Communicating with the sexes: male and female responses to print advertisements.

Its {Nike's} executives have come up with strategies they hope will take advantage of the differences between how women and men conceive of sport, how they shop for clothing and shoes and even what they think of celebrity athletes.

--Edward Wong, New York Times, June 2001

Nike is not alone. Sex (1) is frequently used as a basis for segmentation for a significant proportion of products and services. Such segmentation seems consistent with the attribution of specific skills or personality traits to the two sexes, and the observation that men and women have unique interests and knowledge associated with their respective social roles. From a socialization perspective, men and women have different sorts of communication and interaction with various social agents (Fischer and Arnold 1994; Helgeson and Fritz 1998; Moschis 1985), different worldviews (Brunel and Nelson 2000; Gilligan 1982), and often occupy different social roles and are subjected to different social pressures (Eagly 1987; Eagly and Wood 1991). Such differences in socialization could, at least partially, explain observed sex differences such as female interpersonal sensitivity and male self-orientation. However, while the sexes might differ along the communal/agentic continuum, these differences have blurred significantly in recent years (Hupfer 2002).

From a biological point of view, research suggests that sexual hormones are connected with differences in perceptual-motor skills observed in men and women (Berenbaum 1999). Sex differences might also exist in hemispheric lateralization (the degree to which one hemisphere is relatively dominant for various kinds of processing), such that male brains are more functionally lateralized and female brains are more integrated (Everhart et al. 2001; Gorman, Nash, and Ehrenreich 1992; Saucier and Elias 2001). In addition, some biological personality traits might be reinforced by cultural gender role norms. Women might score higher on "need for affiliation," for example, because they are socialized in a way that encourages them to believe that this is an appropriate gender role trait.

That men and women are different is commonly accepted in most societies, and a host of social and biological factors seem to drive these differences. However, an important marketing/advertising question is whether such differences translate into consistent differences in the processing and evaluation of marketing communications. In comparison with the areas of psychology and sociology, there has been a more limited examination of sex differences in the area of ad response, with the bulk of such research focusing on sexual appeals (Jones, Stanaland, and Gelb 1998; LaTour and Henthorne 1994) or gender role portrayals (Al-Olayan and Karande 2000; Ford and LaTour 1993). The few studies that have examined sex differences in ad/information processing report very mixed findings. For example, some researchers have found that women have higher ad readership for complex messages (Chamblee et al. 1993), exhibit increased discrimination ability (Meyers-Levy and Maheswaran 1991), process ads more comprehensively under conditions of low/ moderate involvement (Meyers-Levy and Sternthal 1991), and use both subjective and objective information (Darley and Smith 1995). In contrast, others report opposing findings: Ads in male magazines were found to be more complex compared with those in female magazines (Whissell and McCall 1997), and no sex differences have been found in motivation to engage in cognitive elaboration (Peracchio and Tybout 1996), preference for lexical complexity (Putrevu, Tan, and Lord 2004), recall/recognition of the ad sponsor (McDaniel and Kinney 1998), or recall of words/pictures (Ionescu 2000, 2002).

Such mixed findings are troubling because they do not generate a reasonable body of evidence on which to base an advertising strategy. Some potential reasons for such contradictory findings are differences in product/task involvement, response contexts, dependent variables, methodologies, and units of analysis (i.e., extent of data pooling). This research addresses some of the above limitations by examining sex differences in ad responses across multiple dependent variables while controlling for product/task involvement. Specifically, it tests whether men and women differ in terms of affect ([A.sub.ad] and [A.sub.b]), purchase intent, number and type of cognitive responses, and memory when exposed to print advertisements. More important, as called for recently by advertising scholars (e.g., Wolin 2003), rather than merely testing for sex differences in ad responses, these differences are examined within the context of established theories.

CONCEPTUAL DEVELOPMENT

Cognitive Abilities and Brain Lateralization

A common belief is that women excel in verbal skills, whereas men show superiority in spatial and mathematical skills. The literature seems to support this view (Halpern 1997; Lee 2000), although the overall differences seem somewhat modest. Through a meta-analysis of 165 studies of sex differences in verbal ability, Hyde and Linn (1988) found a small difference that favored girls ages five to eighteen. In terms of spatial and mathematical skills, males have consistently been found to outperform females (Geary 1996; Maccoby 1998; Maccoby and Jacklin 1974). Consistent with such findings, a recently released 32-nation OECD (Organization for Economic Cooperation and Development) study shows that girls are better at reading and boys are better at math in every country surveyed (Sokoloff 2001).

The human brain is divided into two hemispheres, and lateralization refers to the specialization in the functioning of each hemisphere: The left hemisphere specializes in verbal abilities and the right hemisphere specializes in spatial perception (Hansen 1981). At some point in development, lateralization begins, and one hemisphere becomes dominant in its control of an individual's behavior. Recent clinical and experimental research shows that the two hemispheres are more symmetrically organized (i.e., integrated) in females and more specialized in males (Everhart et al. 2001; Gorman, Nash, and Ehrenreich 1992; Saucier and Elias 2001). The timing and extent of such lateralization might affect the development of specific skills. Since the most consistent (though modest) sex differences in cognitive functioning are found in tasks involving either verbal or spatial skills (Geary 1996; Hyde and Linn 1988; Lee 2000), the differential lateralization might underlie, to some extent, the sex differences in cognitive skills.

The more integrated and symmetrical brain attributed to women suggests that the provision of verbal descriptions might be beneficial to a female audience. On the other hand, the specialized hemispheric brain attributed to men suggests that they might require nonverbal reinforcement (e.g., pictures, graphs, music, etc.) of the verbal product information contained in a message. However, the provision of reinforcing nonverbal information is not likely to disadvantage the more flexible and symmetrical brain processes attributed to women.

H1: Compared with men, women will exhibit more positive affect ([A.sub.ad] and [A.sub.b]) and stronger purchase intention (PI) toward predominantly verbal ads.

H2: No a priori sex differences in affect ([A.sub.ad] and [A.sub.b]) or PI are expected toward ads that provide visual reinforcement of verbal information.

The brain lateralization differences attributed to the sexes are also likely to influence product evaluation and judgment. For example, men and women might attach varying levels of salience to product attributes, use advertised product information in different ways when rendering judgment, and undertake different types of message elaboration. The more functionally lateralized male brain tends to process information on a piecemeal basis. Specifically, specialized processors (i.e., males) are likely to value highly focused information pertaining to one or a few key attributes. An example would be a simple verbal description of one or two key features accompanied by nonverbal reinforcement (pictures, graphs, music, etc.) of these features. On the other hand, the less functionally lateralized female brain tends to process information holistically, considering all the available information. Hence, integrated/symmetrical processors (i.e., females) are likely to value information-rich sources. An example would be the provision of detailed information on multiple features/benefits accompanied by rich/complex visual portrayals. The findings that men encode fewer ad claims (Gilligan 1982) and women interpret verbal and nonverbal cues more accurately than men (Chamblee et al. 1993; Everhart et al. 2001; Goos and Silverman 2002) are consistent with the above arguments.

