When viewers avoid commercials by changing from channel to channel it is called?

'Zapping' refers to the practice of viewers of changing channels while watching television programs. A study was conducted to examine the impact of the zapping behavior on advertising effectiveness by analyzing its relationship with brand choice probability. It also investigated the effect of TV remote, cable TV, VCR ownership and the demographic characteristics of households on zapping frequency. In addition, the study examined the profile of households likely to be heavy zappers. A major finding of the study was that television viewers tend to zap while watching a program instead of during commercial breaks. Results also showed that household zapping behavior is significantly influenced by the availability of TV remote-control units and cable TV. The tendency to zap is greatest among households with several members, with children under 18, with video tape recorders, and with a college-educated member.

Full Text: 

Over the past decade, there have been growing concerns in the advertising industry about the potentially deleterious effects of consumer video technology on the effectiveness of IV advertising. In particular, the growth of cable television programming, the penetration of VCRs, and remote-control-operated TVs all suggest the possibility of significant losses in the effectiveness of TV commercials (e.g., Kaplan, 1985; Heeter and Greenberg, 1g85; Yorke and Kitchen, 1985; Greene, 1988; and Sylvester, 1990). It has recently been reported that 70 percent of homes had remote tuning capability. In homes so equipped, an increase m channel switching of about 75 percent was noted (Sylvester, 1990). These statistics raise fears for advertisers that households will zap out all or portions of commercials, thus reducing the advertising impact on the full audience size that they are paying for. Therefore, it is of interest for managers to find out how much of this channel switching occurs and how it affects the effectiveness of commercials.

To begin with, it is necessary to define zapping within the context of our study. The term zapping has been variously used to connote "channel switching" (Kaplan, 1985), "muting of commercials while on the air," "fast forwarding of video-taped commercials" (e.g., Tauber, 1985), and so on. Here we adopt the broad definition of the term for our study. Accordingly, we define zapping as the phenomenon where individuals viewing television programs switch channels. This broad definition has been similarly adopted by others in several past studies (e.g., Heeter and Greenberg, 1985; Kaplan, 1985). We also study the changing of channels during commercials which we call "commercial zapping." While the idea of a measure of commercial avoidance is appealing, it is difficult to operationalize such a measure given the limitations of available data. In particular, since our data does not allow us to determine why people zapped during a particular commercial (i.e., to avoid commercials or to view another program), we restrict our attention to the more general definition of commercial zapping as well.

While the zapping phenomenon itself has been examined in past studies, no research has yet specifically investigated the relationship and the effect of zapping on brand purchase behavior. In this study, we first examine the impact of TV remote, cable TV, and VCR ownership as well as that of household demographic characteristics in relation to zapping frequency. This provides information about the profile of households likely to be heavy zappers. Once this profile is established, we then turn to the issue of studying the relationship between zapping and brand choice probability in order to measure the impact of zapping on advertising effectiveness.

Zapping can influence advertising effectiveness in two ways. First, if a household zaps out of a program during a commercial break, ads will have no impact because they are not viewed. However, a more interesting question is what happens to the effectiveness of ads when only a portion of the ad is seen as a consequence of zapping. To address this question, we examine channel-switching behavior during the viewing of TV commercials and investigate whether the worth of partial commercial exposures is different from those of commercials that are viewed without interruption. This is accomplished by means of a mathematical model that is based on a multinomial logit analysis which examines the relationship of commercial zapping as well as other explanatory variables to brand-choice response behavior.

Overview of Related Studies

The majority of related prior studies have focused on the description of the zapping phenomenon. Thus, in some of the earliest reported studies on zapping in the United States (e.g., Kaplan, 1985; Heeter and Greenberg, 1985), attention is given to describing the pattern and extent of channel-switching behavior as well as profiling zapping-prone households. Thus focus has been shared by Yorke and Kitchen (1985) who similarly examine zapping behavior in the United Kingdom.

In a more recent study, Olney et al. (1991) examine a hierarchy of effects in which advertising contents influence emotions and attitudes toward an ad. These intervening variables are then linked to a commercial-zapping measure. Thus, the primary focus of the latter study is on the determinants of zapping behavior. However, as in the previous studies, it does not investigate the relationship of zapping to advertising effectiveness.

