This is the author’s version of a work that was submitted/accepted for pub-lication in the following source:

McMeekin, Sascha, Jansen, Elena, Mallan, Kimberley M., Nicholson, Jan,Magarey, Anthea, & Daniels, Lynne(2013)Associations between infant temperament and early feeding practices. Across-sectional study of Australian mother-infant dyads from the NOURISHrandomised controlled trial.Appetite, 60(1), pp. 239-245.

This file was downloaded from: https://eprints.qut.edu.au/54016/

c© Copyright 2012 Elsevier Ltd.

This is the author’s version of a work that was accepted for publication in Appetite.Changes resulting from the publishing process, such as peer review, editing, corrections,structural formatting, and other quality control mechanisms may not be reflected in thisdocument. Changes may have been made to this work since it was submitted for publica-tion. A definitive version was subsequently published in Appetite, Volume 60, 1 January2013. DOI: 10.1016/j.appet.2012.10.005

License: Creative Commons: Attribution-Noncommercial-No DerivativeWorks 4.0

Notice: Changes introduced as a result of publishing processes such ascopy-editing and formatting may not be reflected in this document. For adefinitive version of this work, please refer to the published source:

https://doi.org/10.1016/j.appet.2012.10.005

1 

Associations between infant temperament and early feeding practices: a cross-sectional

study of Australian mother-infant dyads from the NOURISH randomised controlled

trial

Authors: Sascha McMeekin a, Elena Jansen a, Kimberley Mallan a, Jan Nicholson b,c,

Anthea Magarey a,d, Lynne Daniels a,d.

a Institute of Health and Biomedical Innovation, School of Exercise and Nutrition

Sciences, Queensland University of Technology, 60 Musk Avenue, Kelvin Grove,

Queensland 4059, Australia.

b Parenting Research Centre, 232 Victoria Parade, East Melbourne, Victoria 3002,

Australia.

c Centre for Learning Innovation, Queensland University of Technology, Brisbane 4059,

Australia.

d Nutrition and Dietetics, Flinders University, Adelaide, South Australia 5001,

Australia.

2 

Abstract

The purpose of this study was to investigate the association between temperament in

Australian infants aged 2–7 months and feeding practices of their first-time mothers

(n=698). Associations between feeding practices and beliefs (Infant Feeding

Questionnaire) and infant temperament (easy-difficult continuous scale from the Short

Temperament Scale for Infants) were tested using linear and binary logistic regression

models adjusted for a comprehensive range of covariates. Mothers of infants with a

more difficult temperament reported a lower awareness of infant cues, were more likely

to use food to calm and reported high concern about overweight and underweight. The

covariate maternal depression score largely mirrored these associations. Infant

temperament may be an important variable to consider in future research on the

prevention of childhood obesity. In practice, mothers of temperamentally difficult

infants may need targeted feeding advice to minimise the adoption of undesirable

feeding practices.

Key words: feeding practices, feeding beliefs, infant temperament, obesity, feeding

relationship.

3 

Introduction

In 2010, it was estimated that nearly 12 % of children from birth to five years in

developed countries were overweight or obese (de Onis, Blössner, & Borghi, 2010).

These children are at a greater risk of becoming overweight adults (Margarey, Daniels,

Boulton, & Cockington, 2003; Power & Parsons, 2000), with associated increased

morbidity and mortality risk (Reilly et al., 2003). Direct costs of obesity are estimated

to be around 9 % of the total health care costs in the USA, between 1 % and 5 % in

Europe and 2 % in Australia (Minister for Health and Ageing, 2007).

Early maternal feeding practices and beliefs, hereafter referred to as ‘feeding practices’,

have been associated with child weight status (Ventura & Birch, 2008). The feeding

relationship is implicitly bi-directional in that both mother and infant contribute to the

exchange (DiSantis, Hodges, Johnson, & Fisher, 2011). Infants are required to

contribute clear hunger and satiety cues, and the caregiver must accurately interpret and

respond to those cues in order for responsive feeding to occur. If feeding is discordant

to the child’s hunger and satiety the child’s innate self-regulatory capacity may be

hindered (Birch & Deysher, 1986; Kral et al., 2007), and over-feeding, accelerated

weight gain and increased risk of future overweight may result (DiSantis, et al., 2011).

The infant’s role in the feeding relationship, that is clearly expressing hunger and

satiety, may be influenced by their temperament.

Infant temperament is a set of traits that characterises individuals’ behaviour (Goldsmith

et al., 1987; Maccoby & Martin, 1983). These traits have a genetic component that is

4 

influenced by a complex interplay with environmental factors (Li Fein, & Grummer-

Strawn, 2008; Worobey, Islas Lopez & Hoffman, 2009). Temperament has been linked

with weight gain during infancy (Carey, 1985; Darlington & Wright, 2006; Niegel,

Ystrom, & Vollrath, 2007) and later weight status (Agras, Hammer, McNicholas, &

Kraemer, 2004; Pulkki-Raback, Elovainio, Kivimaki, Raitakari, & Keltikangas-

Jarvinen, 2005). The mechanisms by which temperament influences weight are yet to

be clearly elucidated. One possible pathway is via the influence infant temperament

may have on the development of feeding practices adopted by mothers, however very

few studies have explored this relationship.

Evidence supports the notion that negative, or ‘difficult’, temperament dimensions are

associated with both undesirable feeding practices, and feeding difficulties during

infancy. Of note, all studies reviewed here used maternal-report measures of infant

temperament and feeding practices. In a cross-sectional study of six-month-old infants

(n=99) from the United Kingdom (Farrow & Blissett, 2006a), unadaptable and fussy-

difficult infant temperament were associated with maternal reports of feeding

difficulties. A longitudinal follow-up of the above study (n = 62) concluded that infant

difficultness reported by mothers at six months predicted restrictive feeding practices at

two years of age (Blissett & Farrow, 2007). In a US study of older children (n = 55, six

to 12 years) a relationship between infant temperament and maternal feeding practices

was also found, via the association found between the ‘difficult’ temperament

dimension low adaptability, (Horn, Galloway, Webb, & Gagnon, 2011) and the feeding

practice using pressure to eat. These findings, though limited, do appear to support that

5 

infant temperament may be a contributing factor in the development of maternal feeding

practices.

