Detecting pod position effects in the context of multi-segment sport programs: implications from four super bowl broadcasts.

Jeong, Yongick ; Tran, Hai

Introduction

When buying airtime for TV commercials, advertisers often seek premium spots that optimize audiences' attention and facilitate their working memory. In major broadcast events, where various commercials compete for attention, placing ads in the right spot becomes a central task for advertising practitioners. Certain positions, thanks to their advantage in appealing to viewers (Moorman, Neijens, & Smit, 2005), are often bought and sold at higher prices in major sports broadcast events. For the 2013 Super Bowl, for example, the average cost for a 30-second commercial spot was $4 million, but some commercial spots sold for an even higher price because of pod positioning (Smith, 2013).

A block consisting of a group of commercials is called a pod or a break. Advertising scholars have studied the effectiveness of commercials in certain serial positions within a pod (e.g., first, middle, and last position). The findings provide support for the primacy hypothesis, which suggests that the item presented first is more likely to be remembered than those items presented later in the sequence (Newell & Wu, 2003). However, empirical evidence remains limited to laboratory settings, where individual ads only appear in a single commercial pod. Little is known about the composite effects of pod position in natural, multi-segment programs containing multiple pods placed within several sub-broadcasting units (e.g., the first, second, third, and fourth quarters in football broadcasts). We attempt to fill this gap by examining pod position effects in the context of the Super Bowl, a four-hour program with several segments.

Message Order Effect

As indicated in past research, stimuli in the book-end positions (i.e., first and last positions) are more influential in memory-based evaluations than those in the middle (Burke & Srull, 1988; Haugtvedt & Wegener, 1994). The tendency to remember an item in the first position within a sequence is the primacy effect, and the tendency to remember an item in last position is the recency effect (Biswas, Grewal, & Roggeveen, 2010; Gurhan-Canli, 2003; Murphy, Hofacker, & Mizerski, 2006).

Previous studies show that primacy and recency effects are moderated by various determinants, such as the degree of audiences' cognitive status, processing motivation, and message relevance. Primacy effects tend to occur in individuals with lower cognitive sophistication (Krosnick & Alwin, 1987) or those with greater critical thinking toward later information and engaging in a systematic, critical processing of the information. Recency effects are observed when respondents are not highly motivated to process incoming information (Brunel & Nelson, 2003). Using the Elaboration Likelihood Model (ELM) as a framework, Haugtvedt and Wegener (1994) found that when personal relevance was high (high motivation to process), primacy effects dominated, and thus an initial message had a significant impact on final judgments and led to better recall. Conversely, when personal relevance of the issue was low (low motivation to process), recency effects occurred; thus, a later message had a greater impact on final judgments and led to better recall. Petty, Tormala, Hawkins, and Wegener (2001) also observed that those with high motivation were more likely to be persuaded by the first message whereas those with low motivation tended to be influenced by the last message.

Processing Capacity and Order Effect

Simon (1974) suggests that capacity constraints exist in short-term memory because generally people only have a limited amount of resources to process information at any one time. Because excessive arousal and increasing numbers of events will eventually cause cognitive overload (Easterbrook, 1959), individuals' attention and working memory can only accommodate a certain amount of information, then fail to handle the rest as effectively (Burke & Srull, 1988; Webb, 1979). Similarly, Lang (2000) suggests that advertisements that appear later in a sequence can quickly take up the space of consumers' short-term memory, leaving little capacity for additional advertisements and thereby causing incomplete processing. According to the limited capacity model for mediated message processing, a person's motivational and cognitive systems are constantly interacting with media messages without conscious thought (Lang, Sanders-Jackson, Wang, & Rubenking, 2013). Therefore, these interactions affect how a person processes messages and experiences media (Weber, Westcott-Baker, & Anderson, 2013).

According to Krosnick and Alwin (1987), due to the limits in cognitive processing capability, people tend to construct a cognitive framework with stimuli presented in the earlier stages to compare with later ones. Earlier stimuli are likely to be considered as more significant determinants in subsequent judgments. Because stimuli presented in later stages not only compete with other stimuli for audiences' attention and working memory but also are cluttered with thoughts about prior alternatives that hinder extensive cognitive consideration, stimuli presented in earlier stages are more likely to be processed at a cognitively deeper level than those presented later. Therefore, earlier stimuli tend to take over the entire cognitive process. Moreover, people have a tendency to minimize psychological costs by seeking satisfactory or acceptable alternatives as early as possible. Instead of attempting to search for optimal alternatives, people search for acceptable ones earlier in the process (Simon, 1957).

These cognitive benefits acquired via the placement in earlier sequences also translate into advantages in advertising message processing. Webb (1979) found a decline of brand mentions as ad clutter increases among television commercials. As attention to stimuli over a complete list progressively declined, the significance of later stimuli was less heavily weighed during processing. Similarly, Burke and Srull (1988) argue that associative interference processes contribute to primacy effects on the recall and recognition of TV ads because later stimuli inhibit viewers' ability to remember the promoted message effectively. In a study of 622 on-air promos, Eastman and Newton (1999) demonstrate promos that appeared in earlier broadcast segments often receive higher ratings.

Pod Position Effects in Broadcast Media

While the general findings show both primacy and recency effects, advertising research suggests the primacy tendency as a dominant order effect regarding ad performance in broadcast media (Pieters & Bijmolt, 1997; Riebe & Dawes, 2006; Zhao, 1997). For instance, despite the recall advantages of radio commercials that are placed at the start and end of large ad blocks, brand recall is considerably stronger for ads aired at the beginning (Riebe & Dawes, 2006). Another study of TV advertisements obtains a similar finding: when the recall of brands in the middle commercial position was calculated as 100%, recall of brand in the first spot commercial was 129%, while recall of brands in the last spot was 101% (Pieters & Bijmolt, 1997). In the context of the Super Bowl broadcast, Newell and Wu (2003) find support for the primacy effect, in which commercials presented first in the pod and the pods placed at the beginning of the program are better recalled.