H3: Compared with women, men will generate more positive affect ([A.sub.ad] and [A.sub.b]) and stronger PI toward ads that are simple and focus on one or a few key features.

H4: Compared with men, women will have more positive affect ([A.sub.ad] and [A.sub.b]) and stronger PI toward ads that are complex and contain a lot of verbal and visual information. (2)

Aggression, Affiliation, and Social Roles

The notion that men are more aggressive than women has been supported in empirical research across multiple settings, measurement instruments, and age groups. Furthermore, the fact that these differences emerge quite early in life and are found cross-culturally suggests that biological factors may be involved (Costa, Terracciano, and McCrae 2001; Maccoby 1998; Maccoby and Jacklin 1974). In addition, researchers studying personality traits have consistently found that women score higher on need for affiliation than men (Schultheiss and Brunstein 2001), and standard paper-and-pencil psychological tests consistently indicate that women are more anxious, moody, and fearful than men (Costa, Terracciano, and McCrae 2001).

Social role theory suggests that the division of labor between the sexes creates gender-role expectations, which then lead to differences in social behavior and personality. According to this theory, men and women possess or acquire attributes suited for the roles that they typically occupy. The psychological attributes and social behaviors associated with these roles translate into the agentic characteristics frequently observed in men and the communal characteristics frequently observed in women (Eagly 1987; Eagly and Wood 1991; Helgeson and Fritz 1998). Men are more assertive and aggressive because historically they have been more likely to assume positions of power and leadership. Women, on the other hand, have nor played these roles (at least not as often as men), and thus do not develop these characteristics. Hence, the agentic content of the male gender role is assumed to derive from the typical roles men occupy in society, and the communal content of the female gender role is derived from the domestic/maternal role and from occupational roles filled disproportionately by women (e.g., in nursing, teaching, and secretarial work). In sum, the sex differences in aggression, affiliation, and social roles suggest that men would be more persuaded by messages that contain agentic sentiments and women would be better persuaded by messages containing communal elements. Such reasoning leads to the following hypotheses:

H5: Compared with women, men will exhibit more positive affect ([A.sub.ad] and [A.sub.b]) and stronger PI toward ads that feature comparative appeals.

H6: Compared with men, women will have more positive affect ([A.sub.ad] and [A.sub.b]) and stronger PI toward ads that feature harmonious relationships. (3)

Information-Processing Styles

The selectivity hypothesis suggests that, except under high-involvement conditions, sex differences emerge because men are more likely to be driven by overall message themes or schemas, whereas women are more likely to engage in detailed elaboration of message content (Meyers-Levy 1989; Meyers-Levy and Maheswaran 1991; Meyers-Levy and Sternthal 1991). Specifically, men are considered to be selective processors who often do not engage in comprehensive processing of all available information before rendering judgment. Instead, they rely on various heuristics in place of detailed message elaboration. These heuristics involve a cue or cues that are highly available and salient, and convergently imply a particular inference. Such processing implies that men will often base their judgment on a select subset of all available information. Women, on the other hand, are considered to be comprehensive processors who attempt to assimilate all available information before rendering judgment. Unless restricted by memory constraints, women will attempt effortful elaboration of all available information. Therefore, women will give equal weight to self-generared and other-generated information, encode more message claims, and elaborate on specific claims more extensively (Meyers-Levy 1989). It is commonly acknowledged that comprehensive processors generate more cognitive responses compared to heuristic processors (Chaiken 1980; Petty, Ostrom, and Brock 1981). Hence, under low/ moderate levels of involvement, the selectivity hypothesis predicts the following:

H7: Comprehensive-processing women will generate more cognitive responses than heuristic-processing men.

A related but alternative interpretation of the observed sex differences is based on research in the area of cognitive psychology that suggests that there are two types of elaboration that facilitate comprehension in alternative ways. One type of elaboration is item-specific processing, which stresses attributes that are unique or distinctive to a particular message. This sort of processing might occur spontaneously when people receive multiple message cues that are, in context, largely unrelated to each other. The second type of elaboration, relational processing, emphasizes similarities or shared themes among disparate pieces of information. This sort of processing might occur spontaneously when people receive many similar message cues (Einstein and Hunt 1980; Hunt and Einstein 1981).

The bulk of research with respect to the item-specific/relational processing dichotomy has focused on situational factors (e.g., message type, contextual cues, etc.) that might encourage the choice of a particular processing style (Hunt and Einstein 1981; Hunt and Seta 1984; Jun et al. 2003; Malaviya, Kisielius, and Sternthal 1996). However, individual differences such as culture and gender-role prescriptions might predispose consumers to a particular type of processing. Men, who are primarily concerned with self-focused agentic goals, are more likely to focus on those message claims that affect them directly--not all claims. Women, who are driven by relationship-oriented communal goals, are more likely to consider all aspects of the message since they are interested in its global impact. In other words, men are likely to pay attention to only those attributes that have the greatest personal impact, whereas women are likely to consider several attributes in an attempt to decipher the intricate interrelationships between them. Hence, all else being equal, men might undertake item-specific processing and women might engage in relational processing.

In contrast to the selective versus comprehensive processing dichotomy, the item-specific/relational processing dichotomy does not propose differences in depth of processing. Rather, it only suggests differences in style of processing (see Putrevu 2001). For example, item-specific processors are likely to focus on particular attributes of the target brand, whereas relational processors are likely to focus on the relative position of the target brand within the product category. There is no a priori suggestion that one style of processing is superior to the other, however. Which of the two processing styles leads to superior evaluation/judgment depends on the characteristics of the particular message, the context, and the recipient (e.g., the nature of the highlighted attributes, the diagnostic value of individual attributes versus overall product category information, prior knowledge and experience of the consumer). Similarly, the item-specific/relational processing dichotomy does not imply a greater number of cognitive responses for a particular processing style, but it does suggest different types of cognitive responses, as discussed below.

In an ad response context, as item-specific processors, men would value attribute-based messages that bring out the distinctive or unique features of the claim (i.e., the brand's differential advantage) because such attribute-oriented information is consistent with their processing style. In contrast, women, who tend to be relational processors, would find category-oriented information to be more consistent with their processing style. Hence, women would value category-based messages that focus on the common themes of the claim, in addition to its unique features (i.e., the brand's position within the subcategory). When viewing/processing advertisements, men, as item-specific processors, are likely to generate more attribute thoughts, whereas women, as relational processors, are likely to generate more category thoughts. Hence:

H8: Compared with women, men will exhibit more positive affect ([A.sub.ad] and [A.sub.b]) and stronger PI toward attribute-oriented ads.

