Information about http://www.commercialalert.org/chamberlaintoys.pdf
ARTICLE …
Tags: baseline, cali time, chamberlain, elementary schools, food drinks, main outcome measures, media expo, media exposure, ms thomas, objective, participants, prospective analysis, relationship, robinson md, screen time, spear, tv view, tv viewing, variables, yun wang,Pages: 6
Language: english
Created: Thu Mar 23 08:26:52 2006
Display cached document
Page 1
Page 2
Page 3
Page 4
Page 5
Page 6
ARTICLE
Does Children's Screen Time Predict Requests
for Advertised Products?
Cross-sectional and Prospective Analyses
Lisa J. Chamberlain, MD, MPH; Yun Wang, MS; Thomas N. Robinson, MD, MPH
Objective: To examine children's screen media expo- man r=0.16 [TV viewing] and r=0.18 [total screen time];
sure and requests for advertised toys and food/drinks. both P .001). In prospective analysis, children's screen
media time at baseline was significantly associated with
Design: Prospective cohort study. their mean number of toy requests 7 to 20 months later
(Spearman r=0.21 [TV viewing] and r=0.24 [total screen
Setting: Twelve elementary schools in northern Cali- time]; both P .001) and foods/drinks requests (Spear-
fornia. man r=0.14 [TV viewing] and r=0.16 [total screen time];
both P .01). After adjusting for baseline requests and
Participants: Eight hundred twenty-seven third grade sociodemographic variables, the relationship between
children participated at baseline; 386 students in 6 schools screen media exposure and future requests for adver-
were followed up for 20 months. tised foods/drinks remained significant for total TV view-
ing and total screen media exposure. The relationship with
Intervention: None. future requests for toys remained significant for total
screen media exposure.
Main Outcome Measures: Child self-reported re-
quests for advertised toys and foods/drinks. Conclusions: Screen media exposure is a prospective risk
factor for children's requests for advertised products. Fu-
Results: At baseline, children's screen media time was ture experimental studies on children's health- and con-
significantly associated with concurrent requests for ad- sumer-related outcomes are warranted.
vertised toys (Spearman r=0.15 [TV viewing] and r=0.20
[total screen time]; both P .001) and foods/drinks (Spear- Arch Pediatr Adolesc Med. 2006;160:363-368
C
HILDREN ARE IMMERSED IN advertised products.2,4-15 These influences
advertising in media. A re- start very young. In 1 experimental trial,
cent national survey preschool children demonstrated prefer-
found that the average ences for foods seen in ads only briefly.12
child in the United States In another study, 63% of Latino preschool-
spends 6 and a half hours a day using me- ers had asked for a toy seen on TV in the
dia, more than any other waking-time ac- previous 2 weeks and 55% had made a re-
tivity, and the majority of media time is quest for advertised food or drinks.4 Since
spent watching television.1 In the new era parents control family budgets, child re-
of expanded media channel options, niche quests are important forces for family
programming and advertising have singled spending and may negatively impact inter-
out children as a targeted audience.2 The actions between parents and children.2,4,10
average US child will see more than 40 000 In addition, in a small-scale randomized
television commercials a year3 in addi- controlled trial, a screen media reduction
Author Affiliations: Division of tion to frequent product placements in tele- curriculum for third and fourth graders re-
General Pediatrics, Department vision programs and movies. sulted in children making fewer toy pur-
of Pediatrics (Drs Chamberlain Advertising to children is used to change chase requests compared with controls.16
and Robinson) and Stanford
their preferences for advertised products Much of the past research exploring the
Prevention Research Center,
Department of Medicine and to change their requests to parents for effects of television advertising on chil-
(Ms Wang and Dr Robinson), such items. Consistently, studies have dren was conducted in the 1970s. Re-
Stanford University School of shown that advertising influences chil- cently, this area of research has become
Medicine, Palo Alto, Calif. dren's preferences, choices, and requests for more relevant to children's health, with evi-
(REPRINTED) ARCH PEDIATR ADOLESC MED/ VOL 160, APR 2006 WWW.ARCHPEDIATRICS.COM
363
Downloaded from www.archpediatrics.com on April 6, 2006
©2006 American Medical Association. All rights reserved.
