Effects of distraction on the development of satiety

Abstract

Two experiments explored the hypothesis that distraction causes a reduced sensitivity to the physiological and sensory cues that signal when to terminate a meal. In Experiment 1, eighty-eight females ate five 'Jaffa Cakes' either while distracted by a computer game or while sitting in silence. Analysis of the difference in rated hunger, fullness and desire to eat (pre- to post-intake) revealed that distracted participants experienced smaller changes in their desire to eat and fullness than did non-distracted participants. Experiment 2 assessed whether changes in ratings are attenuated because sensory-specific satiety (or a related process) fails to develop. Using a similar procedure, eighty-four females provided desire to eat, pleasantness and intensity ratings for Jaffa Cakes and for two 'uneaten' foods, both before and at three time-points after consuming five Jaffa Cakes. Non-distracted participants reported a reduction in their desire to eat the eaten food relative to the uneaten food (food-specific satiety), whereas distracted participants maintained a desire to eat all foods. Moreover, this difference between distracted and non-distracted participants was evident 5 and 10 min after the eating episode had terminated. The present findings invite speculation that distraction attenuates the development of sensory-specific satiety, and that this effect persists (at least for a brief period) after the distractor has terminated. More generally, this kind of phenomenon warrants further scrutiny because it holds the potential to contribute towards overeating, either by prolonging an eating episode or by reducing the interval between meals.

Full text

Effects of distraction on the development of satiety
Jeffrey M. Brunstrom
1
* and Gemma L. Mitchell
2
1
Department of Experimental Psychology, University of Bristol, 12a Priory Road, Bristol BS8 1TU, UK
2
Department of Human Sciences, Loughborough University, Loughborough LE11 3TU, UK
(Received 26 January 2006 – Revised 10 May 2006 – Accepted 12 May 2006)
Two experiments explored the hypothesis that distraction causes a reduced sensitivity to the physiological and sensory cues that signal when to
terminate a meal. In Experiment 1, eighty-eight females ate five ‘Jaffa Cakes’ either while distracted by a computer game or while sitting in
silence. Analysis of the difference in rated hunger, fullness and desire to eat (pre- to post-intake) revealed that distracted participants experienced
smaller changes in their desire to eat and fullness than did non-distracted participants. Experiment 2 assessed whether changes in ratings are atte-
nuated because sensory-specific satiety (or a related process) fails to develop. Using a similar procedure, eighty-four females provided desire to eat,
pleasantness and intensity ratings for Jaffa Cakes and for two ‘uneaten’ foods, both before and at three time-points after consuming five Jaffa
Cakes. Non-distracted participants reported a reduction in their desire to eat the eaten food relative to the uneaten food (food-specific satiety),
whereas distracted participants maintained a desire to eat all foods. Moreover, this difference between distracted and non-distracted participants
was evident 5 and 10 min after the eating episode had terminated. The present findings invite speculation that distraction attenuates the develop-
ment of sensory-specific satiety, and that this effect persists (at least for a brief period) after the distractor has terminated. More generally, this kind
of phenomenon warrants further scrutiny because it holds the potential to contribute towards overeating, either by prolonging an eating episode or
by reducing the interval between meals.
Distraction: Desire to eat: Sensory-specific satiety: Dietary restraint: Attention: Disinhibition
The potential for food intake to increase when a meal is con-
sumed under conditions of high cognitive load (Ward &
Mann, 2000; Mann & Ward, 2004), or when distracted
(Bellisle & Dalix, 2001; Wansink & Park, 2001; Boon et al.
2002; Bellisle et al. 2004), is well documented. These findings
are important, because larger meal sizes represent an import-
ant predictor of obesity (Levitsky, 2005). While recent studies
have explored the extent to which individuals differ in their
sensitivity to the effects of distraction (Ward & Mann, 2000;
Mitchell & Brunstrom, 2005), the reason why distraction influ-
ences eating behaviour has remained unclear.
Meals that are consumed with others tend to be both larger and
longer (de Castro, 1990, 1994; de Castro & Brewer, 1992) than
those eaten alone. Therefore, one possibility is that distraction
serves to prolong the duration of an eating episode, thereby
increasing the amount of food that is subsequently consumed.
Consistent with this idea, obese participants are more likely to
terminate their meals when a concurrent task, for example a tele-
vision programme, ends (Tuomisto et al. 1998). An alternative
proposition derives from the observation that (in the absence
of food) people can ‘forget’ to eat when engrossed in a task.
One possibility is that a similar phenomenon extends to
occasions when an individual eats during distraction. In this con-
text, distraction limits an individual’s capacity to monitor sig-
nals associated with satiety. Consistent with this idea, it has
been suggested that normal dietary control requires access to a
degree of ‘cognitive resource’ (Boon et al. 2002). Since
distraction-related overeating has been observed most
frequently (although not exclusively) in restrained eaters (e.g.
Ward & Mann, 2000, but see Boon et al. 2002), an added
possibility is that the demands of cognitive restraint further
undermine attempts to monitor these signals.
One way to explore these ideas is to use a crossover design,
incorporating a ‘distraction’ and a ‘no distraction condition’.
However, when tested in this way the participants may suspect
that they should behave or feel differently in the two con-
ditions. This is problematic, because it makes it difficult to
rule out the possibility that their responding is based on
demand characteristics. To address this problem, we tested
two different groups of participants in Experiment 1. In one
group, participants were required to consume a fixed portion
of a palatable snack, at regular intervals, over a 5 min
period. The participants in a second group completed the
same task while performing a highly engaging cognitive
task. Our decision to use a relatively short eating episode
was motivated by a concern to ensure that the participants
attended continuously to the task. In both the no distraction
and the distraction condition, measures of hunger, fullness
and desire to eat were taken both before and after eating. In
so doing, our aim was to determine the effect of distraction
on several measures of satiety. We were also interested to
explore whether the effects of distraction are moderated by
everyday dietary behaviour (dietary restraint and disinhibited
eating).
* Corresponding author: Dr Jeff Brunstrom, fax þ 44 (0)117 928 8588, email Jeff.Brunstrom@Bristol.ac.uk
Abbreviations: DEBQ, Dutch Eating Behaviour Questionnaire; TFEQ, Three Factor Eating Questionnaire.