Purpose and Significance

Through this study the researchers examined the impact of pod positioning on advertising effectiveness. The findings might help inform media strategists in selecting specific quarters or pods in multi-segment sports broadcasting. Newell and Wu (2003) merely studied quarter positioning effects and effects of positioning within a pod. In the current undertaking, the researchers are able to provide greater insight by examining positioning effects of different pods within the same quarter and similarly positioned pods in different quarters.

Hypotheses

Since people only have a limited capacity to process information, it is assumed that cognitive capability progressively declines during the program over various stimuli that compete for the audiences' attention and working memory (Webb, 1979). While earlier stimuli can acquire relatively considerable cognition and are likely to be processed to a cognitively deeper extent, the significance of later ads tends to be weighed less (Burke & Scrull, 1988; Krosnick & Alwin, 1987; Simon, 1957). Researchers of order effects (also known as serial position effects) indicate that the order in which a consumer sees products advertised has an impact on his/her brand preferences. Specifically, there are both primary and recency effects. The primacy effect causes consumers to favor brands they saw advertised first, while the recency effect causes consumers to favor brands they saw advertised most recently (Brunel & Nelson, 2003; Biswas, Grewal, & Roggeveen, 2010; Longinova, 2009; Murphy, Hofacker, & Mizerski, 2006; Scott, 2005).

We utilized Super Bowl broadcasts to determine the impacts of various pod positions on advertising effectiveness in multi-segment TV programs. Based on theoretical frameworks of order effects, cognitive processing capacity, and empirical evidence of pod position effects, we present the following hypotheses:

H1a: Brands advertised in earlier segments (i.e., quarters) are better recognized than those promoted in later segments.

H1b: Commercials presented in earlier segments are more favorably evaluated than those presented in later segments.

H2a: Within each segment, brands advertised in earlier commercial pods are better recognized than those promoted in later pods.

H2b: Within each segment, commercials presented in earlier pods are more favorably evaluated than those presented in later pods.

H3a: When pod placement is consistent across segments, brands advertised in earlier segments are better recognized than those promoted in later segments.

H3b: When pod placement is consistent across segments, commercials presented in earlier segments are more favorably evaluated than those presented in later segments.

Method

Research Design

A quasi-experiment was conducted to gauge audience responses to advertisements aired during four Super Bowl game broadcasts (2002, 2003, 2004, 2006). Data were collected through random sampling of local residents of North Carolina. Due to the constraints of the project, it was only possible to contact residents of North Carolina. As such, the results may not be generalized to the larger population of Super Bowl consumers.

Using a list of random phone numbers of local residents obtained from a local research firm, we contacted 2,553 people via phone and about 60% (n = 1,529) responded during the week following each of the four games. Despite using the random telephone numbers, in order to ensure random selections, the interviewers asked for the person who had the next birthday in each household. If an initial telephone contact yielded an answering machine, busy signal or no answer, the research sample with that number was re-contacted at three different times before being discarded. The sample included 795 females (51.3%) and 734 (48.7%) males, and the age average of 37.4.

Our research design has several advantages. First, respondents were not aware of the study while viewing a game broadcast, thereby avoiding the limitations of lab experiments, which rely on artificial manipulations to detect advertising effects. Second, the researchers examined responses from an actual TV audience, thereby addressing concerns about the external validity in advertising research, which often draws inferences from small student samples (Cornwell & Maignan, 1998; Pieters & Bijmolt, 1997). Third, this multi-year research controls for the effects of circumstantial factors pertaining to the nature of each game.

Multi-Segment Program: The Super Bowl

The popularity of sports has attracted substantial financial investments in the sport industry from corporations (Pyun & James, 2009), which in turn, influence sports broadcasts and the formation of multi-segment programs. These programs are often found in sports broadcasts where the inclusion of individual segments (i.e., sub broadcasting units) is decided by game rules and regulations (e.g., the first, second, third, and fourth quarters in football). Compared to single-segment programs (e.g., dramas, sitcoms, and movies), which are tightly connected from the beginning to the end, each segment in multi-segment programs functions somewhat independently, offering a relatively fresh start with each subunit within the whole program.

We chose Super Bowl broadcasts as a multi-segment program to determine pod position effects. Admittedly, the Super Bowl generates higher levels of emotional excitement, involvement, and attention to commercials than general TV programs. However, this sport event is one of very few programs that allow researchers to examine the audience's response in a natural environment. Super Bowl advertising has been actively researched, including context effects (Nail, 2007), ad placement (Newell & Wu, 2003), ad clutter (Zhao, 1997), ad liking (Tomkovick, Yelkur, & Christians, 2001), corporate social responsibility (Babiak & Wolfe, 2006), ambush marketing (Lyberger & McCarthy, 2001), sport fandom (Tobar, 2006), visual attention for the game (Beasley, Shank, & Ball, 1998) and commercials and sports sponsorships (O'Reilly, Lyberger, McCarthy, Seguin, & Nadeau, 2008).

Independent Variable: Pod Position

We identified the four segments from Super Bowl broadcasts, including the first, second, third, and fourth quarters. We then classified the positions of commercial pods within each quarter based on their sequential order and found five pod sequences, which are often seen in a typical quarter in football broadcasts. It should be noted that we detected a few additional pods offered beyond this five-pod position sequence in the second and fourth quarters of the 2003 and 2004 broadcasts when longer adjustment time was necessary (e.g., higher scoring, interruptions, timeouts). However, we noticed that these additional pods were generally shorter than regular commercial pods, containing a single ad with few on-air promos and being offered only a few times. We decided to exclude these additional pods from data analysis because the fewer number of ads contained in those extended pods creates a pod condition different from other regular pods, while preventing proper data analysis. Thus, we included the first five positions in each quarter to ensure consistency in the number of commercials and pod duration.

Pod position effects were examined in three ways: quarter positions, within-quarter positions, and between (across)-quarter positions. Quarter positions were defined as pod positions in each quarter (e.g., pods in the first, second, third, and fourth quarter). Within-quarter positions were identified by the sequential order of each pod within a quarter (e.g., the first, second, third, fourth, and fifth pod in each quarter). Between (across)-quarter positions compared the same pod in different quarters (e.g., the first pod in the first, second, third, and fourth quarter).