Little is known about the effect of zapping on advertising effectiveness due to the sparseness of research studies linking the two variables. An exception to this is a Gallup and Robinson study by Greene (1988). This study reports that zappers show only a slightly reduced commercial recall of 3 percentage points relative to nonzappers. It then concludes that zapping behavior itself may have a positive impact on advertising recall in that it forces viewing action back to the TV set when the viewer might not otherwise be attentive.

Greene's (1988) contribution is that it relates zapping behavior to advertising effectiveness. However, the study is limited by its use of a survey-based measure of day-after TV commercial recall as well as self-reported measures of zapping from a sample of viewers. As such, the methodology relies on potentially biased self-reported data about viewer zapping behavior. Moreover, it uses recall as the criterion of advertising effectiveness and thus does not directly measure the effect of zapping on purchase behavior.

In contrast to the conclusions reached by Greene (1988), Percy Co. has suggested that zapping is indeed a potentially significant problem for advertisers (see Kneale, 19&8). The Percy study electronically tracked viewing audience and zapping for a sample of 1,000 homes in the New York metropolitan area to measure "commercial ratings." TV set monitoring was accomplished by means of electronic metering and computer devices which provided second-by-second information about TV set status, while viewer presence in the TV viewing room was measured by heat sensors and people meters.

Among its interesting findings, the Percy study noted that while a new specialty commercial, such as the Pepsi ad featuring singer Michael Jackson at the 1988 Grammy music awards, showed a relatively low audience loss (about 2 percent), certain ads, aired during competing popular sports events, showed staggering losses of as much as 43 to 53 percent due to zapping. The study concludes that a "commercial rating" is typically much lower than that of a show. Although the Percy study shows that a significant amount of zapping behavior does exist by means of a very sophisticated measurement methodology, it does not specifically evaluate the effect of this behavior on advertising effectiveness. Nevertheless, the advertising concerns raised by the Percy study results have led to controversy and debate within the TV and advertising industry. Thus, TV networks and some ad agencies have disputed the results of the Percy study, questioned the representativeness of the New York city sample used, and dismissed "zapping" as a minor problem (see Kneale, 1988).

In summary, we know that zapping is a fairly widespread phenomenon and potentially of considerable-concern to advertisers. However, we have very little understanding of the way zapping affects advertising effectiveness. In this study, the main focus is on the impact of zapping on advertising effectiveness. Stewart and Furse (1986) note that most studies that have focused on advertising effectiveness rely on effectiveness criteria such as advertising persuasion, recall, comprehension, and executional factors. Thus, considerations of brand choice and purchase-related effects are typically ignored. A major reason for this oversight has been due to limitations in available data. However, new advances in the collection of single-source data now allow more explicit links to be made between advertising ad purchase behavior (e.g., see Assael and Poltrack, 1991).

Several recent studies have used analytical models based on single-source scanner panel data to study the effects of advertising exposures on household purchase behavior. Tellis (1988) uses tobit models to examine the relationship of brand choice and purchase quantity to advertising exposures and other explanatory variables. Other researchers have also studied advertising exposure effects by applying multinomial logit models to single-source data (e.g. Deighton et al., 1988; Batra et al., 1989; Russel and Winer, 1989; Pedrick and Zufryden, 1991). Although the latter studies suggest useful approaches to model the relationship between advertising exposures and brand-choice behavior, none of these have examined the impact of zapping on the effectiveness of ad exposures. In our study, "Are specifically address and provide insight into this issue by applying multinomial logit analysis to our scanner panel database. Thus, we focus on the effects of commercial zapping as well as other relevant causal variables on a household's brand purchase behavior.

Empirical Results

Data Source. The data used in this study have been previously described by Pedrick and Zufryden (1991) and are from a new data source supplied by the A. C. Nielsen Co. This database contains store-level causal and household-level purchase data along with TV commercial exposure information obtained from individual household TV meters. In addition, complete demographic information is also available for the household panel.

More specifically, the data consists of records of purchase transactions for the yogurt product category for a panel of 584 households over a period of two years. This product category was chosen for this study because advertising exposures had been previously demonstrated to significantly affect brand-choice probabilities in this category by Pedrick and Zufryden (1991). The purchase transactions were recorded with electronic scanning equipment in all major supermarkets in Sioux Fall, South Dakota, an A. C. Nielsen electronic test market. For each supermarket chain, detailed records of causal data at point-of-purchase, such as displays, feature ads, and price promotions, for yogurt brands in the category were also available on a weekly basis for each Universal Product Code (UPC).

