How brands appear in children's movies. A systematic content analysis of the past 25 Years

ABSTRACT

Based on a theoretical framework of product placement characteristics, we present the most comprehensive content analysis of children's movies to date. We analyzed the integration of brands in 250 movies from 1991 to 2015. We found that 64.4% of all movies contained at least one brand placement. Comedies and movies produced in the US contained a higher number of brand placements compared to other genres and non-US movies. Animated movies and movies presenting a nature-, or fantasy-scenery contained a comparatively low number of placements. Brands were inserted mostly visually, and placed in the centre of the screen without explicit evaluation. Placements frequently presented a physical or verbal interaction with a main character. This practice was found to increase over time even though the sole number of brand placements remained stable over the past 25 years. Implications for parents, consumer–advocacy-groups, and advertising researchers are discussed.

KEYWORDS:

• Product placement
• content analysis
• children's movies

Introduction

In the Walt Disney movie ‘Cars’, several clearly branded cars are featured; the three loveable chipmunks Alvin, Theodor, and Simon sing about their favorite snack brand UTZ cheeseballs; and every other child in Emil's group of private detectives wears a clearly branded item of Nike. These brand appearances are not coincidences but a deliberative advertising technique, the so called ‘product or brand placements’ (Karrh 1998). Product placement practices occur in movies (Devlin and Combs 2015; Ferraro and Avery 2000), music videos (Burkhalter and Thornton 2014; Schemer et al. 2008), computer games (Lee et al. 2009), or TV-shows (La Ferle and Edwards 2006). A vast amount of placements are specifically made for children (Moore 2004). Therefore, the gradual proliferation of embedded brands in entertaining content has led to increasing criticism from consumer advocacy groups, who wonder if the audience is always sufficiently capable to distinguish persuasive from editorial content. When it comes to content specifically targeted at children, embedded advertising practices are seen even more critically (Hudson, Hudson, and Peloza 2008; Moore 2004), as children and adolescents are assumed to be more susceptible to persuasive messages compared to adults (Rozendaal, Buijzen, and Valkenburg 2010; Wright, Friestad, and Boush 2005). In fact, several effect studies on product placements have proven significant effects on children's attitude and behaviour toward the embedded brands (Auty and Lewis 2004; Matthes and Naderer 2015, Naderer et al. 2016; Naderer, Matthes, and Zeller 2017). Furthermore, behavioural effects of brand placements stayed unchanged independent of children's level of persuasion knowledge (Naderer et al. 2016) which points toward a strong susceptibility based on the limited cognitive capacities of children (Buijzen, van Reijmersdal and Owen 2010). As young children have more difficulties to critically process and review two contents at the same time compared to adults (De Pauw, Hudders and Cauberghe 2017), the usage of embedded advertising in content specifically targeted at children is discussed as being unfair (Hudson, Hudson, and Peloza 2008).

Against this background, the European Union requires broadcasters to disclose sponsored content in television programmes. Also, broadcasters are not allowed to produce TV-content including placements specifically targeted at children. Similar regulations have also been discussed in the US (Cain 2011). Discussion on regulations of brand placements in the TV-context due to the use of public resources date back to the early 90s (Karrh 1998). However, for movies the statutory regulation follow a ‘hands off’-approach (Synder 1992). This leaves the international movie-market quite unregulated (Halford and Boyland 2013).

Hence, it is of great importance to investigate the current status of product placements in movies specifically targeted at children. Watching movies holds a considerable amount in children's and adolescent's media use, since DVD-use has slightly increased in the past 10 years, and as the use of on-demand services like Netflix is on the rise. In addition, children tend to watch their favourite movies repeatedly and often unsupervised. They have been a stable movie-audience over the years and some of the biggest Hollywood successes, such as ‘Home Alone’, are children's movies (Li-Vollmer 2007).

Despite the importance of this topic, we lack a systematic longitudinal content analysis of movies targeting children. Extant research only provides limited insights about how placements in movies have changed over the years (Devlin and Combs 2015). Sutherland et al. (2010) analyzed just 10 years (1996–2005), and they have only focused on food and restaurant brands, leaving other product categories unobserved. Additionally, their codebook did not cover the theoretically important variables type of placement integration and placement interaction. Galician and Bourdeau (2004), by contrast, covered different product categories and placement interaction. However, they only coded 45 movies which included 15 movies for 1977, 1987, and 1997 each. No content analysis to date has specifically focused on brand integrations in children's movies.

We sampled the top 10 children's movies published in Germany from 1991 to 2015. The German movie market is the biggest European non-English language market. Movies from all over Europe are published here, as well as the popular Hollywood productions (InsideKino.de). Thus, our sample allows a comparison of the placement integration practices between American and European movie-productions. This is relevant because placements are unregulated, highly professionalized, and accepted in the US Findings about the US may thus not be generalizable to other countries. And since previous studies have focused on either one product category (Sutherland et al. 2010), one point in time (Devlin and Combs 2015), or one country (Devlin and Combs 2015; Galician and Bourdeau 2004; Sutherland et al. 2010), we present the most comprehensive content analysis of children's movies to date, with the most ample code book.

Conceptual framework

For a comprehensive content analysis of product placements, a guiding conceptual framework is necessary in order to derive the variables from theory. It further helps us to understand why the coded variables are of theoretical as well as practical importance and how they matter for placement effects. We combine two existing frameworks which we shortly outline in the following. First, drawing on Balance Theory (Heider 1946), Russell and Stern's (2006) Balance Model of Product Placement Effects suggests three pillars: the consumer, the brand, as well as the character who is shown in association with the brand. According to Balance Theory, a consumer's high attachment with the brand and the depicted character, and the positive association between the two of them explains the effectiveness of placements (Russell and Stern 2006, 9). Hence, it is of importance to consider character–brand-interaction, the features of the interacting character itself, as well as the evaluative framework of the interaction. Furthermore, it is crucial to assess brand characteristics such as product category and availability (i.e. fake vs. real brand). Second, we build on Rathmann and Enke's (2011) Model of Product Placement Effects that includes consumer, medium, and brand characteristics as well as composition characteristics (Gupta and Lord 1998). Furthermore, Rathmann and Enke (2011) underline the importance of context involving the statutory and the cultural context. All these facets are crucial in order to understand how placements exert their effects on the consumer.

