A first step in this review was to gather the available quantitative research findings on the relationship between happiness and healthy eating. The second step was to present these findings in an uncomplicated form.
Gathering of Research Findings
In order to identify relevant papers for this synthesis, I inspected which publications on the subject of healthy eating were already included of the Bibliography of World Database of Happiness, in the subject sections ‘Health behaviour’ and consumption of ‘Food’. Then to further complete the collection of studies, various databases were searched such as Google Scholar, EBSCO, ScienceDirect, PsycINFO, PubMed/Medline, using terms such as ‘happiness’, ‘life satisfaction’, ‘subjective well-being’, ‘well-being’, ‘daily affect’, ‘positive affect’, ‘negative affect’ in connection with terms such as ‘food’, ‘healthy food’, ‘fruit and vegetables’, ‘fast food ‘and ‘soft drinks’ in different sequences.
All reviewed studies had to meet the following criteria:
A report on the study should be available in English, French, German or Spanish.
The study should concern happiness in the sense of life-satisfaction (cf. Healthy Eating section). I excluded studies on related matters, such as on mental health or wider notions of ‘flourishing’.
The study should involve a valid measure of happiness (cf. Happiness section). I excluded scales that involved questions on different matters, such as the much-used Satisfaction With Life Scale (Diener et al. 1985).
The study results had to be expressed using some type of quantitative analysis.
Together, I found 20 reports of an empirical investigation that had examined the relationship between healthy eating and happiness, of which two were working papers and one dissertation. None of these publications reported more than one study. Together, the studies yielded 47 findings.
All the papers were fairly recent, having been published between 2005 and 2017. Most of the papers (44.4%) were published in Medical Journals, including the International Journal of Behavioural Medicine, Journal of Health Psychology, The Journal of Nutrition, Health & Aging, The Journal of Nutrition, Health & Aging, The Journal of Psychosomatic Research, The International Journal of Public Health, and Social Psychiatry & Psychiatric Epidemiology.
Together, the studies covered 149.880 respondents and 27 different countries. The publics investigated in these studies, included the general population in countries and particular groups such as students, children, veterans and medical patients. The majority of respondents belonged to a general public group (50%), students made up 27.8%, with children and veterans each forming 11.1%.
Research Methods Used
Most of the studies were cross-sectional 64.4%, longitudinal and daily food diaries accounted for 22% and 10.2% of the total number of studies respectively, and one experimental study accounted for 3.4%.
I present an overview of all the included studies, including information about population, methods and publication in Table 1.
Format of this Research Synthesis
As announced, I applied a new technique of research reviewing, taking advantage of two technical innovations: a) The availability of an on-line findings-archive (the World Database of Happiness) that holds descriptions of research findings in a standard format and terminology, presented on separate finding pages with a unique internet address. b) The change in academic publishing from print on paper to electronic text read on screen, in which links to that online information can be inserted.
Links to Online Detail
In this review, I summarize the observed statistical relationships as +, − or 0 signs.Footnote 2 These signs link to finding pages in the World Database of Happiness, which serves as an online appendix in this article. If you click on a sign, one such a finding page will open, on which you can see full details of the observed relationship; of the people investigated, sampling, the measurement of both variables and the statistical analysis. An example of such an electronic finding page is presented in Fig. 2. This technique allows me to present the main trends in the findings, without burdening the reader with all the details, while keeping the paper to a controllable size, at the same time allowing the reader to check in depth any detail they wish.
Organization of the Findings
I first sorted the findings by the research method used and these are presented in three separate tables. I distinguished a) cross-sectional studies, assessing same-time relationships between diet and happiness (Table 2), b) longitudinal studies, assessing change in happiness following changes in diet (Table 3), and c) experimental studies, assessing the effect of induced changes in diet on happiness (Table 4).
In the tables, I distinguish between studies at the micro level, in which the relation between diet and happiness of individuals was assessed and studies at the macro level, in which average diet in nations is linked to average happiness of citizens.
I present kinds of foods consumed vertically and horizontally two kinds of happiness: overall happiness (life-satisfaction) and hedonic level of affect.
