Early Childhood Research Quarterly 45 (2018) 197–203

Contents lists available at ScienceDirect

Early Childhood Research Quarterly

Cultivating interest in art: Causal effects of arts exposure during early

childhood
Brian Kisida a,∗ , Daniel H. Bowen b , Jay P. Greene c
a
Department of Economics and Truman School of Public Affairs, University of Missouri, 229 Professional Building, Columbia, MO 65203, United States
b
Department of Educational Administration & Human Resource Development, 553 Harrington Tower, 4226 TAMU, College Station, TX 77843, United States
c
Department of Education Reform, College of Education and Health Professions, University of Arkansas, 201 Graduate Education Building, Fayetteville, AR
72701, United States

a r t i c l e i n f o a b s t r a c t

Article history: Despite a growing body of literature examining the effects of arts exposure and participation for youth,
Received 20 March 2017 little is known about the development of attitudes toward art in early childhood. In this study, we
Received in revised form used an experimental research design to investigate the effect of arts exposure on the development
30 November 2017
of children’s attitudes toward art. Applicant groups (n = 26) with students in kindergarten through 2nd
Accepted 12 December 2017
grade (n = 2,253) were randomly assigned to participate in an art museum’s educational program, which
Available online 30 December 2017
included pre-curricular materials, a visit to an art museum with a guided tour and arts-based activities,
and post-curricular classroom materials. We collected original data from students in their classrooms
Keywords:
Arts education
that measured their attitudes toward art museums and art generally, as well as art knowledge. We found
Experimental design that exposure to the arts at an early age produced significant positive effects on the development of stu-
Attitude development dents’ attitudes toward the arts. Our findings demonstrate that arts-based exposure facilitated by schools
Museum education can be an effective strategy for developing positive orientations toward art in young children.
Education policy © 2017 Elsevier Inc. All rights reserved.

1. Introduction cultivation is often differentiated along social class lines. Parents

lacking in resources, including social and cultural capital, are less
The study of arts exposure and participation in early childhood able to provide these experiences for their children (Lareau, 2002;
typically focuses on examining relationships between the arts and Roksa & Potter, 2011). If parents are not able to provide access to
other positive outcomes, such as the relationship between arts the arts for their children, the next line of defense is the schooling
participation and academic skills, neurocognitive outcomes, and system, which has long facilitated valuable arts experiences for stu-
social-emotional development. In other words, researchers tend to dents. Yet the amount that schools provide arts instruction in early
focus on the benefits of arts education, taking for granted the avail- grades has declined in recent years as accountability pressures have
ability of the arts and students’ interest in the arts. In some ways caused schools to increase focus on preparation for standardized
this approach makes sense, given the apparent pervasiveness of assessments in early grades (Bassok, Latham, & Rorem, 2016; U.S.
art in children’s lives. Most children are encouraged to create art Government Accounting Office, 2009). This development is prob-
from the age they can grasp crayons, and many parents enrich their lematic for the future of arts participation in adulthood, as the
children’s artistic experiences through more formal mechanisms, National Endowment for the Art’s (NEA) Survey of Public Partic-
such as music lessons, dance lessons, children’s theater, and other ipation in the Arts (SPPA) found that the most significant predictor
organized arts activities. of adult arts participation is exposure during childhood (Rabkin &
At the same time, however, only some parents choose or are Hedberg, 2011). It is also concerning that noted declines in arts edu-
able to provide enriching arts experiences for their children. Soci- cation are concentrated among historically underserved groups. In
ologists have referred to parental efforts to organize culturally 2008, African American children were 49% less likely to receive arts
enriching activities for their children as “concerted cultivation,” education than they were in 1982, whereas Hispanic children were
and the amount that parents are able to engage in this form of 40% less likely to receive arts education. Moreover, children whose
parents have less than a high school education were nearly 77%
less likely to have received arts education in 2008 than they were
∗ Corresponding author.
in 1982 (Rabkin & Hedberg, 2011). If the many benefits of arts expo-
E-mail address: [email protected] (B. Kisida).
sure are to be realized and experienced, especially by underserved

https://doi.org/10.1016/j.ecresq.2017.12.003
0885-2006/© 2017 Elsevier Inc. All rights reserved.
198 B. Kisida et al. / Early Childhood Research Quarterly 45 (2018) 197–203