The advertising exposure information also provides concurrent TV-tuning data for the households within the panel sample. Thus, a significant advantage of our database is that it provides information not only about the yogurt-commercial exposures that each household had an opportunity to see but also about any channel changes and consequent viewing interruptions during any portion of the transmission of aired commercials. In total, almost 50,000 household yogurt-commercial exposures are available in our data. In this study, an exposure is defined as a household's opportunity to see a TV commercial scheduled at a particular time. Of these, roughly 5 percent were partially interrupted by channel switching and thus classified as zapper exposures. Consequently, these data provided a unique opportunity to extend previous advertising research studies by considering the effects of zapping behavior on brand purchase probability.

Descriptive Results on Zapping Behavior. As a first step in our study, we investigated the profile of zappers. Based on the available single-source data, w. e examined a set of household-related factors that were judged likely to predict household zapping behavior. These factors included the presence of a remote-control device, the number of available channel choices, VCR usage patterns, household income, education, the number of children in the household, and the total number of household members.

To examine the relationship of these factors to zapping behavior, we calculated four measures of household zapping propensity from our single-source scanner panel database. These measures were expressed in terms of the number of zaps per hour of television watched under four situations: (1) during all time periods of television viewing, (2) during prime-time television viewing, (3) during commercial breaks scheduled over all time periods, and (4) during commercial breaks scheduled in prime-time. These measures are represented in columns 1 thru 4 of Table 1. In addition, results in each column are broken down according to the levels of demographic and other explanatory variables that were found to reflect differences in average household zapping behavior.

Table 1
Average Number of Zaps per Hour of TV Viewing
                                                  Zaps during
                            All zaps            commercials only
Demographic description  All time  Prime-time  All time  Prime-time
  of household:          periods     only       periods   only

TV remote control
  Present                  3.34       5.21        1.07      1.28
  Absent                   2.11       3.22        0.72      0.84
Cable TV
  Present                  3.24       5.15        1.06       1.28
  Absent                   1.89       2.69        0.63       0.69
VCR time shifter
  Yes                      3.21       5.04         1.07       1.32
  No                       2.68       4.13         0.87       1.01
Income
  Low                      2.18        3.36        0.72       0.82
  Medium                   2.81        4.41        0.95       1.13
  High                     3.39        5.21        1.07       1.29
Single member household
  Yes                      1.62        2.48         0.53       0.61
  No                       3.00        4.64         0.98       1.16
Children under 18 in
 household
  Yes                      3.08         4.93        1.04       1.27
  No                       2.64         4.00        0.85       0.98
College-educated
 household member
  Yes                      3.07          4.81        0.98      1.18
  No                       2.65          4.09        0.88      1.04
Overall sample average     2.81          4.35        0.92      1.09

The overall sample averages in Table 1 show that zapping levels tend to be much higher in prime-time but that commercial zapping is only slightly higher during prime-time. Thus, it is noted that the difference in the overall mean zapping levels between columns 1 and 2 is highly statistically significant. However, the difference in commercial-zapping levels between columns 3 and 4 is not significant at the 95 percent confidence level.

In general, commercial zaps make up between one-fourth to one-third of overall household zapping in or database. This is an interesting result as it indicates that a majority of household channel-switching occurs during actual programming rather than during commercial breaks. The fact that this pattern is shown to be exaggerated during prime-time, when more and presumably better quality viewing options are available, suggests that a primary cause of zapping may be a desire to seek variety among other programming alternatives rather than a desire to avoid commercials. While interesting, this finding remains controversial as other studies based on self-reported measures of zapping behavior (e.g., Kaplan, 1985) have found that commercial zapping is higher than non-commercial zapping during the beginning of programs and roughly equal during the programs.

To assess the significance of the household-level explanatory variables and their impact on average zapping rates, we estimated dummy variable regression models to predict each of the four household zapping measures described above. The estimated coefficients for these models are presented in Table 2. We now describe and discuss the explanatory variables found to explain zapping behavior.

