Ruut Veenhoven
Conditions of Happiness, Kluwer Academic, 1984, Dordrecht/Boston.
Having established which indicators of happiness are acceptable, the next step was to
take stock of the investigations that actually used them. Therefore I tried to gather all
empirical investigations in the field up to and including 1975. This was not an easy job.
Many relevant reports had not been included in any international reference system and
terms like 'happiness' were not yet currently used in quick indices. A laborious search
yielded some thousand promising titles. All these reports were inspected. Only 150
appeared to concern investigations that had used acceptable indicators of happiness.
Together these publications covered 156 research projects in 245 different samples. Though
probably incomplete, this harvest is richer than that of any earlier literature surveys.
The earliest empirical happiness investigations were performed in the 1910' s among
students and focused on hedonic level. After World War II their number increased rapidly
and the emphasis shifted to overall happiness in national surveys. The increase in
happiness studies parallels the growth of social research as such.
More than half of the investigations were performed in the US and a quarter in western
Europe. In other parts of the world only a few have been carried out as yet. Almost half
of the investigations was based on probability samples in national populations, the other
half concerning mainly regional populations, aged people and students.
Most investigations focused on 'overall happiness', generally measured by single direct
questions using keywords such as 'happiness' and 'satisfaction with life'. Almost
one-third dealt with 'hedonic level', mostly measured by direct questions on general mood
and by summed specific affect scores. Only a few investigations assessed
'contentment'.Part of the haul (66) consists of studies which merely count the number of
happy and unhappy persons in a country. The bulk ( 179) assesses also whether some
characteristics are more frequent among the former than among the latter. Most of these
latter studies make do with zero-order correlations, but quite a few have gone into
statistical specifications, though generally not very far. Eight longitudinal
investigations excepted, all the studies are synchronic ones.
The pile of reports thus gathered was almost impossible to handle, because of the
enormous variation in presentation, language and technical jargon. The reports were
therefore excerpted uniformly in terms of a carefully technical vocabulary. All reports
were considered by two independent excerpters. As far as possible the excerpts were
checked by the original investigators. In some cases they were supplemented by unpublished
information. Abridged versions of these excerpts are printed in a separate volume, named
'Databook of Happiness' which is published together with this book (See part II of the
Databook).
Then the separate findings were drawn from these excerpts and classified by subject matter.
More than four-thousand were involved. The presentation of these findings took 335 pages
of tables. These pages are also printed in the Databook; the correlational findings in
part III and the data on the distributions of happiness in various countries in part IV.
Thus a bookcase full of rather diverse and often chaotic reports was reduced to a
convenient one-inch-thick inventory. This process did not involve any loss of empirical
information. In its turn the inventory served as a basis for the following chapters of
this book.
Though very abundant at first sight, the findings cover a rather limited field While there
is much information about differences in happiness among American college students, we
know little about its variation in Third World countries. Many investigations focused on
the same variables, while many potentially promising factors were left unconsidered. The
usefulness of the findings is limited in another sense as well. The bulk consists of
simple synchronic zero-order correlations, which provide no definite answer to questions
of causality. Much of the common variance with happiness may be due to parallel response
bias and to spuriousness. Yet the data will appear to be sufficiently dependable to allow
the conclusive identification of at least a handful of conditions of happiness in
contemporary society and to make a clean sweep among current beliefs on the matter.