Children's influence in purchase decisions: a social power theory approach

Children's influence in purchase decisions: a social power theory approach

Journal of Business Research 58 (2005) 593 – 601 Children’s influence in purchase decisions: a social power theory approach L.A. Flurrya,*, Alvin C. ...

342KB Sizes 22 Downloads 73 Views

Journal of Business Research 58 (2005) 593 – 601

Children’s influence in purchase decisions: a social power theory approach L.A. Flurrya,*, Alvin C. Burnsb a

Department of Management and Marketing, Louisiana Tech University, P.O. Box 10318, Ruston, LA 71272, USA Department of Marketing, E.J. Ourso College of Business, Louisiana State University, 3127 CEBA, Baton Rouge, LA 70808, USA

b

Received 5 November 2002; accepted 14 August 2003

Abstract To understand children’s influence in family decision-making, this research uses social power theory to develop a conceptual model of children’s influence. The conceptual model hypothesizes that children’s active social power, children’s passive social power, preference intensity, and decision history should aid in the explanation of variations observed in children’s influence. Empirical tests with a random sample of 987 matched pairs of children aged 8 – 11 and their mothers support the model. D 2003 Elsevier Inc. All rights reserved. Keywords: Children; Social power theory; Preference intensity; Purchase decision

1. Introduction

2. Theoretical model of children’s influence

Understanding children’s purchase influence has been identified as an area in great need of research. Calls for research note that while many studies have focused on the influence of adolescents (ages 12– 16) in family decisionmaking, little attention has been given to children in the analytical stage of development (ages 7– 11) (Roedder-John, 1999). There is ample evidence that children exert varying degrees of influence on family decision processes and that this influence varies by product, decision stage, child, parental, and family characteristics. However, theoretical explanations for these variations have not been adequately researched. One approach that may be useful in explaining the differences in children’s influence on purchases is social power theory. Social power theory distinguishes between influence derived from active and passive power, a notion that is conceptually compatible to direct and indirect influence. To explore this approach, this research develops a conceptual model of children’s influence based on social power theory and tests the model with two samples of children and their mothers.

In an attempt to guide this research, a conceptual model was developed based on the study of family power. Conceptually, the model posits that a child’s ability to exert influence is largely determined by his or her active and passive sources of social power, decision history, and preference intensity (see Fig. 1 for interrelationships).

* Tel.: +1-318-257-3597; fax: +1-318-257-4253. E-mail address: [email protected] (L.A. Flurry). 0148-2963/$ – see front matter D 2003 Elsevier Inc. All rights reserved. doi:10.1016/j.jbusres.2003.08.007

2.1. Social power in family research The concept of social power was initially introduced by Lewin (1951) and later developed theoretically by Cartwright (1959), French and Raven (1959), and Wolfe (1959) in a series of studies on group dynamics. This theoretical framework suggests that social power acts as a resource that people may deploy to exert influence on others. These resources form bases of power over others. The power bases considered most important in exercising influence are expert power, legitimate power, referent power, reward power, and coercive power. Every person is thought to possess some combination of these five power bases. To illustrate, the following discussion defines each of the five power bases and identifies how each could be used by a child. Expertise represents the extent to which a person is perceived to be knowledgeable about a particular subject.

594

L.A. Flurry, A.C. Burns / Journal of Business Research 58 (2005) 593–601

of an intentional action; however, sometimes it may be passive, such as when the mere presence of power is influential (French and Raven, 1959). Both active and passive social power contribute to a person’s potential for directing an outcome according to his/her own preference. Thus, social power theory leads us to hypothesize that children exert influence via some combination of active and passive social power.

Fig. 1. Conceptual model of children’s influence in purchase decisions.