Table 2
Zapping Regression Model Coefficients with t-Statistics
in Parentheses
                                 Dependent variables
                                            Zaps during
                               All zaps    commercials only
                          All time  Prime-time  All time
Prime-time
Independent variables ***  periods    only       periods    only
Intercept                   1.32(*)   1.61(*)    0.44(*)   0.41(*)
                            (5.43)    (4.55)     (5.32)    (3.93)
TV remote control present    0.91(*)  1.47(*)     0.25(*)   0.29(*)
                            (5.18)    (5.74)      (4.17)    (3.89)
Cable TV present             1.13(*)  2.14(*)     0.36(*)   0.50(*)
                             (6.20)   (8.06)       (5.77)    (6.40)
VCR time shifter             0.19     0.34         0.10(**)
0.18(*)
                             (0.96)   (1.20)       (1.50)    (2.12)
Medium income                0.23     0.37        0.10
0.15(**)
                             (1.07)   (1.15)      (1.38)     (1.56)
High income                  0.53(*)  0.62(**)    0.16(*)
0.21(*)
                             (2.03)   (1.63)      (1.78)     (1.85)
Single member household     -0.71(*)  -0.97(*)   -0.22(*)
-0.22(*)
                            (-2.58)   (-2.41)    (-2.34)
(-1.89)
Children under 18 in          0.06     0.34        0.08
0.15(*)
  household                 (0.31)     (1.27)     (1.22)     (1.96)
Household contains a        0.29(**)   0.56(*)      0.05      0.08
  college-educated member   (1.58)     (2.11)      (0.83)
(1.06)
(*) Significant at the 95% Confidence Level.
(**) Significant at the 90% Confidence Level.
(***) Coding of Variables: 1 (yes); O (no).

TV Remote. The model coefficients indicating the presence of a TV remote-control unit were highly significant for all four measures of household zapping. Thus, the presence of remote-control devices appears to significantly explain the variance in channel switching. Our results are corroborated by those of several previous studies that have found that the presence of a remote-control device increases the tendency for households to engage in zapping (e.g., Kaplan, 1985; Heeter and Greenberg, 1985; and Sylvester, 1990).

Cable TV. The presence of cable TV in a household was also shown to significantly explain zapping frequency. In particular, we note that households with cable tend to engage in channel switching more frequently than the households without. This result suggests that the number of channel choices may in itself increase the frequency of household zapping. Thus, when there are more channels for the household to view, it is more likely that it will switch to another channel during commercial breaks. This may be because cable typically provides access to several times the number of channels that are available by over-the-air reception. Cable TV systems also frequently provide households with remote-tuning capability which will also tend to increase zapping frequency. Thus, as has been pointed out (Kaplan, 1985), households with cable may be using commercial breaks between programs to check the headlines on CNN or view music videos on stations such as MTV.

VCR Time Shifter. The results of Table 2 show that the use of VCRs represents another potential threat to TV advertising effectiveness. These are consistent with those of Potter et al. (1988). These results are likely due to the fact that VCRs provide the means for a viewer to preempt regular programming by viewing pretaped alternatives (source shifting). In addition, the viewing of recorded TV programs may be deferred to another time period (time shifting) or may not be viewed at all. Moreover, TV commercials may be deleted during the recording process (zapping) or by fast-forwarding (zipping) the VCR during the playback of recorded programs. Our results suggest that households that use their VCRs to record programs for later viewing are likely to zap more often than other households. These households may change channels more often because of a desire to directly control their entertainment to a greater extent. This view is supported by Heeter and Greenberg (1985) who argue that zapping is a manifestation of a certain orientation toward television and thus should not be treated as just behavior toward TV commercials. Interestingly, Heeter and Greenberg (1985) were not able to empirically detect significant increases in zapping levels by VCR users. However, their result may be due to their use of unreliable survey-based measures of household zapping.

Number of Household Members. Our results show strong evidence that multiple person households will zap more than single person households. This may be the to the fact that each individual member within a household contributes to channel changes according to his or her individual preferences. In this case, we would also anticipate that such households will exhibit a higher likelihood of checking programming on other channels during commercial breaks. Note, however, that we found no evidence that household zapping was significantly different between multiple person households of varying sizes. Consequently, the variables allowing for these differences were omitted from the final Table 2 results.