A number of other infant and maternal characteristics may also play a role in the

feeding relationship. Infant gender has been shown to modulate associations between

maternal controlling feeding practices and children’s food intake, which was

significantly associated in female children and not male children (Fisher & Birch,

1999). Child’s weight is another important covariate: variations in maternal feeding

practices (in particular controlling feeding and pressure to eat) have been associated

with child weight and weight gain during infancy (Farrow & Blissett, 2006b; Webber,

Hill, Saxton, Van Jaarsveld, & Wardle, 2009). Mothers’ own weight status has been

implicated in the initiation and duration of breastfeeding, and may therefore influence

feeding practices as well (Donath & Amir, 2000; Li, Jewell, & Grummer-Strawn, 2003).

In addition, mothers with a university education, who breastfeed and are older have

been shown to use more ‘positive child feeding behaviours’ (Hendricks, Briefel, Novak,

& Ziegler, 2006). Maternal mood may also play a role in the feeding practices reported

by mothers, as it has been shown that mothers with depression reported less responsive

feeding practices (Hurley, Black, Papas, & Caufield, 2008). It is therefore appropriate

to consider (i.e., “statistically adjust for”) these infant and maternal characteristics as

covariates in the study of feeding practices.

Although there is emerging evidence to support the proposal that infant temperament

could be associated with maternal feeding practices, the limited number of studies

investigating this relationship typically have varied methods to define both temperament

6 

and feeding practices, small samples, and control for few if any covariates. The aim of

this study was to investigate whether infant temperament is associated with maternal

feeding practices and beliefs in a large sample of Australian first-time mothers of

infants aged 2-7 months, adjusting for a comprehensive range of potential covariates.

Methods

Participants

This cross-sectional study reports a secondary analysis of baseline data from the

NOURISH randomised controlled trial (Australian and New Zealand Clinical Trials

Registry Number [ACTRN] 12608000056392), a primary prevention intervention for

childhood obesity, commencing when infants were aged 2–7 months. The trial protocol

and recruitment have been described in detail elsewhere (Daniels et al., 2009). A total

of 698 first-time mothers and their infants, from Brisbane and Adelaide in Australia,

consented to participate in the NOURISH trial. First-time mothers were included in the

NOURISH study as it was expected they would have less experience in infant feeding

and as a result the intervention would develop rather than modify their feeding

practices. Eligible participants for the study were ≥ 18 years old who delivered a

healthy term infant (birth weight > 2500 g, and ≥ 37 weeks gestation) and were willing

and able to attend intervention sessions. Mothers were excluded if they did not have

facility with English or had a documented history of domestic violence, intravenous

substance abuse, mental health problems or self-reported eating disorders.

7 

Procedures

A two-stage recruitment process was conducted. Stage 1 involved approaching a

consecutive sample of eligible (as above) first-time mothers in hospital within 72 hours

of delivery to obtain consent to be contacted again. Participants agreeing to later

contact at Stage 1 provided basic demographic data, as did approximately half those

who declined to be re-contacted. Three to four months later, at Stage 2, mothers were

re-contacted by mail to request consent for study participation. Prior to allocation when

infants were 2–7 months old, consenting mothers were mailed a questionnaire which

was returned during an appointment to gather basic anthropometric data at a local child

health clinic. Ethical approval was obtained from relevant university and hospital

institutions (Queensland University of Technology HREC 00171 Protocol 0700000752)

(Daniels, et al., 2009).

Measures

Infant temperament. Infant temperament was assessed using a maternal self-report

measure entitled the “Short Temperament Scale for Infants” (STSI) (Sanson, Prior,

Garino, Oberklaid, & Sewell, 1987). The STSI is based on the temperament research

conducted in the New York Longitudinal Study (Thomas, Chess, Birch, Hertzig, &

Korn, 1963) and measurement tools developed by Carey and colleagues (1978). The

STSI has been validated for use in Australian samples of infants aged 4–8 months (n =

2443) (Sanson, et al., 1987). The questionnaire has 12 items assessing three factors; (i)

approach, the infant's level of comfort with new situations or people (4 items, e.g. “My

baby’s first reaction (at home) to approach from strangers is acceptance”); (ii)

cooperation, the adaptability of the infant to new activities (4 items, e.g. “My baby

8 

stays still during procedures like hair brushing and nail cutting”); and (iii) irritability,

the infant's agitation and difficulty to soothe (4 items, e.g. “My baby continues to cry in

spite of several minutes of soothing”). All items were measured on a 6–point scale

from “almost never” (1) to “almost always” (6) and high scores related to greater

approach, cooperation and irritability. The 12-item STSI was administered to mothers

when their infants were aged 2–7 months. An overall continuous easy-difficult

temperament scale ranging from 1–6 was calculated from the three factors by first

recoding individual items such that high scores were always indicative of

temperamental difficulty (i.e. low cooperation, low approach and high irritability), and

generating a mean score across all items (Sanson, et al., 1987). The internal reliability

estimate for the easy-difficult temperament scale (higher scores indicate more difficult

temperament) was acceptable: Cronbach’s α = 0.74.

Feeding practices and beliefs. The “Infant Feeding Questionnaire” (IFQ) (Baughcum et

al., 2001) was self-administered by consenting mothers to assess current self-reported

maternal feeding practices. Original validation of the 20–item questionnaire as a

retrospective measure of feeding practices during the first year of life revealed a seven

factor solution (Baughcum, et al., 2001). The questionnaire required alterations to make

it a suitable measure of feeding practices in an Australian sample with anticipated high

rates of breastfeeding. The tense was changed from past to present, Australian wording

adapted (e.g. “being unsettled” instead of “fussiness”), and a “not applicable” response

category was added for three items that assumed formula feeding was occurring (e.g.

adding cereal to the bottle). These three items, and hence two of the original seven

scales, were subsequently excluded from the questionnaire as the majority of the sample

9 

responded with “not applicable”. The remaining five factors and the reliability

estimates in our sample were: awareness of infant hunger and satiety cues (4 items, e.g.