It follows that a content analysis of brands in children's movies should cover the following six pillars: (1) medium characteristics (i.e. genre, type of production, movie-scenery, and main-characters); (2) brand characteristics (i.e. product category, and fake vs. real brands); (3) composition characteristics (i.e. brand modality, placement centrality, and placement length); (4) character characteristics (i.e. physical and verbal interaction, type of character, placement evaluation); as well as the (5) statutory (i.e. age-rating, country of production), and the (6) cultural (i.e. year of production, country of production) framework. One additional important factor is the consumer. The consumer can, however, not be investigated with a content analysis. All our coded variables can be derived from this model (see Figure 1). Except for consumer characteristics, it covers all highly interrelated aspects that are key to understand how product placements operate and how they exert their persuasive power.

Figure 1. Conceptualization of product placement characteristics.

Placement frequency within different types of movies

Brand placement is a deliberative promotional effort, aiming to change the audience's memory of, evaluation of, and behaviour toward the inserted brand (Balasubramanian, Karrh, and Patwardhan 2006). In previous studies, the genre of the movie as a medium characteristic has been connected to the number of brand placements. Comedies have been found to be the genre with most brand placements, whereas fantasy movies are least likely to contain placements (e.g. Sutherland et al. 2010). The preference for brands in comedies might be based on the companies’ interest to elicit positive brand associations (Pompper and Choo 2008). Previous effect studies base the success of brand placements on the positive context in which the brand was integrated (e.g. van Reijmersdal, Rozendaal, and Buijzen 2012). This can be explained by Emotional Conditioning Theory (Kroeber-Riel 1984): The affective nature of an entertaining context can be transferred to the embedded persuasive message (Balachander and Ghose 2003). In other words, if the movie is enjoyable, this may have a positive impact on the evaluation of the presented brand. As comedies provide a joyful context, companies might be particularly inclined to sponsor placements in this genre.

H1a: Comedies contain more placements compared to other movie-genres.

Children's movies differ from movies aimed at adult viewers. For instance, children's movies are often animated (Devlin and Combs 2015) and thus are less likely to contain brand placements. In non-animated movies, brands can be used as props and are easily inserted into the storyline, since products regularly embedded in the plot like cars, foods, clothes, or technical equipment are usually branded. In animated movies, by contrast, brand integrations have to be particularly animated, and it is therefore less easy to insert brands. Animated movies may thus be less suited to exert placement effects compared to non-animated ones. Also, as animated movies are clearly targeted at young children, regulatory guidelines (Moore 2004) might make it harder to place brands in this context. Hence, we generally assume that brand placements should occur less likely in animated movies compared to non-animated movies.

H1b: Animated movies contain less placements compared to non-animated ones.

Previous studies have concentrated on the top earning movies in general (Devlin and Combs 2015; Galician and Bourdeau 2004; Sutherland et al. 2010) with no specific focus on children's movies. Children's movies, however, can occur in various sceneries and present different main-characters (e.g. talking animals) than movies aimed at adults (Dreier 2007). Brands might be integrated more easily in a world that closely mirrors a human environment, like cities and villages (e.g. ‘Home Alone 2: Lost in New York’), than movies depicting mostly nature-scenes (e.g. ‘The Lion King’), or fantasy-worlds (e.g. ‘The Chronicles of Narnia’). Furthermore, if the main-character in a movie is human, brands are more easily integrated compared to movies showing only animals. As the scenery and the type of main-characters have not previously been investigated, we state the following research question:

RQ1a–b: What type of children's movie regarding (a) scenery and (b) main-characters is most likely to contain placements?

Regarding the statutory framework, it seems important to include the age-rating of a movie. Age-rating is a regulatory measure, which allows parents to estimate whether a specific movie is appropriate for their children to watch. However, we have limited knowledge, whether age-ratings are also considered by marketers, as advertising to young children under the age of 12 is deemed as unethical (Hudson, Hudson, and Peloza 2008; Moore 2004). In fact, children are assumed to be more susceptible to persuasive messages than adults (Rozendaal, Buijzen, and Valkenburg 2010; Wright, Friestad, and Boush 2005). Literature on the development of children's persuasive understanding postulates that older children can more easily understand the intention of advertising messages (Friestad and Wright 1995). It is presumed that during late childhood (10–12 years), children become capable of abstract thought processes, and their ability to understand persuasive intentions increases. After the age of 12, the level of understanding is getting closer to adult's apprehension (Valkenburg and Cantor 2001). Interestingly, Sutherland et al. (2010) found in their content analysis that PG-13 and R-rated movies were more likely to contain placements compared to G- and PG-rated movies. Thus, they postulate that placements in movies are generally aimed at an adult viewership. From an economic standpoint, however, targeting children who are especially susceptible to advertising techniques appears profitable, because these children have a considerable impact on their parents’ purchase decisions and also have money available to spend on their own (Skinner et al. 2002). In our study, we therefore want to investigate whether age-rating of children's movies is connected to the amount of brands presented. In light of the conflicting considerations we generally ask:

RQ1c: Is the age-rating of children's movies associated with the amount of brands presented in a movie?

Regarding the statutory as well as the cultural framework, we include country of production in the present analysis. In the US, audiences evaluate advertising efforts more positively compared to European countries and have more predisposition against placement regulations (Sabour et al. 2016). This is also mirrored by the existing regulations of placements. In the European Union, statutory regulations have been discussed over the past 25 years (Snyder 1992). As a consequence, disclosures for embedded advertising have been implemented for broadcasters, and they are welcomed by the audience (Sabour et al. 2016). In the US, by contrast, there are no such regulations (Cain 2011), and hence, placement practices for US productions might differ from movies produced in Europe. To date, there are no content analyses investigating differences between several countries of regions of origin. However, based on the established placement regulations we assume the following:

H1c: Children's movies produced in the US contain more placements compared to movies produced in Europe.