Within a family, it may be acknowledged that a child possesses detailed information in certain product categories, such as toys and games, apparel, and certain grocery items (e.g., Simmons Market Research Bureau, 1993). Reward power is the ability to bestow something favorable on another party contingent upon desired behavior. From a child, rewards might include good behavior, completion of chores, or a display of affection. The third power base, referent power, is the degree to which one person wishes to identify with another person. Also referred to as attraction power, referent power is exercised when one person conforms to the anticipated preferences of another person to feel closer to him/her. Conformity in referent power may be based on respect for a referent or a simple desire to emulate, but is not done as a source of reward. Fourth, legitimate power is the degree to which a person is perceived to have the right to exert influence or the right of one person to prescribe the behavior or beliefs of another person although he/she cannot implement sanctions. Children have legitimate power when they are perceived to have the right to make a selection based on their vested interest in the product decision. This would likely be true in products that would be personally consumed by the child, including toys, foods, and clothing. Last, coercive power resides in the other party’s perception that punishment will result from noncompliance. Although one does not view children as possessing power to coerce their parents, psychologically, the threat of any negative or bothersome behavior of the child acts as coercion. Accordingly, sanctions used by children could include such behaviors as the child displaying anger or misbehavior, complaining, or even pestering. 2.2. Active and passive social power Social power theory further suggests that the five power bases may be utilized in two ways: actively and passively. Use of power to influence is commonly active, or the result

2.2.1. Children’s active social power Active social power is perceived and directly controlled by the child. To exert active influence, a child must make an assessment of his/her social power capabilities, choose an influence attempt consistent with his/her sources of social power, and exert action toward achieving his/her desired outcome (French and Raven, 1959). Influence attempts are defined as goal-directed actions intended to affect a decision outcome. In the conceptual model, a child’s power bases give rise to influence attempts; that is, a child’s influence attempts should be largely shaped by his/her social power attributes. Descriptive studies have documented that children use a number of different influence attempts, including, but not limited to, asking, pleading, bargaining, persisting, using force, telling, being demonstrative, sugar-coating, threatening, and using pity (Atkin, 1978; Isler et al., 1987; McNeal, 1992; Williams and Burns, 2000). A child’s influence attempts are intended to achieve control over the decision outcome, and it can be seen that the influence attempts that have been documented in children are manifestations of their various active social power bases. For example, a child using active social power to influence an entertainment choice, such as a movie, might promise to do an extra household chore if the parent allows the child to go (reward power exercised through bargaining). To test this theory, it is hypothesized that children who possess more active social power will employ more influence attempts than will children who possess less active social power. In turn, children who exert more influence attempts are hypothesized to have more purchase influence than will children who exert less influence attempts. 2.2.2. Children’s passive social power A child’s influence may also be passive, where there is no evidence of speech or overt actions on the part of the child. In this sense, passive sources of power need only be possessed to have an effect (Corfman and Lehmann, 1987). For a child, a power source is passive if the parent infers its presence and acts in lieu of any overt action on the part of the child. It is influence attributed to the child by the parent or the parent’s perception of a child’s unstated preferences (Wells, 1965). As children age, they influence family purchase decisions in a more passive way, as parents learn their children’s likes and dislikes and make purchase decisions accordingly (Roedder-John, 1999). Thus, it is hypothesized that children whose parents perceive them to have

L.A. Flurry, A.C. Burns / Journal of Business Research 58 (2005) 593–601

more passive social power will be seen by the parents as having more influence in purchase decisions than will children whose parents perceive them to have less passive social power. 2.3. Decision history

595

come will exert more influence attempts than will children with a less intense preference for a particular outcome. In addition, it is hypothesized that children with more preference intensity for a product or service will perceive themselves to have more influence in purchase decisions than will children with less preference intensity for a product or service.

In our model, influence is a perceptual construct, and perceptions are molded by a number of factors, not the least of which is the historical pattern of decision outcomes. Social exchange theory assumes that a person will select an influence attempt based on the expected benefits and costs associated with its use (Turner, 1982). This is particularly relevant to children, in that they learn over time which types of behaviors are appropriate for the situation. Social exchange theory also assumes that a person will repeat rewarding behaviors until they are no longer successful. Decision history is defined as a person’s perception of his or her past success in a similar decision exchange. Corfman and Lehmann (1987) discuss decision history as a predisposition to act that reflects the pattern of prior decision episodes. Persons may have multiple decision histories depending on the context and participants in the decision. These decision histories are an ongoing account of successful and rewarding encounters (Emerson, 1981; Turner, 1982). As such, a person’s perception of his/her decision history in a similar purchase context should mold to some extent his/her perception of his/her potential for influence (Corfman and Lehmann, 1987). Thus, it is hypothesized that children who perceive that they are historically more successful in directing the outcome of similar purchase decisions will believe that they have more influence than will children who perceive that they are historically less successful in directing the outcome of similar purchase decisions.

