CURRENT TOPICS IN NUTRACEUTICAL RESEARCH Vol. 9, No. 1, pp. 41-45, 2011
ISSN 1540 -7535 print, Copyright © 2011 by New Century Health P ublishers, LLC
w ww .n e w c e nt ur y h e al t h p u bl i sh e rs. c o m
All rights of reproduction in any form reserved
SNACKING FREQUENCY, MEN TAL HEALTH, HEALTH BELIEFS AND PHYSICAL HEALTH
Katherine S. Chaplin and Andrew P. Smith
Centre for Occupationa l and Health Ps ychology, Schoo l of Psychology, Cardif f University,
P O Box 901, Cardiff, CF10 3AS, UK.
[Receive d Apri l 4, 2011; Accepte d May 9, 2011]
ABSTRA CT: The aims of the current study were to examine
association s between snackin g frequency and mental health,
healt h related behaviors , health belief s and physica l health.
Snackin g frequency per s e was examined in additio n to
frequency of consumption of specific snack items. One hundred
and thirty six participants (96 females, 40 males, mean age
37 years), including both students and members of the general
public took part in the study. Participants completed a variety
of psychosocial questionnaires and a snacking questionnaire.
Participants who had a snack everyday reported significantly
less depression than those wh o did not. Snacking frequency
positivel y correlate d wit h c h o c o la t e, cris p and biscuit
consumption. Participants who consumed snacks like crisps and
chocolat e reported greater concern s about their current and
future health status and also had less motivation to be healthy.
No difference s were foun d wit h respect to physica l health.
Participants who ate snacks such as chocolate and crisps ate a
lot of fried foods and had more worries about their health. The
participants who consumed snacks such as fruit and yoghurt
had better diets and were optimistic about their future health.
It is not possible to determine from the present results whether
eatin g habit deter m ines health concern s or whether health
determines eating habits.
KE Y W ORDS : Health B eliefs , M ental Health, P hysica l
Health, Snackin g Frequency
Cor r esponding Autho r: Pro f. And rew P. Smith, Centre for
Occupational and Health Ps y cholog y, School of Ps y cholog y,
Cardiff U ni versit y, P.O. Bo x. 901, Cardiff CF11 3YG, UK;
Fa x: +44 29 20874758; E- mail: S mithAP@Cardiff .ac.uk
INT RODUCTION
Snacking or a grazing pattern of eating is becoming more
popular as opposed to eating three straight meals a da y. There
is still a lack of agreement on a definition o f snacking however
this has been addressed in our previous research (Chaplin and
Smith, in press) and for the purposes of this study snacking
wil l be defined as food or drin k eaten betwee n main meal
times (unless stated otherwise). Regardless of how snacking is
defined it contributes si gni ficant l y t o a n i n d i v i d u a l ’s
nutritional intake (Gregory et al, 1990; 1995). It is believed
that snacks contribute 15- 20% of our daily energy intake
(Summerbell et al., 1995), 15-20% of our daily mineral intake
and 13- 17% of our daily vitamin intake.
So me researcher s have argued that snacks provide
empty and extra calories which are a predisposition for obesity
(Booth, 1988). These researchers believe that the current rise
in obesity is related to the rise in snacking. However many of
these claims are based on anecdotal assumptions. Scientific
studies have provided evidence that frequent eating episodes
(snacking) are actually associated with lower bo d y weight than
eating fewer meals per day (Fabry et al., 1964; Fabry et al.,
1966).
Metzner et al. (1977) examined eatin g frequency
and adiposity and found an inverse relationship between the
two. The par ticipants consuming 6 mea ls per day were
significantly thinner than those who ate 2 meals per day were.
Edelstein et al. (1992) foun d a significant differenc e with
respect to wais t hip ratio ; aga in those participants wit h a
higher eating frequency had a lower waist hip ratio than those
with a lower eating frequency.
A study of 7147 adults from America is one of a few
studies to examine the long-term effects of eating frequency
on body weight (Kant , 1995). Baselin e data were collected
between 1971 and 1975 and follo w up data betwee n 1982
and 1984. The out come measures w ere body weight change,
B M I and skin-fold thickness. At baseline significant differences
were found for all thre e of the outcome measure s between
participants who ate 2 meals and fewer and participants who
ate 7 meals or more. However no significant differences were
found for any of the outcome measures at the follow up. It is
not possible however to conclude that eating frequency has
no effect on body weight due to a number of methodological
problems with the stud y.
