World Database of Happiness, Introduction to Bibliography

Chapter 4

All publications entered in this bibliography are identified by 'discipline', 'subject matter' and 'study type'. If a publication reports the results of an empirical study, the 'measures' of happiness used in that study are characterized as well.

4.1 Discipline

Publications are characterized as belonging either to the field of 'philosophy', 'social sciences' or 'medical sciences'. This classification works for the bulk of the studies considered, but there are of course cases of doubt. For example, in studies on the relationship between happiness and longevity both biological and psychological processes are involved. In such cases two disciplines are mentioned. Most of the literature reviews cover more than one discipline.

table 2
Happiness studies by discipline

philosophy                                  88

social sciences                         2927

medical sciences                      125

As can be seen in table 2, the great majority of the studies in this bibliography come from the social sciences. Most of these are psychological studies. There is also quite a number of sociological publications, in particular from the field of Social Indicator Research. Curiously, antropological studies on happiness are non-existant.

4.2 Subject matter

Topics are characterized by means of five key words: 'concept' of happiness, 'measurement' of happiness, 'determinants' of happiness, 'consequences' of happiness and 'views' on happiness.

Of course all publications on happiness say something about the concept. Hence, this description is used only when conceptualisation is a main topic. By means of the subject-index to this bibliography you find these titles ordered by sub-issues (Subject category 1)

Likewise, all empirical studies involve measures of happiness. This description is applied to the ones that focus on measurement problems in the first place, or involve unusual methods of measurement. Specific measurement issues are spelled out in the subject-index to this bibliography (Subject categories 2 and 3)

Many studies focus on causes of happiness: either causal conditions or causal processes. These studies are typified as focusing on 'determinants', even though most of them work with correlational data that often allow an other interpretation as well. The various determinants of happiness appear in detail in the subject-index of this bibliography (Subject categories 4 to 13).

Some studies consider the effect of enjoying life or not: for example whether a positive appreciation of life adds to good health and responsible social behavior. Many studies involve relevant data for the answering of such questions, but in this context only those studies are mentioned that enter the matter explicitly. Studies on consequences of happiness are presented in more detail in subject-category 4h.

Views on happiness
Next to these studies on actual causes and consequences of happiness, there are also studies about what people think about happiness, and how valuable they deem it to be as a goal. These studies are presented in more detail in the section on 'Sociology of happiness ideas' in subject-category 15 of this bibliography
Obviously a study can concern more than one topic. Literature reviews typically cover them all. These cases are typified by more than one description
Table 3 presents an overview. Clearly the most common topic is 'determinants' of   happiness. The subject of consequences is not so much studied as yet.

table 3
Happiness studies by main subject matter

concept                                              193

measurement                                      500

determinants                                      2436

consequences                                      162

beliefs & values on happiness               286

4.3 Study type

The studies in this bibliography are also characterized by the main research technique. Three types are discerned: 'treatise', 'empirical study' and 'literature review'.

A study is characterized as a 'treatise', if it is largely based on theoretical speculation and does not involve empirical checks of these by means of happiness measures. An example of such a treatise is MacKaye's 'Economy of Happiness' which argues that free enterprise is likely to make life of the average man more satisfactory in the long run, but does not provide any direct evidence for that claim. Speculative studies were the rule, until survey data became available in the 1950's.

Empirical study
All studies that involve empirical assessment of happiness are characterized as 'empirical'. Obviously such studies also involve theoretical notions. In some cases the theory is in fact more prominent than the data.
Among the empirical studies four methodological variants are noted that are particularly relevant for the identification of the causes and consequences of happiness. These are: 'longitudiual' studies (including studies, that involve time-sampling), 'experimental' studies, 'cross-national' studies and 'meta-analytic' studies. The other empirical studies are merely 'descriptive' or 'correlational'.

Literature review
Most review studies focus on one particular issue in the study of happiness: on a particular theory, (i.e.comparison theory) or on a particular determinant (i.e. social prestige). These studies are characterized as 'special issue' reviews. Literature reviews that aim at a complete picture of happiness-research are depicted as 'general' reviews.
Of course most publications on happiness involve some reference to the literature. The description 'review' is used only for studies in which literature surveying is central.
Table 4 lists the characterization of the studies in this bibliography. The great majority are empirical studies.

Table 4
Happiness studies by type of inquiry

Treatise                                  100

Empirical study                     2480

            of which:

            - longitudinal                              256

            - experimental                              58

            - cross cultural                             66

            - meta-analysis                             46

Literature review                     362

            of which:

            - special subject                         235

            - general review                           51

4.4 Happiness Measure

The measures of happiness used in empirical studies are described in some detail. It is established whether the temporal perspective is in the present, and if it is, what variant of happiness is measured. Next, it is considered whether the measures used validly measure happiness. Only the studies that do, are deemed acceptable for further meta-analysis, and are included in the further catalogues of the World Database of Happiness ('Catalogue of Correlates' or 'Catalogue of Happiness Distributions in Nations'.

Temporal perspective
Past happiness. Some studies assess happiness retrospectively, for example by inquiring about the happiness in ones youth or about the most happy time in life up to now. These studies are depicted with the term 'past'. Follow-up studies are not depicted as such: though these studies also asses earlier happiness, that happiness was 'present' at the time of its assessment.

Present happiness.
All studies that asses current satisfaction with life are typified as 'present'. The main goal of this bibliography is to gather empirical studies on present happiness.

Expected happiness
Some studies assess how happy people think they will be in the future: mostly in the next five years. These prospective studies are characterized as pertaining to 'expected' happiness. Expected happiness is not the same as present happiness. In fact presently unhappy people tend to console themselves with optimistic projection in the future.