students, then it is imperative that strategies are identified that connections to childhood outcomes, most existing arts education
inculcate children with positive attitudes toward arts experiences studies do not examine these potential moderator effects and this
that will orient them to engage with the arts throughout their lives. is a major need for the field (Menzer, 2015).
Considering the research on early childhood arts activities as a
1.1. Arts participation and early childhood development whole, it is apparent that there is no single definition of what consti-
tutes arts participation. Most existing studies involve interventions
Early childhood arts participation includes a broad array of activ- focused on producing or performing, such as singing, playing an
ities, including music, dance, theater, and the visual arts (National instrument, dancing, or acting. Studies that examine exposure to
Endowment for the Arts, 2004; Menzer, 2015). Existing empirical visual art are uncommon, and the interventions associated with
research on the effects of these activities has largely focused on them often incorporate very specific techniques, such as art ther-
two broad areas: the relationship between arts participation and apy interventions (Drake & Winner, 2013). Summarizing the effects
academic skills and neurocognitive outcomes, and the relation- of arts education in early childhood requires a nuanced consider-
ship between arts participation and social-emotional development ation of interventions that vary in the types of art, program goals,
(Menzer, 2015). A third pathway of research has examined rela- duration of exposure, and the ways that students participate.
tionships between the arts and physiological outcomes (National
Endowment for the Arts, 2004). 1.2. Arts participation and later outcomes
In terms of academic skills, researchers have found positive
relationships between musical skills and literacy (Anvari, Trainor, There is also a growing body of research that has found that
Woodside, & Levy, 2002). Related research has found that arts arts participation has important benefits for older students, which
enrichment in preschool activities is related to school readi- is particularly relevant if early childhood exposure is a precursor
ness skills, higher achievement, and improved vocabulary (Brown, to arts engagement in later years. Two studies that examine the
Benedett, & Armistead, 2010). Younger children, however, are not relationship between arts education and traditional student out-
often assessed using standardized testing methods and such tests comes are particularly worth highlighting. First, Catterall, Dumais,
are typically not developmentally appropriate. As a result, research and Hampden-Thompson (2012) found a relationship between
on the arts and early childhood tends to focus on academic out- high levels of arts-rich school experiences and higher writing and
comes other than test-based student achievement. science scores, higher grade point averages, and higher levels of
In the area of the arts and social-emotional development, a sys- college attendance and graduation. Similarly, Thomas, Singh, and
tematic review of studies published between 2000 and 2015 found Klopfenstein (2015) found that cumulative credits in the arts are
18 studies that identified significant relationships (Menzer, 2015). associated with lower instances of dropout. Causal studies along
Several studies identified relationships between social skills and these lines, however, have been nearly nonexistent.
participation in music-based activities. These include a study that One notable exception was an experimental study of a school-
found children whose parents sing to them have more developed facilitated art museum partnership program, where researchers
social skills (Muñiz et al., 2014), and research showing that students found that arts exposure increased students’ tolerance levels, his-
participating in a music-based education program demonstrated torical empathy (Greene, Kisida, & Bowen, 2014), and their ability
improved social cooperation and interaction (Ritblatt, Longstreth, to think critically about works of art (Bowen, Greene, & Kisida,
Hokoda, Cannon, & Weston, 2013). Related research has found 2014; Kisida, Bowen, & Greene, 2016). Additional studies involv-
that participation in dance activities generates stronger pro-social ing museum programs are rare and tend to be correlational.
behaviors (Lobo & Winsler, 2006). Additionally, some research has For example, a quasi-experimental evaluation of the Solomon R.
found that drama-based activities are positively associated with the Guggenheim’s Literacy Through Art program found correlations
development of social skills (Nicolopoulou, Barbosa de Sa, Ilgaz, & between fifth grade students’ participation and literacy skills (Korn
Brockmeyer, 2009; Schellenberg, 2004). Notably, many of the arts- & Associates, Inc., 2007). A similar evaluation conducted at the
based activities examined in this area included a social component Isabella Stewart Gardner Museum in Boston found that arts par-
as a part of the arts activity (Menzer, 2015). ticipation increased 8th grade students’ critical thinking skills
Other studies have examined relationships between the arts and (Desantis, 2009). To our knowledge, however, no causal studies
emotional regulation. For example, toddlers in an arts integration have examined the effects of an art museum’s educational program
program with multiple art forms were observed to have improve- on younger children.
ments in teacher-rated emotional regulation (Brown & Sax, 2013).
Music participation studies have also found relationships between 1.3. Arts education and cultural capital
better emotional regulation (Gerry, Unrau, & Trainor, 2012) and
improved expressive emotions (Mualem & Klein, 2013). Studies Finally, prior research has shown that exposure to an art
incorporating visual art therapy have found similar results (Drake museum’s educational program has the effect of creating cultural
& Winner, 2013). There is also some evidence that art therapy in consumers with an increased desire to engage with cultural insti-
early childhood may be particularly effective for students dealing tutions and the arts generally (Kisida, Greene, & Bowen, 2014).
with emotional stress (Drake & Winner, 2013) and for students with Importantly, this research found that disadvantaged students had
autism (Kim, Wigram, & Gold, 2008). the least amount of previous arts exposure, and they experienced
Empirical studies have also extensively documented the chronic the greatest benefits from the program. Related research has con-
gaps in childhood social and emotional outcomes by gen- firmed a strong relationship between socioeconomic status and
der, race/ethnicity, and socioeconomic status. However, while children’s museum attendance (Dumais, 2006). Under the sociolog-
researchers have consistently found that engagement in the arts ical framework of cultural capital, arts exposure can be a valuable
is positively correlated with social and emotional development, resource that acts as a gateway to children’s future academic,
these studies have not investigated the extent to which the arts social, and economic success (Bourdieu, 1977; Bourdieu & Passeron,
can reduce these persistent gaps (Menzer, 2015). Prior studies have 1990). A growing body of research has found a positive relationship
also noted significant arts participation gaps along gender lines between measures of cultural capital and academic achievement
that emerge as early as age three, with girls much more likely to (Aschaffenburg & Maas, 1997; DiMaggio, 1982; Dumais, 2002;
participate in the arts (Smyth, 2016). While gender, race/ethnicity, Yamamoto & Brinton, 2010). To date, however, interventions that
and socioeconomic status have long been found to be have strong lead to the development of positive attitudes toward the arts have
 B. Kisida et al. / Early Childhood Research Quarterly 45 (2018) 197–203 199