It has been suggested that there is an issue of causalit y.
2
Snacking frequency and health
Some researchers argue that a decrease in meal frequency is a
result of increased body weight as opposed to the cause. Many
people will skip a meal in an attempt to lo se weight. Another
possible suggestion is dietary induced thermogenesis (DIT);
this is a measure of the heat energy which is produced through
digestion an d metabolis m o f f o o d. H o we v e r, studies
comparing energy expenditure on nibbling and gorging diets
have failed to find any significant differences ( Verboeket-van
de Venne and Westerterp, 1991; 1993, V erboeket - va n de
Venne et al., 1993). These findings suggest that DIT does not
seem to be affected by the frequency o f eating occasions.
Frequent snacking appears to be less efficient than eating
the same number o f calories in larger meals and therefore
tend s to keep the metabolic rate higher and result s in an
increase of total daily expenditure (Drummond et al., 1995).
H owev er m o re researc h is neede d in this are a du e to
conflicting evidence (Belko and Barbieri, 1987; Kinabo and
Durnin , 1990).
It is proposed that those individuals that snack frequently
are mo re p h y sical l y a c t ive . Frequen t sna ck e rs a r e mo r e
physically active which in turn leads to an increase in energy
expenditure , this in turn acts as an appetite stimulant and
they snack more (Drummond et al., 1995).
The current literature has mainly considered physiology and
physical health wit h respect to the relationship between
snackin g frequency and bod y weight / obesit y. The current
stud y aims to examine any association between snacking
frequency and consumption of specific snack foods and mental
health. Th is is bas ed on a methodolog y used to examine
correlation s between breakfas t consumption, menta l health,
physica l healt h and healt h related behavior s (Smith , 1998;
1999; 2003). These studies found that regular breakfast cereal
consumptio n was associat ed with les s stress, depressio n and
emotional distress . Smit h (1999; 2003) als o found regular
breakfast cerea l consumptio n was associated with fewer
physical symptoms. It is of interest whether the same findings
can be obtained between snacking frequency, mental health,
physical health and health related behaviors.
On e study has considere d snacking and menta l health
(Smith, submitted). P articipants who snacked ev er y day scored
lowe r on a ll m e a s u r es of mental health (stress, depression,
emotional distress and anxiety) than those participants who
reported never snacking. However none of these differences
were significant. This result was found in four samples: one
from the general population (aged 20- 60); one from late teens
living at home; one from a student population (aged 18-30)
and one fro m an elderly samp le (aged 65+). Although no
significant benefits of snacking were found no negative effects
of snackin g were identified whic h su ggest s that snacking
should not be avoided. A number of suggestions have been
made about wh y no effect was found after a mid morning
snack. There are questions about whether the type of snack,
the frequency of snacking and the timing of snacks produce
different effects on mental health.
The current study aimed to examine snackin g frequenc y
and physica l health , mental health, healt h relate d behavior s
and health beliefs. In addition to snacking frequency per se
frequency o f consumption of specific snacks was considered.
METHOD
P a r ticipants.
A tota l of 136 participant s were recruit ed fro m two
populations: 82 student s and 55 members of the general
public. The participants consisted of 96 females and 40 males.
The mean age was 37 yea rs (age range was 17 -80 years).
Students were either contacted via a student participant
database and were sent the materialin the post or they were
r e c r uite d fro m Ca r dif f Uni v ersit y t h roug h poster
adv er t i s e m e n t s . The members of the genera l public were
recruited from the general public participant database. They
we re sent the materia l to complete throug h the post. All
participants were paid for pa rticipating in this study.
Procedure.
Par ticipant s from t he databases wer e sent a letter and
informatio n sheet detailin g the stud y and a consent form.
Participants w ere requested to complete the consent form and
a variety o f psyc h osoc ial q ue stion na ir es in additio n to a
snackin g questionnaire and retu rn th e m in the freepost
envelope p rovided. Participants recruited through the poster
campaig n came to the laboratory and were give n the same
informatio n sheet detailin g the study and consent form.
Pa rticipants completed the questionnai res in the laboratory.
Snack food consumption.