dence that food advertising is playing a role in the child- day" and "last Saturday" on the first assessment day and "yes-
hood obesity epidemic.14,17 As the home screen media en- terday" on the second assessment day. Prior to reading these
vironment, children's media use, and advertising have items, the research staff led children through several partici-
changed, new data are needed to explore the role of ad- patory time-estimating exercises.18 This instrument was adapted
from a similar instrument previously used in young adoles-
vertising in today's children's health and behaviors. Pro-
cents with high test-retest reliability (r=0.94).19 For purposes
spective studies are also needed to examine the temporal of this analysis, children's screen media use reports derived 4
relationships between screen media exposure and poten- variables: average weekday television viewing; average Satur-
tial outcomes to establish media as a risk factor for pur- day television viewing; total weekly TV viewing (the sum of
chase requests. Therefore, in a large, ethnically and so- 5 the average weekday viewing and 2 Saturday viewing);
cioeconomically diverse sample, we conducted a and weekly total screen time (the sum of 5 average weekday
prospective study of preadolescents' screen media expo- TV, movies or videos, and video games use and 2 Saturday
sures and their concurrent and future requests for adver- TV, movies or videos, and video games use). Children also re-
tised toys and foods/drinks. We hypothesized that chil- ported if there was a TV set in the room where they regularly
dren who reported more screen media exposure at the slept.
beginning of third grade would report more requests for
advertised toys and food and drinks, both then and dur- Purchase Requests
ing the subsequent 20 months. Establishing screen me-
Children were asked, "In the past week, have you asked some-
dia exposure as a risk factor for purchase requests would
one to buy you any foods or drinks that you have seen on TV?"
provide stronger support for further efforts to reduce screen and "In the past week, have you asked someone to buy you any
time and/or advertising exposure as potential clinical and toys that you have seen on TV?" For each question, those who
public health intervention strategies. responded yes were asked to write the names of up to 4 spe-
cific items they had requested and the responses were scored
METHODS from 0 to 4 for analysis.
All third grade students from 12 public elementary schools in STATISTICAL ANALYSIS
northern California and their parents were eligible to partici-
pate in a randomized controlled trial of obesity prevention (6 The baseline sample for cross-sectional analysis included chil-
schools) and smoking prevention (6 schools). Assessments were dren from all 12 schools. Because the obesity-prevention in-
performed by trained research staff at baseline (September- tervention included screen time reduction, only participants
October 1999), in the spring of third grade (April-May 2000), from the 6 schools randomly assigned to smoking prevention
in the fall of fourth grade (September-October, 2000), and in were included in the prospective analysis sample.
the spring of fourth grade, approximately 20 months after base- To test our hypothesis, we used nonparametric Spearman
line (April-May 2001). At each point, children completed self- rank correlations. First, we examined baseline (fall, third grade)
report questionnaires on 2 non-Monday weekdays. A research screen exposure variables and baseline requests first for toys
staff member read each question out loud. Classroom teachers and then baseline requests for foods/drinks. Then we exam-
did not participate in the assessments. Parents were inter- ined the prospective sample for correlations between baseline
viewed by telephone at baseline by trained interviewers fol- screen exposure and the mean number of requests for toys and
lowing a standardized protocol. foods/drinks over the subsequent assessments (mean of spring
Tracking and confidentiality were maintained by using third grade, fall fourth grade, and spring of fourth grade).