British Journal of Nutrition (2006), 96, 761–769 DOI: 10.1079/BJN20061880
q The Authors 2006

Experiment 1
Method
Participants. Eighty-eight female undergraduate students
were recruited using advertisements posted around the Lough-
borough University campus (mean age 19·52 (
SD 1·69) years).
All were paid £5 (Sterling) for their participation. The Lough-
borough University Ethics Committee gave approval for the
research carried out in this experiment and also in Experiment
2. From the outset, participants were informed that they would
be asked to provide a measure of their height and weight.
However, they were also told that participation in this aspect
of the experiment was optional. This decision was motivated
by a concern to sample individuals from a wide distribution
of eating behaviours, including those who might otherwise
decline to take part because of concerns associated with
their body weight or body image.
Subjective measures and questionnaires. At different
times during the experiment, 100 mm visual-analogue scales
were used to obtain ratings of hunger, fullness and desire to
eat. Specifically, participants were asked, ‘How hungry/full
do you feel right now?’ and ‘How strong is your desire to
eat right now?’ Ends were anchored ‘Not at all hungry/full/
strong’ and ‘Extremely hungry/full/strong’ on the left- and
right-hand sides, respectively.
Dietary behaviour was assessed using the restraint section
of the Dutch Eating Behaviour Questionnaire (DEBQ; Van
Strien et al. 1986) and the disinhibition scale of the Three
Factor Eating Questionnaire (TFEQ; Stunkard & Messick,
1985). The DEBQ-restraint scale was chosen because it ident-
ifies those individuals who successfully limit their dietary
intake (Heatherton et al. 1988) and because it has been used
previously in studies exploring the relationship between
restraint and cognitive performance (e.g. Green & Rogers,
1995). The DEBQ-restraint scale contrasts the TFEQ-disinhi-
bition scale, which identifies a failure to maintain dietary con-
trol and a tendency to overeat (Westenhoefer et al. 1994).
Distractor task. The distractor task was a commercially
available computer game called ‘Pong’ (produced by Atari
Inc.), which was presented on a 14-inch colour television. In
Pong, the player is represented on the screen by a coloured
cursor. The player can move the cursor using a joystick. The
activity of the computer is represented by a similar cursor.
Respectively, the player and computer assume right and left pos-
itions on the screen. During the game, the player and computer
‘bounce’ a virtual ball back and forth. One point is awarded to
a player when their opponent fails to return the ball. The first
player to score twenty-one points wins the game.
Food. In this experiment participants consumed five Jaffa
Cakes (chocolate-covered sponge cake with orange-flavoured
filling). This snack product is widely available in the UK
(McVities, London, UK; 193 kJ (46 kcal) per 8 g cake).
Procedure. Participants were tested between 14.00 and
18.00 hours in individual cubicles in the Ingestive Behaviour
Laboratory at Loughborough University. All were instructed
to abstain from eating for at least 3 h prior to the onset of the
experiment. Compliance with this request was verified using a
questionnaire. On arrival, participants were randomly allocated
to the distraction or the no-distraction condition. They were then
briefed about the procedure. When completing ratings of hunger,
fullness and desire to eat, they were told to ‘not think too hard.
Just go with your instantaneous response’. After confirming
that they understood these instructions the participants com-
pleted the initial ratings of hunger, fullness and desire to eat. Par-
ticipants were then left alone for 5 min. During this period they
were required to eat five Jaffa Cakes, at the rate of one every 60 s.
A timer was used to signal the end of every 60 s interval. Those
participants allocated to the no-distraction condition completed
this procedure while sitting in silence. Those allocated to the dis-
traction condition were instructed to complete this procedure
while playing Pong for the full 5 min duration. At the end of
the fifth 60 s interval (after all five Jaffa cakes had been con-
sumed) participants completed a second set of ratings relating
to hunger, fullness and desire to eat. They were then issued
with the DEBQ-restraint and TFEQ-disinhibition scales. If the
participant consented, a measure of height (cm) and weight
(kg) was also taken. Five participants declined to be weighed;
three in the distraction group and two in the no-distraction
group. All participants were then paid, debriefed and thanked
for their participation.
Data analysis. For each participant, we calculated a set of
‘change’ scores (pre-intake minus post-intake) for our
measures of hunger, fullness and desire to eat. The change
scores associated with each condition were then compared
using independent t tests. In addition, we used regression
analysis to establish whether an effect of distraction is modi-
fied by one of our measures of everyday dietary behaviour.
Specifically, we were interested to determine whether an inter-
action exists, either between dietary restraint and distraction
condition or between dietary disinhibition and distraction
condition.
Results
Group characteristics. One participant in the no-distraction
group failed to eat all five Jaffa Cakes. Consequently, her
data were excluded from the analysis. Distracted and non-dis-
tracted participants were very similar (see Table 1). Across
these groups, BMI (t[77] 1·19, P¼ 0·24) and scores on the
DEBQ-restraint (t[83] 0·90, P¼ 0·37) and the TFEQ-disinhibi-
tion scale (t[83] 0·74, P¼ 0·46) did not differ significantly.
Effects of distraction on the change in hunger, fullness and
desire to eat. The distracted and non-distracted groups had
Table 1. Mean BMI, Dutch Eating Behaviour Questionnaire
(DEBQ)-restraint and Three Factor Eating Questionnaire
(TFEQ)-disinhibition scores, and mean initial hunger, fullness
and desire to eat ratings (mm), for the distraction and no dis-
traction groups in Experiment 1*
(Mean values and standard deviations)
Distraction
(n 44)
No distraction
(n 43)
Group Mean
SD Mean SD
BMI (kg/m
2
) 22·60 2·78 23·41 3·34
DEBQ-restraint 2·76 0·77 2·61 0·86
TFEQ-disinhibition 6·80 4·50 7·35 3·49
Initial hunger 58·14 21·27 61·93 23·87
Initial fullness 26·25 19·46 22·56 20·93
Initial desire to eat 52·57 27·17 59·00 24·37
* For details of procedures, see p. 762.
J. M. Brunstrom and G. L. Mitchell762

similar baseline hunger (t[83] 0·78, P¼ 0·45), fullness (t[84]
0·85, P¼ 0·40) and desire to eat (t[84] 1·16, P¼ 0·25) ratings
(Table 1). The change scores associated with these measures
are shown in Fig. 1. Non-distracted participants reported a sig-
nificantly greater increase in fullness (t[83] 2·50, P¼ 0·014)
and a significantly greater reduction in desire to eat (t[84]
2·09, P¼ 0·04). They also experienced a greater reduction in
hunger. However, this difference narrowly missed significance
(t[83] 1·95, P¼ 0·055).