Previous studies suggest that the number of placements is relatively stable over time (e.g. Galician and Bourdeau 2004). Yet, past investigations covered a small yearly range (i.e. 10 years, Sutherland et al. 2010) or compared only three different years over three decades (Galician and Bourdeau 2004). Thus, studies may not have been able to observe changes. Our study employs a broader time range to examine whether the number of brand placements in children's movies has increased over the past 25 years. This is important in order to understand if children are increasingly targeted with placements.

RQ1d: Has the number of placements in children's movies increased over time?

Brand, composition, and interactive characteristics

Previous content analyses of audio-visual content have frequently focused on the type of brands embedded (e.g. Ferraro and Avery 2000; La Ferle and Edwards 2006). More specifically, Galician and Bourdeau (2004) found Coca Cola and Mercedes to be the most recurrently placed brands in their analysis of the 15 top-grossing movies in 1977, 1987, and 1997. Similarly, Sutherland and colleagues (2010) analyzed 200 movies published from 1996 to 2005. Their analysis, however, focused on beverage-, food-, and restaurant-brands only. They came to the conclusion that Pepsi and Coca Cola are the brands occurring most often in movies. If non-food and food-placements are taken together, cars (Burkhalter and Thornton 2014; Devlin and Combs 2015; Galician and Bourdeau 2004), and media outlets (Ferraro and Avery 2000; La Ferle and Edwards 2006) are the product categories which are most likely to be inserted. Both categories refer to products that are mostly of interest for an adult audience. Hence, it is unclear whether those product categories are also the most prominent ones in children's movies in which other product categories such as toys may occur more often:

RQ2: What kind of product categories will most likely be depicted in children's movies?

As the type of brand integration is described as a predictor for placement effects (Russell and Stern 2006), past content analyses focused on the manner of brand integration (e.g. Ferraro and Avery 2000), and the level of interaction with the inserted brands (e.g. Pervan and Martin 2002). The effects of placement integration concerning brand modality (e.g. Law and Braun 2000), placement centrality (e.g. Matthes et al. 2012), as well as the length of a placement (e.g. Gupta and Lord 1998) have been investigated previously. In the existing content analyses, most of the inserted brands were found to be visual only placements (e.g. La Ferle and Edwards 2006). Even though verbal and audio-visual brand integrations are most successful in creating brand awareness (e.g. Brennan and Babin 2004), inserting brands in a visual manner is most feasible, as brands are commonly used as props for the storyline (Balasubramanian, Karrh, and Patwardhan 2006).

H2a: Children's movies will most likely depict brands visually rather than verbally and audio-visually.

The embedded brands were also most frequently found to be inserted fully visibly and placed centrally (La Ferle and Edwards 2006). In order to create awareness for the embedded brand, placing a brand in the background may not be very successful in creating awareness for the brand (Russell 2002). This is based on the theory of selective attention which assumes that ‘we perceive and remember only those objects and details that receive focused attention’ (Simons and Chabris 1999, 1059). This might be especially true for children under the age of 10, as they have only a limited attention span (Buijzen, van Reijmersdal and Owen 2010). As advertisers pay for placements, they are aiming at an organic tie to the story and thus request a more central brand integration (Pompper and Choo 2008), such as an incorporation in the centre of the screen. However, placing a brand in an overtly obtrusive way such as close-up may backfire (van Reijmersdal 2009). A centre screen placement may thus be – all else equal – better pay off than background placements or overly obtrusive close-up placements.

H2b: Children's movies will most likely depict brands at the centre of the screen as opposed to the background or as a close-up.

Independent of the centrality of the placement, most brand integrations in past content analyses were found to be rather short, as they were shown for no longer than 5 sec (Galician and Bourdeau 2004; La Ferle and Edwards 2006). Past studies show that high brand depiction times can elicit counter arguing and lead to negative brand outcomes due to activated persuasion knowledge (Russell 2002). This suggests that optimal placements should be visible enough to be perceived but they should not be too obtrusive. This is clearly the case for placements under 5 sec (e.g. Russell 2002). In line with existing research, we thus assume the following:

H2c: Children's movies will most likely depict brands under 5 sec within one scene as opposed to longer than 5 sec.

Alongside placement integration, interaction with the placed brands is regarded as important (Russell and Stern 2006), as studies found placement interaction to be a key predictor of placement effects (e.g. Kamleitner and Jyote 2013). Research indicates that physical interaction with the brands occurred only in 23% of all cases; however, if an interaction took place, it was most likely through one of the leading characters (e.g. Burkhalter and Thornton 2014). Similarly, verbal brand integrations were most likely mentioned by one of the leading characters (La Ferle and Edwards 2006). Furthermore, Galician and Bourdeau (2004) point out that the practice of interactive placements has increased over time. Previous studies have also investigated the context in which brand interactions take place. The results show that brands are most likely presented in a positive or neutral context (e.g. Devlin and Combs 2015). A purely negative product association was observed less frequently in movies compared to neutral and positive ones (Devlin and Combs 2015; Galician and Bourdeau 2004). As the goal of placing a brand is to create a positive brand association, companies are interested in integrating their brands in a positive context, or in connection with a likable actor. In their study, Russell and Stern (2006) examined explicit evaluations of TV-series-characters in connection with a product interaction. They found that people who can identify themselves with the presented character are likely to align their product evaluations with the character's evaluation. The same should be true for children's movies. Based on these findings, and the general interest to elicit positive brand responses, we hypothesize that brands are evaluated rather positively than negatively. Furthermore and in line with this, brand interaction is more likely to occur with a main (rather than a secondary) character. The reason is that primary characters are much more likely to generate attention compared to secondary-characters. This should be the case both for physical and for verbal interaction:

H2d–e: For (d) physical and (e) verbal interaction, brands are more likely to be presented with a main-character than with a secondary-character.