not been examined among younger students using experimental were not selected to be in the treatment or control groups were
methods. If exposure to the arts in early childhood increases the encouraged to apply in future rounds.
likelihood of students seeking arts-rich experiences in later child-
hood, then early childhood exposure may lead to benefits that 2.3. Treatment
compound over subsequent years of schooling and adulthood.
Before visiting the museum, teachers of treatment group stu-
dents were sent a packet containing a video orientation for
2. Method teachers and students to watch that covered museum etiquette and
emphasized that the tour would lean heavily on student-driven dis-
2.1. Setting cussions about works of art. Teachers also received a selection of
three images the students would see on their tour of the museum,
In November 2011, Crystal Bridges Museum of American Art information about the themes of the tour, and guiding questions
opened in Bentonville, Arkansas. With a world class art collection to familiarize students with the dialogue-driven nature of the tour
and a permanent endowment of over $800 million, it is widely con- and the types of themes they would discuss.
sidered the first major American art museum to open in 50 years The museum offered a variety of tours, tailored for particular
(Vogel, 2011). The opening of the museum was particularly novel grade groups and encompassing various themes (Kraybill, 2014).
for the community, as there were no other major art museums in Students in grades K-2 participated in Me and My Museum, an
the region. Bentonville had a population of just over 35,000 in 2010, introductory tour that exposed students to a variety of paintings
with the larger metropolitan region approaching a half million res- and sculptures and asked them to interact through conversation,
idents. movement, stories, and hands-on activities. Trained museum edu-
cators conducted these tours which followed a constructivist-based
learning approach. In a typical tour, students were split into small
2.2. Participants groups of 10–15 that focused on four or five paintings or sculp-
tures in the museum. Museum educators facilitated open-ended,
In March of 2012, the museum launched their School Visit Pro- student-centered discussions about the works of art that encour-
gram (SVP). A portion of the museum’s endowment made it possible aged students to think together, engage with the works of art on
for local school groups to participate in the SVP at virtually no a deep level, and seek out their own unique interpretations. When
cost to schools or students. The endowment covered pre- and appropriate, museum educators supplied historical and sociologi-
post-curricular materials, admission to the museum, transporta- cal information to facilitate student understanding.
tion costs, students’ lunches on the day of the visit, and substitute
teachers. 2.4. Data collection
Announcements were sent to area schools informing them of
the opportunity to participate in the museum’s educational offer- Members of the research team visited students in their class-
ings, and teachers were invited to apply through a website that rooms and administered surveys to the treatment and control
collected their school information, grade level, size of the group, groups roughly three weeks, on average, after the treatment group’s
and any other special considerations (e.g., students with disabili- visit to the museum (M = 22.7days, SD = 13.4). Researchers were not
ties or students with limited English proficiency). The application made aware as to whether surveyed groups were in the treatment
portal was open for several weeks. Because the opening of a major or control group. Moreover, to guard against the possibility of social
art museum in an area with limited cultural resources was such a desirability bias, researchers identified themselves as representa-
significant event, and because the cost of participation was covered, tives of the local university and were careful not to suggest they
there was far more demand from local schools than the museum had any affiliation with the museum. Additionally, students were
could accommodate. encouraged to be honest in their responses and assured that all of
In total, 525 school groups representing over 38,000 students in their answers would be confidential. In total, 160 matched appli-
grades K-12 applied during the first year of the program. In order to cant groups (80 treatment and 80 control) representing a total of
fairly allocate the opportunity to participate, available slots were 10,912 students at 123 schools completed surveys. Due to tour can-
awarded through a lottery that we administered in partnership cellations or erroneous application information, we excluded 12 of
with the museum. Through this lottery, we were able to create the matched pairs that were originally part of the lottery.
treatment and control groups to facilitate a rigorous study of the Because participation in data collection was a mandatory com-
causal effects of participation in the art museum’s educational pro- ponent of receiving an immediate or deferred school tour, all of the
gram. remaining treatment groups and control groups completed student
To increase balance between the treatment and control surveys. Of the 160 groups we collected data from, 26 groups (13
groups, we used a stratified randomization procedure, which can treatment and 13 control) consisted of 2,253 students in grades K-
strengthen statistical power while preserving the advantages of 2 and constitute the sample we focus on in this article. Because of
random assignment (Schneider, Carnoy, Kilpatrick, Schmidt, & their age, students in grades K-2 completed a different survey than
Shavelson, 2007). To ensure that the treatment and control groups students in older grades. The surveys were read aloud to students by
were similar on pretreatment characteristics, we paired appli- members of the research team and students’ response options were
cants with similar demographics (e.g., grade, region, percent free or picture-based (thumbs-up and thumbs-down for yes/no responses;
reduced-price lunch, and percent minority) and randomized school a range of happy and sad faces for Likert-based responses).
groups within these pairings. Applicant groups that won the lot-
tery made up our treatment group, and matched applicants who 2.5. Measures
did not win the lottery constituted the control group. As an incen-
tive to participate in the research study, control group applicants The survey contained questions regarding student grade, gen-
that participated in data collection were guaranteed a spot for the der, attitudes toward cultural institutions, and attitudes toward
subsequent semester. Through the random allocation of available art. For ease of interpretation, Likert-based items were coded so
slots, 92 groups were invited to participate, while their matched that strongly disagree/somewhat disagree responses were coded
counterparts were awarded deferred tours. Applicant groups that as zeroes, and somewhat agree/strongly agree were coded as ones.
200 B. Kisida et al. / Early Childhood Research Quarterly 45 (2018) 197–203