unique identification numbers. Parents or guardians provided We performed multivariate linear regression to further ex-
passive informed consent for their children to participate in each amine these relationships, adjusting for covariates. First, all vari-
assessment. During classroom assessments, research assis- ables were centered20 (for example, boys were coded 0.5 and
tants acquired verbal assent for participation from each child. girls were coded -0.5). In model 1, we regressed follow-up re-
Children who declined to participate did another nonpunitive quests on screen media exposure, baseline requests, and their
activity with their classroom teacher. Parents gave active ver- interaction. In model 2, we additionally adjusted for sex and
bal consent for their own participation in telephone inter- ethnicity and all first-order, 2-way interactions. In model 3, us-
views. The study was approved by the participating school dis- ing the subset of participants with parent interviews, we fur-
tricts and the Stanford University Panel for the Protection of ther adjusted for parent education, marital status, and lan-
Human Subjects in Medical Research (Stanford, Calif ). guage spoken at home and all first-order, 2-way interactions.
Statistical significance was defined at a 2-tailed =0.05. Analy-
sis was performed with SAS for Windows, version 9.1 (SAS In-
MEASURES stitute Inc, Cary, NC).
Demographics RESULTS
Children reported their date of birth and sex. Ethnicity was ob-
tained from school district records. Parents reported the main
Of 860 children enrolled in the 12 elementary schools
language spoken at home, their marital status, and the highest at baseline, 836 (97%) completed surveys. Nonpartici-
level of education completed by parents or guardians. pants included 22 parent refusals, 1 child unable to par-
ticipate owing to cognitive difficulties, and 1 child who
Media Use and Home Media Environment was absent on all assessment days. We randomly chose
1 child for analysis from each of 9 sibling sets to avoid
Children reported the time they spent "watching television," violating the assumption of independence, leaving a fi-
"watching movies or videos on a VCR," and "playing video nal baseline analysis sample of 827 children. Of the 827
games" separately for before school and after school "yester- students, 410 were enrolled in 1 of the 6 smoking-
(REPRINTED) ARCH PEDIATR ADOLESC MED/ VOL 160, APR 2006 WWW.ARCHPEDIATRICS.COM
364
Downloaded from www.archpediatrics.com on April 6, 2006
©2006 American Medical Association. All rights reserved.
Table 1. Participant Characteristics at Baseline Table 2. Cross-sectional Association Between Screen Media
Exposure and Requests for Toys and Food/Drinks*
Baseline Prospective
Sample Sample Toy Requests Food/Drink Requests
(12 Schools) (6 Schools)
Weekly TV exposure 0.15 0.16
Sample size 827 392 Weekday TV exposure 0.16 0.17
Boys, % 48 47 Saturday TV exposure 0.10 0.09§
Age, y, mean (SD) 8.45 (0.38) 8.45 (0.39) Weekly total screen time 0.20 0.18
Race/ethnicity, %
Hispanic/Latino 35 39 *Values are expressed as nonparametric Spearman rank correlation
Filipino 27 25 coefficients.
White 17 15 P .001.
Asian/Pacific Islander 15 13 P .01.
African American 4 4 §P .05.
Other 2 4
Primary language spoken at home, %*
English 72 66
Spanish 20 25 Table 3. Prospective Association Between Baseline Screen
Other 8 9 Media Exposure and Subsequent Requests for Toys
Marital status, %* and Foods/Drinks*
Married 74 71
Separated/divorced 17 20
Toy Requests Food/Drink Requests
Single, never married 8 9
Widowed 1 1 Weekly TV exposure 0.21 0.13
Maximum parent education level, %* Weekday TV exposure 0.19 0.15§
High school 8 10 Saturday TV exposure 0.19 0.06
High school graduate or GED 23 23 Weekly total screen time 0.22 0.16§
Some college/technical school 33 31
4-year college graduate 28 29 *Values are expressed as nonparametric Spearman rank correlation
Some/graduate degree 8 7 coefficients.
Screen media use, h, mean (SD) P .001.
Weekly TV exposure 10.8 (10.8) 11.3 (11.2) P .05.