Dependent Variables: Advertising Effectiveness

Advertising effectiveness is measured in terms of cognitive (recall, recognition, attention, exposure, awareness) and affective (attitude, liking) outcomes (Peter, Bonfrer, & Dhar, 2008; Soh, Reid, & King, 2009). Brand recognition is the primary measurement of cognitive effects (Cianfrone & Zhang, 2006; Riebe & Dawes, 2006), while ad liking is often used to gauge the affective aspect of advertising effectiveness (Dahlen, Rosengren, & Torn, 2008; Decrop, 2007; Galloway, 2009). We examined cognitive response through brand recognition and affective response through ad liking.

Brand recognition. This variable was measured through a list of brands. Respondents were asked if, during a game, they remember seeing commercials for each brand. The recognition rates were calculated by the proportion of respondents who recognized the brand to the total number of respondents who viewed the broadcast segment that included the advertised brand. In this process, we included on the list seven to eight competing brands that had not been advertised during the game to detect the presence of a false alarm where respondents accidentally recognize a brand they had seen somewhere else besides the Super Bowl broadcast. The results revealed few cases with false responses, and for the sake of precision, we deleted these data. The correlation between the datasets with and without false responses was .99, and parallel analyses yielded identical results. Although false alarms were not assumed to be significant threats, we used the reduced data without false alarms to avoid possible contaminations.

Advertisement favorability. Previous research has found that ad likability/favorability is positively correlated with ad effectiveness (Newstead & Romaniuk, 2010; Smit, Van Meurs, & Neijens, 2006). The general premise is that if viewers favor an ad they will pay attention to it, which will lead them to respond more positively and develop a positive attitude toward the brand that is being advertised (Smit et al., 2006). We adopted Zhao's (1997) measure, asking respondents to use 7-point Likert scales to evaluate how poor (1) or good (7) the ad was.

Control Variables

Previous researchers have determined that ad frequency is positively associated with advertising effectiveness (Eastman, Schwartz, & Cai, 2005; Matthes, Schemer, & Wirth, 2007; Moorthy & Hawkins, 2005). In our study, ad frequency was treated as a control and measured by the number of advertisements for a certain brand that appeared in a broadcast.

Advertising clutter is believed to reduce attention to the commercials, and at the same time increase uncertainty (Carroll, 2009; Eastman & Newton, 1999; Pieters & Bijmolt, 1997; Zhao, 1997). Ad clutter increases as commercial length decreases. Researchers have found that ad clutter decreases the overall effectiveness of television advertisements (Ha & McCann, 2008; Laroche, Cleveland, & Maravelakis, 2006). To control for this confound, we determined the degree of advertising clutter by counting the number of other ads in a pod, meaning the overall number of commercials in the pod minus one. Providing two or more commercials for a brand were advertised, the number of other ads for the same brand in all pods was added together to measure the total size of clutter for a particular brand.

In addition, Zhao (1997) found that the year of the Super Bowl broadcast significantly influenced advertising effectiveness during three years of Super Bowl games (1992-1994). Thus, considering the likelihood in which the audience's reaction to the ads aired in four Super Bowl programs might have varied from year to year due to various circumstantial factors such as creativity and production quality, we also controlled for the year of the Super Bowl broadcast.

Analysis and Results

Statistical Procedure

Of 272 brands promoted in the four Super Bowl broadcasts, 158 brands promoted in four game quarters were selected for analysis, excluding brands advertised in non-game parts of the broadcast (e.g., pregame, halftime show, postgame, and between quarters). We used multiple regression analysis to control for confounding factors when assessing unique effects of the independent variables (Cohen & Cohen, 1983). The analysis procedure was the same for each dependent variable. Advertising frequency, clutter, and broadcast year were entered into the control blocks. For the year variable, three dummy variables were created for year 2002, 2003, and 2004, and brands from 2006 served as a comparison group. On top of those controls, the independent variable, pod positions, was entered.

Pod Position Effects: Quarter Position Effect

With the first analysis we tested whether there was better recall of the brands promoted in earlier quarters compared to those placed in later quarters (H1a), and whether advertisements appearing in the earlier quarters were evaluated more favorably than ads evaluated in the later quarters (H1b). For this analysis, three dummy variables were created for the first three quarters in the Super Bowl games while the fourth quarter was used as the baseline for comparison. The results of multiple regression analyses are presented in Table 1.

Effects of control variables. From the results we found that the control variables significantly influenced the recognition and favorability of advertising in the Super Bowl games. First, the year variable predicted 6.4% of the variance for brand recognition and 49.5% for ad liking. Second, advertising frequency additionally accounted for 13.0% of the variance in brand recognition, while accounting for 0.5% of the variance in ad liking. Third, advertising clutter contributed an additional 15.3% of the change in recognition and 1.5% of the variance in ad liking. Therefore, the total variance explained by the control block was considerably high with 34.6% for brand recognition and 51.5% for ad favorability.

Pod position effect. Models 1a and 1b were used to assess the impact of game quarter position on brand recognition and ad favorability, respectively. The results supported the first hypothesis (H1a). Those brands promoted in the earlier quarters were recognized at a significantly higher rate than those presented in the later quarters. Based on the beta coefficients, the recognition scores for brands advertised in the first quarter were highest (b = 9.50, = .34, p < .001), followed by brands shown in the second quarter (b = 8.65, = .27, p < .001). However, recognition of brands promoted in the third quarter was not significantly higher than recognition of brands advertised in the fourth quarter. The results failed to support the hypothesis regarding ad favorability (H1b). Ad favorability for the commercials placed within the earlier quarters was not significantly higher than that of ads shown in the later quarters. Overall, the independent variable block predicted an additional 3.4% of the variance in recognition and 7.3% in ad favorability.

Pod Position Effects: Within-Quarter Position Effects

The second set of analysis (Models 2-5) investigated the impact of pod positions on brand recall and ad favorability within each quarter (H2a and H2b). For these analyses, commercial pods were categorized into five positions based on their sequential order within a quarter. Four dummy variables were constructed for the first, second, third, and fourth-positioned pods, and commercial pods in the fifth position were used as a basis for comparison. The results are presented in Tables 2 and 3.