H9: Compared with men, women will show more positive affect ([A.sub.ad] and [A.sub.b]) and stronger PI toward category-oriented ads.

H10: As item-specific processors, men will generate more attribute thoughts than women.

H11: As relational processors, women will generate more category thoughts than men.

STUDY 1

Stimuli

Eight ads (four sets) drawn from different product categories were used to test the various hypotheses in an experimental setting. A common trade-off in any such experimental study is between realism/generalizability and experimental control. To keep the study realistic, actual full-color print ads were selected from national magazines, except in one instance--the purely verbal manipulation (Braun razors) was developed from existing razor ads. Every attempt was made to closely match the product categories for the members of each set. The first set was Braun (H1: purely verbal) and Gillette (H2: verbal plus visual reinforcement) razors; the second set was Chevy Silverado (H3: simple ad execution) and Dodge Dakota (H4: complex, informative ad execution) trucks; the third set was Progresso Chicken Noodle Soup (H5: comparative) and Brita water filters (H6: harmonious); and the fourth set was Nissan Maxima (H8: attribute-oriented copy) and Toyota Camry (H9: category-oriented copy) sedans. The other hypotheses (H7, H10, and H11) were tested for all the target ads.

The Braun ad provides verbal descriptions of the razor's attributes and benefits without any accompanying pictures, whereas the Gillette ad contains similar verbal descriptions along with pictorial depictions of the various attributes/benefits. The Silverado ad shows a truck in a rugged, outdoor setting and the copy reinforces this feature, whereas the Dakota ad highlights (both verbally and pictorially) the truck's benefits in terms of size, warranty, all-wheel drive, safety, engine size, towing capacity, and interior features. The Progresso ad positions the brand in direct competition with another brand, whereas the Brita ad portrays the brand as being in harmony with nature. The Maxima ad highlights six major attributes of the car (i.e., V6 engine, 222 horsepower, 17-inch wheels, leather trim, seven-speaker Bose audio/CD system, and fuel economy), whereas the Camry ad positions the car as the "perfect midsize family car."

The selected print advertisements were pretested on an independent sample of 30 students (15 male and 15 female) drawn from the same overall participant pool. Participants responded to single-item measures of brand familiarity and involvement (1 = low and 7 = high), as well as ad type manipulation measures (e.g., simple [1] versus complex [7] for the Silverado and Dakota ad set; competitive [1] versus harmonious [7] for the Progresso and Brita ad set). The results showed that there were no sex differences in brand familiarity or involvement for any of the test brands (all absolute t values < 1.43, p > .10) and the ad type manipulation was successful for each of the four sets (all absolute t values > 3.12, p < .05).

Sample and Procedure

The sample consisted of 144 undergraduate students (72 male and 72 female) attending a midsized university. Multiple sessions with eight respondents per session ensured dispersed seating and reduced the risk of subject interaction during the experiment. The respondents were told that they would see a few print ads and were asked to view the material at their normal pace. To keep the exposure scenario somewhat realistic and to ensure smooth progress of the experiment, participants were allowed a maximum of 60 seconds for viewing each ad. Once the viewing task was completed, the stimulus material was retrieved and the questionnaire was administered. The participants were allowed approximately 75 seconds to respond, and upon completion, they were exposed to the next ad and so on. Since there were eight test ads and eight participants per session, this procedure ensured that each subject viewed the test ads in a different order, thereby reducing any potential order effects. Also, to avoid any potential confounds, the two members from each set of test ads did not immediately precede or follow each other (i.e., the Nissan Maxima ad did not immediately precede or follow the Toyota Camry ad).

The respondents were first asked to list any thoughts that occurred to them as they read/viewed the test ad. Following this task, attitude toward the ad ([A.sub.ad]) was measured using four sets of seven-point bipolar adjectives: not believable/ believable, not true/true, not sincere/sincere, and dishonest/ honest (a ranged from .88 to .93 for the eight ads); attitude toward the brand ([A.sub.b]) was measured using four sets of seven-point bipolar adjectives: dislike/like, bad/good, unfavorable/ favorable, and useless/useful (a ranged from .85 to .95 for the eight ads); and purchase intent (PI) was measured using three sets of seven-point bipolar adjectives: unlikely/likely, improbable/probable, and impossible/possible (a ranged from .90 to .96 for the eight ads). Since the a's for the measures of [A.sub.ad], [A.sub.b], and PI were acceptable, the respective means were used as indicators of the constructs in all subsequent analyses. Brand familiarity and involvement were each measured using a single seven-point scale (1 = low; 7 = high). Demographic information was collected at the end of the session.

Results

Two experienced judges (one male and one female) who were unfamiliar with the purpose of the study were asked to classify the thoughts listed by the respondents into attribute-oriented thoughts, category-oriented thoughts, and other thoughts. The agreement rate between the two judges was .88, and all disagreements were resolved through discussion. For each respondent, total thoughts were calculated as the sum of attribute-oriented, category-oriented, and other thoughts. Examples of each thought type are as follows: attribute thoughts: "Brita removes 90% of lead in water," "Maxima has 222 horsepower," "Progresso has large pieces of chicken"; category thoughts: "There are no significant differences between the various brands of razors" (reference to Gillette), "Brita is better than other water filters like PUR," "Toyota Camry is a family car," "Dodge trucks are very similar to other pickups"; other thoughts: "I like the waterfall" (reference to Brita), "I have to go shopping" (reference to Progresso), "The setting seems artificial" (reference to Camry). As in the pretest, there were no sex differences in brand familiarity or involvement for any of the test brands (all absolute t values < 1.25, p > . 10). However, both of these variables were used as covariates in the multivariate analyses of covariance (MANCOVAs) to rule out alternative explanations for the results. The covariates were not significant except in two instances (involvement-Brita: F = 2.07, p < .10; familiarity-Camry: F = 2.20, p < .05). The relevant univariate contrasts from the MANCOVAs are summarized in Tables 1 ([A.sub.ad], [A.sub.b], and PI) and 2 (cognitive responses).