Table 1 sures of interest in art museums and art, as well as art knowledge
Treatment/Control Balance of the Grade K-2 Analytic Sample.
outcomes, of student i in matched pair m and school s:
Characteristic Treatment (n = 1150) Control (n = 1103) Difference

Percent female 47.74 48.23 −0.49 ArtOutcomesims = ˛ + ˇ1 Treat i + Matchim ␤2 + Xi ␤3 + Zs ␤4 + εims

Average grade 1.27 1.43 −0.16
School Percent Hispanic 19.76 26.75 −6.98* The indicator variable Treati is equal to one if the student was in
School Percent black 2.46 3.73 −0.83
School percent white 67.52 59.48 8.04
the treatment group randomly assigned to visit the museum for a
School percent FRL 43.41 53.77 −10.36* school tour and is equal to zero otherwise. Because we used a strat-
Average school size 644.86 582.35 62.52 ified randomization procedure within matched applicant group
Town Population 42,692 40,687 2,006 pairs, Matchim is also included in the model as a vector of dummy
Note: School percent FRL = percentage of students receiving free or reduced-price variables that have the statistical effect of estimating within, as
lunch measured at the school level. The reported p-value is from the coefficient on opposed to across, matched pairs. Xi is a vector of student charac-
the treatment indicator when each covariate is regressed on the treatment indica- teristics including an indicator variable for gender and grade level
tor and the matched pair dummy variables (as in the base model for the impact
estimates).
dummy variables. Zs is a vector of school characteristics including
*
p < .05, two-tailed. school-percent Black and school-percent Hispanic, and school-level
percent free- or reduced-lunch eligibility. Finally, ␧ims is a stochas-
tic error term clustered at the applicant group level to account for
Two additional questions assessed students’ knowledge of art. First, the spatial correlation of students nested within applicant groups.
students were asked to correctly identify a portrait when asked For ease of interpretation, all of our results were derived from lin-
to choose between a portrait and a landscape. Students were also ear probability models. However, due to the binary nature of our
asked to identify George Washington from a selection of three pres- outcome variables, we validated our statistical significance lev-
idential portraits, since their tour included a discussion of a portrait els using logistic regression models. Finally, in addition to overall
of Washington by artist Gilbert Stuart. impacts, we tested for the heterogeneous effects across male and
female students. We tested for heterogeneous effects by modifying
2.6. Analysis this equation to include interactions between the binary treatment
variable and indicator variables for male and female students.
Because mere chance determined treatment group status, the
treatment and control groups were largely identical except for
whether they participated in the museum’s program. As a result, 3. Results
any outcomes that differ between the treatment and control groups
can confidently be attributed to having participated in the art Regression estimates for the full sample of grade K-2 students
museum’s school tour and related classroom activities. are provided in Table 2. Though not an outcome of primary inter-
est, the first row compares treatment and control group students on
treatment status to get a sense of how much exposure the control
2.6.1. Treatment-control baseline comparisons group previously received. Roughly eighteen percent of the control
Comparisons between the treatment and control groups on key group reported that they have visited an art museum on a school
variables show that the randomization procedure largely achieved field trip. This result could likely be explained by some fraction
comparable balance on observable characteristics. Table 1 pro- of the control group students actually having had such an experi-
vides key descriptive characteristics of the treatment and control ence, but also because young children who had never been to an art
groups and notes any differences. The groups were not significantly museum were not exactly sure what the question was asking. Addi-
different on percent female, average grade, school-percent black, tional data that we collected from older children, in which we asked
school-percent white, average school size, or the average popula- them to list the names of art museums they had attended, demon-
tion of the town where the students live. On two variables, however, strated that roughly half of the children mistakenly thought that
school percent Hispanic and school percent free- or reduced-price any museum was an art museum. As expected, nearly every student
lunch, there were significant differences. Control group schools in the treatment group reported having visited an art museum on
enrolled slightly more Hispanic students and were more likely to a school field trip since they had just visited one by design.
enroll students eligible for free- or reduced-price lunch. It should
be noted that even when implementing random assignment, some
significant differences may emerge by chance. In our analytic mod- 3.1. Impact analysis
els, we control for both of these school-level variables to improve
the precision of our estimates and adjust for any potential bias. In terms of our outcomes of interest, there were significant
effects across five of the seven outcomes that measured attitudes
2.6.2. Estimating treatment effects toward art museums and art. Students in the treatment group were
With comparable treatment and control groups, the analytic more likely to think trips to art museums were fun, felt happy about
techniques to estimate treatment effects are straightforward. In art museums, and indicated that they would like to visit art muse-
its most basic form, this technique could estimate simple regres- ums with their family. Students in the treatment group were also
sion models to compare mean differences between the treatment significantly less likely to say that looking at art is boring. It is also
and control groups on our outcome measures. However, though noteworthy that even though there were positive treatment effects,
randomization generated comparable experimental groups, these the control group percentages were very high. Even before partic-
groups are not perfectly identical. We improve our basic regression ipating in the art museum’s educational program, young children
model by adding controls for observable characteristics to adjust for had generally positive attitudes toward art. In some cases, ceiling
any differences across the groups and to increase the precision of effects may explain lack of significant effects on some items. Finally,
the treatment impact estimates. Moreover, by adding observable the treatment group was more likely to answer knowledge ques-
characteristics to the regression model, we were able to examine tions correctly, both in terms of identifying a portrait (an art-related
effects of these characteristics on outcomes. This analytical strategy question) and identifying George Washington (a historical-content
yielded the following equation for ArtOutcomes, our binary mea- related question).
 B. Kisida et al. / Early Childhood Research Quarterly 45 (2018) 197–203 201