Influence is defined as the use of power to achieve an outcome (Coleman, 1973). By definition, influence may only be achieved as the result of a reciprocal exchange process between two or more parties (Sprey, 1975). This use of power to exert influence is referred to as a ‘‘circular causal process’’ (Olson et al., 1975), whereby the outcome power attributed to any one party is a function of the perspectives of all parties involved in the decision-making process. Given the reciprocal nature of influence, it is necessary to measure influence from the perspectives of all significant members of the decision-making process (Olson et al., 1975). Further, it should be expected that the perceptions of the parties will be similar, but not identical (French and Raven, 1959). Similarly, it is expected that there will be a relationship between the child’s perception of his/her influence and the parent’s perception of the child’s influence, although the magnitude of influence perceived by both parties may not be identical. For these reasons, children’s influence is measured from both the child’s and the parent’s perspectives and a reciprocal relationship between the child’s perception of his/ her influence and the parent’s perception of the child’s influence is hypothesized.

2.4. Preference intensity

3.1. Research design

Preference intensity is a concept in the family decisionmaking literature that has the potential to affect children’s influence perceptions. In our conceptual model, preference intensity is a motivational construct that reflects the extent to which a person desires to achieve a particular outcome or purchase. The more important the outcome of a purchase decision is to a person, the more he/she will exert effort to influence the decision (Scanzoni and Szinovacz, 1980). According to Corfman and Lehmann (1987), decision outcomes are directed by the individual preferences of the members of the decision-making group. Persons who strongly desire a particular outcome, or product, are more likely to exert influence in a group decision. In fact, Corfman and Lehmann (1987) suggest that preference intensity may be the most important predictor of relative influence. For this reason, it is hypothesized that children with an intense preference for a particular decision out-

The context of this study was a toy purchase, a product category wherein children are expected to exert influence. The research design selected for this research was a crosssectional survey with random sampling. The unit of analysis was a mother – child pair. Children ages 8 –11 were selected as respondents due to their cognitive capabilities and their level of consumer experience. Research suggests that children in the analytical stage are adaptive decision-makers, able to make independent decisions and self-evaluations, and employ influence tactics to negotiate for desired outcomes (McNeal, 1992; Roedder-John, 1999). Mothers were selected as respondents for several reasons. First, research indicates that mothers are more often the recipients of influence attempts than are fathers (Cowan and Avants, 1988; Cowan et al., 1984). Second, mothers are often the purchasing agents for the family, and mothers are

2.5. Children’s influence

3. Research method

596

L.A. Flurry, A.C. Burns / Journal of Business Research 58 (2005) 593–601

considered to be more familiar with their children’s purchase influence attempts. Finally, as this is an initial examination of the conceptual model, it was desirable to avoid the difficulties inherent in accommodating more than two respondents. 3.2. Data collection and sample

Table 1 Sample measures Construct

Sample scale items

Active expert power

. I am a toy expert. . I know a lot about toys. . I have the right to tell my mother what I want her

Active legitimate power

to buy. . I should be able to tell my mother what I want

her to buy for me.