Hypothesis 1 posits a female affinity for verbal ads. Consistent with this expectation, for the purely verbal ad featuring Braun razors, compared with men, women exhibited superior [A.sub.ad] (female M = 5.25, versus male M = 4.21), [A.sub.b] (female M = 4.91, versus male M = 3.85), and PI (female M = 4.54, versus male M = 3.42) (see Table 1). Hypothesis 2 predicts that the two sexes should have equal affinity for ads that contain verbal information and visual reinforcement of such information. This hypothesis is partially supported since men and women do not differ with respect to [A.sub.ad] (male M = 5.63, versus female M = 5.51) or [A.sub.b] (male M = 6.04, versus female M = 5.95) for Gillette razors (F values < .66, p > .10). As predicted by H3, the mate preference for the simple ad (i.e., Chevy Silverado) was confirmed for all three dependent variables: [A.sub.ad] (male M = 5.35, versus female M = 3.76), [A.sub.b] (male M = 5.59, versus female M = 3.66), and PI (male M = 4.84, versus female M = 2.94). Similarly, in line with H4, the female preference for the complex, informative ad (i.e., Dodge Dakota) was also evident across all three dependent variables: [A.sub.ad] (female M = 5.58, versus male M = 4.43), [A.sub.b] (female M = 5.48, versus male M = 4.24), and PI (female M = 4.75, versus male M = 3.47).

Hypothesis 5 posits a male preference for ads that feature comparative appeals. Consistent with this prediction, for the comparison-oriented Progresso ad, compared with women, men exhibited superior [A.sub.ad] (male M = 5.04, versus female M = 3.93), [A.sub.b] (male M = 4.89, versus female M = 3.83), and PI (male M = 4.70, versus female M = 3.62). According to H6, compared with their male counterparts, women were expected to show a stronger affinity for ads featuring harmonious relationships. This expectation is confirmed across all three dependent variables for the harmonious Brita ad: [A.sub.ad] (female M = 5.55, versus male M = 4.38), [A.sub.b] (female M = 6.04, versus male M = 4.71), and PI (female M = 6.11, versus male M = 4.66). This pattern of results is in line with the socialization literature, and suggests that the two sexes prefer ad types that are consistent with their perceived gender roles.

According to the selectivity hypothesis (H7), comprehensive-processing women should generate more cognitive responses (i.e., total thoughts) than heuristic-processing men. Although women generated slightly more total thoughts than men, the observed differences do not reach statistical significance for any of the eight test ads (see Table 2). Women generated marginally more total thoughts for the Toyota Camry ad (female M = 3.58, versus male M = 3.25; F = 3.16, p < .10), whereas in all other cases, the difference in total thoughts was not significant (F values < .70, p > .10). Hence, even though the conditions were ripe for testing the selectivity hypothesis (low {3.26} to moderate {4.35} level of involvement), the results do not support this theory.

Consistent with the prediction in H8, for the attribute-based ad featuring the Nissan Maxima, compared with women, men exhibited higher levels of [A.sub.ad] (male M = 5.72, versus female M = 4.37), [A.sub.b] (male M = 5.61, versus female M = 4.33), and PI (male M = 4.76, versus female M = 3.56) (see Table 1). Furthermore, as predicted by H9, the reverse pattern emerged in relation to the category-oriented ad (i.e., Toyota Camry): Women had more positive [A.sub.ad] (female M = 5.39, versus male M = 4.26), [A.sub.b] (female M = 5.63, versus male M = 4.44), and PI (female M = 5.20, versus male M = 3.66). These results confirm that men prefer attribute-oriented ads, whereas women prefer category-oriented ads. As predicted by H10 and H11, men were expected to generate more attribute thoughts and women were expected to generate more category thoughts. The results (see Table 2) clearly support this prediction: Men generated more attribute thoughts and women generated more category thoughts for all test ads, irrespective of the product category or specific ad execution style (F values > 9.69, p < .01). These results are consistent with the classification of men as item-specific processors and women as relational processors.

Although several of the expected sex differences did emerge, a major prediction of the selectivity hypothesis was not supported by the data: Women did not generate more cognitive responses (total thoughts) than men for any of the eight test ads. However, researchers advocating the selectivity hypothesis have suggested that the sex differences in processing are only likely to occur when message or task factors do not strongly motivate men or women to engage in comprehensive processing (Meyers-Levy and Maheswaran 1991). Although the involvement levels were low/moderate in Study 1, the experimental procedure used (i.e., participants provided their responses almost immediately after having viewed a particular ad) could have engendered deeper processing among all the participants, thus eliminating the processing difference predicted by the selectivity hypothesis. To mirror actual exposure conditions, the experimental procedure controlled the amount of time allocated for viewing and responding to each ad. However, this procedure might not have fully compensated for the contrived experimental setting and could have somehow affected the cognitive response task. Also, since well-known national brands were used as stimuli, the prior preferences of the participants could have confounded the results.

STUDY 2

To address these issues, a follow-up study was conducted using an independent sample of 96 students (48 male and 48 female) drawn from the same overall subject pool. Participants were exposed to a 10-page excerpt that was positioned as a newly created university magazine. In addition to the cover page and instruction page, the excerpt contained ads for two brands that were completely unfamiliar to the participants. The ads for RBC Credit Card and Bay Department Store were selected from a Canadian magazine and appeared on the fifth and eighth pages of the excerpt. The other six pages were devoted to filler material that included two filler ads, two editorials, and two university announcements. The credit card ad was more attribute-oriented and focused on four major features: low interest rate, no annual fee, membership rewards, and security features. The department store ad was more category-oriented, arguing that the store offered a more relaxed and fun-filled retail shopping experience. Participants were given a maximum of eight minutes to view the magazine. After all participants returned the magazine, they responded to the dependent measures. Task involvement was measured using a single seven-point scale (1 = low; 7 = high). Attitude toward the magazine ([A.sub.mag]) was measured using four sets of seven-point bipolar adjectives (dislike/like, bad/good, unfavorable/favorable, and useless/useful {a = .89}); unaided recall was measured by asking participants to list the names of the topics listed in the magazine; cognitive responses were elicited for each of the two target ads using the brand name as the probe cue; and recognition (discrimination ability) was assessed by asking respondents to identify the correct information about the target brand from a list of eight statements (four true and four false).

The respondents exhibited low/moderate levels of task involvement with a sample mean of 3.05. There were no sex differences in task involvement (female, 3.19, and male, 2.92; t = 1.32, p > .10) or [A.sub.mag] (female, 5.02, and male, 4.79; t = 1.17, p > .10). However, both task involvement and [A.sub.mag] were used as covariates to rule out alternative explanations for the results. In the MANCOVA, task involvement was not significant (F = 1.30, p > .10), but [A.sub.mag] was marginally significant (F = 1.76, p < .10). The relevant univariate contrasts from the covariance analysis are shown in Table 3. The unaided recall did not differ between women and men (female M = 5.56, versus male M = 5.29). The ability of the respondents to correctly identify the details about the advertised brand was assessed by calculating signal detection scores as proposed by Grier (1971). These scores measure discrimination ability uncontaminated by the response tendency of participants, and their use has been recommended in assessing ad recognition (cf. Singh and Churchill 1986). The scores were calculated as follows: A' = .5 + {(y - x) (1 + y - x)}/{4y(1 - x)}, where x is the probability of a false alarm (wrong acceptance of a foil statement), and y is the probability of a hit (correct acceptance of a true statement). The signal detection scores did not differ across the sexes for either target ad (F values < 1.13, p > .10).

