Impact of different facets of product involvement and consideration set on teens ’ family purchase decision making : A multivariate analysis

Creating and maintaining strong brand and a band wagon of loyal customers have become increasingly difficult in today's competitive environment due to proliferation of numerous brands in a generic product category. Brand loyalty has been shown to be associated with higher rates of return on investment due to increase in the market share. Children’s influence on family purchase decision depends on a number of parameters and situations. Children exercise various methods to influence their parents’ decision. Their influence varies from products to products. It depends on parents’ education, profession, income, single parent working or both parents working, and type of family. Astonishingly, very few studies have been undertaken to relate the brand loyalty and product involvement behavior of teenagers. The data for this study are gathered from a cross section of teenagers of different socioeconomic backgrounds, from the major metros of Indi during the third quarter of 2014. In this paper aside including only involvement of product brand influence scores, brand trust and the size of the consideration set have been incorporated to predict brand loyalty of teens. The findings of the study reveal that multi-dimensional measure is a better predictor of loyalty behavior. Research findings also reveal that different explanatory variables have diverse influence on the brand loyalty behavior of teens.


INTRODUCTION
Teenagers in the contemporary marketing environment constitute a pivotal market segment and deserve considerable attention from marketers and academicians due to the fact that market is expanding and teens spend vast amount of money for a wide variety of products.It is a reality that children play a central role in influencing family purchasing decisions; this has urged the marketing researchers to track their brand influencing behavior.
It can hardly be denied that the degree of influence exerted by children differs across product categories as well as the stage of the decision making process.The teenage population is increasing exponentially over the *Corresponding author.E-mail: dipaksit@gmail.com.Tel: +91(0)9434210759.
Authors agree that this article remain permanently open access under the terms of the Creative Commons Attribution License 4.0 International License last decade and for this reason the consumer behavior researchers are showing enormous interest to unveil the buying behavior of this growing segment.
The Indian consumer market, which is primarily dominated by young generation, is becoming increasingly sophisticated and brand conscious.A typical upper middle class young consumer is beginning to look beyond the utility aspect of a product to seek intangibles like brand and lifestyle statement associated with the product.This modern consumer wants his purchases to reflect his lifestyle or at least the one he aspires for.As a result of this brand consciousness, the food and beverage segment of the FMCG sector is already witnessing a significant shift in demand from loose to branded products.
India alone is home to 1.136 billion people, out of which an estimated 350 million are in the age bracket of 10-24 years.Their purchasing power has significantly increased, both, in terms of salary and pocket money.An ASSOCHAM survey revealed that the average monthly allowance of urban children in the age group of 10-17 years has gone up from 300 in 1998 to 1,300 in 2008.This segment is very attractive due to its size, increasing spending power, and large exposure to media.Among the existing studies, there is none in our knowledge that documents brand relationships of young consumers in an emerging economy.Finally, young consumers the world over are influenced by peers and family in their brandrelated decisions (Singh et al., 2003).For marketers, it is important to understand the impact of these factors on brand relationships and brand switching intentions.
Teenagers who belong to the age group of 13 to 19 approximately spend $150 billion per year globally.Teens also exert influence on the tune of an additional $150 billion per year globally with "pester power."It is believed from various sources that they indirectly influence another $300 billion per year.That is a total purchasing/influencing power of $600 billion this year.Moreover, teens to a considerable extent influence various products to be consumed and used by their parents and other members of their family to which they belong.
The world is witnessing a rise in the number of young consumers and evidence suggests they are brand conscious.In addition, 57 per cent of the teenagers cite marketing and media in their conversations as compared to 48 per cent adults (Hein, 2007).Teenagers' share of expenditure in the Indian market is worth $2.8 billion (Rana, 2007); young consumers tend to be more involved with material possessions (Belk, 1988).Consumer socialization process begins at home; young consumers see brands which are consumed in the family and are likely to give first preference to the use and purchase of those brands.Even though young consumers start consuming and developing relationships with the new brands they get exposed to, the impact of the brand exposure from their families is likely to be strong.In the Indian context, family has traditionally played a strong role in influencing choices of their progeny.Bravo et al. (2007) argue that family always provides suggestions regarding brands for young consumers.The greater the family's influence on brand choice, the lower the depth of brand relationship (Sahay and Sharma, 2010).
Today's teenage customers have emerged as big-time spenders, who not only have a good amount of pocket money but also know how to supplement the same by means of internships, summer jobs and part-time jobs.It can hardly be denied that the teenage market in India is growing at a fast pace although no systematic effort has been made to study the loyalty behaviour of teenage consumers.Considering a research gap in this area, in this paper an effort is made to discern the product involvement and brand loyalty behaviour of teenagers in India.
There is a voluminous literature dedicated to the study of the concept of brand loyalty.However, most of the research work carried out in this area has focused on the Indian consumers, and studies involving the Indian teenage consumers are quite less in number.It is high time that research pertaining to loyalty involving Indian teenage consumers is strongly encouraged, because the outcome of any such research work would help the marketers to implement innovative changes in their product portfolio and thereby retain the customers.
Entry of multinationals and their aggressive way of garnering market share results in sleepless nights for brand executives.Research pertaining to loyalty involving Indian consumers is the need of the hour, because the outcome of any such research work would help the corporate to implement innovative changes in their product portfolio and thereby retain the customers.There is a significant increase in the spending power of Indian teenagers and their desire to purchase sophisticated products.Availability of more number of multinational brands with unique attributes has forced the oscillating consumers to buy new brands.
Another major consideration for the marketer is to look at the issue of teenager brand loyalty from the perspective of teenagers' level of involvement.The findings of various studies (Laurent and Kapferer, 1985;Bordo, 1993;Leclerc and Little, 1997) indicate a positive relationship between these two dimensions.However, since the pattern of Indian teenagers' involvement has not been fully explored, a study is required to examine the relative influence of the antecedents of teenage involvement on brand loyalty incorporating a few important explanatory variables that have not been addressed by researchers to predict brand loyalty behaviour of teenagers.Keeping in view the gaps in the existing literature this study is conducted with the following objectives: i) To determine the level of involvement of teens with respect to various brands included in our study and to look into the number of brands in their consideration set.ii) To assess the brand loyalty scores for various product categories considered in our study.iii) To develop a Brand Influence Score (BIS) scale of using a seven point Likert type of items.iv) To investigate nomological validity of the measurements by investigating the degree of association between brand loyalty and a set of explanatory variables.v) To explain adequately why results are divergent for a cross section of products included in our study.vi) To integrate the findings above and suggest possible managerial implications based on the findings of the study.
The present study encompasses five broad sections including the introductory section which contained an overview of the teenage market in India as well as changes taking place in the developed markets.Instead of providing research questions, the objectives of the study have been included in the introductory section.