Table 2
Regression Estimates of Treatment Effects on Students’ Interest in Arts Engagement.

Outcomes Treatment Control Difference Effect Size

Have you ever been to an art museum on a school field trip?a 98.90 17.89 81.01** 1.64**
Do you think trips to art museums are fun?a 94.34 88.58 5.76** 0.20**
If your family wanted to take you to an art museum, would you like to go?a 92.54 88.30 4.24** 0.14**
Do you think looking at art is boring?a 14.32 18.83 −4.51* −0.11**
How do you feel about art museums?b 94.20 88.86 5.34** 0.18**
How do you feel about visiting art museums when you are a grown-up?b 88.61 86.77 1.84 0.06
How do you feel about art?b 94.83 93.12 1.71 0.06
Can correctly identify a portrait.c 63.79 57.92 5.86** 0.12**
Can correctly identify George Washington.c 95.30 91.02 4.28** 0.17**

Note: Estimates are obtained from ordinary least squares regression models with robust standard errors clustered by applicant group. All models control for gender, grade
level, school percent FRL, school-percent Black, school-percent Hispanic, and lottery pair. Effect sizes are in terms of standard deviation units. For ease of interpretation,
Likert-based items have been collapsed so that strongly disagree/somewhat disagree responses are coded zero, and somewhat agree/strongly agree are coded one. a Measured
as thumbs up/thumbs down. b Collapsed from 4-category based Likert-scale. c Measured from multiple choice responses to questions: Which of these is an example of a
portrait?; and Can you circle the picture of George Washington?
**
p < .01.
*
p < .05, two-tailed.

Table 3
Regression Estimates of Treatment Effects on Students’ Interest in Arts Engagement, By Gender.

Outcomes Male Students Female Students

Treatment Control Difference Treatment Control Difference

Do you think trips to art museums are fun?a 91.58 85.60 5.99* 96.15 90.64 5.51**
If your family wanted to take you to an art museum, would you like to go?a 90.96 85.74 5.22* 93.31 90.14 3.17*
Do you think looking at art is boring?a 19.62 21.19 −1.57ˆ 9.78 17.53 −7.76** , ˆ
How do you feel about art museums?b 90.47 85.41 5.06* 96.29 90.64 5.65**
How do you feel about visiting art museums when you are a grown-up?b 85.54 82.30 3.23 91.60 91.28 0.31
How do you feel about art?b 91.45 88.37 3.08* 96.73 96.53 0.20
Can correctly identify a portrait.c 63.13 56.15 6.98** 63.46 58.82 4.63
Can correctly identify George Washington.c 94.15 93.20 0.95ˆ 94.66 86.70 7.95** , ˆ

Note: Estimates are obtained from ordinary least squares regression models with robust standard errors clustered by applicant group. All models control for gender, grade
level, school-percent FRL, school-percent Black, school-percent Hispanic, and lottery pair. For ease of interpretation, Likert-based items have been collapsed so that strongly
disagree/somewhat disagree responses are coded zero, and somewhat agree/strongly agree are coded one.
a
Measured as thumbs up/thumbs down.
b
Collapsed from 4-category based Likert-scale.
c
Measured from multiple choice responses.
ˆ
The difference between male and female student outcomes is statistically significant.
**
p < .01.
*
p < .05.