Two judges (one male and one female), who were unaware of the study objectives, coded the cognitive responses into attribute thoughts, category thoughts, and other thoughts. The agreement rate was .86, and disagreements were resolved through discussion. As in Study 1, these three types of thoughts were summed to arrive at total thoughts. There was no difference in total thoughts between women and men (RBC Credit Card: female M = 2.42, versus male M = 2.25; Bay Department Store: female M = 2.63, versus male M = 2.40). Once again, the results do not support the predictions of the selectivity hypothesis: Comprehensive-processing women did not have significantly superior recall, better discrimination ability, or higher total thoughts in comparison with heuristic-processing men. As observed in Study 1, however, men did generate significantly more attribute thoughts than women (RBC Credit Card: male M = 1.13, versus female M = .63; Bay Department Store: male M = 1.17, versus female M = .65), and women did generate significantly more category thoughts than men (RBC Credit Card: female M = 1.15, versus male M = .60; Bay Department Store: female M = 1.25, versus male M = .60). This pattern of results suggests that men and women do use different information-processing styles, but that they do not necessarily differ in terms of the depth/comprehensiveness of processing as the selectivity hypothesis would suggest.

DISCUSSION

There is strong and unequivocal evidence from the two studies that men and women exhibit sharply varying reactions to identical print advertisements. Specifically, women show superior affect ([A.sub.ad] and [A.sub.b]) and purchase intent toward advertisements that are verbal, harmonious, complex, and category-oriented, whereas men exhibit superior affect ([A.sub.ad] and [A.sub.b]) and purchase intent toward advertisements that are comparative, simple, and attribute-oriented. The cognitive-response (Studies 1 and 2), recall, and recognition (Study 2) results do not support the prediction of the selectivity hypothesis that female processing is more comprehensive compared with that of males. However, the types of cognitive responses listed by the two sexes are consistent with the classification of men as item-specific processors and women as relational processors.

It is interesting to note that the data also show significant differences (t values > 2.41, p < .05) in all but two instances when analyzed within the sex of the respondent across each set of test brands (e.g., comparison of men's scores for [A.sub.ad], [A.sub.b], and PI for the Braun versus the Gillette ads, as shown in Table 1). Such an analysis clearly indicates the following: Both sexes prefer the ad containing visual reinforcement as opposed to the purely verbal ad (i.e., for both men and women, Gillette > Braun for [A.sub.ad], [A.sub.b], and PI); men prefer the simple ad execution over the complex one (i.e., for men, Chevy Silverado > Dodge Dakota for [A.sub.ad], [A.sub.b], and PI), whereas women prefer the complex ad execution over the simple one (i.e., for women, Dodge Dakota > CheW Silverado for [A.sub.ad], [A.sub.b], and PI); men favor the comparative ad over the harmonious one (i.e., for men, Progresso > Brita for [A.sub.ad]), whereas women exhibit the reverse pattern (i.e., for women, Brita > Progresso for [A.sub.ad], [A.sub.b], and PI); and men prefer the attribute-oriented message over the category-oriented message (i.e., for men, Nissan Maxima > Toyota Camry for [A.sub.ad], [A.sub.b], and PI), whereas women favor the category-oriented message over the attribute-oriented message (i.e., for women, Toyota Camry > Nissan Maxima for [A.sub.ad], [A.sub.b], and PI). Hence, not only do the sexes differ in their reaction to the same ad; they also show a strong preference for particular types of ad execution styles. Although the above effects were not specifically hypothesized, such within-sex/across-ad findings are fully consistent with the hypotheses. This pattern of results provides further evidence for the varied reactions of men and women to print ads and clearly identifies the different ad execution styles that might be appropriate for male versus female audiences.

From an advertising strategy perspective, this research suggests that men and women are likely to respond more favorably to messages that are in tune with their respective gender-role expectations and information-processing styles. Presumably due to their differing social roles in society, men show a preference for advertising messages that feature competition and engage in brand comparisons, whereas women favor messages that emphasize harmony and show importance to self as well as others. Also, perhaps due to differing levels of brain lateralization, women have a greater affinity for purely verbal information, whereas men benefit from visual reinforcement of verbal material. Furthermore, men prefer simple ads that focus on one or a few key attributes, whereas women prefer complex ads that contain rich verbal and visual information. Thus, advertisers would do well to create gender-specific ad campaigns that feature differing levels of hard versus soft sell, as well as differing levels (and types) of verbal and visual information. The observation that ads in male magazines are more wordy and complex than corresponding ads in female magazines (see Whissell and McCall 1997) suggests that advertising practitioners could improve on their ad execution styles targeting the two sexes.

The relational processing style of women and the item-specific processing style of men suggest that advertisers should present category-oriented messages to a female audience and attribute-oriented messages to a male audience. Specifically, ads targeting men should emphasize only those features that are unique to the advertised brand and highlight its differential advantage. In contrast, ads targeting women should focus on features that are common to the product category and highlight how the advertised brand fits in (or compares) with other brands belonging to the category. Once again, a gender-specific ad campaign seems appropriate.

This research does have some limitations. First, the samples were comprised of undergraduate students, which somewhat limits the generalizability of the results. However, the objective of this research was to test theory applications, and in such situations, homogenous samples are acceptable. Specifically, in theory application testing, homogenous samples help enhance the validity of statistical conclusions since the results can be clearly attributed to treatments and not to the heterogeneity of the sample (Calder, Phillips, and Tybout 1981). Second, some caution is warranted in interpreting the within-sex/across-ad results that emerged from Study 1. Since actual magazine ads were used, there were a number of differences among the ads, thus making it difficult to completely rule out competing explanations. A related issue in Study 1 is that the target brands in one set (Brita and Progresso) did not belong to the same product category. Although this is less than ideal, it is not a major concern since all of the hypothesized male/female comparisons are within (not across) brands. Finally, the contrived experimental setting might have influenced the responses. This is more of a concern in Study 1, where the response task was repetitive. Nevertheless, the debriefing session following the data collection in Study 1 indicated that although the participants might have anticipated the response task, none of them guessed the hypotheses. The replication of results in Study 2 further alleviates this concern.