Average weekday TV exposure 1.6 (1.7) 1.7 (1.8) §P .01.
Saturday TV exposure 1.4 (1.8) 1.4 (1.8)
Weekly total screen time 22.8 (24.7) 24 (26)
TV in child's bedroom, % 70 70
Children's purchase requests, mean (SD)
smaller prospective sample with complete parent data for
No. of toy requests per week 0.95 (1.29) 1.15 (1.41) model 3 was nearly identical to the full prospective sample
No. of food requests per week 0.61 (1.14) 0.77 (1.29) for all variables.
Spearman rank correlation coefficients between base-
Abbreviation: GED, general equivalency diploma; TV, television. line screen exposure variables and baseline requests for
*From parent survey data, N = 684 to 691 (baseline sample), N = 315 to toys and food/drinks are reported in Table 2. All cor-
320 owing to missing data.
relations were statistically significant (range, 0.09-
0.20), confirming the hypothesis that screen exposure
is cross-sectionally related to children's requests for ad-
prevention schools. Of the 410 students, there were 14 vertised products.
parent refusals, 2 child refusals, 2 children were unable To test whether the amount of screen media expo-
to complete the surveys, 1 child was absent, and 5 chil- sure is a prospective risk factor for future requests for
dren from sibling sets were removed, which resulted in toys or food/drinks, we examined Spearman correla-
a potential prospective sample of 386 students. There were tions between baseline screen media exposure and the
an additional 39 children with no follow-up data: 22 chil- average toy and food/drink requests at 7, 12, and 20
dren had moved and 17 parents refused participation, leav- months (Table 3). All screen media exposure variables
ing 347. Overall, 347 (85%) of 410 students in the 6 smok- were significantly associated with the mean of future re-
ing-prevention schools participated in at least 1 follow-up quests for toys (r =0.19-0.24) and weekly TV viewing;
assessment and were included in the final prospective weekday TV viewing and weekly total screen viewing time
analysis sample. Two hundred ninety (84%) of these chil- were significantly associated with the frequency of fu-
dren also had complete parent interview data for parent ture requests for food/drinks (r =0.14-0.16), confirming
education, marital status, and the language spoken at the hypothesis that screen exposure is related to subse-
home and were included in the model 3 analysis sample. quent requests for advertised products. Having a TV in
Characteristics of the baseline and prospective analy- the bedroom was not significantly correlated with re-
sis samples are presented in Table 1. The sample was quests for advertised products cross-sectionally at base-
ethnically and sociodemographically diverse. Children line or prospectively (r=0.5-0.6).
reported watching more than 10 hours of TV and more As a secondary analysis, we further examined these
than 22 hours of total screen time per week. Children prospective relationships, adjusting for demographic fac-
reported making nearly 1 request per week for toys and tors using multivariate linear regression (Table 4). Af-
more than 1 request every 2 weeks for food/drinks. The ter adjusting for baseline requests for toys (model 1),
(REPRINTED) ARCH PEDIATR ADOLESC MED/ VOL 160, APR 2006 WWW.ARCHPEDIATRICS.COM
365
Downloaded from www.archpediatrics.com on April 6, 2006
©2006 American Medical Association. All rights reserved.