Relationship between everyday dietary behaviour and the
effects of distraction. We conducted a set of regression ana-
lyses to determine whether changes in hunger, fullness or
desire to eat are predicted by an interaction between distrac-
tion condition and dietary restraint or dietary disinhibition.
In every case, the associated interaction term failed to reach
significance (P values in the range 0·232–0·412).
Discussion
The main aim of Experiment 1 was to assess the effect of dis-
traction during a meal on changes in hunger, fullness and
desire to eat. The data indicate that the presence of a distractor
task attenuates the changes that otherwise occur in these
measures. Further, this phenomenon appears to be universal,
as little evidence was found to indicate variation associated
with DEBQ-restraint or TFEQ-disinhibition scores (although
see later for further comment on this issue). Previously, dis-
traction has been associated with a reduction in hunger
(Herman et al. 1999). Here, the converse was found –
hunger is preserved. The reason for this discrepancy is
unclear. However, it seems likely that the effects of distraction
are dependent on the context in which it occurs. In relation to
this idea, it is important to note that in the study by Herman
et al. (1999) distraction was not contemporaneous with food
intake. Rather, appetite was simply assessed before and after
watching a video. Indeed, these authors acknowledge that
when attention is directed to a concurrent task during a
meal, then subjective hunger and eating behaviour are likely
to interact in a complex way.
There are at least two reasons why the distraction task
might influence the way that ratings change. Previously, it
has been argued that the potency of direct controls (i.e.
those that arise from contact with pre-absorptive receptors
that are located from the tongue to the end of the small intes-
tine) can be modulated by other indirect controls (e.g. cogni-
tive) (Smith, 1996). If this is the case, then it follows that the
effects of distraction could be general. That is, the distractor
task might inhibit an individual’s opportunity to attend to
the visceral sensations associated with the development of
satiety, and this stalls the general decline in hunger and
desire to eat that normally occurs during a meal.
An alternative possibility is that the effects of distraction
are more specific. Normally, sensory-specific satiety (Rolls,
1986) occurs when the same food is consumed repeatedly.
However, this process might be arrested when food is experi-
enced in conjunction with distraction. Consistent with this
idea, distraction appears to delay the development of salivary
habituation to a food cue (Epstein et al. 1997, 2005). This is
important, because this kind of habituation is functionally
quite similar to sensory-specific satiety. Both result from
repeated exposure to the sensory properties of a food, irrespec-
tive of energy content (e.g. Rolls et al. 1988; Epstein et al.
1993), and the decrease in salivation that indicates habituation
also tends to be accompanied by a reduction in ratings associ-
ated with the hedonic properties of the food (Wisniewski et al.
1992). In Experiment 1, we compared changes in desire to eat
one food (Jaffa Cakes). Consequently, it remains unclear
whether the distraction/no distraction condition differentially
effects changes in the desirability of all foods or only the
specific food that is consumed. Experiment 2 incorporated a
more comprehensive set of comparisons involving other
types of food, thereby facilitating an assessment of the poten-
tial ‘specific’ and ‘general’ effects of the distractor task.
A ‘standard’ sensory-specific satiety paradigm involves
participants eating to satiety and then completing pleasantness
ratings before and at regular intervals after a meal has termi-
nated. Participants rate both the food that forms the meal and a
set of uneaten foods, the particular food that is consumed is
counterbalanced across participants, and sensory-specific sati-
ety is assumed to occur when ratings for the eaten food decrease
relative to ratings for the uneaten foods. Here, we decided to
deviate from this basic design in two ways. First, in Experiment
2 the eaten food was not counterbalanced. Instead, as in Exper-
iment 1, participants always consumed Jaffa Cakes. This
30
(a) (b) (c)
20
10
0
No
distraction
Distraction No
distraction
Distraction No
distraction
Distraction
Change in hunger (mm)
30
20
10
0
Change in fullness (mm)
30
20
10
0
Change in desire to eat (mm)
Fig. 1. Mean change in hunger (a), fullness (b) and desire to eat (c) for the distracted and the non-distracted participants in Experiment 1. For details of pro-
cedures, see p. 762. Values are means with their standard errors depicted by vertical bars.
Distraction and satiety 763

approach enabled us to maintain a degree of consistency across
experiments. However, it also limits the extent to which we can
assume that the present findings generalise to all eaten and unea-
ten food pairings. To ensure that this distinction is transparent,
we will reflect this constraint by referring to the present results
in terms of ‘food-specific’ rather than ‘sensory-specific’ satiety.
Second, rather than allowing ad libitum access to food, as in
Experiment 1 (five Jaffa Cakes), we chose to give participants
a single, fixed, portion size. This reasoning was motivated by
the possibility that distraction simply delays the onset of sen-
sory-specific satiety. This is important, because under ad libitum
conditions we might expect to find little difference between dis-
tracted and non-distracted participants. This is because both
groups will experience a similar degree of food-specific satiety
– it will simply take distracted participants longer to arrive at this
state. A related problem with ad libitum eating is that any com-
parison across groups will be confounded by differences in meal
duration, making it difficult to disentangle the effects of distrac-
tion and meal duration. To address these problems we measured
the development of food-specific satiety after participants
achieved only partial satiety. Also, by controlling meal size
we were able to compare the development of food-specific sati-
ety at a common time-point in both groups.
A further motivation here was a more general concern to
characterise the effects that distraction might have on
hunger and pleasantness. Previous studies have focused
solely on the effects of a distractor within or immediately
after a period of distraction. Consequently, it remains unclear
whether the effects of distraction persist for a period after an
eating episode has terminated. To explore this issue, in Exper-
iment 2, measures of general and food-specific satiety were
assessed, both before, and at three time-points after consuming
a fixed quantity of Jaffa Cakes.
Experiment 2
Overview
Participants were allocated randomly to a distraction or a no-
distraction group. Three identical sets of 100 mm ratings were
issued on four separate occasions – at baseline, and then
immediately, 5 and 10 min after consuming five Jaffa Cakes.
Each set included a rating of pleasantness, intensity and
desire to eat. One set was used to assess the Jaffa Cakes.