The survey population was defined as mother – child pairs where the children were enrolled in the fourth and fifth grades in public schools located in the northern and central regions of a southern state. Thirteen schools in these two regions were randomly selected and agreed to participate in the study. All children enrolled in the fourth and fifth grades in the 13 schools were asked to take home a packet to their mothers that included an introductory letter outlining the purpose of the research, a consent form detailing both the mothers’ and the children’s rights as research participants, a questionnaire for the mother, and a questionnaire for the child. The cover letter provided instructions for the completion and return of the questionnaires. Mothers were told that it was extremely important for the two questionnaires to be completed independently and that they were not to influence their child’s responses in any way. Incentives were offered to mothers and children to encourage the return of the questionnaires. A total of 987 usable questionnaire packets were analyzed for this study (45% response rate). To examine the potential biasing effects of nonresponse error, the principal of each school system was sent a questionnaire that requested basic demographic statistics regarding the fourthand the fifth-grade student populations. When these data were compared to those of the respondents, no significant differences were found between the sample and the population statistics. The total sample of 987 was then randomly divided into two data sets, one each used for hypotheses tests and validation. 3.3. Measurement 3.3.1. Active social power Adaptations of Swasy’s (1979) social power scales were used to measure the child’s perception of his or her reward power (four items; a1=.83, a2=.86), expert power (five items; a1 =.82, a2 =.83), coercive power (four items; a1 =.80, a2=.81), legitimate power (five items; a1=.79, a2=.77), and referent power (five items; a1=.69, a2=.71). Children were instructed to think about when they wanted to buy a toy and to answer each item on a four-point Likert scale. Sample items for each scale are listed in Table 1. Active power bases are represented in the structural model as n1 –n5 (see Fig. 2). 3.3.2. Direct influence attempts Adaptations of Williams and Burns’ (2000) scales were used to measure the child’s perception of his or her influence attempts. The seven scales were ask nicely (six items, ‘‘e.g., I

Active referent power

. My mother wants to understand what I like

about the things that we buy for me. . My mother cares what I think about the things

she buys for me. Active reward power

. I can reward mother when she buys what I want

her to buy. . I can make my mother feel good when she buys

what I want her to buy for me. Active coercive power . If mother does not buy what I want, I can usually get her to change her mind by misbehaving. . I can get my mother to buy what I want by doing something that she would not want me to do. Passive expert power . I think that my child is a toy expert. . My child knows a lot about toys. . I should listen when my child tells me what he or Passive legitimate she wants to buy. power . My child has the right to influence me when I make purchase decisions. Passive referent power . My child’s feelings influence what I will buy. . I usually want to buy things that my child likes. Passive reward power . My child has the ability to reward me in some manner when I buy what he/she wants me to buy. . I like to buy what my child wants because he or she may give me something nice for doing it. . I may give in and buy what my child wants if he Passive coercive power or she threatens to misbehave. . Sometimes I may buy what my child wants to keep him or her from getting upset.

ask in a nice way’’; a1=.88, a2=.86, respectively, for the two samples), bargain (four items, e.g., ‘‘I say that I will help clean the house’’; a1=.69, a2=.66), show affection (five items, e.g., ‘‘I say that she is the best mom in the whole world’’; a1=.84, a2=.85), just ask (three items, e.g., ‘‘I just ask’’; a1=.63, a2=.63), display anger (four items, ‘‘e.g., ‘‘I get mad’’; a1=.78, a2=.79), beg (four items, e.g., ‘‘I ask over and over’’; a1=.88, a2=.88), and con (three items; e.g., ‘‘I say that I don’t already have one, when I do’’; a1=.65, a2=.62). Children were instructed to think about what they did when they really wanted their mother to buy a toy for them and to respond to a five-point frequency scale. To examine the possibility of a higher order factor structure, a single-item indicator was computed for each of the seven scales and then subjected to factor analysis. Factor analyses revealed that when aggregated, two higher order dimensions of influence attempts emerged. These represented positive behaviors, including ask nicely, bargain, just ask, and show affection, and negative behaviors, including display anger, beg, and con. These two aggregate dimensions explained an average of 58% of the variance in total influence attempts. The coefficient alphas for the

L.A. Flurry, A.C. Burns / Journal of Business Research 58 (2005) 593–601

597

Fig. 2. Path estimates: Samples 1 and 2. Path coefficients are reported in order: sample 1, sample 2. * Significant path ( P < .05).

aggregate positive and negative influence attempts were (a1=.67, a2=.67) and (a1=.69, a2=.67), respectively. The two dimensions of influence attempts are represented in the structural model as g1 and g2 (see Fig. 2).