“I know when my baby is hungry”, Cronbach’s α = 0.72), using food to calm infant’s

fussiness (2 items, e.g. “When your baby gets upset, is feeding him/her to first thing you

do?”, Cronbach’s α = 0.66), concern about infant over-eating and becoming overweight

(3 items, e.g. “Do you worry that your baby is feeding too much?”, Cronbach’s α =

0.66), concern about infant under-eating or becoming underweight (4 items, e.g. “I am

worried that my baby will become underweight”, Cronbach’s α = 0.76), and feeding on

demand vs. schedule (2 items, e.g. “Do you let your baby feed whenever s/he wants

to?”, Cronbach’s α = 0.83) 1. Items were measured on a 5-point Likert scale from

“never” (0) to “always” (4). Internal consistency on all five scales in our sample was

higher than reported in the original validation study (Baughcum, et al., 2001).

Covariates. Data collected at first contact with participants in hospital included infant

gender, infant birthweight (from hospital records), maternal age, and maternal education

level (dichotomised into university education vs. no university education). Scores on

the Edinburgh Post Natal Depression Scale (EPDS) were calculated as the sum of

responses to the 10 items (Cox, Holden, & Sagovsky, 1987). Each item was scored

from 0–3 giving a maximum total score of 30. For descriptive purposes only, mothers

scoring between 10 and 13 were classified as “borderline”, and those scoring above 13

were labelled “depressive” based on approximate thresholds advised in the validation

study (Cox, et al., 1987). Breastfeeding status was categorised into three levels: 1)

exclusive breastfeeding, breastmilk only, 2) any breastfeeding, including combination

feeding with formula and fully breastfeeding with or without some complementary

10 

foods, 3) no breastfeeding, formula feeding (no breastmilk) with or without some

complementary foods. These data were dummy coded into two variables for the

analyses. Trained study staff collected anthropometric measures of both mother and

child using standard protocols to the nearest 0.01 kg and 0.01 m for weights and heights

respectively. Three infant nude weight and recumbent length measures were averaged.

Infant weight gain (g/wk) was calculated by subtracting the infant’s birthweight from

their weight at 2–7 months and dividing by the infant’s exact age in weeks. Z-scores for

weight for age and birthweight were calculated using the software program WHO

Anthro version 3.0.1 and macros (World Health Organisation, 2006). Raw change in

weight for age Z-score from birth to baseline was calculated, for descriptive purposes,

by subtracting birthweight Z-score from weight-for-age Z-score at 2–7 months.

Maternal weight and height were measured once without shoes. Maternal body mass

index (BMI) was calculated from these height and weight measures (kg/m2) (World

Health Organisation, 1995). Maternal perception of infant weight status, measured with

the item “Do you think your baby is?”, was considered for inclusion as a covariate;

however, as the response category “normal weight” was overrepresented and just small

percentages of mothers classed their infants as “underweight” or “overweight” this

variable was only included to describe the sample (see Table 1).

Analytic strategy

Statistical analysis was conducted using SPSS 18.0.3. Tests for normality were first

conducted for the easy-difficult temperament scale and the potential covariates,

revealing that just the EPDS score had a non-normal distribution. The square root

transformation of the EPDS data was therefore used in all further analyses to improve

11 

the sensitivity of the measure. Using food to calm and awareness of cues were found to

be normally distributed, whereas concern about over- and underweight were positively

skewed and feeding on demand was negatively skewed. Non-parametric bivariate

analyses were therefore used to explore the association between potential covariates

with each of the feeding practices. Spearman’s correlations were used for continuous

variables, and Independent samples Mann-Whitney U tests for categorical variables

(Field, 2009). Statistical significance was set at P < 0.05. The covariates found to have

a significant association with at least one of the feeding practices (outcome variables)

were included in the regression analyses that followed.

Square root, log and reciprocal transformations were performed on the three non-

normally distributed IFQ factors, which failed to improve the distribution in any

instance. These three IFQ factors were then split into roughly equal tertiles, from which

the ‘outlier’ tertile (upper tertile for concern about over- and underweight, lower tertile

for feeding on demand vs. schedule) was separated from the other two tertiles to create

binary variables that were entered as the outcome variable into hierarchical binary

logistic regression models. This method is in accordance with that used by Burdette et

al. (2006). Using food to calm and awareness of cues were left as continuous scores

and entered into hierarchical linear regression models. Step 1 of each of the 2-step

linear and binary logistic regression models included infant gender, maternal age at

delivery, maternal education, maternal depression score (EPDS), breastfeeding status,

infant weight gain (g/wk) and maternal BMI. The easy-difficult temperament scale

(independent variable) was incorporated into each model in Step 2. Two multivariate

outliers were detected based on Mahalanobis distance values > 26.12, P < 0.001.

12 

Exclusion of these two participants did not affect the results and therefore the full

dataset was used in the analyses. Calculation of Cross’s distance values showed there

were no influential data points in the sample. Post-hoc univariate analysis of variance

were run to test the moderating effects of infant gender and feeding mode combined

with infant temperament on maternal feeding practices with the covariates included.

Results

The characteristics of the sample are shown in Table 1. Forty-four percent of the

sample consented to study participation at Stage 2 of recruitment, excluding those who

could not be contacted or became ineligible. Differences between consenters and non-

consenters at Stage 2 were established from the data collected at Stage 1. Compared

with mothers who either declined consent or could not be recontacted (at Stage 2),

mothers who consented to participate were older (M = 30.1 ± 5.3 vs. M = 27.4 ± 5.6; P

< 0.001), more likely to have completed a university degree (58 % vs. 33 %; OR = 2. 9;

CI95% = 2.4 to 3.5; P < 0.001), and were more likely to report that they intended to

breastfeed their baby exclusively (88 % vs. 75 %; OR = 1.8, CI95% = 1.3 to 2.5; P <

0.001).