Effects of control variables. The control variables explained significant percentages of the total variance for brand recognition and ad favorability within each game quarter. In the first quarter, the control block explained 47.9% of the variance in brand recognition and 61.8% of the change in ad favorability. The corresponding numbers in later quarters were 45.6% (second quarter), 55.9% (third quarter), and 48.2% (fourth quarter) for brand recognition and 59.0% (second quarter), 57.9% (third quarter), and 52.3% (fourth quarter) for ad favorability, respectively.

Pod position effect within the first quarter. Models 2a and 2b were included to examine the impact of the sequential order of pod positions on brand recognition and ad favorability in the first game quarter. The results failed to support the proposed hypotheses concerning brand recognition (H2a, see Table 2) and ad liking (H2b, see Table 3). The brands promoted in the earlier positions were not more likely to be recognized than those shown later in the first quarter. Similarly, ad favorability measures for the commercials placed within the earlier positions were not higher than those shown at later positions in the first quarter. Overall, the independent variable block predicted 2.1% of the change in recognition and 7.3% in favorability.

Pod position effect within the second quarter. Models 3a and 3b were used to test the pod effects on advertising effectiveness within the second quarter. The results failed to support our predictions regarding brand recognition and ad favorability (see Tables 2 and 3). The ads placed in the earlier positions in the second quarter were not more effective than those shown later in that quarter. Overall, the independent variable block explained an additional 3.4% of the variance in brand recognition and 3.5% in ad favorability.

Pod position effect within the third quarter. Models 4a and 4b were included to investigate the impact of commercial pod within the third quarter. Again, the results failed to support the hypotheses (see Tables 2 and 3). The independent variable block explained an additional 2.0% of the variance in brand recognition and 1.6% in ad favorability.

Pod position effect within the fourth quarter. Models 5a and 5b were used to examine the position effect of commercial pods on advertising effectiveness in the fourth quarter. The hypothesized effects on both brand recognition and ad favorability were not observed. The independent variable block predicted an additional 13.5% of the change in recognition and 13.6% in favorability.

Pod Position Effects: Between (Across)-Quarter Position Effects

The third set of analyses (Models 6-10) was conducted to compare brand recognition and ad favorability of similarly placed commercials across four quarters (H3a and H3b). For these analyses, commercial pods were categorized under five positions based on their sequential order within each quarter. Then, unlike the previous analyses, three dummy variables were constructed for pods placed in the first, second and third quarters while those placed in each position of the fourth quarters were used as a comparison basis. The results are displayed in Tables 4 and 5.

Effects of control variables. The control variables accounted for a significant percentage of the total variance for brand recognition and ad favorability when comparing similarly situated ads between the four game quarters. Across the five positions, the control block explained 45.5% (2nd position) to 56.7% (1st position) of the variance in brand recognition and 56.2% (4th position) to 63.6% (3rd position) of the variance in ad favorability.

Pod position effect of the first position between (across) quarters. Models 6a and 6b were examined to compare brand recognition and ad favorability in the first pod across the four game quarters. We found support for the hypothesis concerning brand recognition (H3a). Based on the beta coefficients, the recognition scores for brands advertised in the first position of the first quarter (b = 12.89, = .29, p <.05) and second quarter (b = 12.93, = .30, p < .05) were significantly higher than brands presented in the same position of the fourth quarter. Interestingly, recognition for those brands shown in the first position of the third quarter was not significantly higher than that of the fourth quarter (H3a, see Table 4). However, the finding failed to confirm the hypothesis regarding ad liking (H3b, see Table 5). The favorability scores for the commercials placed in the first position of the earlier quarters were not higher than those of the first position of the later quarters. The independent variable block explained an additional 9.3% of the change in brand recognition and 0.6% for ad favorability.

Pod position effect of the second position between (across) quarters. Models 7a and 7b were used to examine the brand recognition and ad favorability in the second pod position alternated in different game quarters. There was no significant effect for either brand recognition or ad liking (H3a/b). The brands and commercials promoted in the second position of the earlier quarters were not more effective than those in the later quarter. The independent variable block contributed an additional 5.0% of the change in brand recognition and 0.3% in ad favorability.

Pod position effect of the third position between (across) quarters. Models 8a and 8b were used to test the pod position effect of ads placed in the third position across the four game quarters. The results failed to confirm our predictions regarding the measures of brand recognition and ad favorability. The independent variable predicted 4.2% of the variance in brand recognition and 3.9% in ad favorability.

Pod position effect of the fourth position between (across) quarter. Model 9a and 9b were included to investigate the impact of commercial placement in the fourth pod position between game quarters on brand recognition and ad favorability. Again, our hypotheses were not supported. The independent variable block explained 5.6% of the variance in brand recognition and 7.3% in ad favorability.

Pod position effect of the fifth position between (across) quarter. The final models (10a and 10b) were included to test whether the effectiveness of commercial placement in the fifth pod position varied across the four quarters. As the results indicate, our hypothesis regarding brand recognition (H3a) was partially supported. Recognition for those brands promoted in the fifth position of the first quarter (b = 16.66, = .48, p < .01) was significantly higher than that of the fourth quarter. Meanwhile, brand recognition for those commercials placed in the fifth positions of the second and third quarters was not significantly higher than those of the fourth quarter. Consistent with other ad favorability measures, the results did not provide evidence to support the hypothesis concerning ad favorability (H3b). Overall, the independent variable block predicted 15.8% of the explained change in brand recognition and 2.0% for ad favorability.

Discussion and Conclusions

We attempted to detect the pod position effects in the context of multi-segment sport broadcasts. Using the Super Bowl as a natural viewing setting for quasi-experiments, we analyzed audience reactions to the brands and commercials presented in four game broadcasts and examined brand recognition and ad favorability as indicators of advertising success. Based on the literature relating to the primacy order effect and the concept of processing capacity, we expected primacy effects in the way the audience responded to ads.

