Birds of a Feather Flock Together: Correlation in BMI Categories in Long‐Term Relationships

• DOI: http://doi.org/10.1111/1467-8462.12181
• Author: Heather Brown
• Published date: 01 December 2016
• Date: 01 December 2016

Birds of a Feather Flock Together: Correlation in BMI

Categories in Long-Term Relationships

Heather Brown*

Abstract

This is the f‌irst article to explore partner

correlations across Body Mass Index (BMI)

categories for couples in long-term, stable

relationships using data from the Household,

Income and Labour Dynamics in Australia

Survey 2006–2011. The moderating effects of

three hypotheses on this correlation are tested:

(i) selection; (ii) relationship stability; and

(iii) social obligation or environment. The

results show gender-asymmetric effects in the

correlation in BMI categories. Selection

explains the majority of the correlation for

men and all of the correlation in BMI

categories for women. Some evidence of

variables related to stability and social

obligation are signif‌icantly associated with

weight outcomes for men.

1. Introduction

The leading causes of death in dev eloped

countries are non-communica ble diseases

such as cardiovascular dis ease and diabetes.

It is estimated that approximat ely 80 per cent

of deaths from coronary hea rt disease and

cerebrovascular disease ar e caused by behav-

ioural risk factors associated with ob esity

such as physical inactivity an d an unhealthy

diet (World Health Organiza tion 2011). These

same behavioural risk facto rs are potentially

inf‌luenced by personal rela tionships such as

being in a couple. Many activit ies, such as

eating or watching televi sion, are more

enjoyable with others than o n one’sown.

Additionally, partners are like ly to inf‌luence

individual perception rega rding social norms

associated with weight (B urke and Heiland

2007; Etil

e 2007; Oswald and Powdthavee

2007; Blanchf‌lower, Oswald and Van

Landeghem 2009). For examp le, if one

partner is obese, over time the othe r partner

may change their perception of an a cceptable

weight. Thus, a partner’s Body Mass Index

(BMI) category gains impor tance over time.

Using data on older adults, Wilson (2 012)

found that transition into marriage leads to

signif‌icant weight gain for both genders.

Additionally, after contro lling for selection

into marriage, Averett, Sikor a and Argys

(2008) found that, compared to unmarried or

cohabiting women, married w omen begin at a

lower BMI before marriage but then gain

weight after marriage occurs. Mar riage may

thus promote weight gain, but why? This

article will explore the route s through which

being in a long-term relationship may affect

BMI and if partner’sweightismoreimportant

at certain points along the BMI spect rum.

* Institute of Health and Society, Newcastle University,

Newcastle Upon Tyne NE2 4AA United Kingdom; email

<heather.brown@ncl.ac.uk>. The Household, Income

and Labour Dynamics in Australia (HILDA) Survey was

initiated, and is funded, by the Australian Government

through the Department of Social Services (DSS).

Responsibility for the design and management of the

survey rests with the Melbourne Institute of Applied

Economic and Social Research, The University of

Melbourne. Data collection for waves 9–12 is being

undertaken by Roy Morgan Research, a private market

research company, and The Nielsen Company collected

waves 1–8. More information about the HILDA Survey

can be found at <http://www.melbourneinstitute.com/

hilda/>. The f‌indings and views reported in this article,

however, are those of the author and should not be

attributed to either DSS or the Melbourne Institute.

The Australian Economic Review, vol. 49, no. 4, pp. 453–70

°

C2016 The University of Melbourne, Melbourne Institute of Applied Economic and Social Research

Published by John Wiley & Sons Australia, Ltd

This article adds to the literature by

empirically testing the factors associated with

correlated BMI for couples in stable partner-

ships, which is def‌ined as being in a relation-

ship

1

with the same person for approximately

3 years prior to 2006 and not separating

from this partner over the sample period

(2006–2011). It is widely accepted that

partners tend to exhibit similar health-related

behaviours. For example, there has been a

number of studies f‌inding shared outcomes

in partner smoking (Clark and Etil

e 2006;

Christakis and Fowler 2008; Cutler and

Glaeser 2008), alcohol consumption (Leonard

and Mudar 2003; Rosenquist et al. 2010) and

obesity (Christakis and Fowler 2007; Kano

2008; Clark and Etil

e 2011; Brown, Hole and

Roberts 2014). Additionally, there has been a

number of studies (Jeffery and Rick 2002;

Sobal, Rauschenbach and Frongillo 2003; Jo

2004; Averett, Sikora and Argys 2008; The

and Gordon-Larsen 2009; Wilson 2012)

that have looked at how marriage transition

affects the likelihood of being obese. The

evidence is mixed and varies by gender. Sobal,

Rauschenbach and Frongillo (2003) found that

only women gain weight after marriage, whereas

Jo (2004) found the opposite, where men are

more likely to become obese after marriage, but

not women. In terms of the impact of cohabiting

relationship on body weight, The and Gordon-

Larsen (2009) found, using a logistic model, that

the likelihood of being obese and partner

correlations in obesity are highest for partners

that have lived together for 2 or more years (bo th

married and cohabiting). Understanding why

individuals in stable relationships tend to gain

weight is important for the development of

policy to reduce the burden of obesity on the

health care system.

Many of the studies mentioned in the

paragraph above focused on specif‌ic segments

of the population: young people aged under

35 years (Jo 2004; Averett, Sikora and Argys

2008; The and Gordon-Larsen 2009), a work-

site health promotion project in Minneapolis-St

Paul, USA (Jeffery and Rick 2002) or

older adults (Wilson 2012). In this article,

data from the Household, Income and Labour

Dynamics in Australia (HILDA) Survey, a

general population survey, covering the period

2006–2011, are used. Looking at the whole

adult population in stable unions will give a

better overall picture on how marriage may

lead to weight gain in all adults.

The empirical analysis uses a random-

effects generalised ordered probit model,

estimated separately for men and women, to

explore the correlation in partner BMI across

healthy-weight, overweight and obese individ-

uals. Further analysis investigates the moder-

ating effect of matching, relationship stability

and social obligation on this correlation. First,

one may choose a partner whose preferences

increase the likelihood of them having a

correlated BMI category. Second, individuals

who are not planning on trying to f‌ind a new

partner may f‌ind it too costly to maintain a

healthy BMI and may choose to match their

BMI to their partner. Finally, social obligations

associated with stable relationships, such as

eating more and regular meals, may contribute

to the correlation in partner BMI. These

hypotheses will be expanded upon in the next

section.

2. Theoretical Models of Stable

Relationships and Body Weight

2.1 Selection

The proportion of people who never enter a co-

residential relationship is very small. In 2001,

only 6.1 per cent of Australian men and 4.2 per

cent of Australian women over 60 years had

never married (de Vaus 2004). Thus, weight is

not likely to affect the likelihood of being in

a relationship but may affect mate quality

and age at f‌irst co-residential relationship

(Averett, Sikora and Argys 2008). Oreff‌ice and

Quintana-Domeque (2010) examine the role of

anthropometric characteristics when choosing

a partner. They f‌ind strong evidence that

individuals with a similar BMI are more likely

to enter into a relationship than individuals

with a widely dissimilar BMI. This is consis-

tent with Becker’s (1991) marriage model in

which the gains from marriage may be

maximised by choosing a partner with similar

characteristics, such as weight, or observable

454 The Australian Economic Review December 2016

°

C2016 The University of Melbourne, Melbourne Institute of Applied Economic and Social Research