Table 4. Relationships of Baseline Media Exposure and Future Requests for Advertised Toys and Foods
and Drinks Adjusting for Demographic Variables*
No. of Toy Requests No. of Food/Drink Requests
Parameter (SE) P Value Parameter (SE) P Value
Model 1
Weekly TV exposure 0.005 (0.031) .88 0.011 (0.004) .005
Weekday TV exposure 0.048 (0.031) .12 0.066 (0.024) .006
Saturday TV exposure 0.006 (0.005) .20 0.044 (0.023) .053
Weekly total screen time 0.008 (0.002) .001 0.006 (0.002) .001
Model 2
Weekly TV exposure 0.009 (0.036) .80 0.019 (0.005) .001
Weekday TV exposure 0.074 (0.047) .06 0.103 (0.030) .001
Saturday TV exposure 0.010 (0.006) .11 0.079 (0.026) .002
Weekly total screen time 0.010 (0.003) .002 0.011 (0.002) .001
Model 3
Weekly TV exposure 0.008 (0.063) .90 0.022 (0.007) .001
Weekday TV exposure 0.068 (0.056) .22 0.134 (0.042) .002
Saturday TV exposure 0.008 (0.009) .40 0.100 (0.046) .03
Weekly total screen time 0.012 (0.004) .004 0.010 (0.003) .003
*Model 1 adjusts for baseline level of requests made for toys and foods/drinks and their interaction with media exposure. Model 2 adjusts for baseline level of
requests made for toys and foods/drinks, sex, ethnicity, and all first-order, 2-way interactions between variables. Model 3 adjusts for baseline level of requests
made for toys and foods/rinks, parental education (ordinal), marital status (married/not married), language spoken in the home (English/other), and all first-order,
2-way interactions between variables.
weekly total screen time was significantly associated with foods/drinks. In this prospective study, the screen me-
future toy requests. This result was consistent when fur- dia exposure was antecedent to the outcome of re-
ther adjusting for ethnicity/race and sex (model 2) and quests; therefore, we are able to define media exposure
parent education, marital status, and language spoken at as a true risk factor22,23 for future requests for toys and
home (model 3). food/drinks.
Television viewing exposure variables alone were not In our sociodemographically diverse sample, we found
statistically significantly related to future requests for toys. that third graders reported an average of nearly 11 hours
For food/drink requests, baseline hours of weekly TV per week of TV watching and nearly 23 hours per week
viewing, weekday TV viewing, and weekly total screen of total screen media use. The media consumption was
time remained significantly associated with future food/ similar24 to less25 than other reports in the literature. They
drink requests after adjusting for baseline food requests also reported requesting an average of about 1 adver-
(model 1), and all screen time exposure variables were tised toy per week and 2 foods or drinks every 3 weeks,
significantly related to future food/drink requests after which is consistent with the literature.4 Our primary pro-
adjusting for ethnicity/race and sex (model 2) and par- spective analysis demonstrated that TV and other screen
ent education, marital status, and language spoken at media exposure are true risk factors for future requests
home (model 3). As in the bivariate analyses, a TV in the for advertised products. In multivariate analysis, we also
bedroom was not significantly correlated with requests explored these relationships adjusted for baseline re-
for advertised products. quests and sociodemographic factors. Only the relation-
ship between the hours of Saturday TV viewing and re-
COMMENT quests for advertised foods or drinks became
nonsignificant in prospective analysis, although the mag-
We hypothesized that in a racially/ethnically and socio- nitude of the correlation was about the same. The rela-
demographically diverse sample of third and fourth grade tionships between screen media exposure and future food/
public school children those who report more screen me- drink requests remained statistically significant in the
dia exposure at the beginning of third grade would re- multivariate analysis. These results indicated that, even
port more requests for advertised toys and food and drinks, after adjustment for baseline product requests and demo-
both concurrently at baseline and subsequently, through graphic variables, an extra 1 hour per day in total weekly
the end of fourth grade. Our findings confirmed our hy- TV viewing at baseline was associated with an average
pothesis. Past research provided evidence that advertis- extra request for an advertised food/drink about every 6
ing exposure influences children's concurrent or imme- to 13 weeks (0.08-0.15 requests per week) 7 to 20 weeks
diate preferences, choices, and requests.4-9,11-13,15,16,21 Our later, and an extra 1 hour per day of total screen media
findings confirm those past findings, using contempo- exposure was associated with an average extra request
rary data from a more diverse sample, and extend them for an advertised food/drink about every 13 to 24 weeks
by following up children prospectively. Our prospec- (0.08-0.04 requests per week) 7 to 20 months later. Be-
tive findings demonstrate that baseline screen media ex- cause our media time self-reports are slightly lower than
posure predicts future requests for advertised toys and other reported samples, it is possible that these effects
(REPRINTED) ARCH PEDIATR ADOLESC MED/ VOL 160, APR 2006 WWW.ARCHPEDIATRICS.COM
366
Downloaded from www.archpediatrics.com on April 6, 2006
©2006 American Medical Association. All rights reserved.