The results indicate that there was better recognition of brands advertised in the pods placed in earlier quarters than those promoted in the pods positioned in later quarters. The brand recognition scores were highest in the first quarter, and then progressively waned over the course of the game. These findings support previous arguments concerning the strong primacy tendency to remember commercials in the earlier positions of an advertising sequence (Brunel & Nelson, 2003; Pieters & Bijmolt, 1997; Riebe & Dawes, 2006). Along with cognitive capacity, this finding can be explained within the context of the Super Bowl game broadcast, which induces high involvement among viewers. According to previous researchers on context effect, commercial breaks interrupting a program have a considerable impact on advertising effectiveness, especially when the audience is highly involved (Moorman, Neijens, & Smit, 2005). In a high involvement condition, due to the significant amount of contextual information and the cognitive effort required to process information, general viewers tend to process ads presented in early stages at a deeper cognitive level (Brunel & Nelson, 2003; Krosnick & Alwin, 1987). Because people only have a limited capacity to process information, excessive arousal and an increasing number of consecutive commercials are likely to cause cognitive overload (Easterbrook, 1959; Webb, 1979). Later stimuli, therefore, inhibit viewers' ability to remember advertised information (Burke & Srull, 1988).

The strong primacy tendency of pod position effects, however, was not detected in within-quarter conditions. The position of the commercial pods did not have a significant impact on brand recognition and/or ad favorability within all four quarters, even in the first quarter when the first commercial pod of the game was launched. This unexpected observation can be explained by the characteristics of multi-segment TV broadcasts. Unlike uni-segment programs, a multi-segment program consists of several sub-segments that are physically divided and work somewhat independently. In this situation, viewers might have maintained their levels of involvement and emotional arousal throughout each quarter and paid a similar amount of attention to the pods placed in the same quarter regardless of their sequential orders. Therefore, ad performance was not altered within each quarter while ad effectiveness was significantly impacted by the order of game quarters.

We also observed that, compared to ad placements in later quarters, placement in the first and fifth commercial pods of the earlier game quarters made the brands more recognizable. However, the impact of commercial pod position on brand recognition was not detected with those brands similarly situated in the second, third, and fourth commercial pods across the four quarters. This finding, to some extent, suggests both primacy and recency effects in multi-segment broadcast programs. The levels of audience attention to commercials were noticeably different between the book-end (i.e., first and fifth) positions and middle positions (i.e., second, third, and fourth) across the quarters.

Equally noteworthy is the complete lack of relationship between commercial pod placement and ad favorability. We failed to find any pod position effects on how viewers evaluated the commercials. In the absence of such an association, it can be assumed that ad favorability is not significantly influenced by ad positioning in multi-segment programs. Instead, viewers' evaluations of commercials might have been affected by other factors such as ad length, frequency, creativity, quality, and/or their brand loyalties/familiarities.

Practical Implications

In general, advertising is considered the most effective promotional tool in televised sports events among other venues such as sponsorships, athlete endorsement and stadium signage (Cianfrone & Zhang, 2006). However, promotion via the television medium is not trouble-free for sports marketers due to its high cost and limited controls given to advertisers in scheduling. Among the few controls, the selection of pod positions is one efficient tool allowed to advertisers for promoting their products. By examining three pod placements, thus, our findings provide useful practical implications on media planning, particularly in Super Bowl broadcasts.

Overall, the pod position effects support the primacy tendency. Advertising effectiveness peaks in the earlier game quarters and then gradually decreases as the game progresses. However, this premise of the primacy effect was not supported with the order of commercial pods within each quarter in the light of advertising effectiveness. Thus, based on these two determinations, our findings suggest that placing commercials in the earlier quarters would be more effective than placing ads in earlier commercial pods within any quarter. In addition, we observed that the impact of quarter order was only discernable in the first and last pod positions. Therefore, we can predict that placing ads in the first and last commercial breaks of the first quarter would yield the most favorable results if the goal of an advertising campaign is to increase the awareness of brand among sports viewers.

In addition, we observed the complete absence of recency effect of the pod position in the Super Bowl broadcasts. In fact, brand recognition efficiency was least effective in the fourth quarter. Although this finding deviates from those reported in the previous studies, this seemingly unexpected observation is, indeed, consistent with the practical viewpoint in which the fourth quarter positions are similar to a general TV buy in terms of memory-based ad evaluation (Blackshaw, 2009). Thus, we recommend avoiding the spots in the fourth quarter if the goal is brand awareness.

While the performance of brand recognition is closely associated with pod positions, the success of ad favorability seems to be completely separate from the pod positions. Instead, it can be assumed that audience evaluations are likely to be influenced by other determinants such as ad creativity, audience brand loyalty/familiarity, and celebrity endorsements. Thus, if the goal of an advertising campaign is to increase the ad favorability, it is suggested that marketers place their ads in later quarters where pods are more reasonably priced and emphasize on the quality of their commercials while keeping away from the earlier pods that cost significantly more.

Theoretical and Methodological Implications

In addition to practical implications, there are theoretical and methodological implications gleaned from this analysis. First, this research represents a modest step toward expanding the order effect research in advertising of sports broadcast. Second, our analysis of quasiexperimental data collected in four years carries methodological implications for advertising effectiveness research. While previous studies have often been conducted with a limited number of stimuli and in a short duration of time (e.g., single commercial break), the current undertaking examines pod position effects within an entire program that contained ads for multiple brands continuously and consecutively over four hours. Manipulations in a laboratory setting are often criticized for altering true advertising effects and producing less generalizable findings. Therefore, compared to those conventional methods, our research design appears naturalistic in several aspects, including a wide range of commercial lengths, repetitions, and qualities, various product types, different pod sizes and lengths, more representative television viewers, and varied natural viewing environments. Consequently, our findings would have comparatively higher external validity.

Limitations and Suggestions for Future Research

The findings of our study cannot be taken without reservations. The most obvious limitation is the focus on the special sport events, Super Bowls, which admittedly are unusual circumstances (Moorman, Neijens, & Smit, 2005). These special circumstances might have evoked different emotional arousal, involvement level, physiological status, and attention to commercials that may not be representative of those induced by general broadcasts. Additionally, considering that over time the Super Bowl has become more and more commercialized, the data analyzed in this study (2002-2004, 2006) might not reflect the same effects as this year's Super Bowl.