It should be noted that the dependent variables measured in this research are outcome measures, not process measures. Hence, despite the lack of support for the selectivity hypothesis, one cannot discount this theory purely based on the findings of this research. Nevertheless, other scholars have cast serious doubts on the portrayal of women as comprehensive processors (Hupfer 2002), and recent tests of sex differences in the general motivation to process information have produced null results (Peracchio and Tybout 1996) that are similar to those reported here. Furthermore, contrary to the prediction of the selectivity hypothesis that there will be a primacy effect for men and a recency effect for women (Meyers-Levy 1989, p. 241), researchers have found a primacy effect for both sexes (Dube and Morgan 1996). In view of this body of evidence, perhaps a reexamination of the selectivity hypothesis is warranted in future research. It is clear from this research, however, that the item-specific/relational processing dichotomy provides a reasonable alternative explanation for the differences observed between the sexes. Future research could confirm or question these findings by replicating this research using more representative samples, additional brand exemplars, and other sorts of media. Another avenue for future research would be to design appropriate ads that could efficiently test the within-sex/across-ad differences found in this research. It might also be fruitful to explore whether contextual or individual difference factors moderate the processing style differences observed in this study.

In modern times, there have been tremendous socio-cultural changes, such as more enlightened attitudes about equality between the sexes, higher female participation in education and the workforce, more women in leadership roles, more stay-at-home dads, and so forth--changes that have blurred the traditional gender divide. There is little doubt that gender identities are constantly evolving and recent societal changes are responsible for more agentic content in female gender roles and more communal content in male gender roles. Notwithstanding such sociocultural trends, however, there are numerous circumstances where advertisers who are targeting one sex in particular would be wise to tailor their advertising strategies according to the different processing styles found in male versus female audiences.

The author thanks Venkata Naga Shylaja Putrevu for help in data collection. The author also thanks James Gentry, Kenneth Lord, the Editor, and three anonymous reviewers for helpful comments on earlier drafts of this manuscript.

TABLE 1
MANCOVA UnivariateTests (Study 1):[A.sub.ad], [A.sub.b], and PI

                                              Male         Female

                                          Mean     SD     Mean   SD

Braun
Attitude toward the ad ([A.sub.ad])       4.21      .71   5.25   .99
Attitude toward the brand ([A.sub.b])     3.85      .90   4.91  1.20
Purchase intent (PI)                      3.42     1.13   4.54  1.30

Gillette
Attitude toward the ([A.sub.ad])          5.63      .96   5.51   .76
Attitude toward the brand ([A.sub.b])     6.04      .69   5.95   .80
Purchase intent (PI)                      6.11      .92   5.09  1.71

Silverado
Attitude toward the ([A.sub.ad])          5.35      .87   3.76  1.07
Attitude toward the brand ([A.sub.b])     5.59     1.11   3.66  1.08
Purchase intent (PI)                      4.84     1.57   2.94  1.27

Dodge Dakota
Attitude toward the ([A.sub.ad])          4.43      .76   5.58   .90
Attitude toward the brand ([A.sub.b])     4.24      .84   5.48   .95
Purchase intent (PI)                      3.47     1.23   4.75  1.15

Progresso
Attitude toward the ([A.sub.ad])          5.04      .95   3.93  1.10
Attitude toward the brand ([A.sub.b])     4.89     1.04   3.83  1.24
Purchase intent (PO                       4.70     1.25   3.62  1.37

Brita
Attitude toward the ([A.sub.ad])          4.38      .91   5.55   .90
Attitude toward the brand ([A.sub.b])     4.71      .96   6.04   .93
Purchase intent (PI)                      4.66     1.15   6.11  1.10

Nissan Maxima
Attitude toward the ([A.sub.ad])          5.72      .86   4.37   .98
Attitude toward the brand ([A.sub.b])     5.61      .97   4.33  1.01
Purchase intent (PI)                      4.76     1.09   3.56  1.33

Toyota Camry
Attitude toward the ([A.sub.ad])          4.26      .92   5.39   .90
Attitude toward the brand ([A.sub.b])     4.44      .82   5.63  1.00
Purchase intent (PI)                      3.66     1.18   5.20   .88

                                        F value   p value

Braun
Attitude toward the ad ([A.sub.ad])      51.11     <.01
Attitude toward the brand ([A.sub.b])    34.57     <.01
Purchase intent (PI)                     30.37     <.01

Gillette
Attitude toward the ([A.sub.ad])           .66     <.42
Attitude toward the brand ([A.sub.b])      .58     <.45
Purchase intent (PI)                     20.26     <.01

Silverado
Attitude toward the ([A.sub.ad])         94.32     <.01
Attitude toward the brand ([A.sub.b])   110.16     <.01
Purchase intent (PI)                     62.34     <.01

Dodge Dakota
Attitude toward the ([A.sub.ad])         68.16     <.01
Attitude toward the brand ([A.sub.b])    67.57     <.01
Purchase intent (PI)                     39.58     <.01

Progresso
Attitude toward the ([A.sub.ad])         39.73     <.01
Attitude toward the brand ([A.sub.b])    29.36     <.01
Purchase intent (PO                      22.96     <.01

Brita
Attitude toward the ([A.sub.ad])         60.08     <.01
Attitude toward the brand ([A.sub.b])    68.77     <.01
Purchase intent (PI)                     57.00     <.01

Nissan Maxima
Attitude toward the ([A.sub.ad])         74.41     <.01
Attitude toward the brand ([A.sub.b])    58.48     <.01
Purchase intent (PI)                     32.89     <.01

Toyota Camry
Attitude toward the ([A.sub.ad])         52.49     <.01
Attitude toward the brand ([A.sub.b])    59.41     <.01
Purchase intent (PI)                     75.38     <.01

Note: MANCOVA = multivariate analysis of covariance.

TABLE 2
MANCOVA Univariate Tests (Study 1): Cognitive Responses

                       Male      Female

                    Mean   SD   Mean   SD    F value  p value

Braun
Attribute thoughts   .60   .69   .26   .53     9.69     <.01
Category thoughts    .26   .48   .67   .75    15.63     <.01
Other thoughts      2.29  1.18  2.25  1.34      .10     <.75
Total thoughts      3.17   .98  3.19  1.25      .02     <.89

Gillette
Attribute thoughts  1.46  1.09   .61   .55    35.24     <.01
Category thoughts    .29   .52  1.21   .71    76.82     <.01
Other thoughts      1.61  1.22  1.65  1.09      .05     <.81
Total thoughts      3.36  1.17  3.47  1.06      .31     <.58

Silverado
Attribute thoughts  1.60  1.35   .71   .64    25.44     <.01
Category thoughts    .51   .56  1.33   .77    50.10     <.01
Other thoughts      1.29  1.13  1.50  1.13     1.25     <.27
Total thoughts      3.40  1.29  3.54  1.14      .30     <.58

Dodge Dakota
Attribute thoughts  1.17  1.15   .56   .58    16.17     <.01
Category thoughts    .53   .61  1.19   .83    28.83     <.01
Other thoughts      1.78  1.37  1.74  1.18      .05     <.82
Total thoughts      3.47  1.20  3.49  1.02      .02     <.97