H2f: If brands are evaluated, they are more likely evaluated positively rather than negatively.

Character characteristics

When it comes to character characteristics, investigating the development of placement interaction is of key importance. In past studies, such plot-placements were found to lead to more positive brand outcomes in adults (Kamleitner and Jyote 2013; Russell 2002; Russell and Stern 2006) and children (Naderer et al. 2016; Naderer, Matthes, and Zeller 2017) compared to screen-placements. This process is largely driven by the recipient's attachment to the presented character. Interactive placements are successful because they can transfer meaning from the interacting character to the brand (Russell and Stern 2006). Thus, interactive placements might enhance brand outcomes based on an Evaluative Conditioning reasoning, which postulates that a positive evaluation of the interacting character can be transferred onto the brand (e.g. Russell and Stern 2006). Furthermore, based on the Learning Approach, interactive placements present the specific usage of the embedded product (Bandura, Grusec, and Menlove 1966), which brings to mind the potential usefulness of the product or service for the viewers. Watching a movie-character interacting with a product may thus lead to second-hand learning (Kamleitner and Jyote 2013), which in turn can increase the perceived value of the embedded brand. Past content analyses found high levels of prominent interactive placements. Brands were most frequently portrayed with main-characters (Burkhalter and Thornton 2014; Devlin and Combs 2015). More importantly, Galician and Bourdeau (2004) found that the number of highly integrated plot-placements became increasingly common in recent movies, though their study only compared three different years. To date, the examination of placement interaction has mostly dealt with the sole occurrence of this practice (i.e. how frequently does it occur). We lack knowledge about how placement interaction has changed over time. That is, are there specific characteristics of medium, or statutory, and cultural background that help to predict when placement interactions are more likely to occur? Therefore, we ask:

RQ3a–e: How likely do placement interactions occur in (a) more recent movies compared to older movies; (b) non-animated movies compared to animated movies; (c) movies rated appropriate for children from the age of 12 compared to movies appropriate for all age groups; (d) comedies compared to other movie-genres; and (e) US productions compared to European productions?

In previous studies, interaction has been connected to other placement characteristics (Devlin and Combs 2015). Placements depicting an interaction with a character are more frequently in the centre of the frame and are depicted longer than placements that show no character product interaction. The reason is that characters are usually at the centre of attention. This should also be true for children's movies. We further assume that it makes a difference whether the inserted brand shows an actual product or some kind of service or institution. Previous studies found institutions and services to be placed more prominently and show stronger character associations than placements for products (La Ferle and Edwards 2006). This might be explained by the necessity of presenting the use of a service or institution in order to make clear what the service or institution stands for. Hence, showing people interacting with a service or institution makes clear what kind of services are provided. We therefore investigated whether placement interactions are connected to other placement characteristics such as placement length, centrality, and brand category. Based on prior studies, we thus assume:

H3a--c: (a) Length, (b) placement in the centre of the screen, and (c) presenting a branded institution or service opposed to a branded product are predictors for placement interaction.

Method

Sample

We used InsideKino.de, a database which provides a ranking of the 100 most successful cinema movies in Germany for each year, 1991–2015. We defined children's movies through three criteria: the age-rating (Sutherland et al. 2010), the presented content (Van de Voord 2007), and the main-characters depicted. Age-rating as sole criterion is not sufficient. As children relate strongly to animated characters (Dreier 2007), animated movies (with the exception of animated movies inappropriate for children such as ‘The Simpsons Movie’, etc.) were considered. Because children relate strongly to characters of their own age (Arendt et al. 2010), movies depicting children, and adolescents as leading characters such as ‘The Chronicles of Narnia’ were sampled. Popular teen-movies such as ‘American Pie’ were not considered, due to child-inappropriate content. Since the genre of family-movies reaches out to parents and children alike (Arendt et al. 2010), movies which focus on the story of a particular family were also added to the sample, such as ‘Mrs. Doubtfire’ or ‘Home Alone’. The identification of main-characters was reliable (Alpha = 0.77–0.82), and the identification of the movies was straight-forward and perfectly reliable.

We considered the German system for age-rating (FSK) which distinguishes between movies appropriate for all age groups (FSK0), movies appropriate for children from the age of 6 (FSK6), the age of 12 (FSK12), the age of 16 (FSK16), and movies for adults from the age of 18 (FSK18). To depict the target group of all children up to the age of 12, movies classified as FSK0, FSK6, or FSK12 were considered (equivalent of G-, PG- and PG13-rated movies). For the remaining movies, we watched the trailer to assess whether the content fulfilled our defined criteria for a children's movie. Only the 10 top-ranked children's movies per year were considered (N = 250).

Coding procedures

Brand appearances were characterized similar to previous content analyses (i.e. brand name mentioned explicitly; brand name or logo depicted and clearly recognizable which means at least shown for a second; brand depicted visually as well as mentioned verbally; Ferraro and Avery 2000; La Ferle and Edwards 2006; Siegert et al. 2007; Sutherland et al. 2010, Alpha = 0.84). As a coder cannot distinguish between deliberate and unintentional placements, all brand appearances were coded (e.g. Sutherland et al. 2010). Our coding scheme was based on the pre-defined movie-chapters (preset on a DVD), which are based on coherent story lines within the movie. Hence, all placements of one brand within a chapter were considered as a single depiction, similar to the scene-coding scheme applied in previous content analyses (e.g. La Ferle and Edwards 2006).

The initial coding procedure started with three coders, a fourth coder was later included in the coding procedure, hence, there were two coding rounds and we report a range of Krippendorfs Alpha. After the coders coded several movie-scenes together, we randomly selected two movie-chapters from 12, respectively, eight individual movies to determine inter-coder reliability, a total of 98 min of the same movie-material was then coded separately by each coder. This study is part of a larger data-set not fully detailed here.