Another weakness is our collection of advertising responses in the week following each Super Bowl game. Despite the popular use of this measurement to determine the immediate impact on advertising performance, its drawback lies in the inability to detect sleeper effects that influence advertising success. In addition, because telephone surveys were conducted in the week following the game, respondents might have talked to other people about the ads or might have accessed other media (e.g., USA Today--Ad Meter) to obtain reviews before the interview. An additional limitation is the impact of brand equity on advertising effectiveness. Researchers have found that brand equity significantly impacts the effectiveness of advertising (Hsu, 2012). However, through this study it was not possible to separate the impact of brand equity on advertising effectiveness. Finally, this study did not consider commercial creativity and ad quality as a factor that might have affected viewers' liking of commercials.

With its strengths and weaknesses, this study is among the first to explore pod position effects in the context of multi-segment broadcasts. A direct extension of this research may examine other aspects of advertising effectiveness such as recall, broader attitude toward ads, different measures of advertising likability, and purchase intention as dependent variables. Future research may also consider the possible relationship between commercial pod positions and audiences' involvement/arousal within a sub-segment to develop a more nuanced understanding of sequential order effects. In addition, subsequent studies may examine the ad content, viewers' involvement, physiological status, degree of excitation transfer, and brand familiarity as well as the level of brand equity as variables moderating pod position effects on advertising performance. Future studies regarding pod positioning should correct limitations in this study including the data used. Additionally, the data in this study was collected only in one state. In order to increase generalizability of the findings, subsequent researchers should collect data from multiple regions. Finally, the findings presented in this study need to be replicated in multi-segment sport broadcasts other than the Super Bowl and other types of television programs (e.g. movies, reality programs, and game shows) to examine whether similar pod position effects will occur.

Contribution statement: Advertising scholars have studied the impacts of a commercial pod (blocks consisting of a group of commercials) on advertising effectiveness by determining serial positions of ads within a pod (e.g., first, middle, and last position). In spite of useful implications regarding order effects in TV advertising, however, less is known about the composite effects of pod position in more prolonged broadcasting conditions, such as multi-segment programs that contain multiple pods placed within several sub-broadcasting units (e.g., the first, second, third, and fourth quarters in football broadcasts). We attempted to detect these pod position effects on advertising effectiveness in multi-segment sport broadcasts. Thus, the major contribution of this study comes from its practical implications in sports marketing, particularly for marketers who use multi-segment sport programs to promote their brands. The findings of this study also help inform broadcasters in establishing ad rates based on the position of the pods in which the ads are embedded.

References

Babiak, K., & Wolfe, R. (2006). More than just a game? Corporate social responsibility and Super Bowl XL. Sport Marketing Quarterly, 15, 214-222.

Beasley, F. M., Shank, M. D., & Ball, R. W. (1998). Do Super Bowl viewers watch the commercials? Sport Marketing Quarterly, 7, 33-40.

Biswas, D., Grewal, D., & Roggeveen, A. (2010). How the order of sampled experiential products affects choice. Journal of Marketing Research, 47(3), 508-519.

Blackshaw, P. (2009). Maximizing Super Bowl advertising ROI in a paid vs. earned media environment. Retrieved from http://blog.nielsen.com/nielsenwire/consumer/maximizing-super-bowl-advertising-roi-in-a-paid-vs-earned-media- environment/

Brunel, F. F., & Nelson, R. N. (2003). Message order effects and gender differences in advertising. Journal of Advertising Research, 43, 330-341.

Burke, R. R., & Srull, T. K. (1988). Competitive interference and consumer memory for advertising. Journal of Consumer Research, 15, 55-68.

Carroll, D. M. (2009). Commercial programming at a single-sport cable channel: Strategies and practices at Golf Channel. International Journal of Sport Communication, 2, 484-499.

Cianfrone, B. A., & Zhang, J. J. (2006). Differential effects of television commercials, athlete endorsements, and venue signage during a televised action sports event. Journal of Sport Management, 20, 322-344.

Cohen, J., & Cohen, P. (1983). Applied multiple regression/correlation analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Lawrence Erlbaum.

Cornwell, T. B., & Maignan, I. (1998). An international review of sponsorship research. Journal of Advertising, 27(1), 1-21.

Dahlen, M., Rosengren, S., & Torn. F. (2008). Advertising creativity matters. Journal of Advertising Research, 48, 392-403.

Decrop A. (2007). The influence of message format on the effectiveness of print advertisements for tourism destinations. International Journal of Advertising, 26, 505-525.

Easterbrook, J. (1959). The effect of emotion on cue utilization and the organization of behavior. Psychological Review, 66(3), 183-201.

Eastman, S. T., & Newton, G. D. (1999). Hitting promotion hard: A network response to channel surfing and new competition. Journal of Applied Communication Research, 27, 73-85.

Eastman, S. T., Schwartz, N. C., & Cai, X. (2005). Promoting movies on television. Journal of Applied Communication Research, 33, 139-158.

Galloway, G. (2009). Humor and ad liking: Evidence that sensation seeking moderates the effects of incongruity-resolution humor. Psychology & Marketing, 26(9), 779-792.

Gurhan-Canli, Z. (2003). The effect of expected variability of product quality and attribute uniqueness on family brand evaluations. Journal of Consumer Research, 30(1), 105-114.

Ha, L., & McCann, K. (2008). An integrated model of advertising clutter in offline and online media. International Journal of Advertising, 27, 569-592.

Haugtvedt, C .P., & Wegener, D. T. (1994). Message order effects in persuasion: An attitude strength perspective. Journal of Consumer Research, 21(1), 205-218.

Hsu, K. (2012). The advertising effects of corporate social responsibility on corporate reputation and brand equity: Evidence from the life insurance industry in Taiwan. Journal of Business Ethics, 109, 189-201.

Krosnick, J. A., & Alwin, D. F. (1987). An evaluation of a cognitive theory of response-order effects in survey measurement. Public Opinion Quarterly, 51, 201-219.

Lang, A., Sanders-Jackson A., Wang, Z., & Rubenking, B. (2013). Motivated message processing: How motivational activation influences resource allocation, encoding, and storage of TV messages. Motivation & Emotion, 37, 508-517.

Lang, A. (2000). The information processing of mediated messages: A framework for communication research. Journal of Communication, 50, 46-70.