would be even larger if examined in other samples re- consumerism is the outcome of interest, it would be ideal
porting greater exposure, making these results even more to assess and catalog advertising as direct or product place-
compelling. ment on TV, video games, movie, videotape, and DVD
When adjusting for baseline toy requests and demo- in movies. However, from a practical standpoint and in
graphic variables, however, only baseline total screen me- the current policy environment, the easiest way for par-
dia exposure continued to be statistically significantly re- ents to reduce exposure to screen advertising is to re-
lated to future toy requests. An extra hour per day of total duce screen time. Therefore, we believe that assessing total
screen media exposure at baseline was associated with screen time was the most reasonable method for draw-
an average extra request for an advertised toy about ev- ing relevant practical and policy conclusions. In addi-
ery 12 to 18 weeks (0.06-0.08 requests per week) 7 to tion, this study does not take into consideration that new
20 months later. requests may have displaced requests for similar prod-
For both the toy and food/drink multivariate analy- ucts; however, in analyzing requests over time it is clear
sis, we did not find any consistent evidence of signifi- that overall requests increased. Despite its limitations, the
cant differences by sex, race/ethnicity, parent educa- current study does document that screen media expo-
tion, marital status, or language spoken in the home. In sure is a true prospective risk factor for subsequent con-
addition, we found no evidence for baseline requests or sumeristic behavior, adding to the evidence supporting
demographic variables as significant moderators of the behavioral and policy interventions to reduce children's
associations between baseline screen exposure and sub- exposure to screen media and advertising, whether imple-
sequent requests for toys. mented at the individual family level, institutional level,
Although both toy and food/drink requests were sig- or the population level through legislation and changes
nificantly predicted by TV viewing and total screen me- in social norms.
dia exposure more than 6 months earlier, it is interest- One compelling reason to study the effects of media
ing to speculate on potential reasons for the difference and advertising on children is to further explore their role
between toy and food/drink requests in the multivariate in the current obesity epidemic.14 Advertising and con-
analysis. First, it is possible that excessive power was lost sumerism have been identified as potential targets for both
because of the addition of variables that may not have individual and population-based strategies to prevent obe-
sufficiently added to the fit of the models. Additional sity.2,26 A causal relationship has been established be-
power was sacrificed by using a smaller sample in model tween children's exposure to food advertisements and food
3 because it required complete parent survey data. It is choice15 and a school-based intervention to reduce screen
also possible that advertising trends and preferences for time resulted in reduced toy requests.16 Investigators in
toys are more transient or seasonal than those for food/ Australia found that overweight children were more sus-
drinks. If individual toy popularity and toy advertising ceptible to food advertising than their lean counter-
change more frequently with time and age, the link be- parts.11 Our study contributes support that reducing chil-
tween TV exposure at 1 point and requests for toys more dren's exposure to screen media may reduce their requests
than 6 months later may be weaker. The media variable for advertised foods and drinks, which are predomi-
that did remain statistically significant, total screen me- nantly high in calories and low in nutritional density.14
dia exposure, included video game, movie, videotape, and Potential policy actions should be grounded in em-
DVD use in addition to TV. More total screen media time pirically derived evidence. Further prospective studies
might also produce greater overall exposure to market- could be designed to further establish advertising di-
ing for toy products, thus creating a more durable and rected at children as a risk factor for obesity and con-
sustained level of requests. Alternatively, this group of sumerism and help identify biological, psychological,
children who use greater amounts of multiple forms of and/or social factors that may moderate an individual's
screen media may represent a subset within our popu- susceptibility to advertising and marketing messages
lation more heavily geared to consumerism and/or whose and/or mediate their effects on behavioral and physi-
families are more likely to respond positively to such re- ological outcomes. The finding that screen media expo-
quests, themselves leading to additional subsequent re- sure is a true risk factor for subsequent requests for ad-
quests. vertised products provides rationale for studies examining
It is always possible that the lack of association in toy the effects of individual- and population-level interven-
requests could be influenced by biased sampling, in which tions to reduce screen media exposure in general, and
there was a differential in sampling between those pre- advertising in particular, for their impacts on child health.