3.3.3. Passive social power Adaptations of Swasy’s (1979) social power scales were used to measure the parent’s perception of her child’s reward power (three items; a1=.84, a2=.83), expert

598

L.A. Flurry, A.C. Burns / Journal of Business Research 58 (2005) 593–601

power (four items; a1=.66, a2=.66), coercive power (four items; a1=.69, a2=.73), legitimate power (four items; a1=.71, a2=.71), and referent power (five items; a1=.78, a2=.78). Mothers were instructed to think about occasions when they bought a toy for their child although the child did not specifically request the toy and to respond to the standard five-point Likert scale for each item. Passive power bases are represented in the structural model as n8 –n12 (see Fig. 2). 3.3.4. Decision history Decision history was measured by three items generated by the authors. These items represented the child’s perception of his or her general success in obtaining desired outcomes. Children were instructed to think about all of the times in the past that they asked their mother to buy a toy and to respond to a four-point Likert scale (a1=.70, a2=.71). Decision history is represented in the structural model as n7 (see Fig. 2). 3.3.5. Preference intensity Preference intensity was measured by a derivation of Deighton et al.’s (1989) value of the object scale. This six-item Likert scale was modified to be specific to a toy purchase and was answered by children (a1=.87, a2=.85). Preference intensity is represented in the structural model as n6 (see Fig. 2). 3.3.6. Child’s perception of influence Adaptations of Beatty and Talpade’s (1994) relative influence scales were used to measure the child’s perception of his/her influence in a toy purchase. Children responded to two scales representing dimensions of influence: the child’s perception of his/her initiation influence (four items; a1=.72, a2=.76) and the child’s perception of his/her search/decision influence (five items; a1=.70, a2=.73). Children were instructed to think about all of the times that they wanted their mother to buy a toy for them and to respond to each item on the fivepoint frequency scale. Children’s perception of their own influence is represented in the structural model as g3 (see Fig. 2). 3.3.7. Parent’s perception of child’s influence Adaptations of Beatty and Talpade’s (1994) relative influence scales were also used to measure the mother’s perception of the child’s influence in a toy purchase. Similarly, mothers responded to two scales: the parent’s perception of the child’s initiation influence (four items; a1=.79, a2=.76) and the parent’s perception of the child’s search/decision influence (five items; a1=.72, a2=.76). Parents were instructed to think about all of the times that their child wanted a toy and to respond to each item on a five-point frequency scale. Parent’s perception of their children’s influence is represented in the structural model as g4 (see Fig. 2).

4. Results The sample for the initial examination of the structural model was 491 mother – child pairs. The structural model was estimated in LISREL 8.30 with a correlation matrix as input. An aggregated model was estimated where constructs were represented using summed scale indicators. To account for the effect of measurement error, the lambda loadings were set to the square root of coefficient alpha for the unidimensional scales and, correspondingly, the error terms were set to 1 a. Research has shown that path estimates generated from single-indicator models, which incorporate random measurement error and full latent variable models, are virtually identical in direction, significance, and strength (Cohen et al., 1990; Netemeyer et al., 1990). Finally, the lambda loadings and the error terms were set free for the two influence constructs, where each construct was measured by two summed scale indicators. The overall fit indices for the structural model were above the advocated range [v2(df ) = 129.44 (65)]. Both the absolute fit statistics (GFI=.97, AGFI=.93, RMSEA=.04) and the relative fit indices (CFI=.96, IFI=.96, NFI=.92) suggested that the structural model fit the model adequately (Hu and Bentler, 1999). Fig. 2 provides the path estimates for the structural model. Results suggest that children use four bases of social power to exert active influence in toy purchases. Specifically, children used expert, referent, and reward powers to enact positive influence attempts, such as showing affection, asking nicely, just asking, and bargaining [c(1,1)=.20, t = 2.06; c(1,3)=.38, t = 5.28; c(1,4)=.29, t = 4.17, respectively] and coercive power to enact negative influence attempts, such as displaying anger, begging, and conning [c(2,5)=.45, t = 7.97]. Children did not, however, perceive themselves to actively employ legitimate power [c(1,2) = 0.08, t = 1.25]. Two bases of passive social power were found to be associated with the parents’ perception of children’s influence. Parents’ perceptions of their children’s passive expert and legitimate powers were positively related to their perception of their children’s influence [c (4,8) = 0.61, t = 8.17; c (4,9) = 0.14, t = 2.30, respectively]. In contrast, parents’ perceptions of their children’s passive referent, reward, and coercive powers were not significantly related to their perception of their children’s influence [c(4,10) = 0.10, t = 1.95; c(4,11) = 0.02, t = 0.33; c(4,12)=.03, t = 0.43, respectively]. Preference intensity was significantly related to children’s use of negative influence attempts and to children’s perceptions of their influence [c(2,6)=.19, t = 3.44; c(3,6)=.42, t = 7.84, respectively]. However, preference intensity was not significantly related to positive influence attempts [c(1,6)=.17, t = 1.82]. Decision history was significantly related to a child’s perception of his/her influence [c(3,7)=.25, t = 4.49]. Both positive influence attempts and negative influence attempts were related to a child’s per-