The others were used to assess two different ‘uneaten’
foods, against which the Jaffa Cakes were compared. Each
time that the food-specific measures were issued the partici-
pants also provided a more general indication of their current
appetite (hunger, fullness and desire to eat). In this way, two
sets of measures were obtained, one that related to general
aspects of appetite and one that assessed the momentary
pleasantness of, and desire to eat, specific foods.
Method
Participants. Eighty-four female undergraduate students were
recruited using advertisements posted around the Loughborough
University campus (mean age 21·24 (
SD 0·54) years). All were
paid £5 (Sterling) for their participation. As in Experiment 1,
we were concerned to sample individuals from a wide distribution
of eating behaviours. Participants were informed that they would
be asked to provide a measure of their height and weight, but that
involvement in this aspect of the procedure was voluntary.
Subjective measures and questionnaires. Measures of gen-
eral appetite (hunger, fullness and desire to eat) were obtained
using the same 100 mm rating scales as in Experiment 1. For
the food-specific measures the participants completed the fol-
lowing ratings; ‘How pleasant is the taste of the food in your
mouth?’ (end anchors ‘very unpleasant’ and ‘very pleasant’),
‘How intense is the taste of the food in your mouth?’ (end
anchors ‘very weak’ and ‘extremely strong’) and ‘How
strong is your desire to eat more of this food right now?’
(end anchors ‘not at all strong’ and ‘extremely strong’). As
in Experiment 1, participants also completed the DEBQ-
restraint and the TFEQ-disinhibition scales.
Foods. Again, as in Experiment 1, participants ate five
Jaffa Cakes. The uneaten foods were bacon-flavoured corn
snacks (‘Frazzles’; Walkers Snack Foods Ltd, Leicester, UK,
1975 kJ (470 kcal) per 100 g) and mandarin fruit segments
(Tesco Stores Ltd, Cheshunt, UK, 315 kJ (74 kcal) per
100 g). A 200 ml glass of water was also provided to allow
participants to clean their palate between samples. This was
removed during the main eating episode.
Procedure. Participants were tested between 14.00 and
18.00 hours in individual cubicles in the Ingestive Behaviour
Laboratory at Loughborough University. All were instructed
to abstain from eating for at least 3 h prior to the onset of the
experiment. Compliance with this request was verified using a
questionnaire. On arrival, the participants rated their hunger,
fullness and desire to eat (general measures). They were then
presented with a tray containing the three food samples. Partici-
pants were instructed to taste each sample and to rate their indi-
vidual pleasantness, intensity and desire to eat after consuming
each one in turn (specific measures). The bacon-flavoured
snack (one corn chip ¼ approximately 3 g, 59 kJ (14 kcal))
was always tasted first, followed by the mandarin fruit segment
(one segment ¼ approximately 8 g, 25 kJ (6 kcal)) and the Jaffa
Cake sample (a quarter portion of a single cake ¼ approxi-
mately 2 g, 50 kJ (12 kcal)). Participants were instructed to
take a sip of water after consuming each sample.
After completing these ratings, the procedure for the 5 min
eating episode was explained. This was identical to Experiment
1. Half of the participants ate while sitting in silence. The other
half ate while playing Pong continuously. At the end of the
eating episode, all of the participants completed a second set
of general appetite ratings followed by a second assessment of
the three food samples. When they had finished this task they
were left alone to sit in silence. This procedure was then repeated
5 and 10 min after the end of the eating episode.
Finally, the participants completed the DEBQ-restraint
scale and the TFEQ-disinhibition scale. At this point a
measure of height (cm) and weight (kg) was taken from
those participants who consented to this aspect of the pro-
cedure (n 79). The participants were then paid, debriefed
and thanked for their assistance with the experiment.
Data analysis. In this experiment, sets of ratings were
issued before, and then immediately, 5 and 10 min after the
eating episode. For each of the general and food-specific rat-
ings, change scores were calculated based on the difference
between baseline ratings (t
base
) and those taken after the
eating episode (t
0min
,t
5min
and t
10 min
). In this way, for each
type of rating and each participant, three scores were derived
J. M. Brunstrom and G. L. Mitchell764

that reflected how the ratings changed after eating relative to
baseline.
To analyse changes in the general measures of appetite
(hunger, fullness and desire to eat), each set of change
scores was submitted to a mixed-model ANOVA. Respect-
ively, ‘distraction’ (no distraction/distraction) and ‘time’
(change at t
0min
,t
5min
and t
10 min
) were included as a between-
and a within-subject factor.
To assess evidence for food-specific satiety we used a
similar mixed-model ANOVA. Sensory-specific satiety is nor-
mally evidenced by a reduction in the pleasantness of an eaten
food relative to uneaten foods. Accordingly, in our analysis we
explored the extent to which pleasantness change scores are
predicted by an interaction between ‘food type’ (eaten and
uneaten), ‘time’ and ‘distraction’. To determine whether
these variables predict our other food-specific measures, the
same three-way ANOVA was also applied to the change
scores relating to intensity and desire to eat.
At each time-point, linear regression analyses were also
conducted to assess whether changes in ‘general’ and ‘food-
specific’ appetite are predicted by an interaction between
distraction group and everyday dietary behaviour. For the gen-
eral measures of appetite (hunger, fullness and desire to eat),
the change scores associated with each time-point were entered
into a separate regression model. For the food-specific measures
(pleasantness, intensity and desire to eat), a single measure of
food-specific satiety was derived at each time-point by subtract-
ing the change score associated with the uneaten food from the
change score associated with the eaten food. These scores
were also entered into separate regression models.
Results
Group characteristics. Table 2 shows the mean BMI, TFEQ-
disinhibition score and DEBQ-restraint score associated
with the distracted and the non-distracted groups, separately.
It also includes their mean baseline ratings of hunger, fullness
and desire to eat. A comparison across groups revealed no sig-
nificant differences in any of these measures (all P. 0·05). For
each of the three foods, Table 3 shows the baseline ratings of
pleasantness, intensity and desire to eat. Again, across groups,
no significant differences were found (all P. 0·05).
Effect of distraction on changes in hunger, fullness and
desire to eat. Relative to baseline, Fig. 2 shows the change
in hunger, fullness and desire to eat at the three time-points
after the 5 min eating episode.