Measures

Medium Characteristics. Medium characteristics such as movie-title and ranking of the movie were coded based on the information provided on InsideKino.de. We retrieved the information on length of the movie, genre (action/adventure, comedy, family, miscellaneous), type of production (animated, or non-animated), and main-characters based on the international movie database imdb.com (Devlin and Combs 2015, Alpha = 1.00). As children's movies often depict characters and worlds not associated to our world, we assessed whether most characters were humans, animals, or fantasy figures such as monsters (Alpha = 0.93). We looked at whether the movie depicted a world similar to human environment, a scenery largely showing nature, or a fantasy-world (Alpha = 0.70).

Brand Characteristics. Based on previous studies (Ferraro and Avery 2000; La Ferle and Edwards 2006), we differentiated several product categories (see Table 3, Alpha = 0.84--0.97). In an open category, we recorded the accurate brand name of the product, institution or service depicted. We differentiated between real brands, spoofs (fake brands closely based on existing brands; e.g. a laptop picturing a pear instead of an apple; Sutherland et al. 2010), and fake brands (Alpha = 0.89–0.92). We chose to include fake brands into the analysis to assess the overall importance of branding in children's movies. Furthermore, fake brands such as Hogwarts from the Harry Potter movies can turn into an important merchandise brand, and therefore have a considerable value to the market.

Composition Characteristics. Type of brand integration comprises brand modality, placement length (seconds a brand was shown per chapter), and placement centrality. Brand modality differentiated between verbal, visual, and audio-visual brands (Alpha = 0.79--0.90; Gupta and Lord 1998). Placement length was coded in segments of 1–3 sec; 4–20 sec; 21–40 sec; 41–60 sec; 61–180 sec; and longer than 180 sec (Alpha = 0.79--0.80). Placement centrality distinguished background-placements, which were presented on the second image plane; central placement, which were shown on the first image plane, and close-ups which depicted the brand on the first image plane on more than 50% of the screen (Alpha = 0.71--0.79; Ferraro and Avery 2000).

Character Characteristics. Brand interaction was assessed by verbal interaction and physical interaction. We also coded the characters’ evaluation of the brand. Verbal interaction was only coded if brand modality was verbal or audio-visual and differed between verbal interaction with the main-character, verbal interaction with a secondary-character, and verbal interaction with more than one character within a movie-chapter (Alpha = 0.77--0.81; La Ferle and Edwards 2006). Physical interaction occurred if a character was shown carrying, consuming, or using a branded product (Devlin and Combs 2015). Again we differentiated between no interaction, interaction with the main-character, interaction with a secondary-character, and interaction with more than one character within a movie-chapter (Alpha = 0.77--0.82; e.g. Pervan and Martin 2002). Evaluation of the brand assessed whether the brand was not evaluated at all, evaluated positively, negatively, or in a mixed, respectively, neutral way (Alpha = 0.83–1.00). Evaluation had to be explicit by using adjectives (e.g. ‘great’), attributes such as ‘favourite’, or exclamations such as ‘Wow’.

Statutory and Cultural Framework. Year of release and age-rating was coded based on the labelling provided on the DVD cover, insidekino.de, or the international movie database imdb.com (Devlin and Combs 2015, Alpha = 1.00).

Sample characteristics

Out of the children's movies analyzed (N = 250), 83.3% (n = 209) contained some kind of branding (M = 17.08; SD = 19.42). In the vast majority of the movies (64.4%, n = 161), at least one product placement could be assigned to a real brand (M = 9.12; SD = 13.31). The mean ranking in the Top100 for the 10 most successful children's movies over the last 25 years was 23.26 (SD = 17.86). The movie's ranking was considered as a control in all analyses. The average length for children's movies was 5800.24 sec (SD = 994.46). We recoded this variable to 1 = 4000–4999 sec; 2 = 5000–5999 sec; 3 = 6000–6999 sec; 4 = 7000–7999 sec; 5 = 8000–8999 sec; 6 = 9000–9999 sec; and 7 = longer than 10,000 sec. Movie-length was also considered as a control in all analyses.

The most common movie genres were adventure-, or action-movies (53.2%, n = 133), comedies (30.8%, n = 77), family-movies (10.4%, n = 26), and 5.6% (n = 14) were coded as miscellaneous. About half of the movies were animated (53.2%, n = 133) and the main-characters were humans (64.6%, n = 162), 22% (n = 55) presented animals as main-characters, and 13.2% (n = 33) fantasy figures. Concerning the scenery of the movie, 64.4% (n = 161) pictured a world mirroring our world, 18% (n = 45) displayed mostly nature, and 17.6% (n = 44) showed fantasy-worlds. The majority of the movies was classified as appropriate for all age groups (57.2%, n = 143); 34.4% (n = 86) were rated as FSK6, and 8.4% (n = 21) were rated as FSK12. The majority of the movies (60%, n = 150) was produced or co-produced in the US. The rest was produced in European countries (n = 99), one movie was a sole production of Australia.

Results

We coded a total of 4,270 brand placements. Nearly half of those branded products, institutions or services portrayed fake brands (43.5%, n = 1,858), 3.1% (n = 133) depicted spoofs, and 53.4% (n = 2,279) presented placements of real brands. The following analyses and hypotheses tests will be conducted based on the occurrence of real brands. The brands which were most commonly placed were Nike (90 placements in 28 movies), Mercedes (73 placements in 32 movies), and Coca Cola (70 placements in 36 movies).

Placement frequency within different types of movies

To assess in which children's movies real brands were most likely to be inserted, we ran negative binomial regressions for count data. The dependent variable was the number of real brand depictions (see Table 1). We inserted ranking in the top 100, movie-length as controls. For medium characteristics, we inserted genre (dummy coded with comedies as reference group), type of production (dummy coded with animation = 1, non-animated = 0), and kind of main-characters (dummy coded with humans as reference group), as well as kind of scenery (dummy coded with real world-scenarios as reference group). Regarding statutory and cultural characteristics, we inserted age-rating (dummy coded, movies rated as FSK12 were inserted as a reference group), year of publication, as well as country of origin (dummy coded with US production = 1, European production = 0). Independent of the inserted reference groups for multi-nominal variables, the results remained constant and no additional differences occurred.