L.A. Flurry, A.C. Burns / Journal of Business Research 58 (2005) 593–601

ception of his or her influence in a toy purchase decision [b(3,1)=.25, t = 4.67; b(3,2)=.24, t = 4.49, respectively]. A significant reciprocal relationship was found between a child’s perception of his or her influence and the parent’s perception of the child’s influence [b(4,3)=.21, t = 3.39; b(3,4)=.19, t = 2.95, respectively]. To assess how much of the variation in children’s influence can be attributed to these factors, the R2 was examined. Results indicate that the relationships in the conceptual model explain 52% of the variance in children’s perceptions of their own influence and 59% of the variance in mother’s perceptions of their children’s influence, both suggesting that the conceptual model provides a good theoretical explanation of variations in children’s influence. Finally, 27% of the variance in positive influence attempts and 24% of the variance in negative influence attempts were explained by the relationships in the conceptual model.

5. Validation To validate the aforementioned results, a second sample of 496 mother – child pairs was analyzed with similar procedures. The model exhibited adequate fit [v2(df) = 131.47 (65); GFI=.97; AGFI=.92; RMSEA=.05; CFI=.96; IFI=.96; NFI=.92] (Hu and Bentler, 1999). Fig. 2 illustrates the path estimates for the structural model obtained by the validation sample. Results of the second sample supported the findings that children used expert and referent power to enact positive influence attempts [c(1,1)=.24, t = 2.69; c(1,3)= .41, t = 5.32, respectively] and coercive power to enact negative influence attempts [c(2,5)=.43, t = 7.44]. Both positive influence attempts and negative influence attempts were found to be significantly related to a child’s perception of his or her influence [b(3,1)=.28, t = 4.99; b(3,2)=.19, t = 3.57, respectively]. In addition, parents perceived their children to have passive expert and legitimate powers to exert influence [c(4,8)=.46, t = 6.90; c(4,9)=.19, t = 3.36, respectively]. Preference intensity was significantly related to negative influence attempts [c(2,6)=.16, t = 2.94]. Both preference intensity and decision history were significantly related to a child’s perception of his/her influence [c(3,6)=.49, t = 9.24; c(3,7)=.14, t = 2.73, respectively]. Finally, a significant reciprocal relationship was found between a child’s perception of his/her influence and the parent’s perception of the child’s influence [b(4,3)=.25, t = 4.07; b(3,4)=.15, t = 2.31, respectively]. Examination of the R2 estimates also supported the findings of the first sample. In the validation sample, 48% of the variance in children’s perceptions of their own influence and 39% of the variance in the mother’s perceptions of their children’s influence were explained by the relationships in the conceptual model. In addition,

599

25% of the variance in positive influence attempts and 21% of the variance in negative influence attempts were explained by a child’s active social power and preference intensity.