Change in hunger and fullness: changes in hunger differed
significantly across time (F[2,164] 12·79, P, 0·001). How-
ever, the effect of time on change in fullness narrowly
missed significance (F[2,164] 2·89, P¼ 0·059). Changes in
hunger and fullness were generally larger in the no-distraction
group. However, these differences were modest and they
failed to reach significance (hunger, F[1,82] 0·32, P¼ 0·58;
fullness, F[1,82] 0·30, P¼ 0·59). Likewise, the interactions
between distraction and time were not significant (hunger,
F[2,164] 1·46, P¼ 0·24; fullness, F[2,164] 0·079, P¼ 0·92).
Change in desire to eat: the main effect of distraction con-
dition was significant (F[1,82] 4·32, P¼ 0·041), indicating that
the non-distracted participants experienced a greater overall
reduction in desire to eat than did the distracted participants.
‘Time’ was also identified as a significant factor, suggesting
that in both groups, the reduction in desire to eat changed
during the course of the experiment (F[2,164] 3·49,
P¼ 0·033). However, the interaction between time and distrac-
tion was not significant (F[2,164] 1·25, P¼ 0·29), indicating
that these effects operate independently (Fig. 2(c)).
Effect of distraction on food-specific satiety. Change in
pleasantness: Fig. 3 shows the change in pleasantness associ-
ated with both the uneaten foods (Fig. 3(a, b)) and the eaten
food (Fig. 3(c)). Across foods, the effect of distraction con-
dition did not reach significance (F[1,82] 0·32, P¼ 0·25). Like-
wise, the main effect of time narrowly missed significance
(F[2,164] 2·95, P¼ 0·055). However, a comparison across
foods revealed a main effect of food type (F[1,82] 39·45,
P, 0·001).
The greatest reduction in pleasantness occurred in the eaten
food. This pattern of results is consistent with evidence
for food-specific satiety. However, the critical interaction
between food type and distraction condition failed to reach
significance (F[1,82] 2·85, P¼ 0·095).
Table 2. Mean BMI, Dutch Eating Behaviour Questionnaire
(DEBQ)-restraint and Three Factor Eating Questionnaire
(TFEQ)-disinhibition scores, and mean initial hunger, fullness
and desire to eat ratings (mm), for the distraction and the no
distraction groups in Experiment 2*
(Mean values and standard deviations)
Distraction
(n 42)
No distraction
(n 42)
Group Mean
SD Mean SD
BMI (kg/m
2
) 22·86 2·74 22·77 2·67
DEBQ-restraint 2·66 0·89 2·86 0·76
TFEQ-disinhibition 7·64 3·10 7·90 3·40
Initial hunger 64·00 22·58 60·55 18·52
Initial fullness 20·98 17·61 24·95 13·67
Initial desire to eat 64·24 24·79 62·79 17·55
*For details of procedures, see p. 764.
Table 3. Mean ratings of pleasantness, intensity and
desire to eat (mm) the three food samples at baseline
for the distraction and the no distraction groups in Exper-
iment 2*
(Mean values and standard deviations)
Distraction
(n 42)
No distraction
(n 42)
Group Mean
SD Mean SD
Bacon-flavoured corn snack
Pleasantness 73·50 18·09 72·19 22·66
Intensity 69·83 21·27 71·36 21·30
Desire to eat 66·64 23·85 63·31 25·47
Mandarin fruit segment
Pleasantness 73·39 21·35 75·31 19·78
Intensity 57·52 17·47 57·83 18·69
Desire to eat 55·12 28·81 60·43 26·22
Jaffa Cake
Pleasantness 69·33 22·04 74·90 23·24
Intensity 62·26 18·44 68·17 17·35
Desire to eat 55·69 26·96 59·93 27·59
* For details of procedures, see p. 764.
Distraction and satiety 765

Change in intensity: Fig. 4 shows the changes in intensity
that occurred in the eaten food (Fig. 4(c)) and the two uneaten
foods (Fig. 4(a, b)). Taken together, the change in the intensity
of the foods differed significantly across time (F[2,162] 3·11,
P¼ 0·047). However, the data indicate little evidence for a
clear trend. In both groups the intensity of the two uneaten
foods did not differ greatly from baseline. With regard to
the eaten food, a different pattern of results was observed,
and this difference was confirmed by a significant interaction
between food type and time (F[2,164] 6·86, P¼ 0·002). Gener-
ally, the distracted group experienced a modest increase in the
intensity of the eaten food over time, whereas the no-distrac-
tion group tended to produce lower intensity ratings through-
out. However, in relation to this difference, it is noteworthy
that the interaction between condition and food type failed
to reach significance (F[1,82] 1·65, P¼ 0·20), as did the
main effect of condition (F[1,82] 1·44, P¼ 0·23) and the
main effect of food type (F[1,82] 1·54, P¼ 0·22).
Change in desire to eat: the changes in desire to eat were
not dissimilar to the changes in pleasantness. Fig. 5 shows
that participants experienced a slightly greater decline in
their desire to eat the eaten food (Fig. 5(c)) compared with
the uneaten foods (Fig. 5(a, b)). Consistent with this obser-
vation, food type emerged as a significant predictor in an
analysis of change scores (F[1,82] 51·46, P, 0·001).
In terms of the effects of distraction, Fig. 5(a, b) shows that
both distracted and non-distracted participants experienced the
same modest changes in their desire to eat the non-eaten
foods. By contrast, a relatively marked reduction was observed
with regard to the eaten food, but only in non-distracted partici-
pants. This food-specific effect was confirmed by a significant
interaction between food type and distraction (F[1,82] 5·77,
P¼ 0·019). Further t tests confirmed that the decline in the
desire to eat Jaffa Cakes was significantly greater in the no-dis-
traction group at t
5min
(t[82] 2·67, P¼ 0·018) and at t
10 min
(t[82]
2·83, P¼ 0·012) (at t
0min
the same difference failed to reach sig-
nificance (t[82] 1·50, P¼ 0·274)).
Finally, ANOVA also revealed a significant interaction
between time and food type (F[2,164] 3·09, P¼ 0·048) and
between time and distraction condition (F[2,164] 7·40,
P¼ 0·001). The main effect of distraction condition (F[1,82]
3·27, P¼ 0·074) and the main effect of time (F[2,164] 1·33,
P¼ 0·27) both failed to reach significance.
Correlation between changes in pleasantness and desire to
eat. In order to ascertain the extent to which changes in plea-
santness and desire to eat are related, a correlation coefficient
was calculated at each time-point, and in each group separ-
ately. All of these coefficients were significant (all
P, 0·001, r values in the range 0·357– 0·616).