Progresso
Attribute thoughts   .92   .88   .43   .60    14.53     <.01
Category thoughts    .51   .61  1.15   .76    29.48     <.01
Other thoughts      1.60  1.13  1.60  1.12     0        <.99
Total thoughts      3.00  1.15  3.14  1.21      .46     <.50

Brita
Attribute thoughts  1.15   .85   .31   .46    51.97     <.01
Category thoughts    .51   .61  1.19   .73    39.28     <.01
Other thoughts      1.51  1.11  1.81  1.04     1.91     <.17
Total thoughts      3.18  1.11  3.32  1.05      .60     <.44

Nissan Maxima
Attribute thoughts  1.72  1.36   .79   .71    27.84     <.01
Category thoughts    .36   .51  1.25   .75    66.25     <.01
Other thoughts      1.29  1.05  1.39  1.26      .40     <.53
Total thoughts      3.36  1.29  3.43  1.33      .70     <.79

Toyota Camry
Attribute thoughts  1.06  1.02   .42   .58    22.10     <.01
Category thoughts    .65   .61  1.63   .83    63.26     <.01
Other thoughts      1.54  1.26  1.53  1.16     0        <.99
Total thoughts      3.25  1.15  3.58  1.10     3.16     <.08

Note: MANCOVA = multivariate analysis of covariance.

TABLE 3
MANCOVA Univariate Tests (Study 2): Unaided Recall, Recognition, and
Cognitive Responses

                          Male      Female

                       Mean   SD   Mean   SD    F value  p value

Unaided recall         5.29  1.22  5.56  1.50      .25    <.62
Recognition (signal
 detection scores)
RBC Credit Card         .76   .12   .74   .11      .64    <.43
Bay Department Store    .76   .10   .78   .11     1.13    <.29

Cognitive responses
 (thought listing)
RBC Credit Card
  Attribute thoughts   1.13   .57   .63   .67    16.99    <.01
  Category thoughts     .60   .57  1.15   .74    14.51    <.01
  Other thoughts        .52   .62   .65   .64      .95    <.33
  Total thoughts       2.25   .89  2.42  1.16      .31    <.58
Bay Department Store
  Attribute thoughts   1.17   .69   .65   .70    14.60    <.01
  Category thoughts     .60   .54  1.25   .70    27.25    <.01
  Other thoughts        .63   .64   .73   .84      .22    <.64
  Total thoughts       2.40  1.16  2.63  1.10      .58    <.45

Note: MANCOVA = multivariate analysis of covariance.

NOTES

(1.) The term sex is used in a purely biological context and the term gender is used in broader sociocultural contexts, per the American Psychological Association guidelines.

(2.) Hypotheses 3 and 4 were developed in relation to brain lateralization (organization) theory. However, results supportive of these hypotheses are also consistent with two distinct information-processing dichotomies: the selectivity hypothesis (Meyers-Levy 1989) and the item-specific/relational processing dichotomy (Einstein and Hunt 1980; Hunt and Einstein 1981). Since these two theories have somewhat different advertising implications, it is important to identify which one is more appropriate. This and other related issues are addressed in H7 through H11.

(3.) Hupfer (2002) argues that it may not be appropriate to ascribe agentic qualities to men and communal qualities to women. She provides a variety of recent anecdotal and advertising examples suggesting that the traditional boundaries between the sexes are becoming increasingly blurred. Still, although gender roles are indeed changing, it is important to recognize that even today, men and women are socialized differently and have different worldviews (Brunel and Nelson 2000).

REFERENCES

Al-Olayan, Fahad S., and Kiran Karande (2000), "A Content Analysis of Magazine Advertisements from the United States and the Arab World," Journal of Advertising, 29 (Fall), 69-82.

Berenbaum, Sheri A. (1999), "Effects of Early Androgens on Sex-Typed Activities and Interests in Adolescents with Congenital Adrenal Hyperplasia," Hormones and Behavior, 35 (February), 102-110.

Brunel, Frederic F., and Michelle R. Nelson (2000), "Explaining Gendered Responses to 'Help-Self' and 'Help-Others' Charity Ad Appeals: The Mediating Role of World-Views," Journal of Advertising, 29 (Fall), 15-28.

Calder, Bobby J., Lynn W. Phillips, and Alice M. Tybout (1981), "Designing Research for Application," Journal of Consumer Research, 8 (December), 197-207.

Chaiken, Shelly (1980), "Heuristic Versus Systematic Information Processing and the Use of Source Versus Message Cues in Persuasion," Journal of Personality and Social Psychology, 39 (November), 752-756.

Chamblee, Robert, Robert Gilmore, Gloria Thomas, and Gary Soldow (1993), "When Copy Complexity Can Help Ad Readership," Journal of Advertising Research, 33 (May/June), 23-28.

Costa, Paul T., Antonio Terracciano, and Robert R. McCrae (2001), "Gender Differences in Personality Traits Across Cultures: Robust and Surprising Findings," Journal of Personality and Social Psychology, 81 (August), 322-331.

Darley, William K., and Robert E. Smith (1995), "Gender Differences in Information Processing Strategies: An Empirical Test of the Selectivity Model in Advertising Response," Journal of Advertising, 24 (Spring), 41-56.

Dube, Laurette, and Michael S. Morgan (1996), "Trend Effects and Gender Differences in Retrospective Judgments of Consumption Emotions," Journal of Consumer Research, 23 (September), 156-162.

Eagly, Alice H. (1987), Sex Differences in Social Behavior: A Social-Role Interpretation, Hillsdale, NJ: Erlbaum.

--, and Wendy Wood (1991), "Explaining Sex Differences in Social Behavior: A Meta-analytic Perspective," Personality and Social Psychology Bulletin, 17 (June), 306-315.

Einstein, Gilles O., and R. Reed Hunt (1980), "Levels of Processing and Organization: Additive Effects of Individual-Item and Relational Processing," Journal of Experimental Psychology: Human Learning and Memory, 6 (September), 588-598.

Everhart, D. Erik, Janet L. Shucard, Teresa Quatrin, and David W. Shucard (2001), "Sex-Related Differences in Event-Related Potentials, Face Recognition, and Facial Affect Processing in Prepubertal Children," Neuropsychology, 15 (July), 329-341.

Fischer, Eileen, and Stephen J. Arnold (1994), "Sex, Gender Identity, Gender Role Attitudes, and Consumer Behavior," Psychology and Marketing, 11 (March/April), 163-182.

Ford, John B., and Michael S. LaTour (1993), "Differing Reactions to Female Role Portrayals in Advertising," Journal of Advertising Research, 33 (September/October), 43-52.