Table 1. Negative binomial regression for number of real brand occurrences per movie.

	B	SE	Wald χ^2
Movie-length	0.20	0.12	2.50
Movie-ranking	0.01	0.01	1.03
Genre family-movie	−0.54	0.35	2.34
Genre action/adventure-movie	−0.50*	0.24	4.47
Genre miscellaneous	−0.96*	0.46	4.38
Production animated movie	−1.34***	0.36	13.93
Nature-scenery	−1.22***	0.29	18.23
Fantasy-scenery	−1.55***	0.32	22.79
Main-characters animals	0.16	0.36	0.19
Main-characters fantasy figures	0.92*	0.42	4.94
Age-rating FSK0	1.05*	0.41	6.51
Age-rating FSK6	0.76*	0.38	3.87
US production	0.62**	0.22	7.71
Year of publication	0.02	0.02	2.34

Note: p < 0.05*; p < 0.01**; p < 0.001***; N = 250.

As expected (hypothesis 1a), comedies contained more placements compared to all other genres (action/adventure: b = −0.50, SE = 0.24, Wald χ² = 4.47, p < 0.05; miscellaneous: b =−0.96, SE = 0.46, Wald χ² = 4.38, p < 0.05) except for family movies (b = −0.54, SE = 0.35, Wald χ² = 2.34, n.s.). But neither the length of the movie nor the ranking of the movie was connected to the number of brand placements. As assumed in hypothesis 1b, animated movies (b = −1.34, SE = 0.36, Wald χ² = 13.93, p < 0.001) were significantly less likely to contain brand placements than non-animated ones. To answer research questions 1a and b, movies set in nature-scenery (b = −1.22, SE = 0.29, Wald χ² = 18.23, p < 0.001) and fantasy-scenery (b =−1.55, SE = 0.32, Wald χ² = 22.79, p < 0.001) showed less placements than movies in real world-sceneries. However, movies depicting fantasy figures as main characters depicted significantly more brands compared to movies depicting human main characters (b = 0.92, SE = 0.42, Wald χ² = 4.94, p < 0.05).

Answering research question 1c, movies rated as appropriate for all age-groups (b = 1.05, SE = 0.41, Wald χ² = 6.51, p < 0.05) and children from the age of 6 (b = 0.76, SE = 0.38, Wald χ² = 3.87, p < 0.05) were more likely to contain brand occurrences than movies rated for children from the age of 12. In line with hypothesis 1c, US movies were more likely to present branding than European ones (b = 0.62, SE = 0.22, Wald χ² = 7.71, p < 0.01). Answering research question 1d, no significant effect of year of publication was found (b = 0.02, SE = 0.02, Wald χ² = 2.34, n.s.; see Table 1).

In additional analyses, we also investigated possible interaction effects of year of publication with type of production, country of production, and the different genres. The idea is to check whether there is an increase for one specific genre or region over time. However, we did not observe any interaction effects of this kind.

Brand, composition, and interactive characteristics

In the following, we report descriptive results for placements of real brands (N = 2279). Answering research question 2, the most widespread product categories were vehicles, technical equipment, and shoes. A chi-square goodness-of-fit test indicated a significant difference between the observed product categories (χ² = 1949.26, df = 20, p < 0.001). The majority of brand placements was presented only visually followed by verbal only placements, and audio-visual placements (χ² = 2,411.27, df = 2, p < 0.001), thus supporting hypothesis 2a. The vast majority of the visual brands was inserted in the centre of the frame, followed by placements in the background, and close-up, lending support to hypothesis 2b (χ² = 2,345.60, df = 2, p < 0.001). Most brand placements were depicted between 1–3 sec, or 4–20 sec (χ² = 2,331.83, df = 4, p < 0.001). Under 10% were depicted longer than 21 sec, providing partial support for hypothesis 2c. For more details on the results see Tables 2 and 3.

Table 2. Frequencies for product category.

	Real brand occurrences
Product category	Number	Percent
Accessories	28	1.2
Clothing/shoes	249	10.9
Clubs/associations	108	4.7
Convenience products	44	1.9
Cosmetic	26	1.1
Drinks	179	7.9
Food	175	7.7
Furniture	6	0.3
Health products/drugs	13	0.6
Media outlets	180	7.9
Restaurants/bars	80	3.5
Schools/universities	65	2.9
Services	178	7.8
Sport equipment	59	2.6
Supermarkets/stores	59	2.6
Technical equipment	251	11.0
Tobacco products	3	0.1
Tourism facilities	91	4.0
Toys	63	2.8
Vehicles	402	17.6
Miscellaneous	20	0.9
Total	2279	100.0

Note: χ² = 1949.26, df = 20, p < 0.001.

Table 3. Frequencies for composition characteristics.

		Real brand occurrences
		Number	Percent
Placement modality	visual	1860	81.6
	verbal	298	13.1
	audio-visual	121	5.3
	χ² = 2,411.27, df = 2, p < 0.001
Placement length	1–3 sec	901	45.5
	4–20 sec	940	47.5
	21–40 sec	114	5.8
	41–60 sec	15	0.8
	61–180 sec	11	0.6
	No visual placement (missing values)	298
	χ² = 2,331.83, df = 4, p < 0.001
Placement centrality	Background	290	14.6
	Frame centre	1665	84.0
	Close-up	26	1.3
	No visual placement (missing values)	298
	χ² = 2,345.60, df = 2, p < 0.001
Total	2279	100.0

In 42% of the cases, no interaction with a branded product was shown visually. If an interaction took place, brand placements primarily showed an interaction with main-characters followed by multiple characters, and secondary-characters (χ² = 177.02, df = 2, p < 0.001). Hence, hypothesis 2d is supported. As expected (hypothesis 2e), out of the 419 verbal and audio-visual brand indications, more than half were named by a main-character (χ² = 142.63, df = 2, p < 0.001). In 92% of the cases, there was no explicit evaluation of the placed brand. However, if the brand was evaluated, it was most likely evaluated positively (χ² = 106.84, df = 2, p < 0.001), therefore supporting hypothesis 2f. For more details on the results see Table 4.