Overall snack food consumptio n on a weekly basis was
measured using a 5 point likert scale where 0 = never and 4 =
ever yda y.
Consumption of 20 specifi c snack foods wer e measure d u sin g
a 7 poin t likert scale whe re 0 = ne v er and 6 = 3 to 4 times a da y.
Consumption of each item wa s categorised into thre e groups:
(1)
The non-consumer – neverconsumed the snack.
(2)
The weekly consume r – c o n s u m e d the snack b et w e e n
once to four times a week.
(3)
The daily consumer – consumed the snack between
once to four times per day.
Mental health.
The following questionnaires were used to measure mental
health:
Stress:
P erceive d S tres s Sca le (Cohen a n d
Williamson, 1988)
Depression :
The Hospital Anxiety and D e p r e ssi o n
Scale (Zigmond and Snaith, 1983)
Emotional Distress: Emotional distress scale of P rofile of
Fatigue Related States (Ray et al.,
1992)
Anxiety :
S p i e l b e r g e r Trait - St a t e Anx ie t y
Inventory (Spielberger et al., 1971)
Snacking frequency and health 3
Health related behaviors (Cohen et al., 1991).
Par ticipant s wer e aske d about their dietar y habits (for
example how often they ate fried food), these were measured
usin g a 5 poin t likert sc a le ranging from 0 (neve r) to 4
(every day). Pa rticipants also recorded their consumption of a
variety of food items (relating to both meal and snackfoods)
using a 6 point likert scale (1 = never and 6 = more than once
a da y) . Informatio n wa s als o collecte d abou t alcohol
consumption , smokin g a n d activitie s an d exercis e by
questionnaire (based on that used by Cohen et al., 1991).
Health attitudes and beliefs.
Participants beliefs about their health status was measured
using the Health Orientation Scale (Snell et al, 1991). This is
a validated questionnaire which provides informatio n on a
variety o f health related beliefs.
Physical health.
A single measure of physical health was used. This was the
Cohen-Hobermann index of physical symptoms (Cohen and
Hobermann, 1983). This is a validated measuring instrument
of physical health.
one way Analysis of Variance rev ealed a significant diffe rence
between snacking frequencies (F (3, 136) = 3.91, p =.01). A
significan t differenc e was found betwee n participants who
snacke d most days and participants who snacke d everyday
(t(93) = 3.35, p = .001); participants who snacked once to
twice p e r week and participants who snacked everyday ( t(69)
= 2.74, p<.01) and participants who snacked never/less than
once a week and participants who snacked everyday (t(58) =
2.21, p<.05). The participants wh o snacke d ev ery da y were
less likely to feel depressed.
Greater frequency of biscuit consumption was negatively
associated with sco res on the State – Trait Anxiety Inventory
(r = - .174, p<.05) , with participants wh o frequently ate
biscuits reporting lo wer levels of anxiety.
No significan t correlation s were foun d with respect to
emotional distres s and stres s for snacking frequency o r
frequency of consumption of specific snacks. Although some
significan t correlations were found there are no more than
would be expected by chance.
Correlations between snacking frequency and health related
behaviors.
Increased snacking on a daily basis was positively correlated
with cris p (r = .223, p<.01), sweet / chocolate ( r = .453,
Statistical analysis.
Pearso n’s product moment correlation coefficient was used
p<.001), biscuit (r = .377, p<.001), cake (r = .240, p = .005)
to asses s any associations between snackin g frequency and
and dairy bas ed puddin g consumptio n (r = .253 , p < . 0 0 5 ) .
health beliefs, physical health and mental health. A one -way
In addition snacking frequency on a daily basis was negatively
associated with green vegetab l e consumption ( r = - . 1 8 7 ,
ANO VA a nd independen t t -tests we re used to assess any
p<.05). Increased snackin g was negativel y associate d wi t h
differences between snackin g frequency and menta l health
howoften participants consumed alcohol (r = -.209, p <.05).
measures.
Pa rticipants who ate fruit as a snack were more likely to eat
yoghurt (r = .211, p<.05) , breakfas t cereal (r = .222, p =
RESULT S
.01), salad / raw vegetables (r = .342, p<.001) and cheese (r
= .190, P<.05). The participants who ate cake as a snack were
Correlations between snacking frequency and mental health.
more likely to eat crisps (r = .222, p =.01), biscuits (r = .312,
Table 1 sho ws par ticipants scores on the menta l health
p<.001) , chocolate (r = .207, p<.05) and frie d food s ( r =
measures based on their snacking habits.