The hierarchical linear and binary logistic regression analyses, shown in Table 2 and 3,

identified that the easy-difficult temperament scale added significantly to the variance in

four of the five feeding practices with the full models explaining 12-19 % of the total

variance in these feeding practices (Adjusted R2 for linear regression, Nagelkerke R2 for

binary logistic regression). Inclusion of the easy-difficult temperament variable elicited

13 

a change in variance of in the awareness of cues model (R2 = 0.09, P < 0.001),

indicating mothers of infants with a more difficult temperament reported less awareness

of infant cues. Mothers of infants with a more difficult temperament were also more

likely to use food to calm (R2 = 0.02, P = 0.001). High concern regarding both

overweight (OR = 1.75, CI95% = 1.23 to 2.49, P = 0.002) and underweight (OR = 2.00,

CI95% = 1.41 to 2.84, P < 0.001) was more likely in mothers of infants with a more

difficult temperament. Feeding on demand vs. schedule was not associated with

temperament. The covariate maternal depression was independently associated with all

feeding practices. No significant interaction effects were found with infant gender (P =

0.430–0.815) and breastfeeding status (P = 0.197–0.615) when coupled with infant

temperament.

Discussion

To our knowledge the current study is one of the largest to investigate the association

between infant temperament and maternal feeding practices and beliefs. The results

showed that infant temperament was associated with maternal feeding practices and as

such is likely to play a role in the early feeding relationship. Mothers of infants scoring

high on “difficult” temperament reported less awareness of infant hunger cues, more use

of food to calm, and high concern about over- and underweight.

The association of ‘difficult’ temperament (low approach, low cooperation, high

irritability) (Sanson, et al., 1987) and lower maternal awareness of infant hunger and

satiety cues is potentially a manifestation of both maternal and infant characteristics and

14 

behaviours. By definition, infants with a difficult temperament react negatively to new

situations, unfamiliar people and novel activities. Mothers of temperamentally difficult

infants may therefore have trouble distinguishing cues that specifically signal hunger

and satiety from the distress cues given in response to the broad range of stimuli

mentioned above. As a consequence of the lack of clarity around an infant's signals of

hunger and satiety and mother's recognition and interpretation of these cues, discordant

feeding is more likely in infants with a difficult temperament and may contribute to the

erosion of the child’s innate ability to regulate intake (DiSantis, et al., 2011).

Alternatively, maternal characteristics may influence the way the mother perceives her

child’s behaviour as well as her capacity to appropriately interpret and respond to cues.

For example, a mother with depression may both perceive her infant as more difficult

and be less responsive to infant cues resulting in nonresponsive feeding styles (Ganley,

1989). Our finding that maternal depression independently influences awareness of

cues lends support to this notion.

Higher temperamental difficulty was associated with greater use of food to calm, a

feeding practice that, at least in older children, has been postulated to compromise self

regulation of intake by providing incidental learning for the child to eat for reasons

other than hunger (DiSantis, et al., 2011). Two studies report links between

temperament and specific feeding practices that the authors argue are a practical

manifestation of using food to calm. The first study, conducted in the USA with “low-

income black mothers” and their three-month-old infants, examined associations

between infant temperament and complementary feeding before four months of age.

This study found that early introduction of solids was associated with high “distress-to-

15 

limitations” in infants (e.g. “When placed on his/her back, how often did the infant fuss

or protest?”), supporting the authors’ hypothesis that “a fussy infant temperament may

lead parents to use food as a soothing technique” (Wasser, et al., 2011). In a study of

older infants mothers of children with “distress-prone temperaments” (n = 40266, 18-

month-olds), both externalising (e.g. “the child gets upset easily”) and internalising (e.g.

“too fearful and anxious”), were more likely to feed sweet foods and drinks. The

authors of this study proposed a similar theory to Wasser and colleagues (2011) to

explain their findings. They suggested that mothers respond to “difficult” temperament

traits by attempting to calm and soothe with sweets foods or drinks, as sweets foods

“can have a calming effect as a result of...neurological mechanisms” (Vollrath, Tonstad,

Rothbart, & Hampson, 2010). Our results provide empirical support for the theories

developed to explain these feeding responses to difficult infant temperament.

Overall, the majority of mothers in the NOURISH sample perceived their infant as

having a normal weight status, consistent with the mean raw weight-for-age Z-score

change from birth to baseline of –0.43, which represents healthy weight gain as infants

are closely tracking growth percentiles (Ong, et al., 2006). Also consistent with these

data is the relatively low concern regarding intake and weight status, on the two relevant

IFQ scales (Baughcum, et al., 2001), as indicated by means of 1 from a 5-point Likert

scale where 5 indicates high concern. However, follow up data (not shown) indicates

that when infants reached 14 months of age the mean factor scores for the concern about

over- and underweight had risen to 1.5 and 2.0 respectively. Perceptions and concern

are likely to be important drivers of feeding practices (Farrow & Blissett, 2006b; Teng,

Bartle, Sadeh, & Mindell, 2012). For example, mothers may be more likely to

16 

encourage their child to feed more frequently, finish the bottle or eat more (i.e. pressure

their child to eat) if they think their child is, or might become, underweight. Such

discordant feeding practices are linked theoretically with reduced self regulation of

intake and increased weight status in older children (Birch, 2006; DiSantis, et al., 2011).

In our study high concern about the child's weight (both over- and underweight) was

associated with higher temperamental difficulty in infants. This concern may arise from

the feeding difficulties commonly reported by mothers of infants high in temperamental

difficulty. As mentioned previously, Farrow and Blissett (2006a) found that feeding

difficulties were associated with inadaptable and ‘difficult’ infant temperament.

Similarly, Galler et al. (2004) showed that Barbadian mothers (n = 226) who rated their

infant as having a difficult temperament were also more likely to rate them as difficult

to feed. It is not surprising that infants that are identified as “difficult” generally are also

perceived as “difficult” feeders. The association between difficult infant temperament

and greater maternal concern for their child’s intake and weight may be mediated by

perceived feeding difficulty, though temporal relationships between these variables

cannot be established from the data available.

Feeding on demand vs. schedule was not associated with infant temperament. Feeding

on a schedule or on demand is likely to be influenced by factors such as mother’s

working commitments, socio-cultural norms (Dettwyler & Fishman, 1992) and advice

of health professionals and significant others. Mothers who return to work soon after

the birth of their child would have a reduced capacity to breastfeed, and more

specifically would have difficulty maintaining a demand feeding regime (Gjerdingen &

Debra, 1991; Van Esterik, 2002).