disposed to the effects of toy advertising vs food/drink
advertising. However, we had data for 85% of the eli- Accepted for Publication: December 15, 2005.
gible population-based sample and there were no sig- Correspondence: Lisa J. Chamberlain, MD, MPH, Divi-
nificant differences in age, sex, marital status, language sion of General Pediatrics, Stanford University School of
spoken at home, or educational level of parents for those Medicine, 750 Welch Rd, Suite 325, Palo Alto, CA 94304
with and without complete data. Therefore, we believe (lchamberlain@stanford.edu).
that sample bias is an unlikely explanation. Author Contributions: Dr Robinson had full access to all
A limitation of our study is that we assessed screen the data in the study and takes responsibility for the in-
media exposure as a proxy for advertising exposure and tegrity of the data and the accuracy of the data analysis.
not advertising directly. It would be methodologically dif- Funding/Support: The study was supported in part by grant
ficult to measure actual exposure and attention to ad- R01 HL62224 from the National Heart, Lung, and Blood
vertising in a population-based study. In a study where Institute, National Institutes of Health, Bethesda, Md.
(REPRINTED) ARCH PEDIATR ADOLESC MED/ VOL 160, APR 2006 WWW.ARCHPEDIATRICS.COM
367
Downloaded from www.archpediatrics.com on April 6, 2006
©2006 American Medical Association. All rights reserved.
Acknowledgment: We would like to acknowledge vertisements for foods on food consumption in children. Appetite. 2004;42:
221-225.
Helena C. Kraemer, PhD, for statistical guidance and
12. Borzekowski DL, Robinson TN. The 30-second effect: an experiment revealing
K. Farish Haydel, BA, Michelle Fujimoto, RD, Sally the impact of television commercials on food preferences of preschoolers. J Am
McCarthy, BA, Connie Watanabe, MS, Janine Bishop, Diet Assoc. 2001;101:42-46.
MPH, and the participating children, parents, teachers, and 13. Jeffrey DB, McLellarn RW, Fox DT. The development of children's eating habits:
schools for their support, guidance, and participation. the role of television commercials. Health Educ Q. 1982;9:174-189.
14. Story M, French S. Food advertising and marketing directed at children and ado-
lescents in the US. Int J Behav Nutr Phys Act. 2004;1:3.
REFERENCES 15. Gorn GJ, Goldberg ME. Behavioral evidence of the effects of televised food mes-
sages on children. J Consum Res. 1982;9:200-205.
1. Roberts D, Foehr U, Rideout V. Generation M: Media in the Lives of 8-18 Year 16. Robinson TN, Saphir MN, Kraemer HC, Varady A, Haydel KF. Effects of reducing
Olds. Menlo Park, Calif: The Henry J. Kaiser Family Foundation; 2005. Available television viewing on children's requests for toys: a randomized controlled trial.
at: http://www.kff.org/entmedia/upload/Executive-Summary-Generation-M- J Dev Behav Pediatr. 2001;22:179-184.