3.2. Moderating effects that students in the treatment group developed more positive
attitudes toward art museums and art. Because we employed an
We have provided the effects for male and female students in experimental methodology, we can be confident that the experi-
Table 3. One noteworthy aspect of disaggregating outcomes by gen- ence caused the impacts we observed. When disaggregating the
der is that female students in the control group started with more results by gender, we found that both male and female students
positive attitudes toward art than male students. Across all six out- experienced gains in positive attitudes about art museums. The sig-
comes that measure arts attitudes, female control group students nificant decrease in thinking that looking at art is boring, however,
indicated more positive pre-treatment responses. In terms of the was driven by female students. The estimated effect for males was
effects of the treatment, the most striking difference is whether stu- not significant on this outcome. Responses by the control group
dents thought that looking at art is boring. The treatment effect on further demonstrated that on average, female students began with
this outcome was driven entirely by female students. On the other more positive attitudes toward the arts, which is consistent with
hand, male control group students were more likely than female prior research that shows that girls participate in the arts more than
student to think that looking at art is boring, and they appear to have boys from an early age (Smyth, 2016). It is interesting that even for
been unaffected as a result of the treatment. Similarly, the effects very young students, initial interest in the arts seems to have played
on identifying George Washington were also driven by female stu- a role in the differential response to the program. Related research
dents, though this result is likely explained by the fact that most suggests that males and females focus on different aspects of arts
male students could already correctly identify Washington. experiences, with males showing more of an interest in the “facts”
behind works of art while females are more focused on the art-
4. Discussion work itself (Moulard, Rice, Garrity, & Mangus, 2014). As a result, it
could be that the nature of this particular program increased male
Using an experimental research design, we examined if young students’ interest in visiting an art museum without generating a
children develop a taste for art when they are exposed to an corresponding interest in looking at art in the way it was structured
art museum’s educational program. Though the intervention was by the museum.
modest, this experience was the first time most of these students This study provides a unique contribution to the research on
had visited an art museum. As a result, even a relatively limited early childhood arts participation. Most existing studies involving
amount of exposure produced significant differences between the young children have focused on music and performance art and
treatment and control groups. Across multiple measures, we found have typically involved more sustained arts participation. As such,
202 B. Kisida et al. / Early Childhood Research Quarterly 45 (2018) 197–203

it is difficult to situate our analysis among existing studies of arts we would be better equipped to make the claim that the art com-
participation effects given the variation in the types, durations, and ponents of the program, rather than the experience as a whole,
goals of different arts-based educational programs. Additionally, generated the effects we observed. Future research could explore
existing studies have primarily examined what have been referred the precise aspects of arts experiences that develop more positive
to as the “instrumental benefits” of the arts—outcomes focused on attitudes among young children.
the benefits of arts education in other domains, rather than intrin- We also cannot confidently say that these short-term changes
sic arts outcomes (McCarthy, Ondaatje, Zakaras, & Brooks, 2004). in attitudes will lead to increased arts exposure in later life and
Our investigation of the attitudinal effects of a relatively brief arts- related instrumental benefits. Students were surveyed on aver-
based experience, where students were encouraged to engage with age three weeks after they visited the museum, yet some were
works of visual art through a constructivist approach, focused on surveyed more than two months later. When we tested the rela-
students’ increased interest in engaging with visual art works and tionship between our outcome variables and the time between
art museums. Future research is needed to determine if other forms the program and survey administration, we found no evidence to
of arts participation would generate similar attitudinal changes. suggest the effects diminish over time. While this provides some
support that the effects may endure, we cannot confidently rule
4.1. Implications out the possibility that the effects will fade over a longer period
of time. Future research should examine the longer-term effects of
There are important implications from this research. A large arts experiences on the dispositions of students.
body of evidence suggests that arts experiences generate positive Finally, the typical tradeoff in experimental research between
benefits in early childhood, ranging from academic skills to social- high internal validity and low external validity applies in our sit-
emotional development. In later years, arts participation is related uation. The treatment and control groups both self-selected to
to better schooling outcomes (Catterall et al., 2012), higher gradu- participate in the intervention, which strengthens the compara-
ation rates (Thomas et al., 2015), and better outcomes later in life bility of our groups and improves internal validity. This aspect of
(Aschaffenburg & Mass, 1997; DiMaggio, 1982). Finally, the most the study, however, comes at the expense of external validity. We
significant predictor of adult arts participation is exposure during cannot make strong claims that these effects would hold for the
childhood (Rabkin & Hedberg, 2011). Yet, childhood arts exposure general student population in the region, students in other regions,
has experienced declines, and much of this decline has been a or students exposed to different types of arts experiences.
result of less exposure through school-facilitated arts educational
activities. This decline has been particularly true for historically
underserved students, who are especially reliant on schools for arts 5. Conclusion
exposure (Dumais, 2006). Our research has shown that exposure in
early childhood is an important catalyst that sparks an interest in From a policy perspective, this research demonstrates that
students to engage with the arts, which could lead to compounded schools can play an important role in providing students with expe-
downstream effects. riences that inculcate a taste for art. Exposure to the arts within
It is also worth noting that trained museum educators provided schools, however, has been decreasing (Rabkin & Hedberg, 2011),
the bulk of this arts experience in a museum setting. The effects of and the amount that U.S. schools are facilitating visits to cultural
such an intervention may hinge upon the expertise and novel expe- institutions has also declined (Blair, 2008; Ellerson, 2010). Given
riences that come with museum visits. Around the United States, the growing evidence that students exposed to the arts in early
partnerships with cultural organizations are an increasingly pop- childhood and beyond see improved outcomes across a broad range
ular strategy for supplementing in-school arts education (Bowen of measures, arts advocates and policymakers should take steps to
& Kisida, 2017). Given the ongoing challenges that K-12 schools reverse these declines. Future work that rigorously examines the
face with providing arts learning opportunities, our results suggest influence of the arts on children and the short-and long-term effects
these partnerships are valuable pursuits. would be an especially valuable contribution to guide those efforts.