Laroche, M., Cleveland, M., & Maravelakis, I. (2006). Competitive advertising interference and ad repetition effects: Comparing high-share and low-share brands. International Journal of Advertising, 25, 271-307.

Loginova, O. (2009) Exposure order effects and advertising competition. Journal of Economic Behavior & Organization, 71, 528-538.

Lyberger, M. R., & McCarthy, L. (2001). An assessment of consumer knowledge of, interest in, and perceptions of ambush marketing strategies. Sport Marketing Quarterly, 10(3), 130-137.

Matthes, J., Schemer, C., & Wirth, W. (2007). More than meets the eye: Investigating the hidden impact of brand placements in television magazines. International Journal of Advertising, 26(4), 477-503.

Moorman, M., Neijens, P. C., & Smit, E. G. (2005). The effects of program responses on the processing of commercials placed at various positions in the program and the block. Journal of Advertising Research, 45, 49-59.

Moorthy, S., & Hawkins, S. A. (2005). Advertising repetition and quality perception. Journal of Business Research, 58, 354-60.

Murphy, J., Hofacker, C., & Mizerski, R. (2006). Primacy and recency effects on clicking behavior. Journal of Computer-Mediated Communication, 11, 522-535.

Nail, J. (2007). Visibility versus surprise: Which drives the greatest discussion of Super Bowl advertisements? Journal of Advertising Research, 47, 412-419.

Newell, S. J., & Wu, B. (2003). Evaluating the significance of placement on recall of advertisements during the Super Bowl. Journal of Current Issues & Research in Advertising, 25, 57-67.

Newstead, K., & Romaniuk, J. (2010). Cost per second the relative effectiveness of 15- and 30-second television advertisements. Journal of Advertising Research, 50, 68-76.

O'Reilly, N., Lyberger, M., McCarthy, L., Seguin, B., & Nadeau, J. (2008). Mega-special-event promotions and intent to purchase: A longitudinal analysis of the Super Bowl. Journal of Sport Management, 22, 392-409.

Peter, J., Bonfrer, A. & Dhar. S. (2008). The effect of competitive advertising interference on sales for packaged goods. Journal of Marketing Research, 45, 211-225.

Petty, R. E., Tormala, Z. L., Hawkins, C., & Wegener, D. T. (2001). Motivation to think and order effects in persuasion: The moderating role of chunking. Personality and Social Psychology Bulletin, 27, 332-344.

Pieters, R. G. M., & Bijmolt, T. H. A. (1997). Consumer memory for television advertising: A field study of duration, serial position, and competition effects. Journal of Consumer Research, 23, 362-372.

Pyun, D. Y., & James, J. (2009). Enhancing advertising communications: Developing a model of beliefs about advertising through sport. International Journal of Sport Communication, 2, 1-20.

Riebe, E., & Dawes, J. (2006). Recall of radio advertising in low and high advertising clutter formats. International Journal of Advertising, 25, 71-86.

Scott, T. (2005). Serial position effects in recall of television commercials. Journal of General Psychology, 132, 151-163.

Simon, H. A. (1957). Models of man. New York, NY: Wiley.

Simon, H. A. (1974). How big is a chunk? Science, 183(4124), 482-488.

Smith, A. (2013). Super Bowl ad price hits record $4 million. Retrieved from http://money.cnn.com/2013/01/03/news/companies/cbs-super-bowl/

Smit, E. G., Van Meurs, L., & Neijens, P. C. (2006), Effects of advertising likeability: A 10-year perspective. Journal of Advertising Research, 46, 73-83.

Soh, H., Reid, L. N., & King, K.W. (2009). Measuring trust in advertising. Journal of Advertising, 38, 83-103.

Tobar, D. A. (2006). Affect and purchase intentions of Super Bowl XL television spectators: Examining the influence of sport fandom, age, and gender. Sport Marketing Quarterly, 15, 243-252.

Tomkovick, C., Yelkur, R., & Christians, L. (2001). The USA's biggest marketing event keeps getting bigger: An in-depth look at Super Bowl advertising in the 1990s. Journal of Marketing Communications, 7, 89-108.

Webb, P. H. (1979). Consumer information processing in a difficult media environment. Journal of Consumer Research, 6, 325-336.

Weber, R., Westcott-Baker, A., & Anderson, G. (2013). A multilevel analysis of antimarijuana public service announcement effectiveness. Communication Monographs, 80, 302-330.

Zhao, X. (1997). Clutter and serial order redefined and retested. Journal of Advertising Research, 37(5), 57-73.

Yongick Jeong, PhD, is an assistant professor and the advertising area head in the Manship School of Mass Communication at Louisiana State University. His research interests include advertising, new and entertainment media, and measuring advertising effectiveness.

Hai Tran, PhD, is an assistant professor in the College of Communication at DePaul University. His research interests include multimedia effects, agenda setting, information processing, and research methodology.

Table 1
Quarter Position Effects

                                        Model 1a              1b
                                   Brand Recognition      Ad Liking

1. Independent Variables
  1st Quarter                        9.50 (.34) ***      -.05 (-.003)
  2nd Quarter                        8.65 (.27) ***       1.06 (.05)
  3rd Quarter                          3.53 (.10)        -.59 (-.03)
2. Incremental/Total [R.sup.2]
  Quarter Order (%)                     10.1 ***              .4
  Ad Frequency (%)                      13.0 ***              .5
  Ad Clutter (%)                        15.3 ***             1.5
  Advertising Year (%)                   6.4 *             49.5 ***

Total Model (%)                         44.8 ***           51.9 ***

Note: Cell entries in section 1 are regression coefficients
(standardized beta coefficients are shown in parentheses).
(*: p < .05, **: p < .01, ***: p < .001)

Table 2
Pod Position Effects (Within-Quarter Position)--Brand Recognition

                                   Model 2a           3a
                                  1st Quarter     2nd Quarter