Presentation of the Findings
The observed quantitative relationships between diet and happiness are summarized using 3 possible signs: + for a positive relationship, − for a negative relationship and 0 for a non-relationship. Statistical significance is indicated by printing the sign in bold. See Appendix. Each sign contains a link to a particular finding page in the World Database of Happiness, where you can find more detail on the checked finding.
Some of these findings appear in more than one cell of the tables. This is the case for pages on which a ‘raw’ (zero-order) correlation is reported next to a ‘partial’ correlation in which the effect of the control variables is removed. Likewise, you will find links to the same findings page at the micro level and the macro level in Table 2; on this page there is a time-graph of sequential studies in Russia from which both micro and macro findings can be read.
Several cells in the tables remain empty and denote blanks in our knowledge.
Advantages and Disadvantages of this Review Technique
There are pros and cons to the use of a findings-archive such as the World Database of Happiness and plusses and minuses to the use of links to an on-line source in a text like this one.
Use of a Findings-Archive
Advantages are: a) efficient gathering of research on a particular topic, happiness in this case, b) sharp conceptual focus and selection of studies on that basis, c) uniform description of research findings on electronic finding pages, using a standard format and a technical terminology, d) storage of these finding pages in a well searchable database, e) which is available on-line and f) to which links can be made from texts. The technique is particular useful for ongoing harvesting of research findings on a particular subject.
Disadvantages are: a) the sharp conceptual focus cannot easily be changed, b) considerable investment is required to develop explicit criteria for inclusion, definition of technical terms and software,Footnote 3 c) which pays only when a lot of research is processed on a continuous basis.
Use of Links in a Review Paper
The use of links to an on-line source allows us to provide extremely short summaries of research findings, in this text by using +, − and 0 signs in bold or not, while allowing the reader access to the full details of the research. This technique was used in an earlier research synthesis on wealth and happiness (Jantsch and Veenhoven 2019) and is described in more detail in Veenhoven (2019). Advantages of such representation are: a) an easy overview of the main trend in the findings, in this case many + signs for healthy foods, b) access to the full details behind the links, c) an easy overview of the white spots in the empty cells in the tables, and d) easy updates, by entering new sign in the tables, possibly marked with a colour.
The disadvantages are: a) much of the detailed information is not directly visible in the + and – signs, b) in particular not the effect size and control variables used, and c) the links work only for electronic texts.
Differences with Traditional Reviewing
Usual review articles cannot report much detail about the studies considered and rely heavily on references to the research reports read by the reviewer, which typically figure on a long list at the end of the review paper that the reader can hardly check. As a result, such reviews are vulnerable to interpretations made by the reviewer and methodological variation can escape the eye.
Another difference is that the conceptual focus of many traditional reviews in this field is often loose, covering fuzzy notions of ‘well-being’ rather than a well-defined concept of ‘happiness’ as used here. This blurs the view on what the data tell and involves a risk of ‘cherry picking’ by reviewers. A related difference is that traditional reviews of happiness research often assume that the name of a questionnaire corresponds with its conceptual contents. Yet, several ‘happiness scales’ measure different things than happiness as defined in “Healthy Eating” section, e.g. much used Life Satisfaction Scale (Neugarten et al. 1961), which measures social functioning.
Still another difference is that traditional narrative reviews focus on interpretations advanced by authors of research reports, while in this quantitative research synthesis I focus on the data actually presented. An example of such a difference in this review, is the publication by Connor & Brookie (Conner et al. 2015) who report no effect of healthier eating on mood in the experimental group, while their data show a small but significant gain in positive affect and a small but insignificant reduction of negative effect (Table 3), which together denote a positive effect on affect balance.
Difference with Traditional Meta-Analysis
Though this research synthesis is a kind of meta-analysis, it differs from common meta-analytic studies in several ways. One difference is the above- mentioned conceptual rigor; like narrative reviews many meta-analyses take the names given to variables for their content thus adding apples and oranges. Another difference is the direct online access to full detail about the research findings considered, presented in a standard format and terminology, while traditional meta-analytic studies just provide a reference to research reports from which the data were taken. A last difference is that most traditional meta-analytic studies aim at summarizing the research findings in numbers, such as an average effect size. Such quantification is not well possible for the data at hand here and not required for answering our research questions. My presentation of the separate findings in tabular schemers provides more information, both of the general tendency and of the details.