Table 4 . Frequencies for interactive characteristics.

		Real brand occurrences
		Number	Percent
Verbal interaction	Int. main-character	254	60.6
	Int. secondary-character	95	22.7
	Int. multiple characters	70	16.7
	No verbal placement (missing values)	1860
	χ² = 142.63, df = 2, p < 0.001
Physical interaction	Int. main-character	603	45.5
	Int. secondary-character	221	16.7
	Int. multiple characters	500	37.8
	No interaction (missing values)	955
	χ² = 177.02, df = 2, p < 0.001
Brand evaluation	positive	127	69.0
	negative	22	12.0
	Mixed/neutral	35	19.0
	No evaluation (missing values)	2095
	χ² = 106.84, df = 2, p < 0.001
Total	2279	100.0

Contextualization of interactive characteristics

In order to assess the development of brand interaction within the defined media as well as the statutory and cultural characteristics, while also considering brand and composition characteristics, we conducted a logistic hierarchical linear model (i.e. multilevel analysis, Raudenbush and Bryk 2002). The advantage of multilevel analysis is that we can explain the individual-level variation while statistically controlling the variation across levels of analysis. We inserted interaction with a placement as our outcome variable (0 = no physical or verbal interaction; 1 = physical or verbal interaction with a character). Level1 variables included brand and composition characteristics. Level2 variables included medium as well as the statutory and cultural characteristics. For brand and composition characteristics, we inserted length of placement and centrality of placement (0 = background placements; 1 = central and close-up placements), and brand category (0 = branded product; 1 = institution and service). We inserted year of publication, ranking and length of the movie, making of the movie, age-rating (FSK12 was again used as a reference group), movie-genre comedy (1 = comedy, 0 = action/adventure-movies, family-movies, miscellaneous), as well as country of origin as independent variables for the medium as well as the statutory and cultural characteristics. We conducted the analysis stratified based on the movies in our sample. No connections of medium characteristics to occurrence of placement interactions are found. Yet, the statutory and cultural framework predicted the occurrence of brand interactions with real brands: US productions showed more brand interactions with real brands (b = 0.56, p < 0.05), and brand interactions became more frequent over time (b = 0.04, p < 0.05). Brands shown for a longer time (b = 0.52, p < 0.001), and those inserted centrally (b = 2.92, p < 0.001) are more likely to present a brand interaction supporting hypotheses 3a−b. In contrast to hypothesis 3c, there was significantly less interaction with institutions and services compared to branded products (b = −0.26, p < 0.05; see Table 5).

Table 5. Multilevel-model as a logistic regression for brand interaction with real brand occurrences stratified per movie.

		B	SE B
Level1 variables	Brand category services and institutions	−0.26*	0.13
	Placement length	0.52***	0.09
	Placement centrality	2.92***	0.19
Level2 variables	Movie-length	0.07	0.11
	Movie-ranking	0.01	0.01
	Production animated movie	−0.06	0.28
	Genre comedy	0.10	0.19
	Age-rating FSK0	−0.08	0.39
	Age-rating FSK6	0.11	0.38
	Year of publication	0.04*	0.01
	US production	0.56*	0.23

Note: R² = 0.42; p < 0.05*; p < 0.01**; p < 0.001***; N = 1.981; Groups = 154.

An overview of all hypotheses and research questions regarding their results is provided in Table 6.

Table 6. Hypotheses and research questions overview.

H1a	Comedies contain more placements compared to other movie-genres.		Partially verified.
H1b	Animated movies contain less placements compared to non-animated ones.		Verified.
RQ1a–b	What type of children's movie regarding (a) scenery and (b) main-characters is most likely to contain placements?		Movies depicting human environment sceneries and main characters that are fantasy figures most likely contain brand placements.
RQ1c	Is the age-rating of children's movies associated with the amount of brands presented in a movie?		Yes, movies appropriate for children from the age of 12 contain less brand placements compared to movies appropriate for all children and children from the age of 6.
H1c	Children's movies produced in the US contain more placements compared to movies produced in Europe.		Verified.
RQ1d	Has the number of placements in children's movies increased over time?		No, the number of brand placements stayed constant over the analyzed time span of 25 years.
RQ2	What kind of product categories will most likely be depicted in children's movies?		Vehicles, technical equipment and, clothing/shoes.
H2a	Children's movies will most likely depict brands visually rather than verbally and audio-visually.		Verified.
H2b	Children's movies will most likely depict brands at the centre of the screen as opposed to the background or as a close-up.		Verified.
H2c	Children's movies will most likely depict brands under 5 sec within one scene as opposed to longer than 5 sec.		Partially verified.
H2d–e	For (d) physical and (e) verbal interaction, brands are more likely to be presented with a main-character than with a secondary-character.		Verified.
H2f	If brands are evaluated, they are more likely evaluated positively rather than negatively.		Verified.
RQ3a–e	How likely do placement interactions occur in (a) more recent movies compared to older movies; (b) non-animated movies compared to animated movies; (c) movies rated appropriate for children from the age of 12 compared to movies appropriate for all age groups; (d) comedies compared to other movie-genres; and (e) US productions compared to European productions?		US-produced movies and more current movies depict brand interactions more likely.
H3a–c	(a) Length, (b) placement in the centre of the screen, and (c) presenting a branded institution or service opposed to a branded product are predictors for placement interaction.		Partially verified.