Geary, David C. (1996), "Sexual Selection and Sex Differences in Mathematical Abilities," Behavioral and Brain Sciences, 19 (June), 229-247.

Gilligan, Carol (1982), In a Different Voice: Psychological Theory and Women's Development, Cambridge: Harvard University Press.

Goos, Lisa M., and Irwin Silverman (2002), "Sex Related Factors in the Perception of Threatening Facial Expressions," Journal of Nonverbal Behavior, 26 (Spring), 27-41.

Gorman, Christine, Madeleine J. Nash, and Barbara Ehrenreich (1992), "Sizing Up the Sexes," Time, 139 (January 20), 42-49.

Grier, J. Brown (1971), "Nonparametric Indexes for Sensitivity and Bias: Computing Formulas," Psychological Bulletin, 75 (June), 424-429.

Halpern, Diane F. (1997), "Sex Differences in Intelligence," American Psychologist, 52 (October), 1091-1102.

Hansen, Flemming (1981), "Hemispheral Lateralization: Implications for Understanding Consumer Behavior," Journal of Consumer Research, 8 (June), 23-36.

Helgeson, Vicki S., and Heidi L. Fritz (1998), "A Theory of Unmitigated Communion," Personality and Social Psychology Review, 2 (July), 173-183.

Hunt, R. Reed, and Gilles O. Einstein (1981), "Relational and Item-Specific Information in Memory," Journal of Verbal Learning and Verbal Behavior, 20 (October), 497-514.

--, and Catherine E. Seta (1984), "Category Size Effects in Recall: The Role of Relational and Item-Specific Information," Journal of Experimental Psychology: Learning, Memory and Cognition, 10 (July), 454-464.

Hupfer, Maureen (2002), "Communicating with the Agentic Woman and the Communal Man: Are Stereotypic Advertising Appeals Still Relevant?" Academy of Marketing Science Review (March), available at www.amsreview.org/articles/ hupfer03-2002.pdf (accessed July 15, 2003).

Hyde, Janet S., and Marcia C. Linn (1988), "Gender Differences in Verbal Ability: A Meta-analysis," Psychological Bulletin, 104 (July), 53-69.

Ionescu, Marcos D. (2000), "Sex Differences in Memory Estimates for Pictures and Words," Psychological Reports, 87 (August), 315-322.

--(2002), "Sex Differences in Memory Estimates, Revisited," Psychological Reports, 91 (August), 167-172.

Jones, Marilyn Y., Andrea J. S. Stanaland, and Betsy D. Gelb (1998), "Beefcake and Cheesecake: Insights for Advertisers," Journal of Advertising, 27 (Summer), 33-51.

Jun, Sunkyu, Sanjay Putrevu, Yongjin Hyun, and James W. Gentry (2003), "The Influence of Editorial Context on Consumer Response to Print Advertisements," Journal of Current Issues and Research in Advertising, 25 (Fall), 1-11.

LaTour, Michael S., and Tony L. Henthorne (1994), "Ethical Judgments of Sexual Appeals in Print Advertising," Journal of Advertising, 23 (September), 81-90.

Lee, Deeborah S. (2000), "Gender Differences in Academic Achievement and Ability: A Longitudinal Study," Ph.D. diss., University of Michigan.

Maccoby, Eleanor E. (1998), The Two Sexes: Growing Up Apart, Coming Together, Cambridge: Harvard University Press, Belknap Press.

--, and Carol N. Jacklin (1974), The Psychology of Sex Differences, Stanford: Stanford University Press.

Malaviya, Prashant, Jolita Kisielius, and Brian Sternthal (1996), "The Effect of Type of Elaboration on Advertisement Processing and Judgment," Journal of Marketing Research, 33 (November), 410-421.

McDaniel, Stephen R., and Lance Kinney (1998), "The Implications of Recency and Gender Effects in Consumer Response to Ambush Marketing," Psychology and Marketing, 15 (July), 385-403.

Meyers-Levy, Joan (1989), "Gender Differences in Information Processing: A Selectivity Interpretation," in Cognitive and Affective Responses to Advertising, Patricia Cafferata and Alice Tybout, eds., Lexington, MA: Lexington Books, 219-260.

--, and Durairaj Maheswaran (1991), "Exploring Differences in Males' and Females' Processing Strategy," Journal of Consumer Research, 18 (June), 63-70.

--, and Brian Sternthal (1991), "Gender Differences in the Use of Message Cues and Judgments," Journal of Marketing Research, 28 (February), 84-96.

Moschis, George P. (1985), "The Role of Family Communication in Consumer Socialization of Children and Adolescents," Journal of Consumer Research, 11 (March), 898-913.

Peracchio, Laura A., and Alice M. Tybout (1996), "The Moderating Role of Prior Knowledge in Schema-Based Product Evaluation," Journal of Consumer Research, 23 (December), 177-192.

Petty, Richard E., Thomas M. Ostrom, and Timothy C. Brock (1981), Cognitive Responses in Persuasion, Hillsdale, NJ: Lawrence Erlbaum.

Putrevu, Sanjay (2001), "Exploring the Origins and Information Processing Differences Between Men and Women: Implications for Advertisers," Academy of Marketing Science Review (October), available at www.amsreview.org/articles/ putrevu10-2001.pdf (accessed March 6, 2003).

--, Joni Tan, and Kenneth R. Lord (2004), "Consumer Responses to Complex Advertisements: The Moderating Role of Need for Cognition, Knowledge, and Gender," Journal of Current Issues and Research in Advertising, 26 (Spring), 9-24.

Saucier, Deborah M., and Lorin J. Elias (2001), "Lateral and Sex Differences in Manual Gesture During Conversation," Laterality, 6 (July), 239-245.

Schultheiss, Oliver C., and Joachim C. Brunstein (2001), "Assessment of Implicit Motives with a Research Version of the TAT: Picture Profiles, Gender Differences, and Relations to Other Personality Measures," Journal of Personality Assessment, 77 (August), 71-86.

Singh, Surendra N., and Gilbert A. Churchill (1986), "Using the Theory of Signal Detection to Improve Ad Recognition Testing," Journal of Marketing Research, 23 (November), 327-336.

Sokoloff, Heather (2001), "Canada in Top Five in Schools Study," National Post (December 4), A1 and A18.

Whissell, Cynthia, and Lorna McCall (1997), "Pleasantness, Activation, and Sex Differences in Advertising," Psychological Reports, 81 (October), 355-367.

Wolin, Lori D. (2003), "Gender Issues in Advertising--An Oversight Synthesis of Research: 1970-2002," Journal of Advertising Research, 43 (March), 111-129.

Sanjay Putrevu (Ph.D., State University of New York at Buffalo) is an associate professor of marketing at Bryant College.