17 

Although not a primary aim of the study, our data also provide information on the

independent associations between a range of covariates and feeding practices in a large

sample. Of most interest, mothers with higher scores on the EPDS (Cox, et al., 1987)

were more likely to use food to calm, were less aware of infant cues, more concerned

for their child’s intake and weight and more likely to feed on demand, largely mirroring

the associations between infant temperament and feeding practices. Similar to the

associations of difficult infant temperament, maternal depression has been found to be

associated with reported feeding difficulty and potentially undesirable feeding practices

in other studies. Farrow & Blissett (2006a) found that prenatal and postnatal maternal

cognitions (e.g. higher levels of prenatal emotional deprivation entitlement core beliefs,

and postnatal lower maternal self-esteem) were implicated in the development of

feeding difficulty in six-month-old infants. In 18-month-olds it was found that mother’s

scoring high in negative affectivity (anxiety and depression) during pregnancy were

more likely to feed their child caloric drinks at night (Vollrath, et al., 2010). Our

findings that, in the same sample, maternal depression and difficult infant temperament

show a similar pattern of association with maternal feeding practices is inherently

plausible. However, an alternative factor may be that mothers with depression are more

likely to perceive their infant’s temperament as more difficult. In a sample of 139

women from the USA, depressed mothers reported higher difficult temperament scores

when infants were both two and six months-of-age than their non-depressed

counterparts (McGrath, Records, & Rice, 2008). The previously mentioned study

conducted in Barbados found maternal despair and anxiety at seven weeks and six

months were both associated with reported ‘difficult’ temperament traits in infants

18 

(decreased adaptability, reduced approach, negative mood and increased sensory

threshold) at six months of age (Galler, et al., 2004). Mothers who rated their infant as

having a difficult temperament were also more likely to rate them as difficult to feed

(Galler, et al., 2004). Unlike our study, this study did not examine the relationship

between maternal mood and feeding practices. Further studies are needed with analyses

designed to specially evaluate the complex, multi-directional relationships between

maternal depression (both pre- and post partum), infant temperament ratings and

maternal feeding practices to better understand the relationship between these variables.

Strengths of this study include a large sample of very young infants which has enabled

contemporaneous assessment of early developmental stages in feeding and related

feeding practices without the limitations of recall bias. Temperament was assessed

using a tool that was validated in a representative Australian sample and the concurrent

use of the modified IFQ performed well in this sample of first-time mothers, compared

with its reliability as a retrospective measure. Also, the analyses adjusted for a

comprehensive range of maternal and infant covariates.

There are also limitations that should be considered. The generalisability of these

results is affected by the sample of primiparous, mostly well educated Caucasian

mothers and a high prevalence of exclusive breastfeeding. Maternal self-report is

subject to potential social desirability bias in responses which may limit the accuracy of

data (Harden, Klein, & Durbin, 2005). However, questionnaires were self-administered,

labelled with study ID only and completed at baseline prior to any personal contact with

19 

study staff. Furthermore, the alternative of direct observation has limitations in terms of

assessing usual behaviour (Haycraft & Blissett, 2008) and was not feasible in such a

large sample. Finally, these analyses are cross-sectional and as such causal

relationships cannot be determined and require longitudinal exploration.

Conclusions

Infant temperament was found to be associated with the feeding practices of first-time

mothers. Overall, the results suggest that mothers with infants of more difficult

temperament tend to adopt feeding practices that have been associated with indicators of

later obesity risk. As such, feeding practices are a potential mediator for the previously

demonstrated associations between infant temperament and weight outcomes and

further analysis of this relationship is recommended. Mothers of temperamentally

difficult infants may benefit from additional guidance and support with feeding to

minimise the adoption of undesirable feeding practices. The findings that maternal

depression has an independent influence on feeding practises suggests that the provision

of support for infant feeding concurrent with treatment of postnatal depression may also

assist the feeding relationship. Longitudinal prospective studies in a range of

population groups are required to extend our understanding of the relationships between

temperament and infant feeding practices. It is recommended that both temperament

and maternal depression be included as covariates in future studies of feeding practices,

child eating, and obesity risk.

Conflict of interest

20 

The authors have indicated that they have no personal financial relationships relevant to

this article to disclose.

Acknowledgements

NOURISH was funded 2008-2011 by the Australian National Health and Medical

Research Council (grant 426704). Dr Kimberley Mallan has occupied the Heinz

Postdoctoral Fellowship funded by HJ Heinz. Additional funding was provided by

Meat & Livestock Australia (MLA), Department Health South Australia, Food

Standards Australia New Zealand (FSANZ) and Queensland University of Technology,

and NHMRC Career development Award to JMN (grant 390136). We acknowledge the

NOURISH investigators: Professors Diana Battistutta, Ann Farrell, Geoffrey Cleghorn

and Geoffrey Davidson. We express appreciation for Professor Karen Thorpe’s

contribution to the manuscript. We sincerely thank all our participants, recruiting staff

and study staff including Dr Carla Rogers, Jo Meedeniya, Gizelle Wilson, Chelsea

Mauch.

21 

References

Agras, W. S., Hammer, L. D., McNicholas, F., & Kraemer, H. C. (2004). Risk factors

for childhood overweight: A prospective study from birth to 9.5 years. The

Journal of Pediatrics, 145(1), 20-25.

Baughcum, A. E., Powers, S. W., Johnson, S. B., Chamberlin, L. A., Deeks, C. M., Jain,

A., & Whitaker, R. C. (2001). Maternal feeding practices and beliefs and their

relationships to overweight in early childhood. Journal of Developmental &

Behavioral Pediatrics, 22(6), 391-408.

Birch, L., & Deysher, M. (1986). Caloric compensation and sensory specific satiety -

evidence for self regulation of food-intake by young-children. [Article]. Appetite,

7(4), 323-331.

Birch, L. L. (2006). Child feeding practices and the etiology of obesity. Obesity, 14(3),

343-344.