Relationship with Dutch Eating Behaviour Questionnaire
and Three Factor Eating Questionnaire scores.Asin
Experiment 1, a set of regression analyses was used to deter-
mine whether the effects of distraction are more or less evi-
dent in those individuals who score highly on measures of
dietary restraint and disinhibited eating. Accordingly, for
each general measure (changes in hunger, fullness and
desire to eat) and each food-specific measure (changes in plea-
santness, intensity and desire to eat), an appropriate interaction
40
(a) (b) (c)
30
20
10
0
05100 055
Time after eating episode (min)
10 10
Change (mm)
Fig. 2. Mean change in hunger (a), fullness (b) and desire to eat (c), for the distracted (B) and the non-distracted participants (W), 0, 5 and 10 min after the eating
episode in Experiment 2. For details of procedures, see p. 764. Values are means with their standard errors depicted by vertical bars.
10
(a) (b) (c)
0
–10
–20
–30
–40
Change in pleasantness (mm)
–50
00055510 10 10
Time after eating episode (min)
Fig. 3. Mean change in pleasantness for the bacon-flavoured corn snack (a), mandarin fruit segment (b) and Jaffa Cake (c), for the distracted (B) and the non-
distracted participants (W), 0, 5 and 10 min after the eating episode in Experiment 2. For details of procedures, see p. 764. Values are means with their standard
errors depicted by vertical bars.
J. M. Brunstrom and G. L. Mitchell766

term was calculated to establish the extent to which the effect
of distraction is moderated by these measures of dietary beha-
viour. Since there were six different measures (three general
and three food-specific), three time-points (t
0min
,t
5min
and t
10 min
) and two assessments of behaviour (restraint and
disinhibition), a total of thirty-six separate regression analyses
(6 £ 3 £ 2 ¼ 36) were conducted.
For the measures of general appetite, all interaction terms
failed to reach significance (P values in the range 0·055–
0·970). For the food-specific measures, all interactions failed
to reach significance (P values in the range 0·061–0·970), with
the exception of one – 10 min after the eating episode the differ-
ence in change in desire to eat between the eaten and uneaten
foods was predicted by an interaction between the TFEQ-disin-
hibition scores and distraction group (t 2·35, b 2 4·123,
SE 1·756,
P¼ 0·021). Based on the number of separate tests performed, this
isolated case is easily attributable to chance.
Discussion
Consistent with Experiment 1, participants experienced a
greater reduction in their general desire to eat when they
were required to eat without distraction. This effect was evi-
dent 10 min after the distractor task had terminated and it
increased over this period (at 0 min the difference between dis-
tracted and non-distracted groups failed to reach significance).
A trend was also observed indicating that distraction limits the
changes in hunger and fullness that are otherwise found in
non-distracted participants. However, unlike in Experiment
1, the effects of distraction failed to reach significance.
The reason why this aspect of Experiment 1 was not replicated
is unclear. Procedurally, it may be relevant that participants
completed fewer types of ratings in Experiment 1. In Exper-
iment 1, changes were expected in both general ratings of
hunger and fullness and in specific ratings related to the
Jaffa Cakes. By contrast, in Experiment 2 we anticipated
that some of our ratings would remain relatively constant (rat-
ings associated with the uneaten foods). One possibility is that
measures of hunger and fullness do not assess abstract motiva-
tional constructs. Rather, they either reflect directly, or at least
are affected by, our momentary disposition to consume a par-
ticular food or foods (Booth, 1981, 1994; Kramer et al. 1992).
If this is the case then we might anticipate relatively modest
changes in rated hunger and fullness in Experiment 2. This
is because, unlike in Experiment 1, participants were encour-
aged to think about their appetite for several types of foods,
the majority of which (the uneaten foods) remained relatively
unaffected by the eating episode.
A further aim was to determine whether food-specific
satiety develops to the same extent in distracted and non-
distracted participants. Owing to various methodological
constraints (outlined earlier) it was not possible to incorporate
the same measures and design features that are typical in studies
of sensory-specific satiety. Nevertheless, with regard to ratings
of desire to eat, both groups experienced a greater decline in
their desire to eat the eaten relative to the uneaten foods. Impor-
tantly, a comparison across groups revealed that this decline was
greater in individuals who were not distracted during eating,
particularly when assessed 5 and 10 min after the distraction
task had terminated.
10
(a) (b) (c)
–10
–20
–30
–40
0 5 10 10
Time after eating episode (min)
005510
–50
0
Change in intensity (mm)
Fig. 4. Mean change in intensity for the bacon-flavoured corn snack (a), mandarin fruit segment (b) and Jaffa Cake (c), for the distracted (B) and the non-dis-
tracted participants (W), 0, 5 and 10 min after the eating episode in Experiment 2. For details of procedures, see p. 764. Values are means with their standard
errors depicted by vertical bars.
10
0
(a) (b) (c)
–10
–20
–30
–40
0 5 10 10
Time after eating episode (min)
005510
–50
Change in desire to eat (mm)
Fig. 5. Mean change in desire to eat the bacon-flavoured corn snack (a), mandarin fruit segment (b) and Jaffa Cake (c), for the distracted (B) and the non-dis-
tracted participants (W), 0, 5 and 10 min after the eating episode in Experiment 2. For details of procedures, see p. 764. Values are means with their standard
errors depicted by vertical bars.
Distraction and satiety 767

Typically, sensory-specific satiety is evidenced by changes
in ratings of pleasantness rather than by changes in ratings
of desire to eat. However, in the present study changes in
desire to eat were consistently correlated with changes in plea-
santness, indicating a close correspondence between both
measures. Other aspects of the data are also consistent with
the notion that food-specific satiety occurred. For example,
sensory-specific satiety appears to generalise from an eaten
food to an uneaten food that has a similar flavour (Johnson
& Vickers, 1993). In the present study a degree of generalis-
ation was also observed. Specifically, one of the uneaten
foods, the mandarin fruit segments, was similar to the Jaffa
Cakes, because both had an ‘orangey flavour’. Generally, the
pleasantness of this uneaten food was found to decline more
than the pleasantness associated with the other uneaten food
(bacon-flavoured corn snack).