Household Income. Significant increases in commercial zapping, particularly during prime-time, were found for households with higher income. A probable reason for this is the likely correlation of household income with a number of other household characteristics that are likely to influence zapping levels For instance, households with higher incomes may be more technologically oriented and thus more likely to zap. Furthermore, television is more likely to be competing with other forms of entertainment in the case of higher income household members. This would imply that these households may be more selective about what they watch on TV. Heeter and Greenberg (1985) note that most studies do not find any significant effect of household income on the level of zapping. However, the more recent study by the Percy Co. found, as we do, that income levels were positively correlated with the extent of zapping (see Kneale, 1988).

Children in Household. Increases in commercial zapping propensity were found to be related to the presence of children under 18 within a household. This effect was shown to be particularly significant during prime-time television viewing, as shown in Table 2.

Household Education. Table 2 shows positive increases in overall zapping levels for households containing a college-educated member. This result is consistent with the notion that household education level is expected to be correlated with household income and thus may influence the selectivity of TV viewing.

It should be noted that many other variables may influence household zapping proclivity. For example, several previous studies have concluded that the variations in household zapping depend on various situational factors that may affect a household's propensity to switch channels. In particular, Olney et al. (1989) and Heeter and Greenberg (1985) suggest that the level of boredom experienced by a viewing household may contribute to channel-switching behavior. Olney et al. (1989) also suggest that the desire to seek more variety in TV programming may contribute to the rate at which households zap. In addition, Kaplan (1985) notes that annoyance with commercials may result in zapping by a household and points to the desire to avoid commercial messages as a reason for this behavior. However, we did not explicitly consider such situational factors in this study because they cannot directly be measured from our single-square data.

Impact of Zapping on Advertising Effectiveness. In the second part of our study, we examine the impact of zapping on advertising effectiveness. When a commercial is zapped, part of the message is lost and, consequently, this is likely to affect the effectiveness of the commercial. Here we focus on the measurement of the effect of partial advertising exposures as compared to the effect of noninterrupted exposures. Thus, we examine the relationship between zapping of commercials and subsequent observed brand-choice behavior using our single-source scanner database.

More specifically, we want to see if a household's probability of purchasing a particular brand is affected by whether or not previous exposures for that brand were interrupted by zapping. Our scanner database provides us with records of retail grocery environments as well as household yogurt-brand purchase and ad-exposure patterns. Thus, we have used this data to estimate an econometric model which separately estimates the impact of interrupted versus noninterrupted ad exposures. This is accomplished while controlling for changes in the retail environment over time (e.g., price, proportions, etc.) and for differences in brand loyalty across households.

This analysis is accomplished by modeling household brand-choice probability and its relationship to zapping with a multinormal logit model. This brand choice model has been widely used in the marketing literature (e.g., see Guadagni and Little, 1983; Gupta, 1988; Tellis, 1988; Pedrick and Zufryden, 1991). Thus, the brand purchase probability [P.sub.k]([x.sub.kit]) for a particular brand k, by household i, given environment [X.sub.kit] on shopping trip t can be expressed mathematically as:

[Mathematical Expression Omitted]

where: [V.sub.k]([X.sub.kit]) = Deterministic

component of

individual

Household i's

utility for Brand k S = Set of alternatives

among which a

brand choice is

made by household i [X.sub.kit] = Vector of causal

variables such as

household i brand k

loyalty, brand k

price, promotions,

coupons, and

advertising exposures

We specified the vector of causal variables [X.sub.kit] in the brand-choice model to depend on point-of-purchase marketing-mix variables, household brand loyalty, and media advertising exposures. Thus, as in Pedrick and Zufryden (1991), point-of-purchase marketing variables in the model included brand k shelf price and promotions using bath major and minor feature ads, store coupon discount size, and an index of manufacturer coupon availability. In addition, both short- and long-term household loyalty variables were also used in the model since each captures a separate aspect of consumer heterogeneity. Long-term household brand k loyalty is measured by a household's average brand k purchase rate from a data sample covering a period prior to the study. This variable captures the heterogeneity of brand preferences across households. Conversely, the short-term loyalty variable measures changes in household preferences over time as well as the carry-over effects of past marketing variables. We operationalized this variable as the number of brand k purchases made in the prior four-week period (i.e., during period t - 1) preceding the current purchase occasion.