1. Independent Variables
  1st Commercial Break            1.61 (.04)      4.55 (.10)
  2nd Commercial Break            6.46 (.15)     -3.84 (-.08)
  3rd Commercial Break           -1.47 (-.04)     3.35 (.07)
  4th Commercial Break            1.14 (.03)      1.96 (.05)
2. Incremental/Total [R.sup.2]
  Commercial Break Order (%)          2.1             3.4
  Ad Frequency (%)                 31.3 ***        39.2 ***
  Ad Clutter (%)                      .8              .7
  Advertising Year (%)              15.8 *            5.7

Total Model (%)                    50.1 ***        49.1 ***

                                      4a              5a
                                  3rd Quarter     4th Quarter

1. Independent Variables
  1st Commercial Break            -.96 (-.02)     9.66 (.19)
  2nd Commercial Break           -1.11 (-.03)     6.72 (.14)
  3rd Commercial Break            6.70 (.14)       .60 (.01)
  4th Commercial Break             .20 (.01)      .17 (.003)
2. Incremental/Total [R.sup.2]
  Commercial Break Order (%)          2.0            13.5
  Ad Frequency (%)                 42.7 ***        41.9 ***
  Ad Clutter (%)                      3.0             2.1
  Advertising Year (%)              10.1 *            4.1

Total Model (%)                    57.8 ***        61.7 ***

Note: Cell entries in section 1 are regression coefficients
(standardized beta coefficients are shown in paren- theses).
(*: p < .05, **: p < .01, ***: p < .001)

Table 3
Pod Position Effects (Within-Quarter Position)--Ad Liking

                                   Model 2b           3b
                                  1st Quarter     2nd Quarter

1. Independent Variables
  1st Commercial Break            .02 (.001)     -1.04 (-.04)
  2nd Commercial Break            -.84 (-.03)    -1.23 (-.04)
  3rd Commercial Break           -7.10 (-.25)     1.68 (.06)
  4th Commercial Break            3.33 (.11)      3.71 (.14)
2. Incremental/Total [R.sup.2]
  Commercial Break Order (%)          7.3             3.5
  Ad Frequency (%)                    5.8            6.9 *
  Ad Clutter (%)                      .4              1.8
  Advertising Year (%)             55.6 ***        50.2 ***

Total Model (%)                    69.1 ***        62.5 ***

                                      4b              5b
                                  3rd Quarter     4th Quarter

1. Independent Variables
  1st Commercial Break           -4.10 (-.15)     5.27 (.20)
  2nd Commercial Break           -5.94 (-.22)    -2.18 (-.09)
  3rd Commercial Break           -3.20 (-.11)      .31 (.01)
  4th Commercial Break           -5.24 (-.20)      .42 (.01)
2. Incremental/Total [R.sup.2]
  Commercial Break Order (%)          1.6            13.6
  Ad Frequency (%)                   8.9 *            8.3
  Ad Clutter (%)                       0              .2
  Advertising Year (%)             49.1 ***        43.8 ***

Total Model (%)                    59.5 ***        65.9 ***

Note: Cell entries in section 1 are regression coefficients
(standardized beta coefficients are shown in parentheses).
(*: p < .05, **: p < .01, ***: p < .001)

Table 4
Pod Position Effects (Between-Quarter Position)--Brand Recognition

                              Model 6a           7a            8a
                               1st Pod        2nd Pod        3rd Pod

1. Independent Variables
  1st Quarter               12.89 (.29) *   14.05 (.25)    5.11 (.12)
  2nd Quarter               12.93 (.30) *    1.81 (.04)    8.69 (.20)
  3rd Quarter                4.17 (.10)      3.38 (.06)    7.88 (.17)
2. Incremental/Total
[R.sup.2]
  Commercial Break (%)           9.3             5             4.2
  Ad Frequency (%)            46.4 ***        41.2 ***      36.6 ***
  Ad Clutter (%)                 3.8             0             5.7
  Advertising Year (%)           6.4            4.3          11.8 *

Total Model (%)               66.0 ***        50.5 ***      58.2 ***

                                 9a               10a
                               4th Pod          5th Pod

1. Independent Variables
  1st Quarter                13.34 (.32)    16.66 (.48) **
  2nd Quarter                9.88 (.26)       13.03 (.32)
  3rd Quarter                7.07 (.18)       1.19 (.03)
2. Incremental/Total
[R.sup.2]
  Commercial Break (%)           5.6            15.8 *
  Ad Frequency (%)            39.2 ***         35.6 ***
  Ad Clutter (%)                 1.2              .2
  Advertising Year (%)         11.9 *            12.3

Total Model (%)                 57.9           63.8 ***

Note: Cell entries in section 1 are regression coefficients
(standardized beta coefficients are shown in parentheses).
(*: p < .05, **: p < .01, ***: p < .001)

Table 5
Pod Position Effects (Between-Quarter PositAd Liking

                              Model 6b          7b             8b
                              1st Pod        2nd Pod        3rd Pod

1. Independent Variables
  1st Quarter                .83 (.03)      2.01 (.06)    -4.72 (-.18)
  2nd Quarter               -1.85 (-.07)   -.05 (-.002)    2.13 (.08)
  3rd Quarter               -.83 (-.03)     .36 (.01)      .57 (.02)
2. Incremental/Total
[R.sup.2]
  Commercial Break (%)           .6             .3            3.9
  Ad Frequency (%)              8.9            6.8            9.0
  Ad Clutter (%)                0.1            7.0             .2
  Advertising Year (%)        54.0 ***       48.8 ***       54.5 ***

Total Model (%)               63.5 ***       62.9 ***       67.5 ***

                                 9b              10b
                               4th Pod         5th Pod

1. Independent Variables
  1st Quarter                4.65 (.19)     -1.58 (-.07)
  2nd Quarter                3.01 (.13)      -.35 (-.01)
  3rd Quarter                -.67 (-.03)     4.19 (.13)
2. Incremental/Total
[R.sup.2]
  Commercial Break (%)           4.0             2.0
  Ad Frequency (%)               5.5             1.8
  Ad Clutter (%)                 1.3             .1
  Advertising Year (%)        49.4 ***        53.0 ***

Total Model (%)               60.2 ***        56.9 ***

Note: Cell entries in section 1 are regression coefficients
(standardized beta coefficients are shown in parentheses).
(*: p < .05, **: p < .01, ***: p < .001)