Appendix 2: Statistics used
SymbolExplanation
D%DIFFERENCE in PERCENTAGES
Type: descriptive statistic only.
Measurement level: Correlate level: dichotomous, but nominal or ordinal theoretically possible as well. Happiness level: dichotomous
Range: [-100; +100]

Meaning: the difference of the percentages happy people at two correlate levels.
DMrDIFFERENCE IN MEAN RIDITS
Type: test statistic
Measurement level: Happiness ordinal
Range: [0; +1]

Meaning:
Mr < .50: average happiness in this subgroup lower than in the larger population
Mr = .50: average happiness in this subgroup the same as in the larger population
Mr > .50: average happiness in this subgroup higher than in the larger population

'Ridit analysis' compares the distribution of happiness scores in subgroups to its distribution in the entire sample ("Relative to an Identified Distribution")

Testing for significance can be performed through a "BROSS Confidence Interval" (BCI). If all values the BCI for a subgroup are above/below 0.500, the subgroup is significantly more/less happy than the larger population.
GGOODMAN & Kruskal's GAMMA
Type: test statistic
Measurement level: Correlate: ordinal, Happinessl: ordinal
Range: [-1; +1]

Meaning:
G = 0 no rank correlation
G = +1 strongest possible rank correlation, where high correlate values correspond to high happiness ratings.
G = -1 strongest possible rank correlation, where high correlate values correspond with low happiness ratings.
tcKENDALL'S TAU-C (Also referred to as Stuart's tau-c)
Type: test statistic
Measurement level: Correlate: ordinal, Happiness: ordinal
Range: [-1; +1]

Meaning:
tc = 0 no rank correlation
tc = 1 perfect rank correlation, where high values of the correlate correspond
with high happiness ratings.
tc = -1 perfect rank correlation, where high values of the correlate correspond
with low happiness ratings.
Source:
Ruut Veenhoven, World Database of Happiness, Collection of Correlational Findings, Erasmus University Rotterdam.
https://worlddatabaseofhappiness.eur.nl