Blissett, J., & Farrow, C. (2007). Predictors of maternal control of feeding at 1 and 2

years of age. International Journal of Obesity, 31(10), 1520-1526. doi:

10.1038/sj.ijo.0803661

Burdette, H. L., Whitaker, R. C., Hall, W. C., & Daniels, S. R. (2006). Maternal infant-

feeding style and children's adiposity at 5 years of age. Arch Pediatr Adolesc

Med, 160(5), 513-520. doi: 10.1001/archpedi.160.5.513

Carey, W. B. (1985). Temperament and increased weight gain in infants. Journal of

Developmental & Behavioral Pediatrics, 6(3), 128-131.

Carey, W. B., & McDevitt, S. C. (1978). Revision of the Infant Temperament

Questionnaire. Pediatrics, 61(5), 735-739.

22 

Cox, J. L., Holden, J. M., & Sagovsky, R. (1987). Detection of postnatal depression:

Development of the 10-item Edinburgh Postnatal Depression Scale. British

Journal of Psychiatry, 150, 782-786.

Daniels, L. A., Magarey, A., Battistutta, D., Nicholson, J. M., Farrell, A., Davidson, G.,

& Cleghorn, G. (2009). The NOURISH randomised control trial: Positive feeding

practices and food preferences in early childhood - a primary prevention program

for childhood obesity. Bmc Public Health, 9. doi: 38710.1186/1471-2458-9-387

Darlington, A. S., & Wright, C. M. (2006). The influence of temperament on weight

gain in early Infancy. Journal of Developmental & Behavioral Pediatrics, 27(4),

329-335.

de Onis, M., Blössner, M., & Borghi, E. (2010). Global prevalence and trends of

overweight and obesity among preschool children. The American Journal of

Clinical Nutrition, 92(5), 1257-1264. doi: 10.3945/ajcn.2010.29786

Dettwyler, K., & Fishman, C. (1992). Infant feeding practices and growth. Annual

Review of Anthropology, 21, 171-204.

DiSantis, K. I., Hodges, E. A., Johnson, S. L., & Fisher, J. O. (2011). The role of

responsive feeding in overweight during infancy and toddlerhood: a systematic

review. Int J Obes, 35(4), 480-492.

Donath, S., & Amir, L. (2000). Does maternal obesity adversely affect breastfeeding

initiation and duration? Journal of Paediatrics and Child Health, 36(5), 482-486.

doi: 10.1046/j.1440-1754.2000.00562.x

Farrow, C., & Blissett, J. (2006a). Maternal cognitions, psychopathologic symptoms,

and infant temperament as predictors of early infant feeding problems: A

23 

longitudinal study. International Journal of Eating Disorders, 39(2), 128-134. doi:

10.1002/eat.20220

Farrow, C., & Blissett, J. (2006b). Does Maternal Control During Feeding Moderate

Early Infant Weight Gain? Pediatrics, 118(2), e293-e298. doi: 10.1542/peds.2005-

2919

Field, A. (2009). Discovering statistics using SPSS (Third ed.). Dubai: Oriental Press.

Fisher, J. O., & Birch, L. L. (1999). Restricting Access to Foods and Children's Eating.

Appetite, 32(3), 405-419.

Galler, J. R., Harrison, R. H., Ramsey, F., Butler, S., & Forde, V. (2004). Postpartum

maternal mood, feeding practices, and infant temperament in Barbados. Infant

Behavior & Development, 27(3), 267-287. doi: 10.1016/j.infbeh.2003.11.002

Ganley, R. M. (1989). Emotion and eating in obesity: A review of the literature.

International Journal of Eating Disorders, 8(3), 343-361.

Gjerdingen, D. K., & Debra, F. (1991). Predictors of health in new mothers. Social

Science &amp; Medicine, 33(12), 1399-1407. doi: 10.1016/0277-9536(91)90285-

k

Goldsmith, H. H., Buss, A. H., Plomin, R., Rothbart, M. K., Thomas, A., Chess, S., . . .

McCall, R. B. (1987). Roundtable: What is temperament? Four approaches. Child

Development, 58(2), 505-529.

Haycraft, E. L., & Blissett, J. M. (2008). Maternal and paternal controlling feeding

practices: reliability and relationships with BMI. Obesity, 16(7), 1552-1558.

Hayden, E. P., Klein, D. N., & Durbin, C. E. (2005). Parent reports and laboratory

assessments of child temperament: A comparison of their associations with risk

24 

for depression and externalizing disorders. Journal of Psychopathology and

Behavioral Assessment, 27(2), 89-100. doi: 10.1007/s10862-005-5383-z

Hendricks, K., Briefel, R., Novak, T., & Ziegler, P. (2006). Maternal and Child

Characteristics Associated with Infant and Toddler Feeding Practices. Journal of

the American Dietetic Association, 106(1, Supplement), 135-148. doi:

10.1016/j.jada.2005.09.035

Horn, M. G., Galloway, A. T., Webb, R. M., & Gagnon, S. G. (2011). The role of child

temperament in parental child feeding practices and attitudes using a sibling

design. Appetite, 57(2), 510-516. doi: 10.1016/j.appet.2011.06.015

Hurley, K. M., Black, M. M., Papas, M. A., & Caufield, L. E. (2008). Maternal

Symptoms of Stress, Depression, and Anxiety Are Related to Nonresponsive

Feeding Styles in a Statewide Sample of WIC Participants. The Journal of

Nutrition, 138(4), 799-805.

Kral, T. V. E., Stunkard, A. J., Berkowitz, R. I., Stallings, V. A., Brown, D. D., & Faith,

M. S. (2007). Daily food intake in relation to dietary energy density in the free-

living environment: a prospective analysis of children born at different risk of

obesity. The American Journal of Clinical Nutrition, 86(1), 41-47.

Li, R., Fein, S. B., & Grummer-Strawn, L. M. (2008). Association of Breastfeeding

Intensity and Bottle-Emptying Behaviors at Early Infancy With Infants' Risk for

Excess Weight at Late Infancy. Pediatrics, 122(Supplement 2), S77-S84. doi:

10.1542/peds.2008-1315j

Li, R., Jewell, S., & Grummer-Strawn, L. (2003). Maternal obesity and breast-feeding

practices. The American Journal of Clinical Nutrition, 77(4), 931-936.