Perceived changes in happiness.
There are also studies that inquire about retrospective and prospective changes in happiness, in particular studies that use life-graph methods. These studies must be distinguished from longitudinal studies that assess actual change in happiness through time. There are only a few studies of this kind. See table 5.

Table 5
Empirical happiness studies by temporal perspective on happiness

Past happiness                                         156

Present happiness                                 2424

Expected happiness                                 103

Perceived changes in happiness                67

Component measured
As noted, the prime aim of this project is to select studies on present happiness that provide data for meta-analysis. Meta-analysis makes sense only if based on studies that measure the same thing; in this case, when they measure happiness as defined above. Therefore all indicators of happiness were carefully inspected. In the first place it was assessed considered what variants of happiness were measured. Secondly, it was whether these indicators meet specific validity demands for these variants. Studies that involve acceptable indicators are included in this database for further analysis; either in the 'Catalogue of Happiness Correlates' or in the 'Catalogue of National Happiness Distributions'. The selection is based on a test for 'face validity'. This involves close reading of questions, instructions and further devices. This procedure is described in more detail in chapter 4 of 'Conditions of Happiness'.

Accepted measures

Indicators of overall happiness
Overall happiness can be assessed by direct questioning only; not by questions that tap essentially different matters supposed to be related to happiness (see below). Questions can be framed in different ways: in closed questions, in open-ended questions and in focused interviews. In the latter two cases clear instructions for content-analysis of responses are required
Overall happiness cannot be assessed by peer ratings, because peers mostly do not know precisely what the subject has on his mind and rather imagine how they themselves would feel if they were in his shoes.

Indicators of hedonic level
Hedonic level of affect can be assessed in three ways: by direct questioning, by projective tests and by ratings on the basis of non-verbal behaviour. Again the method of direct questioning is to be preferred: in particular methods where the individual is asked several times during a certain period how pleasant he feels at that given time (time sampling). Though generally less dependable, indirect methods can sometimes suffice. Some projectives tests at least seem to be reasonably valid. Ratings by others on the basis of non-verbal behaviour will also suffice, provided that rating instructions are sufficiently specific. Unlike cognitive judgements, affective conditions do show up fairly reliably in non-verbal behaviour.

Indicators of contentment
Contentment can be measured by means of direct questions only. Like overall happiness, it cannot validly be assessed by indirect questions or by peer-ratings. Direct questions must again be specific. They probably work best when preceded by an enumeration of one's major aspirations. Questions can be framed in various formats.

Mixed indicators
Finally, there are several acceptable indicators that cover two or more of these happiness variants. The majority of these consist of single direct questions which by wording or answer formats refer to both overall happiness and hedonic level. As long as they do not labour specific deficits these questions are accepted.
Some indicators work with multiple questions. Characteristically these questions cover both overall happiness and one or both of the discerned components. When all items meet the demands outlined above, such composite indicators are accepted.
A last format to be mentioned in this context is the focused interview of which the 'depth interview' is a variant. Such interrogations tend to cover all three happiness variants. A lack of clear reports on the themes of enquiry and on rating procedures makes it difficult to assess their face validity.

Rejected indicators
Several currently used indicators of subjective well being appear not to fit the concept of happiness as defined here. Findings yielded by such indicators are not included in the further catalogues of the World Database of Happiness.

Scales involving non-happiness items
Many currently used indicators consist of lists of questions, part of which refer to happiness as such and part to related, but essentially distinct matters. Many scales in gerontological research for example, mix questions about happiness with items on 'social participation', 'future orientation' and 'health perception'. i.e. Neugartens' LSI-A ). Likewise, current scales in community mental health add items like 'nervousness', 'vigor' and 'emotional stability', f.e. Dupuy's GWB.  Such indicators are rejected because it is not sure that high social participation, future orientation and vigor always mark a high appreciation of life. There are always socially active, future orientated and vigorous people who are profoundly dissatisfied with their life. Scores on indicators of this kind are difficult to compare and in correlation analysis such indicators often produce contamination.

Summed life-aspect satisfactions
Another currently used method is first asking separate question on satisfaction with various domains of life, such as 'work', 'marriage' and 'leisure' and next combine the responses in a sumscore. This method has several drawbacks. Firstly, it does not adequately reflect the individuals 'overall evaluation': such sumscores tap selected aspects of life only and it is the investigator, who awards weights rather than the subject. Secondly, not all aspect- satisfactions apply equally to everybody; how about the marriage-satisfaction of the unmarried and the work-satisfaction of the unemployed? Thirdly, the significance of life-aspects such as work and marriage is not the same accross time, culture and social categories. Comparisons are therefore impossible with such indicators.
These objections apply not only to sumscores of domain satisfaction i.e. Andrews & Witheys' , but also the 'semantic-differential scales' which involve the ratings of ones life on various evaluational criteria such as 'boring/interesting', 'lonely/friendly' and 'hard/easy'. Such a scale is part of the much used Index of Wellbeing by Campbell .

Comparison with others and the past
Several investigators have asked their subjects how happy they think they are compared to others, rather than how they feel themselves. Such items are rejected. Even if one is happier than one's neighbour, one can still be unhappy. For the same reason the item 'I have been happier than I am now' is deemed unacceptable. Being less happy than before does not imply that one is unhappy. Likewise, the item 'If I could live my life over, I would change almost nothing' in Diener's Satisfaction with Life Scale is not acceptable. Enjoying one''s present life does not mean that one does not appreciate other ways of life.

Table 6 shows that less than half of the available empirical studies on present happiness meets the validity demands applied here.

Table 6
Empirical studies on present happiness by indicators used

                                            included              not included

overall happiness                     1126                            804

hedonic level                              415                            197

contentment                                 55                              16