4.2. Limitations Acknowledgements

Our study has important limitations. We do not know which We would like to thank Anne Kraybill, Director of Education
aspects of the program caused the treatment group to have more at the Crystal Bridges Museum of American Art, as well as our
favorable attitudes toward art museums and art. The effects could research assistants Josiah Foster, Kim Torres, LaShunta Williams,
be driven by exposure to the art itself, the museum setting, the Charles Denney, Irene Spalter, Aleza Greene, and Lydia D. Clark.
classroom curricular materials, or some combination of these activ- We are also grateful to the hundreds of students and teachers who
ities. It is important to consider that the tour was a structured graciously participated in this study. This research did not receive
educational experience. The results also suggest that learning about any specific grant from funding agencies in the public, commercial,
art likely played a significant role, as students retained factual infor- or not-for-profit sectors.
mation about the art they encountered during their museum visit.
This also suggests that viewing the art itself was a memorable and
thought-provoking experience for the students. References
Our study incorporated a passive control group–students who
Anvari, S. H., Trainor, L. J., Woodside, J., & Levy, B. A. (2002). Relations among
were not randomly selected to participate experienced “business
musical skills, phonological processing and early reading ability in preschool
as usual.” Their teachers may have replaced the museum experi- children. Journal of Experimental Child Psychology, 83(2), 111–130.
ence with some other enriching activity, but we have no data to Aschaffenburg, K., & Maas, K. (1997). Cultural and educational careers: The
dynamics of social reproduction. American Sociological Review, 62(4), 573–587.
confirm the nature of their activities. As a result, it could be the
Bassok, D., Latham, S., & Rorem, A. (2016). Is kindergarten the new first grade?
case that the treatment effects we observed are partially driven by AERA Open, 1(4), 1–31.
the mere fact that any field trip experience is more fun than school. Blair, E. (2008). Museum field trips tailored to teach to the test. National Public Radio.
In an ideal situation, a second control group condition would have Retrieved June 3, 2014 (http://www.npr.org/templates/story/story.
php?storyId=98593843)
placed students in a related educational experience of a different Bourdieu, P., & Passeron, J. C. (1990). . Reproduction in education society and culture
type, such as visiting a history museum. With such a condition, (Vol. 4) London, UK: Sage.
 B. Kisida et al. / Early Childhood Research Quarterly 45 (2018) 197–203 203