25 

Maccoby, E., & Martin, J. (1983). Socialization in the context of the family: Parent-

child interaction. In P. H. Mussen (Ed.), Handbook of Child Psychology (5 ed.).

USA: John Wiley and Sons.

Magarey, A. M., Daniels, L. A., Boulton, T. J., & Cockington, R. A. (2003). Predicting

obesity in early adulthood from childhood and parental obesity. Int J Obes Relat

Metab Disord, 27(4), 505-513.

McGrath, J. M., Records, K., & Rice, M. (2008). Maternal depression and infant

temperament characteristics. Infant Behavior and Development, 31(1), 71-80.

Minister for Health and Ageing. (2007). Tackling Childhood Obesity in Australia

Summit. Sydney: Queensland Government.

Niegel, S., Ystrom, E., & Vollrath, M. E. (2007). Is difficult temperament related to

overweight and rapid early weight gain in infants? A prospective cohort study.

Journal of Developmental and Behavioral Pediatrics, 28(6), 462-466.

Ong, K. K., Emmett, P. M., Noble, S., Ness, A., Dunger, D. B., & and the ALSPAC

Study Team. (2006). Dietary Energy Intake at the Age of 4 Months Predicts

Postnatal Weight Gain and Childhood Body Mass Index. Pediatrics, 117(3), e503-

508. doi: 10.1542/peds.2005-1668

Power, C., & Parsons, T. (2000). Nutritional and other influences in childhood as

predictors of adult obesity. Proceedings of the Nutrition Society, 59(02), 267-272.

doi: doi:10.1017/S002966510000029X

Pulkki-Raback, L., Elovainio, M., Kivimaki, M., Raitakari, O. T., & Keltikangas-

Jarvinen, L. (2005). Temperament in childhood predicts body mass in adulthood:

the cardiovascular risk in Young Finns Study. Health Psychology, 24(3), 307-315.

26 

Reilly, J. J., Methven, E., McDowell, Z. C., Hacking, B., Alexander, D., Stewart, L., &

Kelnar, C. J. H. (2003). Health consequences of obesity. Archives of Disease in

Childhood, 88(9), 748-752. doi: 10.1136/adc.88.9.748

Sanson, A., Prior, M., Garino, E., Oberklaid, F., & Sewell, J. (1987). The structure of

infant temperament: Factor analysis of the revised infant temperament

questionnaire. [DOI: 10.1016/0163-6383(87)90009-9]. Infant Behavior and

Development, 10(1), 97-104.

Teng, A., Bartle, A., Sadeh, A., & Mindell, J. (2012). Infant and toddler sleep in

Australia and New Zealand. Journal of Paediatrics and Child Health, 48(3), 268-

273.

Thomas, A., Chess, S., Birch, H., Hertzig, M., & Korn, S. (1963). Behavioral

Individuality in Early Childhood. New York: New York University Press.

Van Esterik, P. (2002). Contemporary trends in infant feeding research. Annual Review

of Anthropology, 31, 257-278.

Ventura, A., & Birch, L. (2008). Does parenting affect children's eating and weight

status? International Journal of Behavioral Nutrition and Physical Activity, 5(1),

15.

Vollrath, M. E., Tonstad, S., Rothbart, M. K., & Hampson, S. E. (2010). Infant

temperament is associated with potentially obesogenic diet at 18 months.

International Journal of Pediatric Obesity, 1-7. doi:

doi:10.3109/17477166.2010.518240

Wasser, H., Bentley, M., Borja, J., Goldman, B. D., Thompson, A., Slining, M., &

Adair, L. (2011). Infants perceived as "fussy" are more likely to receive

27 

complementary foods before 4 months. Pediatrics, 127(2), 229-237. doi:

10.1542/peds.2010-0166

Webber, L., Hill, C., Saxton, J., Van Jaarsveld, C. H. M., & Wardle, J. (2009). Eating

behaviour and weight in children. International Journal of Obesity, 33, 21+.

World Health Organisation. (1995). Physical status: the use and interpretation of

anthropometry. Geneva: World Health Organization.

World Health Organisation. (2006). WHO child growth standards: Length/height-for-

age, weight-for-age, weight-for-length, weight-for height and body mass index-

for-age: Methods and development. Geneva.

Worobey, J., Islas Lopez, M., & Hoffman, D. J. (2009). Maternal Behavior and Infant

Weight Gain in the First Year. Journal of Nutrition Education and Behavior,

41(3), 169-175.

Footnote

28 

1 The five factors from the IFQ (Baughcum, et al., 2001) have been shortened to:

feeding on demand, using food to calm, awareness of infant cues, concern about

overweight and concern about underweight.

29 

a High scores indicate greater maternal depressive symptoms (0-30), Edinburgh

Postnatal Depression Scale (Cox, et al., 1987). These data were square root transformed

for the analyses.

b Exclusive breastfeeding, breastmilk only; Any breastfeeding, including combination

feeding with formula and fully breastfeeding with or without some complementary

foods; No breastfeeding, formula feeding (no breastmilk) with or without some

complementary foods.

Table 1. Sample characteristics

Variable N % (n) or mean ± SD Maternal characteristics

Maternal age at delivery (years) 698 30.1 ± 5.3 Maternal education 698 University education 58 (406) Maternal depression a 669 4.7 ± 3.9 Borderline/depressive 12 (79) Breastfeeding status b 631 Exclusive breastfeeding 45 (314) Any breastfeeding 23.7 (150) No breastfeeding 26.5 (167) Maternal BMI (kg/m2) 692 26 ± 5 Feeding practices and beliefs Awareness of hunger/satiety cues c 675 3.2 ± 0.5 Using food to calm fussiness c 671 1.8 ± 0.8 Concern over-eating/overweight d 670 High concern 219 (33) Concern under-eating/underweight d 675 High concern 238 (35) Feeding on demand vs. schedule e 664 Feeding on schedule 246 (37)

Infant characteristics Age (months) 696 4.3 ± 1.0 Baby gender 698 Female 51 (354) Infant weight gain (g/wk) 697 197 ± 46 Infant raw change in weight for age Z-score from birth 693 –0.43 ± 1.00 Maternal perception of child weight status 660 Underweight 5 (32) Normal weight 87 (573) Overweight 8 (55) Temperament d 666 2.5 ± 0.6