6. Discussion Children were found to apply expert, referent, and reward bases of active social power to enact positive influence attempts, while they applied coercive power to enact negative influence attempts. They employed these to gain more control in the decision-making process, suggesting that children perceive that exhibiting knowledge about the product, rewarding parents with ‘‘good’’ behaviors and selecting items that parents would approve of are the best ways to positively influence purchase decisions and that sanctioning parents via negative actions may also be a useful influence technique. Children did not, however, perceive themselves to have legitimate power. It could be that children aged 8– 11 are not yet old enough to believe that they have ‘‘rights’’ in decision-making. With respect to children’s passive social power, parents felt that their children had both legitimate and expert powers. That is, parents perceived their children to be influential by nature of their legitimate right to participate in the decision-making process and due to their knowledge and/or expertise in the product category. Parents did not, however, attribute passive referent, reward, or coercive power to their children. These findings could be explained by parents’ unwillingness to acknowledge their children’s ability to influence purchase decisions by less socially acceptable means. Parents may feel that if they were to recognize their children as having reward and coercive powers, they would then have to admit to being susceptible to manipulation by their children. This would then make parents appear to be lacking in parenting style or skills. Another consideration is that a child’s active or passive use of social power may vary according to the product context. This research examined children’s influence in a toy purchase, a product category in which children would logically consider themselves to be powerful and thereby able to call upon several sources of active social power. On the other hand, parents indicated that their children’s passive social power stemmed from expert and legitimate sources, possibly a concession to the fact that children would be the primary user of the product. Decision history was also found to explain variations in children’s influence. Children’s purchase influence is embedded in a series of decision episodes over time and the general character of these affects a child’s perception of his/ her influence. Children may build confidence following successive favorable decision outcomes. This confidence from past successes teaches a child to expect future success in influencing purchase decisions.

600

L.A. Flurry, A.C. Burns / Journal of Business Research 58 (2005) 593–601

Another significant contributor to children’s influence was preference intensity. A child’s intensity of desire was found to explain both the child’s choice of influence attempts and the child’s assessment of his/her influence. Although preference intensity did not affect a child’s use of positive influence attempts, it did relate to the child’s use of negative influence attempts. Children may be more likely to revert to employing less socially acceptable behaviors when they strongly desire to purchase a particular item. It was also supported that children with strong preferences for a product are more likely to consider themselves influential in the decision-making process. Finally, the interdependent or reciprocal nature of child – parent influence judgments was demonstrated. This finding supports the notion that influence is a reciprocal phenomenon, whereby children’s assessments of their influence are related to the parents’ assessments of children’s influence and vice versa. Despite the limitations of a single product context, the findings of this research contribute to theory application and testing in several ways. First, this research offers an explanation of why variation has been observed in children’s influence by using a theoretically driven approach. The gain in explanatory power is noteworthy. For example, past research has been able to explain approximately 18% of the variation in influence (Ahuja and Stinson, 1993; Jenkins, 1979). In contrast, the proposed conceptual model used social power theory to construct a more comprehensive model, resulting in over 39– 59% of the variance explained in both children’s and parents’ assessments of children’s influence. Thus, while other explanatory factors remain unidentified at this time, the constructs used in the present research are certainly promising. Another contribution of this research is the examination of children’s social power resources from both an active and passive perspective. Past research has noted that the distinction between active and passive influence was an area in need of explication (McNeal, 1992; Roedder-John, 1999). The present study demonstrated that children’s influence is derived from both active social power and passive social power. An interesting phenomenon evident in the findings is that the child’s active power perceptions and his/her mother’s passive perceptions of his/her social power both independently contribute to the magnitude of purchase influence attributed to the child. The separation of the child’s social power into active and passive resources seems warranted and a fruitful approach. This research was also able to show that children were capable of critically evaluating their social power bases and matching their power resources with appropriate influence attempts to yield the greatest return. This finding is theoretically significant, and it is the first research to examine children’s perceptions of their social power and the relationship between their bases of social power and their actions employed in an influence attempt. In doing so, this research has filled a gap in our understanding of

the power bases of children by extending French and Raven’s (1959) bases of social power to children in a purchase context. In conclusion, our research has demonstrated that social power and other salient constructs (e.g., decision history and preference intensity) help us to understand purchase decision-making influence perceptions for children aged 8 –11, a group largely overlooked in recent research (Roedder-John, 1999). Future research should look to extend this research by testing the veracity of the conceptual model with children of other ages, other product categories, other populations, other family purchase decision-makers (such as fathers), and with different types of families such as single parents, stepchildren, or families representing diverse cultural backgrounds.