Media-in-the-Lives-of-8-18-Year-olds.pdf. 17. Henry J. Kaiser Family Foundation. The role of media in childhood obesity. Feb-
2. Kunkel D, Wilcox B, Cantor J, Palmer E, Linn S, Dowrick P. Report of the APA ruary 2004. Available at: http://www.kff.org/entmedia/upload/The-Role-Of-Media-
Task Force on advertising and children. Section: psychological issues in the in- in-Childhood-Obesity.pdf. Accessed January 21, 2006.
creasing commercialization of childhood. February 20, 2004. Available at: http:// 18. Robinson TN. Reducing children's television viewing to prevent obesity: a ran-
www.apa.org/releases/childrenads.pdf. domized controlled trial. JAMA. 1999;282:1561-1567.
3. Kunkel D. Children and Television Advertising. Thousand Oaks, Calif: Sage Pub- 19. Robinson TN, Killen JD. Ethnic and gender differences in the relationships
lications; 2001. between television viewing and obesity, physical activity and dietary fat intake.
4. Borzekowski DL, Poussaint AF. Latino American Preschoolers and the Media. J Health Educ. 1995;26:S91-S98.
Philadelphia, Pa: Annenberg Public Policy Center; 1998. Annenberg Public Policy 20. Kraemer HC, Blasey CM. Centring in regression analyses: a strategy to prevent
Center Report Series 24. errors in statistical inference. Int J Methods Psychiatr Res. 2004;13:141-151.
5. Goldberg ME, Gorn GJ. Children's reactions to television advertising: an experi- 21. Coon KA, Tucker KL. Television and children's consumption patterns: a review
mental approach. J Consum Res. 1974;1:69-75. of the literature. Minerva Pediatr. 2002;54:423-436.
6. Galst J, White M. The unhealthy persuader: the reinforcing value of television 22. Kraemer HC, Kazdin AE, Offord DR, Kessler RC, Jensen PS, Kupfer DJ. Coming
and children's purchase-influencing attempts at the supermarket. Child Dev. 1976; to terms with the terms of risk. Arch Gen Psychiatry. 1997;54:337-343.
47:1089-1096. 23. Kazdin AE, Kraemer HC, Kazdin AE, Kessler RC, Kupfer DJ, Offord DR. Contri-
7. Isler L, Popper E, Ward S. Children's purchase requests and parental re- butions of risk-factor research to developmental psychopathology. Clin Psychol
sponses: results from a diary study. J Advert Res. 1987;27:28-39. Rev. 1997;17:375-406.
8. Goldberg M. A quasi-experiment assessing the effectiveness of TV advertising 24. Wiecha JL, Sobol AM, Peterson KE, Gortmaker SL. Household television ac-
directed to children. J Mark Res. 1990;27:445-454. cess: associations with screen time, reading, and homework among youth. Am-
9. Gorn G, Florsheim R. The effects of commercials for adult products on children. bul Pediatr. 2001;1:244-251.
J Consum Res. 1985;11:962-967. 25. Roberts D, Foehr U, Rideout V, Brodie M. Kids & Media: the New Millennium.
10. Goldberg ME, Gorn GJ. Some unintended consequences of TV advertising to Menlo Park, Calif: Kaiser Family Foundation; 1999. Report 1536.
children. J Consum Res. 1978;5:22-29. 26. Koplan J, Liverman C, Kraak V. Preventing Childhood Obesity: Health in the Balance.
11. Halford JC, Gillespie J, Brown V, Pontin EE, Dovey TM. Effect of television ad- Washington, DC: Institute of Medicine; 2005.
"When is the ideal time for us to have an-
other baby?" a mother will ask.
"Ideal for whom?" I generally reply.
"Well, I'd like him to want the baby--
and to see it as his."
This is wishful thinking. No first child ever
wants the invasion of a second child. Par-
ents should decide for themselves when they
feel they can handle another.
--From Touchpoints by T. Berry Brazelton, 2004
(REPRINTED) ARCH PEDIATR ADOLESC MED/ VOL 160, APR 2006 WWW.ARCHPEDIATRICS.COM
368
Downloaded from www.archpediatrics.com on April 6, 2006
©2006 American Medical Association. All rights reserved.