30 

c Measured on a 5-point Likert scale from 0= “Never” to 4= “Always”, Infant Feeding

Questionnaire (Baughcum, et al., 2001)

d Outlier tertile separated from other two tertiles to create a binary variable, Infant

Feeding Questionnaire (Baughcum, et al., 2001)

e High scores indicate a more difficult temperament, Short Temperament Scale for

Infants (Sanson, et al., 1987)

31 

* p < 0.05

a Measured on a 5-point Likert scale from 0 = “Never” to 4 = “Always”, Infant Feeding

Questionnaire (Baughcum, et al., 2001)

b Values taken from final model

c High scores indicate greater maternal depressive symptoms (data has been square root

transformed), Edinburgh Postnatal Depression Scale (Cox, et al., 1987)

d Exclusive breastfeeding, breastmilk only; Any breastfeeding, including combination

feeding with formula and fully breastfeeding with or without some complementary

foods; No breastfeeding, formula feeding (no breastmilk) with or without some

complementary foods.

e Infant weight gain per week from birth to 2–7-months-of-age

Table 2. Hierarchical linear regression analyses for feeding practices and beliefs and infant temperament. Feeding practices and beliefs a Awareness of cues Using food to calm N 598 596 β 95% CI β 95% CI Step 1 b R2

Adj = 0.05* R2Adj = 0.10*

Infant gender (male) 0.00 –0.08-0.09 0.09* 0.02-0.27 Maternal age at delivery (years) –0.05 –0.01-0.00 –0.01 –0.01-0.01 Maternal education (no university) –0.04 –0.12-0.05 0.04 –0.06-0.20 Maternal depression c –0.13* –0.14- –0.03 0.09* 0.01-0.17 Breastfeeding (BF) status d Exclusive BF vs any BF/no BF 0.08 –0.02- 0.18 –0.15* –0.38- –0.09 No BF vs any BF/exclusive BF –0.02 –0.14-0.08 0.17* 0.13- 0.47 Infant wt gain (g/wk) e 0.03 0.00-0.00 0.03 0.00-0.00 Maternal BMI (kg/m2) –0.02 –0.01-0.01 0.00 –0.01-0.01 F (df) 5.16 (8, 589)* 9.55 (8, 587)* Step 2 R2

Adj = 0.15* R2Adj = 0.12*

R2 = 0.09* R2 = 0.02* Temperament f –0.32* –0.38- –0.23 0.14* 0.09-0.31 F(df) 64.47 (1, 588)* 11.97 (1, 586)* F (df) 12.25 (9, 588)* 9.98 (9, 586)*

32 

f High scores indicate a more difficult temperament, Short Temperament Scale for

Infants (Sanson, et al., 1987)

33 

Abbreviations: B, unstandardised regression coefficient; Wχ2,Wald χ2-test; OR, Odds ratio; 95% CI, 95% CI for Odds ratio

*p < 0.05

a Dichotomised as per Table 1, Infant Feeding Questionnaire (Baughcum, et al., 2001)

b Values taken from final model

c High scores indicate greater maternal depressive symptoms (data has been square root transformed), Edinburgh Postnatal Depression

Scale (Cox, et al., 1987)

Table 3. Hierarchical binary logistic regression for feeding practices and infant temperament Feeding practices and beliefs a Concern about overweight Concern about underweight Feeding on demand vs. schedule N 596 598 588 Variables B Wχ2 OR 95% CI B Wχ2 OR 95% CI B Wχ2 OR 95% CI Step 1b Nagelkerke R2 = 0.15 Nagelkerke R2 = 0.15 Nagelkerke R2 = 0.13 Infant gender (male) –0.11 0.32 0.89 0.60-1.33 0.13 0.44 1.14 0.77-1.68 0.35 3.15 1.41 0.96-2.07 Maternal age at delivery (years) 0.03 2.15 1.03 0.99-1.07 0.03 2.73 1.03 0.99-1.07 –0.02 0.69 0.99 0.95-1.02 Maternal education (no uni) 0.10 0.22 1.10 0.73-1.66 0.62 8.91* 1.85 1.24-2.77 –0.14 0.47 0.87 0.59-1.29 Maternal depression c 0.48 14.50* 1.61 1.26-2.06 0.35 8.19* 1.41 1.12-1.80 –0.25 4.22* 0.78 0.62-0.99 Breastfeeding (BF) status d Exclusive BF vs any BF/no BF 0.00 0.00 1.00 0.63-1.60 0.17 0.56 1.19 0.76-1.87 –0.71 9.67* 0.49 0.32-0.77 No BF vs any BF/exclusive BF

0.26 0.92 1.30 0.76-2.21 –0.25 0.90 0.78 0.46-1.31 0.70 7.88* 2.01 1.23-3.27

Infant weight gain (g/wk) e 0.01 25.95* 1.01 1.01-1.02 –0.01 31.41* 0.99 0.98-0.99 0.00 1.24 1.00 1.00-1.01 Maternal BMI (kg/m2) 0.06 10.54* 1.06 1.02-1.10 –0.03 2.57 0.97 0.94-1.01 0.07 14.75* 1.08 1.04-1.12 Step 2 Nagelkerke R2 = 0.17 Nagelkerke R2 = 0.19 Nagelkerke R2 = 0.13 Goodness of fit χ2(df) =78.89(9)* Goodness of fit χ2(df) = 86.28(9)* Goodness of fit χ2(df) = 60.21(9)* Temperament f 0.56 9.56* 1.75 1.23-2.49 0.69 15.06* 2.00 1.41-2.84 –0.01 0.00 0.99 0.71-1.39

34 

d Exclusive breastfeeding, breastmilk only; Any breastfeeding, including combination feeding with formula and fully breastfeeding with or

without some complementary foods; No breastfeeding, formula feeding (no breastmilk) with or without some complementary foods.

e Infant weight gain per week from birth to 2–7-months-of-age

f High scores indicate a more difficult temperament, Short Temperament Scale for Infants (Sanson, et al., 1987)

35