Discussion

This study provides extensive insights into how brands appear in children's movies. If, based on our results regarding medium characteristics, one is to formulate a practical recommendation for parents who want their children to watch movies without inserted persuasive messages, we would suggest to look for movies set in nature-, or fantasy-scenery, which are animated, are not classified as a comedy, and are not produced in the US. As a novel result giving insight into the statutory framework, we find that movies rated appropriate for all age groups and targeted at children from the age of 6 contain significantly more placements than children's movies aimed at children from the age of 12. One may speculate that movies targeting older children are less attractive to advertisers and marketers because effects are assumed to be weaker in comparison to younger children. Independent of the interpretation of this finding, this knowledge is highly relevant to parents and educators who need to be aware that children at all age groups are targeted heavily with placements.

Regarding the cultural and statutory framework, we have not found an increase of the number of brand placements per movie over time. This suggests that the 1990s were already fully professionalized (Karrh 1998) when it comes to the practice of placing products in children's movies. One may thus speak of a ceiling effect. As another key finding, placements are more likely to occur in US movies, pointing to key markets as well as cultural differences for placements. That is, in Europe, placements are more regulated and consumers are more critical compared to the US, making brand integrations less attractive to advertisers and marketers.

Regarding brand characteristics, children's movies present the product category of vehicles most frequently. Due to their prominent role in a plot and the car industry's financial power, car placements can heavily reduce production costs, and are thus frequently used (Galician and Bourdeau 2004). This finding makes clear that brand appearances in children's movies are not only child-specific. Brand integrations may target parents watching the movie together with their children. Or, they may be used to develop positive brand associations from a very young age, thus treating children as the consumers of the future (Skinner et al. 2002). Also, Coca Cola has been identified as one of the most frequent brands in children's movies. Since effect studies point out that beverage brand placements in children's movies have the potential to influence consumption behaviour (Auty and Lewis 2004), this result is of great importance.

Regarding composition characteristics, we found brands are mostly shown in interaction with the main-characters. Additionally, brands are very frequently shown with more than one character within a sequence, making the interaction even more prominent. This reveals how advertisers strategically insert their brands. We additionally observed that the practice of brand interactions increased over time for real brands. Rephrased, in recent movies, real brands are more likely to be shown together with a movie-character than in the 1990s. This points to an enhanced understanding and increasing professionalization of persuasive tactics by marketers and advertisers. In fact, effect studies have proven interactive placements to be very successful in creating positive brand outcomes (e.g. Kamleitner and Jyote 2013). As children might be particularly susceptible to brand placements showing a character–product-interaction (Naderer et al. 2016;Naderer, Matthes, and Zeller 2017), it certainly seems to be important to also critically discuss this type of brand integration. One could suggest to implement brand regulations for movies aimed at children, similar to the already existing regulations for European broadcasters. These regulations may include disclosures, but should also consider the type of integration regarding prominent brand interactions with the main-character of a movie.

Limitations

We did not assess the most popular children's movies within the US for each year. A comparison with the yearly movie-ranking of the US box office, however, suggests that the most popular children's movies within the US largely match the most seen US productions within Germany. As with all content analyses, we cannot make any assertions on potential placement effects. However, our study, building on a comprehensive theoretical conceptualization, forms an important basis for future effects studies. Our study sheds some light on the development of product placement strategy over time and points to cultural differences with regard to the US and Europe. Future research, however, should investigate placement practices with regard to other regions such as South-America, Africa, or Asia to get a more comprehensive overview of this advertising practice around the world. The present study underlines the importance of fake brands in children's movies. Yet, future research should investigate what impact the occurrence of fake brands in movies have with regard to world-views and traits such as materialism, or life-satisfaction (Opree, Buijzen, and Valkenburg 2012).

Implications

From the perspective of marketers and advertisers, our findings reveal how a typical placement in a children's movie looks like. In accordance with studies demonstrating persuasive effects (e.g. Kamleitner and Jyote 2013), children's movies tend to include placements that are inserted visually, are shown rather short, and in the centre of the frame, most of the time without explicit positive evaluation. This suggests that placements are presented in ways which should lead to memory traces but not persuasion knowledge. In other words, long placements with explicit positive evaluation would most likely increase the awareness about the persuasive nature of the placements, thus potentially impeding persuasive effects. Also, the interaction with a character is an important factor and a powerful strategy to elicit attitudinal and behavioural effects (Matthes and Naderer 2015; Naderer et al. 2016). Interestingly, marketers seem to either target children's parents or aim to influence children's long-term brand preferences because child-specific products were not the most frequently placed product categories.

From the perspective of parents and educators, it is relevant to note that children over 12 are targeted to a lesser extent compared to younger ones. This is alarming as younger children are especially susceptible to advertising effects. Also, the integration tactics seem to have professionalized over time because interactions with real brands have significantly increased. Such brand interactions are likely to elicit powerful effects, even though the sole number of brand placements in children's movies stayed stable over the past 25 years. Clearly, more research is needed to understand how children react to the persuasive tactics revealed in our content analysis. A recent effect study (Naderer et al. 2016), for instance, observed a limited impact of parental mediation strategies to protect children from the persuasive effects of brand placements. Scholars need to further investigate effective mediation and disclosure strategies, in order to ensure a fair confrontation with embedded advertising for the youngest and most vulnerable viewers.

Acknowledgments

We would like to thank Agnes Obereder and Karoline Adam for their tremendous help with the coding procedure. We want to thank Büchereien Wien (Public Library Vienna) for providing the majority of the analysed movies. We also want to thank Raffael Heiss for helping with the data analysis.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This project was funded by the Central Bank of the Republic of Austria (Oesterreichische Nationalbank [OeNB]) [grant number 16319].

Notes on contributors

Brigitte Naderer

Brigitte Naderer (PhD, University of Vienna) is a senior researcher at the University of Vienna, Vienna, Austria. Her research interests include persuasive communication, empirical methods, and advertising effects on children.

Jörg Matthes

Jörg Matthes (PhD, University of Zurich) is full professor of advertising research at the University of Vienna, Vienna, Austria. His research interests include advertising effects, public opinion formation, and empirical methods.

Ines Spielvogel

Ines Spielvogel (MA, University of Vienna) is a junior researcher at the University of Vienna, Vienna, Austria. Her research interests include persuasive communication as well as advertising effects on children and adolescents.