Alternatively, differential effects of distraction on ratings of
pleasantness and desire to eat might indicate that distraction
influences ‘liking’ and ‘wanting’ in different ways. Despite
the fact that separate brain substrates have been identified
for each of these processes (Berridge, 1996), their independent
operation is difficult to establish using the kind of subjective
measures that were employed here (Hetherington & Rolls,
1996). Nevertheless, this general proposition represents a
potentially interesting area for future research.
General discussion
Several aspects of the findings warrant further scrutiny. First,
one possibility is that the effects of distraction extend beyond
10 min. Indeed, after a prolonged period of distraction, they
might even extend for a few hours. This possibility merits
serious consideration, because this kind of process has the
potential to reduce the time interval between meals, thereby
increasing average daily energy intake. Consistent with this
idea, observations of everyday dietary behaviour reveal that
meal frequency is often higher when meals are consumed
during distraction (Stroebele & de Castro, 2004). Furthermore,
sensory-specific satiety effects have been observed up to 2 h
after eating cessation (Weenen et al. 2005), indicating that
effects of distraction might also operate over a similar
period of time.
Second, in the present study the game ‘Pong’ was selected
because it is highly engrossing. As such, it has the capacity to
distract players away from sensations and cognitions associ-
ated with the food that they are consuming. However, an
alternative proposition is that Pong also changes levels of
stress. This is relevant, because cortisol secretion is a major
component of the stress response (Koo-Loeb et al. 2000)
and it has been implicated as a potential mediator of increased
energy intake in males (Epel et al. 2001). However, in relation
to this account of the present data, it is worth noting that more
stressful events such as a cold-stress test (placing a limb in
ice-cold water) do not appear to have a significant effect on
cortisol levels over short periods of time, such as the 5 min
required to eat the Jaffa Cakes (Gluck et al. 2004). Instead,
they tend to peak some time after (20– 40 min) the onset of
the stressor (Brandenberger et al. 1980; Holl et al. 1984).
More generally, isolating the relative effect of stress and cog-
nitive distraction has proved very difficult in the past, and
stress induction does not appear to be a reliable predictor of
overeating (Shapiro & Anderson, 2005). Furthermore, in
response to other recent findings, it has been suggested that
stress-induced eating should, in any case, be attributed to
the cognitive, distracting effects of the stressor rather than to
its negative effect on arousal or its effect on perception of
self-image (Lowe & Kral, 2006). Notwithstanding these
points, in the context of the present design, we are unable to
disambiguate the cognitive distracting and the potential
stress-inducing effect of Pong. Instead, further research is
needed in order to establish its role as a distractor with absol-
ute certainty.
Third, it has been suggested that the effects of distraction are
particularly evident in individuals who restrict their dietary
intake (Ward & Mann, 2000). By contrast, the data in the present
study indicate that the effects of distraction are general, affecting
all participants alike. In relation to this difference, it might be rel-
evant that the distractor task was extremely engrossing. In this
context, both restrained and unrestrained eaters may be dis-
tracted passively from monitoring their intake. On the other
hand, less engrossing tasks might offer greater opportunity for
varying levels of engagement (for evidence of this phenomenon
see Mitchell & Brunstrom, 2005). A further extension of the pre-
sent research might involve an assessment of individual differ-
ences under conditions where participants are exposed to
different kinds of distractor task. It might also be relevant that
we obtained a measure of dietary restraint shortly after the par-
ticipants had consumed a prescribed portion of food. Although
the DEBQ-restraint subscale has excellent test–retest reliability
(Banasiak et al. 2001), it is possible that responses are affected
when the scale is issued under these conditions. More generally,
decisions about the assessment of dietary restraint could prove to
be critical. In relation to this idea it should be noted that Ward &
Mann (2000) chose to code their restraint scores as high or low
based on a pre-specific cut-off. They also used a different
measure of restraint, and this was issued several weeks prior to
the beginning of their study.
Finally, the finding that distraction arrests a process akin to
sensory-specific satiety is highly relevant to researchers with
an interest in meal size. This is because sensory-specific satiety
is regarded as an important determinant of meal termination
(Hetherington, 1996). However, it is important to emphasise
that we have focused solely on the effects of distraction in the
context of a fixed meal, and this limits the extent to which paral-
lels can be drawn between the present findings and those from
ad libitum eating paradigms. In future, studies should seek to
consider the effects of distraction in a context that has greater
ecological validity. Here, our aim has been to explore basic prin-
ciples, and it remains to be established whether these generalise
to a normal meal. In particular, our decision to limit the eating
episode to a fixed 5 min period and to provide food in a tightly
controlled context was motivated by a concern to be able to
provide unambiguous evidence for an effect of cognitive distrac-
tion. A challenge for future studies will be to explore how these
methods can be applied to study eating behaviour under normal
conditions.
In summary, to the authors’ knowledge, the present research
represents the first of its kind to explore the effects of distrac-
tion both within a meal and over a short period after meal ter-
mination. The results indicate that distraction has a potent
effect on the changes in desire to eat that normally occur
during a meal. Moreover, these effects appear to operate in
J. M. Brunstrom and G. L. Mitchell768

a food-specific manner and they persist after the distraction
period has ended. One possibility is that distraction has a simi-
lar effect in everyday activities such as watching television
and so on. Given the continued interest in the relationship
between these behaviours and obesity in children and adoles-
cence (e.g. Hancox & Poulton, 2006), this phenomenon merits
further consideration.
References
Banasiak SJ, Wertheim EH, Koerner J & Voudouris NJ (2001) Test –
retest reliability and internal consistency of a variety of measures
of dietary restraint and body concerns in a sample of adolescent
girls. Int J Eat Disord 29, 85 – 89.
Bellisle F & Dalix AM (2001) Cognitive restraint can be offset by
distraction, leading to increased meal intake in women. Am J
Clin Nutr 74, 197 – 200.
Bellisle F, Dalix AM & Slama G (2004) Non food-related environ-
mental stimuli induce increased meal intake in healthy women:
comparison of television viewing versus listening to a recorded
story in laboratory settings. Appetite 43, 175 – 180.
Berridge KC (1996) Food reward: brain substrates of wanting and
liking. Neurosci Biobehav Rev 20, 1 – 25.
Boon B, Stroebe W, Schut H & Ijntema R (2002) Ironic processes in
the eating behaviour of restrained eaters. Br J Health Psychol 7,
1–10.
Booth DA (1981) How should questions about satiation be asked?