Advertising effects were measured by the number of yogurt-category exposures household i had the opportunity to see for brand k prior to a current purchase occasion in a given four-week time period t. in contrast to Pedrick and Zufryden (1991), to look for differences in the effectiveness of advertising due to zapping, we included different model coefficients for zapped and nonzapped ad exposures. An exposure was defined to be a zapped exposure if a household eIther zapped into or out of an ad during its broadcast but viewed at least five seconds of the ad.

The choice model coefficients are estimated using purchase records from the spoonable yogurt category. Key estimation results for three nested choice models are presented in Table 3. In the first, which is used as a base case, brand choice is modeled as a function of household brand loyalty and point-of-purchase causal variables including price, promotions, and coupons. Results corresponding to this base case have been previously reported in Pedrick and Zufryden (1991) and are omitted here.

[TABULAR DATA 3 OMITTED]

The major focus of our study is on the impact of zapping on brand-choice probability. To explore this impact, the second model adds the effects of advertising exposures, while the third model considers exposures to zapped and nonzapped ads. Thus, model 3 allows zapped ads to have a different advertising coefficient from that of the nonzapped advertising exposures. It is noted that the significance of individual t-statistics support the successive addition of explanatory variables for each of the nested models. Moreover, a likelihood ratio test also shows that allowing zapped ads to differentially affect brand choice significantly improves the fit of the model. In this case, the likelihood ratio statistic for the improved fit due to allowing zapped ads to differentially affect brand choice is 8 (i.e., -2 times the difference between model log likelihood values - 2046 and - 2042). This value is significant at the 99.5 percent confidence level.

The most interesting model result is the relatively large coefficient for zapped ad exposures. In fact, the elasticity for these zapped exposures was found to be almost seven times as large as for exposures that were not subjected to zapping. Thus, commercials interrupted by channel switching appear to have a more significant and stronger relationship to household brand-purchase probabilities than do noninterrupted ads.

Conclusions

In this study, wee focused on the profile of zappers as well as the impact of zapping behavior on advertising effectiveness by means of a new single-source data base.

Among the more significant findings in our research is the result that the majority of household channel switching occurs during actual programming rather than during commercial breaks. As this pattern was shown to have even greater emphasis during prime time, when more viewing options are available this suggests that the desire to seek variety among programming may be a more salient determinant of zapping behavior than a desire to avoid commercials.

In our examination of the determinants of household zapping behavior, our study indicated that the presence of a TV remote-control unit and cable TV in a household both had a significant impact on household zapping behavior. In addition, the household profile that tended to have the greater propensity to zap was ore that included multiple-person households, higher income, children under 18, use of VCRs, and a college-educated member.

The latter results suggest some of the conditions, including audience types, when media strategies or putting in extra effort to produce zap-resistant ad copy may pay off for advertisers. For example, a number of media strategies have recently been tried by advertisers to prevent viewers from avoiding commercials by switching to another channel. These include techniques such as "roadblocking" (i.e., placing commercials on all major channels at a given time), "commercial wraparound" (splitting a 30-second commercial into two 15s), as well as various creative strategies designed to enhance the attention-getting characteristics of ads and thereby reduce commercial avoidance (e.g., see Pahwa 1990). These types of strategies may be needed to reduce losses to zapping in the case of ads which are targeting households with higher zapping propensities. For instance, our study suggests that each households may be characterized by higher incomes, multiple members with children, and as subscribers to cable TV.

With respect to the relationship of zapping to advertising effectiveness, a surprising finding was that zapped commercials were significantly more effective than noninterrupted ads with respect to their impact on brand-purchase behavior. It is hypothesized that a reason for this phenomenon is the potential heightening of viewer attention to the TV set at the time of a zap. This is likely to lead to more active processing of advertising around the time of the zap and consequently to greater effectiveness for those ads. From a managerial point of view this suggests that the potential negative impact of lost viewing audience due to zapping may at least be partially offset by the improved effectiveness of ads which are zapped sometime after they have already begun.

It should be noted that our current research cannot definitively answer the a question of why this improvement occurs since we are not able to directly measure the attention levels of individual household members with our current database. Thus, if the reported improvement in ad effectiveness is in fact due to the increased attention viewers are paying to the TV while zapping, then creative copy strategies that are specifically designed to increase viewer attention may further improve advertising effectiveness. This issue provides an interesting avenue for future research about zapping behavior and its impact on advertising effectiveness.