.265, p<.01). The majority o f this
a n a ly s is involv ed c o r r e l a t i n g
Table 1. Snacking frequency and mental health scores. (Higher scores = mo re mental health general snackin g wit h specifi c
problems).
snack types and th is should be
Snackin g Frequency
taken into consideration when
Most days
Eve ryday
Never/less than
Once or
considering the findings.
once a week
twice a week
N=
15
26
50
45
Co rrela tio n s b e twe e n s n a c k i n g
Mean s. e
Mean s. e
Mean s. e
Mean s. e
frequency and health beliefs.
Emotio na l Distress 26.73 (2.54)
31.35 (2.63) 33.6 (2.40) 31.36 (2.79)
No sign if ic an t c o r relations
State Anxiety
33.87 (2.51)
37.19 (1.49) 37.96 (1.32) 34.93 (1.86)
were foun d betwee n snacking
Percei v ed Stres s
19.54 (1.80)
22.65 (1.55) 22.52 (1.11) 19.64 (1.38)
frequency (both on a daily and
HAD- Depression
10.6
(0.79)
10.65 (0.54) 10.88 (0.42) 9.13 (0.28)
weekly basis) and health beliefs.
Motivation
to
avoid
Two significant correlations were found between snacking
unhealthiness (MUNH) and motivation for health (MH) were
frequency and measures of mental health. Increased snacking
foun d to negativel y correlat e wit h th e frequenc y of
on a weekly b a s is wa s negativel y a s s o c i a t ed wit h lower
consumption of chips (MUNH r = - .276, p = .001; MH r =
pa r t i c i p a n t s ’ sco res on the depressio n component of the
-.265; p = .002) and soft drinks (MUNH r = -.293, p = .001;
MH r = -.269, p = .002). Health -esteem confidence (HEC)
Hospital Anxiety and Depression Index(r = -.181, p< .05). A
4
Snacking frequency and health
was found to be negatively correlated with the frequency of
chip (r = - .176, p < .05) consumption. Health anxiety (HAX)
positivel y correlated wit h frequency o f co n sum pti o n of biscuits
(r = .212, p <.02). Health expectations and optimism (HEO)
was found to negativel y correlat e wit h frequency of eating
chip s (r = -.184, p <.05). In additio n to this frequency of
crisp consumption was also found to negatively correlate with
health internal control (HILC) (r = - .199, p = .021) and health
status (HS) (r = -.197, p =.023).
In contras t breakfas t frequenc y was found to positively
correlate with health -esteem confidence (HEC) (r = .271, p
= .001) , motivatio n to avoid unhealthines s (MUNH ) (r =
.216, p = .012) and motivation for health (MH) (r = .223, p
< .01). Fruit frequency was found to positively correlate with
motivation to avoid unhealthiness (MUNH) (r = .185, p <.05),
motivation for health (MH) (r = .246, p < .01) and health
expectations and optimism (HEO) (r = .206, p = .018).
Snacking frequency and physical health.
In contrast no significant differences were found between
physic a l healt h and consumption of specifi c snac k foods.
Pa rticipants who snacked more frequently on foods which are
traditionally viewed as bein g unhealthy held more negative
beliefs about their health, however their health status was the
same as participants who snacked on typically healthy foods.
DISCUSSIO N
Increased snackin g was positivel y associated with greater
consumptio n of chocolate , crisp s and biscuits ; however’
snacking was negatively associated with drinking habits. This
would appear to support the claims that snacking relates to
unhealthy foods which provide empty calories (Booth, 1988).
H o we ve r t h o s e participants w h o reported eating fruit as snacks
also ate mo re yoghurt , salad s and ra w vegetables . It would
appear tha t there are two distinc t groups of snackers ; those
that snack healthy and t h o s e that do not. This difference could
be due to differences in definitions o f snacking however it is
more likely that these differences a re the result of differing
attitudes and beliefs about health status.