References Ahuja RD, Stinson KM. Female-headed single parent families: an exploratory study of children’s influence in family decision-making. Adv Consum Res 1993;20:469 – 74. Atkin CK. Observation of parent – child interaction in supermarket decision-making. J Mark 1978;42(October):41 – 5. Beatty SE, Talpade S. Adolescent influence in family decision making: a replication with extension. J Consum Res 1994; 21(September): 332 – 41. Cartwright D. A field theoretical conception of power. In: Cartwright D, editor. Studies in social power. Ann Arbor (MI): Institute for Social Research, 1959. p. 183 – 220. Cohen P, Cohen J, Teresi J, Marchi M, Velez CN. Problems in the measurement of latent variables in structural equations causal models. Appl Psychol Meas 1990;14(2):183 – 96. Coleman JS. The mathematics of collective action. Chicago (IL): Aldine, 1973. Corfman KP, Lehmann DR. Models of cooperative group decision-making and relative influence: an experimental investigation of family purchase decisions. J Consum Res June 1987;14:1 – 13. Cowan G, Avants SK. Children’s influence strategies: structure, sex differences, and bilateral mother – child influence. Child Dev 1988;59: 1303 – 13. Cowan G, Drinkard J, MacGavin L. The effects of target, age and gender on use of power strategies. J PersSoc Psychol 1984;47(6):1391 – 8. Deighton J, Romer D, McQueen J. Using drama to persuade. J Consum Res 1989;16(December):335 – 43. Emerson RM. Social exchange theory. In: Rosenberg C, Turner C, editors. Social psychology: sociological perspectives. New York: Basic Books, 1981. p. 30 – 65. French Jr J, Raven B. The bases of social power. In: Cartwright D, editor. Studies in social power. Ann Arbor (MI): Institute for Social Research, 1959. p. 150 – 65. Hu L, Bentler PM. Cutoff criteria for fit indexes in covariance structure analysis: conventional criteria versus new alternatives. Struct Equ Model 1999;6(1):1 – 55. Isler L, Popper ET, Ward S. Children’s purchase requests and parental responses: results from a diary study. J Advert Res 1987;October/ November:28 – 39. Jenkins RL. The influence of children in family decision-making: parents’ perceptions. Adv Consum Res 1979;October/November:6:413 – 8. Lewin K. Field theory in social science. New York: Harper, 1951. McNeal JU. Kids as customers: a handbook of marketing to children. New York: Lexington Books, 1992. Netemeyer RG, Johnston MW, Burton S. Analysis of role conflict and

L.A. Flurry, A.C. Burns / Journal of Business Research 58 (2005) 593–601 role ambiguity in a structural equations framework. J Appl Psychol 1990;75(2):148 – 57. Olson DH, Cromwell RE, Klein DM. Beyond family power. In: Cromwell RE, Olson DH, editors. Power in families. New York: Wiley, 1975. p. 235 – 40. Roedder-John D. Consumer socialization of children: a retrospective look at twenty-five years of research. J Consum Res 1999;26(3):183 – 213. Scanzoni J, Szinovacz M. Family decision making: a developmental sex role model. Beverly Hills (CA): Sage, 1980. Simmons Market Research Bureau. Kids speak out: insights into today’s youth market. New York: Simmons Market Research Bureau, 1993. Sprey J. Family power and process: toward a conceptual integration. In:

601

Cromwell RE, Olson DH, editors. Power in families. New York: Wiley, 1975. p. 61 – 79. Swasy JL. Measuring the bases of social power. Adv Consum Res 1979;6:340 – 6. Turner JH. The structure of sociological theory. Homewood (IL): Dorsey Press, 1982. Wells WD. Communicating with children. J Advert Res 1965;5:2 – 14. Williams LA, Burns AC. Exploring the dimensionality of children’s direct influence attempts. Adv Consum Res 2000;27:64 – 71. Wolfe DM. Power and authority in the family. In: Cartwright D, editor. Studies in social power. Ann Arbor (MI): Institute for Social Research, 1959. p. 99 – 117.