Bourdieu, P. (1977). Cultural reproduction and social reproduction. In J. Karabel, & McCarthy, K. F., Ondaatje, E. H., Zakaras, L., & Brooks, A. (2004). Gifts of the muse:
A. Halsey (Eds.), Power and ideology in education (pp. 487–511). New York, NY: Reframing the debate about the benefits of the arts. Santa Monica, CA: RAND
Oxford University Press. Corporation.
Bowen, D. H., & Kisida, B. (2017). The art of partnerships: Communities Menzer, M. (2015). The arts in early childhood: Social and emotional benefits of arts
collaborating to enhance arts education. Kappan, 98(7), 8–14. participation: A literature review and gap-analysis (2000–2015). National
Bowen, D. H., Greene, J. P., & Kisida, B. (2014). Learning to think critically: A visual Endowment for the Arts.
art experiment. Educational Researcher, 43(1), 37–44. Moulard, J. G., Rice, D. H., Garrity, C. P., & Mangus, S. M. (2014). Artist authenticity:
Brown, E. D., & Sax, K. L. (2013). Arts enrichment and preschool emotions for How artists’ passion and commitment shape consumers’ perceptions and
low-income children at risk. Early Childhood Research Quarterly, 28(2), 337–346. behavioral intentions across genders. Psychology and Marketing, 31(8),
Brown, E. D., Benedett, B., & Armistead, M. E. (2010). Arts enrichment and school 576–590.
readiness for children at risk. Early Childhood Research Quarterly, 25(1), Mualem, O., & Klein, P. S. (2013). The communicative characteristics of musical
112–124. interactions compared with play interactions between mothers and their
Catterall, J. S., Dumais, S. A., & Hampden-Thompson, G. (2012). The arts and one-year-old infants. Early Child Development and Care, 183(7), 899–915.
achievement in at-risk youth: Findings from four longitudinal studies. Muñiz, E. I., Silver, E. J., & Stein, R. E. K. (2014). Family routines and
Washington, DC: National Endowment for the Arts. social-emotional school readiness among preschool-age children. Journal of
Desantis K. (2009). Report to the education department of the Isabella Stewart Developmental & Behavioral Pediatrics, 35, 93–99.
Gardner Museum on the 8th grade School Partnership Program Visual National Endowment for the Arts. (2004). Imagine! Introducing your child to the
Thinking Strategies Adaptation 2008–2009. Retrieved from http://www. arts. Washington, DC: National Endowment for the Arts. Retrieved August 1,
vtshome.org/research/research-studies. 2017 (https://www.arts.gov/sites/default/files/imagine.pdf)
DiMaggio, P. (1982). Cultural capital and school success: The impact of status Nicolopoulou, A., Barbosa de Sa, A., Ilgaz, H., & Brockmeyer, C. (2009). Using the
culture participation on the grades of U.S. high school students. American transformative power of play to educate hearts and minds: From Vygotsky to
Sociological Review, 47(2), 189–201. Vivian Paley and beyond. Mind, Culture, and Activity, 17(1), 42–58.
Drake, J. E., & Winner, E. (2013). How children use drawing to regulate their Rabkin, N., & Hedberg, E. C. (2011). Arts education in America: What the declines
emotions. Cognition & Emotion, 27(3), 512–520. mean for arts participation (Research report #52). Washington, DC: National
Dumais, S. A. (2002). Cultural capital, gender, and school success: The role of Endowment for the Arts.
habitus. Sociology of Education, 75(1), 44–68. Ritblatt, S., Longstreth, S., Hokoda, A., Cannon, B. N., & Weston, J. (2013). Can music
Dumais, S. A. (2006). Early childhood cultural capital, parental habitus, and enhance school-readiness socioemotional skills? Journal of Research in
teachers’ perceptions. Poetics, 34(2), 83–107. Childhood Education, 27(3), 257–266.
Ellerson, N. M. (2010). A cliff hanger: How America’s public schools continue to feel the Roksa, J., & Potter, D. (2011). Parenting and academic achievement:
impact of the economic downturn.. Retrieved May 16, 2014. (http://www.aasa. Intergenerational transmission of educational advantage. Sociology of
org/uploadedFiles/Policy and Advocacy/files/CliffHangerFINAL%281%29.pdf) Education, 84(4), 299–321.
Gerry, D., Unrau, A., & Trainor, L. J. (2012). Active music classes in infancy enhance Schellenberg, E. G. (2004). Music lessons enhance IQ. Psychological Science, 15(8),
musical, communicative and social development. Developmental Science, 15(3), 511–514.
398–407. Schneider, B., Carnoy, M., Kilpatrick, J., Schmidt, W. H., & Shavelson, R. J. (2007).
Greene, J. P., Kisida, B., & Bowen, D. H. (2014). The educational value of field trips. Estimating causal effects using experimental and observational designs (report
Education Next, 14(1), 78–86. from the Governing Board of the American Educational Research Association
Kim, J., Wigram, T., & Gold, C. (2008). The effects of improvisational music therapy Grants Program). Washington, DC: American Educational Research Association.
on joint attention behaviors in autistic children: A randomized controlled Smyth, E. (2016). Arts and cultural participation among children and young
study. Journal of Autism and Developmental Disorders, 38, 1758–1766. people: Insights from the growing up in Ireland study. Retrieved from http://
Kisida, B., Greene, J. P., & Bowen, D. H. (2014). Creating cultural consumers: The www.artscouncil.ie/uploadedFiles/Arts-and-cultural-particiption-GUI.pdf. The
dynamics of cultural capital acquisition. Sociology of Education, 87(4), 281–295. Economic and Social Research Institute.
Kisida, B., Bowen, D. H., & Greene, J. P. (2016). Measuring critical thinking: Results Thomas, M. K., Singh, P., & Klopfenstein, K. (2015). Arts education and the high
from an art museum field trip. Journal of Research on Educational Effectiveness, school dropout problem. Journal of Cultural Economics, 39(4), 327–339.
9(S1), 171–187. U.S. Government Accounting Office. (2009). Access to arts education: Inclusion of
Korn, R., & Associates, Inc. (2007). Teaching literacy through art, final report: additional questions in education’s planned research would help explain why
Synthesis 2004–2005 and 2005–2006 studies. New York, NY: Solomon R. instruction time has decreased for some students (GAO-09-286). Washington, DC:
Guggenheim Museum. U.S. Government Accountability Office.
Kraybill, A. E. (2014). Inside the black box: What happens on a one-time field trip? Vogel, C. (2011, June 16). A billionaire’s eye for art shapes her singular museum.
Journal of Museum Education, 39(3), 1–10. New York Times. Retrieved from http://www.nytimes.com/2011/06/17/arts/
Lareau, A. (2002). Invisible inequality: Social class and childrearing in black design/alice-walton-on-her-crystalbridges-museum-of-american-art.
families and white families. American Sociological Review, 67(5), 747–776. html?pagewanted=1& r=1&ref=arts&.
Lobo, Y. B., & Winsler, A. (2006). The effects of a creative dance and movement Yamamoto, Y., & Brinton, M. C. (2010). Cultural capital in East Asian systems: The
program on the social competence of Head Start preschoolers. Social case of Japan. Sociology of Education, 83(1), 67–83.
Development, 15(3), 501–519.