Snackin g frequenc y per se was not found to correlate with
any of the healt h attitudes and beliefs . Participants who
snacke d more frequently on “health y ” snack foods, for example
breakfas t cerea l a nd fruit , ha d h ig h e r health - esteem
confidence, motivation to avoid unhealthiness, motivation to
remain healthy and higher health expectations and optimism
for th e ir futur e healt h status . Greater consumptio n of
“unhealthy” snacks, for example crisps, biscuits, chips and soft
drinks, was associated with negative feelings or worries about
current and future healt h status . T h is would sugges t that
participants do not view frequent snacking as enough to cause
worry about their health however the types of snacks which
they eat can. Participants seem awa re of the disadvantages of
snackin g on foods such as crisps and chocolate which are
believed to be unhealthy and the advantages of snacking on
foodssuch as fruit and breakfast cereal which are accepted as
being healthier foods. This may explain w h y so m e participants
wh o snack frequently hav e a lowe r body weight than those
who do not (F ab ry et al., 1964; Fabry et al., 1966). However
at this stage it is not possible to say whether pa rticipant s ’ beliefs
about their health status are determined by the snack foods
they eat or whether snack food choice is determined by their
beliefs about their health.
Snackin g frequenc y showe d little associatio n wit h mental
healt h wit h th e exceptio n of depression . However it is
important to note that the number of significant correlations
found is no mo re than would be expected by chance . This
supports earlier research done by Smith (in press submitted)
that also found no relationship between frequency of
c o n su m in g snacks and mental health. In addition the
current stud y found no significant associations between
consumption of specific snack foods and mental health
measures. Similarly to the Smith study no negativ e effect s of
snackin g were identified for mental health which suggests
that snacking should not be avoided.
The current result s found no differences between t h e
participants with respect to actual health status. No negative
effect s of snacking w e re identifie d with respect to physical
healt h which again suggest s that snackin g shou ld not be
avoided. Only one measure o f current health was used in this
study and more research is needed to further examin e the
relationship between frequency of snacking and a variety of
measure s of physical healt h to try and establis h whether
snacking displays any of the negative effects that participants
reporte d worryin g about.
Thes e findings are in contras t to the results of previous
studies looking at breakfast cereal consumption. Smith (1998;
1999; 2003) identifie d significan t differences between daily
breakfast cereal consumers, regular breakfast cereal consumers
and irregular breakfas t cerea l consumers for all measures of
mental and physica l health. For each outco me measure the
daily breakfas t cereal consumers reported the bes t health.
Smit h (1999) took th is a stag e further and foun d tha t the
same pattern of result s was found betwee n daily cornflake
consumers , o cca si on a l cornflake consumers and irregular
cornflake consumers . It wou ld appea r that breakfas t cereal
consumption and snackin g have differing effect s on both
menta l an d physical health. Based o n th is it would be
interesting to know whether different measures of mental and
physical health to the ones used in the breakfast research would
identify any relationship with snacking frequency.
A CKNO WLEDGEMEN T:
The researc h was supported by fro m a Ph D studentship
from the BBS RC and industrial sponsorship from Kellogg ’s
Company Ltd.
REFERENCES
Belko A Z and Barbieri T E (1987) “Effe ct of meal size and
frequency on thermic effect of food”, Nutrition Research, 7 ,
237- 342.
Snacking frequency and health 5
Booth DA (1988) “Mechanisms from models – actual effects
from re a l life: the zero -calorie drink-break optio n”, Appetite,
11(Supplement) , 94- 102.
Chaplin KS and Smith AP (in press) “Definitions and
perceptions of
snacking” , Current Topics in
Nutraceutical Research.
Cohen S and W illiamson G (1988) “Pe rcei v ed s tress in a
probability sample of the United St at es ”. In: S pacapan, S. and
Oska mp, S. , eds, Th e Social Ps ycholog y o f Health (Ne wbur y
Park, CA: S age), pp. 31-67.
Cohen S, Tyre ll DA J and Smith AP (1991 ) “Ps y chological
stress and susceptibility to the common cold ”, New England
Jo u r nal o f Medicine, 325, 606-612.
Drummond S, Kirk T and de Looy A (1995) “S n a c k i n g :
Implications in body compositio n and energy balanc e”, British
Food Jou r nal, 97, 12- 15.