References

Assael, Henry, and David F. Poltrack. "Using Single Source Data to Select TV Programs Based on Purchase Behavior." Journal of Advertising Research, 31, 4 (1991) 9-17.

Batra, R.; W. R. VanHonacker; and D. Kim. "Effects of Advertising on Brand Purchases: Some Results from Single-Source Data." Paper presented at the ORSA/TIMS Marketing Science Conference, Duke University, Durham, North Carolina, 1989.

Deighton, J.; C. Henderson; and S. Neslin. "Advertising Framing Effects in Field Data." Working paper, Amos Tuck School of Business, Dartmouth College.

Greene, William F. "Maybe the Valley of the Shadow Isn't So Dark After All." Journal of Advertising Research 28, 5 (1988): 11-15.

Guadagni, Peter M., and John D. C. Little. "A Logit Model of Brand Choice Calibrated on Scanner Panel Data." Marketing Science 2 (1983): 203-38.

Gupta, Sunil. "Impact of Sales Promotion on When, What and How Much to Buy." Journal of Marketing Research 25, 4 (1988): 342-55.

Heeter, Carrie, and Bradley S. Greenberg. "Profiling the Zappers." Journal of Advertising Research 25, 2 (1985): 15-19.

Kaplan, Barry M. "Zapping--The Real Issue Is Communication." Journal of Advertising Research 25, 2 (1985): 9-12.

Kneale, Dennis. "Zapping of TV Ads Appears Pervasive: Study Details Viewing Habits, to the Second." Wall Street Journal, April 25, 1988.

Olney, Thomas J.; Morris B. Holbrook; and Rajeev Batra. "Consumer Response to Advertising: The Effects on Ad Content, Emotions, and Attitude Toward the Ad on Viewing Time." Journal of Consumer Research 17, 4 (1941): 440-53

Pahwa, Ashok. "Boom Generation More Receptive to Quality TV Ads." Marketing News, September 17, 1990.

Pedrick, James H., and Fred S. Zufryden. "Evaluating the Impact of Advertising Media Plans: A Model of Consumer Purchase Dynamics Using Single-Source Data." Marketing Science 10, 1 (1991): 111-30.

Potter, W. James; Edward Forrest; Barry Sapolsky; and William Ware. "Segmenting VCR Owners." Journal of Advertising Research 28, 2 (1988): 29-39.

Russel, G. J., and R. S. Winer. "Uncovering the Determinants of the Lagged Impact of Advertising: Consumption Feedback Versus Advertising Carry-Over." Paper presented at the ORSA/TIMS Marketing Science Conference, Duke University, Durham, North Carolina, 1989.

Stewart, David W., and David H. Furse. Effective Television Advertising. Lexington, MA: Lexington Books, 1986.

Sylvester, Alice K. "Controlling Remote." Marketing and Media Decisions, February 1990.

Tauber, Edward. "Editorial: Zapping." Journal of Advertising Research 25, 2 (1985): 5.

Tellis, Gerard J. "Advertising Exposure, Loyalty and Brand Purchase: A Two Stage Model of Choice." Journal of Marketing Research 15, 2 (1988): 134-44.

Yorke, David A., and Philip J. Kitchen (1985). "Channel Flickers and Video Speeders." Journal of Advertising Research 25, 2 (1985): 21-25.

1. FRED S. ZUFRYDEN is the Ernest W. Hahn Professor of Marketing at the University of Southern California. He received his Ph.D. in business administration from the University of California at Los Angeles 2 JAMES H. PEDRICK is the director of market response analysis for Information Resources Inc. in San Francisco, CA 3. AVU SANKARALINGAM is a doctoral student in the marketing department at the Graduate School of Business, University of Southern California.

The authors wish to acknowledge and thank the A. C. Nielsen Marketing Research Division for the data used in this study. Thanks also go to Gerry Tellis, Mike Kamins, and Valerie Folkes for their comments and suggestions on an earlier draft of this paper.

What refers to changing channels to avoid commercials?

What is the difference between zipping and zapping?

What is zipping in TV?

What is fast forwarding through commercials called?