Appetite 2, 237 –244.
Booth DA (1994) Psychology of Nutrition, p. 67. London and Bristol,
PA: Taylor & Francis.
Brandenberger G, Follenius M, Wittersheim G, Salame P, Simeoni M
& Reinhardt B (1980) Plasma-catecholamines and pituitary-adrenal
hormones related to mental task demand under quiet and noise con-
ditions. Biol Psychol 10, 239 –252.
de Castro JM (1990) Social facilitation of duration and size but not
rate of the spontaneous meal intake of humans. Physiol Behav
47, 1129–1135.
de Castro JM (1994) Family and friends produce greater social facili-
tation of food-intake than other companions. Physiol Behav 56,
445–455.
de Castro JM & Brewer EM (1992) The amount eaten in meals by
humans is a power function of the number of people present. Phy-
siol Behav 51, 121 – 125.
Epel E, Lapidus R, McEwen B & Brownell K (2001) Stress may add
bite to appetite in women: a laboratory study of stress-induced cor-
tisol and eating behavior. Psychoneuroendocrinology 26, 37 – 49.
Epstein LH, Mitchell SL & Caggiula AR (1993) The effect of subjec-
tive and physiological arousal on dishabituation of salivation. Phy-
siol Behav 53, 593 – 597.
Epstein LH, Paluch R, Smith JD & Sayette M (1997) Allocation of
attentional resources during habituation to food cues. Psychophy-
siology 34, 59 –64.
Epstein LH, Saad FG, Giacomelli AM & Roemmich JN
(2005) Effects of allocation of attention on habituation to
olfactory and visual food stimuli in children. Physiol Behav
84, 313– 319.
Gluck ME, Geliebter A, Hung J & Yahav E (2004) Cortisol,
hunger, and desire to binge eat following a cold stress test in
obese women with binge eating disorder. Psychosom Med 66,
876–881.
Green MW & Rogers PJ (1995) Impaired cognitive-functioning
during spontaneous dieting. Psychol Med 25, 1003 – 1010.
Hancox RJ & Poulton R (2006) Watching television is associated
with childhood obesity: but is it clinically important? Int J Obes
30, 171–175.
Heatherton TF, Herman CP, Polivy J, King GA & McGree ST (1988)
The (mis)measurement of restraint - an analysis of conceptual and
psychometric issues. J Abnorm Psychol 97, 19 – 28.
Herman CP, Ostovich JM & Polivy J (1999) Effects of attentional
focus on subjective hunger ratings. Appetite 33, 181 – 193.
Hetherington MM (1996) Sensory-specific satiety and its importance
in meal termination. Neurosci Biobehav Rev 20, 113 – 117.
Hetherington MM & Rolls BJ (1996) Sensory-specific satiety: theor-
etical frameworks and central characteristics. In Why We Eat What
We Eat, pp. 267 –290 [ED Capaldi, editor]. Washington, DC:
American Psychological Association.
Holl R, Fehm HL, Voigt KH & Teller W (1984) The midday surge in
plasma-cortisol induced by mental stress. Horm Metab Res 16,
158–159.
Johnson J & Vickers Z (1993) Effects of flavor and macronutrient
composition of food servings on liking, hunger and subsequent
intake. Appetite 21, 25 –39.
Koo-Loeb JH, Costello N, Light KC & Girdler SS (2000) Women
with eating disorder tendencies display altered cardiovascular,
neuroendocrine, and psychosocial profiles. Psychosom Med 62,
539–548.
Kramer FM, Rock K & Engell D (1992) Effects of time of day and
appropriateness on food-intake and hedonic ratings at morning
and midday. Appetite 18, 1 –13.
Levitsky DA (2005) The non-regulation of food intake in humans:
hope for reversing the epidemic of obesity. Physiol Behav 86,
623–632.
Lowe MR & Kral TVE (2006) Stress-induced eating in restrained
eaters may not be caused by stress or restraint. Appetite 46, 16–21.
Mann T & Ward A (2004) To eat or not to eat: implications of the
attentional myopia model for restrained eaters. J Abnorm Psychol
113, 90–98.
Mitchell GL & Brunstrom JM (2005) Everyday dietary behaviour and
the relationship between attention and meal size. Appetite 45,
344–355.
Rolls BJ (1986) Sensory-specific satiety. Nutr Rev 44, 93 – 101.
Rolls BJ, Hetherington M & Burley VJ (1988) Sensory stimulation
and energy density in the development of satiety. Physiol Behav
44, 727–733.
Shapiro JR & Anderson DA (2005) Counterregulatory eating beha-
vior in multiple item test meals. Eat Behav 6, 169–178.
Smith GP (1996) The direct and indirect controls of meal size. Neuro-
sci Biobehav Rev 20 , 41 –46.
Stroebele N & de Castro JM (2004) Television viewing is associated
with an increase in meal frequency in humans. Appetite 42, 111 – 113.
Stunkard AJ & Messick S (1985) The three-factor eating question-
naire to measure dietary restraint, disinhibition and hunger. J Psy-
chosom Res 29, 71 –83.
Tuomisto T, Tuomisto MT, Hetherington M & Lappalainen R (1998)
Reasons for initiation and cessation of eating in obese men and
women and the affective consequences of eating in everyday situ-
ations. Appetite 30, 211 –222.
Van Strien T, Frijters JER, Vanstaveren WA, Defares PB &
Deurenberg P (1986) The predictive-validity of the Dutch
Restrained Eating Scale. Int J Eat Disord 5, 747 – 755.
Wansink B & Park SB (2001) At the movies: how external cues and per-
ceived taste impact consumption volume. Food Qual Pref 12, 69 –74.
Ward A & Mann T (2000) Don’t mind if I do: disinhibited eating
under cognitive load. J Pers Soc Psychol 78, 753 – 763.
Weenen H, Stafleu A & de Graaf C (2005) Dynamic aspects of liking:
post-prandial persistence of sensory specific satiety. Food Qual
Pref 16, 528 –535.
Westenhoefer J, Broeckmann P, Munch AK & Pudel V (1994) Cog-
nitive control of eating behavior and the disinhibition effect. Appe-
tite 23, 27 –41.
Wisniewski L, Epstein LH & Caggiula AR (1992) Effect of food change
on consumption, hedonics, and salivation. Physiol Behav 52, 21–26.
Distraction and satiety 769