Edelstein S L, Barrett- Conner EL, Wingard D L and Cohn
BA (1992 ) “Increas e d meal frequency a s s o c ia t ed with
d ecr ea se d cholestero l concentrations; Rancho Bernardo,
CA , 1984- 1987” , A meric a n Jou r n a l o f Clinical Nutrition,
55 , 664- 669.
Fabr y P, Fodor J, Hejl Z, Braun T and Zvolankova K (1964)
“ Th e frequency o f meals : it s relatio n to overweight,
hypercholesteremia and decreased glucose tolerance”, Lancet,
2, 614- 615.
Fabr y P, Hejda S, Cerny K, Osancova K and Pechar J (1966)
“Effect of meal frequency in schoolchildren. Changes in the
weight height proportion and skinfold thicknes s”, A merican
Jo u r nal o f Clinica l Nutrition,18, 358- 361.
Gr ego ry J, Collins DL, Davies PSW, Hughes JM and Cla rke
PC (1995) Nationa l D iet and Nutritio n S u r v ey : Child ren
Aged 1½ to 4½ Years. Vol 1, Report of the Diet and Nutrition
Su rv ey. (London: H. M. Statione ry Office).
Gregor y J, Foster K, Tyle r H and Wisema n M (1990) The
Dieta ry and Nutritional S u rvey of British Adults. (London:
H. M . Stationery Office).
Kant AK (1995) “Frequency of eating occasions and weight
chan ge in NHANES I e pi d e mi o lo g i c follow- u p stud y”,
Internationa l Jo u r na l o f Obesit y, 19, 468- 474.
Kinabo J LD and Durnin JVGA (1990) “Effect of meal
frequency on the thermic effect of food in w o m en”, E uro pean
Jo u r nal o f Clinica l Nutrition, 44, 389- 395.
Metzne r H L, Lamphiear D E, Wheele r N C and Lar kin FA
(1977) “The relationship between the frequency of eating
and adiposit y in adult men and wome n in the Tecumseh
Communit y Healt h Study”, A meric a n Jo u r n a l of Clinical
Nutrition , 30, 712-715.
6 Snacking frequency and health
Ray C , W e ir W RC, P hillip s C an d Culle n S (1992)
“Development of a measure of symptoms in chronic fatigue
syndrome: the profile of fatigue related symptoms (PFRS)”,
Psychology and Health, 7, 27- 43.
Smith AP (1998) “Breakfast and mental health ”, I nter national
Jou r nal of Food Sciences and Nutrition, 49, 397-402.
Smith AP (1999) “Breakfast cereal consumptio n and subjective
reports of health”, Inter national Jo u r nal of Food Sciences and
Nutrition , 50, 445- 449.
Smith AP (2003) “Breakfast cereal consumptio n and subjective
reports of health by young adult s”, Nutritional N euroscience,
6, 59-61.
Smit h AP (in press) “ S nac kin g hab it, m en tal hea lt h a n d
cognitive performance” , Current Top ics in N u t r a c e u t i c a l
Research.
S p ie lb e r g e r CD , Gorsuc h RL a nd Lushen e RE (1970)
“Manual for the State-trait Anxiet y Inventor y”. (Pa lo Alto:
Consulting Ps ychologists Press).
S ummerbell CD, M o o d y RC, S hanks J, Stock M J and Geissler
C (1995) “Source s of energ y fro m meals vs snacks in 220
people in four age groups ”, E uro pe an Jour n a l of Clinical
Nutrition, 49, 33- 41.
Verboeket -v an de Venne WPHG, Westerterp KR and Kester
ADM (1993) “Effect of the pattern of food intake on human
energy metabolis m”, British Journal of Nutrition, 70, 103- 115.
Verboeket -v an de Venne WPHG and Westerterp KR (1993)
“ Fr e q u e n c y o f feeding , w eigh t reductio n a nd energy
metabolis m”, Internation a l Journa l o f Obesit y, 17, 31-36.
Verboeket -v an de Venne WPHG and Westerterp KR (1991)
“Influence of feeding frequency on nutrient utilisation in man:
consequences for energ y metabolis m” , E ur opea n Jo u r nal of
Clinica l Nutrition, 45, 161- 169.
Zigmond AS and Snaith RP (1983) “The hospital anxiety and
dep ressio n scal e”, Acta P sychiatrica Scandinavica, 67, 361370.