LITERATURE REVIEW
Product involvement and brand loyalty are two important components believed to explain a considerable proportion of consumer decision making behaviour.Several empirical studies have been reported in various marketing literature to establish relationship between consumer involvement with products and Brand Loyalty (Quester and Lim, 2003;Douglas, 2006; as cited by Sritharan et al., 2008).The findings of their studies in general postulate hat consumers who are more involved with a product category exhibit greater loyalty towards the brand.A few researchers in the field of consumer behaviour view that loyalty is a process of repurchasing which happens due to situational restrictions, lack of feasible alternatives, or out of expediency (Sadasivan et al., 2011).The researchers' interest to study the consumer involvement behaviors and brand loyalty has gained momentum in recent years after the publication of two articled by Quester and Lim (2003).Quester and Lim (2003), in their empirical observation, explained that the relationship between the product involvement and brand loyalty is found to involve different aspects of involvement for different product categories considered in their study.Knox and David (2003) also support the findings of Quester and Lim by integrating classical theory of involvement, brand loyalty, and commitment (Traylor, 1981).Even in a grocery product purchase setting the outcome of the study corroborates the relationship between involvement and brand loyalty.In a similar study, Yi and Hoseong (2003) conducted a research to investigate the moderating role of product involvement and brand loyalty.The study further confirms that the consumer loyalty was highly affected by their level of involvement.In Indian context, Jain and Sharma (2002) observed that differences in consumer involve-ment with the product depends on large number of products and brand related factors viz.consumer risk perception and hedonic value of the product, brand awareness and so on.The findings of the study reveal that consumer involvement differs across different type of products.Sahay and Sharma (2010), in a very recent study, reported that strong association has been observed between brand name and loyalty.The research indicated a positive as well as significant association among different faces of brand loyalty for cosmetics brands.Another current study conducted by Sridhar ( 2007) reveals that users of cell phone are highly brand loyal.Buyers of cell phones in their repeat purchase stick to the same brand once they find the brand satisfying all their needs and desires.In marketing terminology the phenomenon can be explained by the concept of risk importance which signifies that consumers in general are risk averse and try to avoid the psychological stress due to mispurchase of the desired brand.It is quite normal for consumers to favour a user friendly cell phone due to the fact that they do not have to pass through new learning and adoption process.In the context of store image study it is also revealed that involvement plays a dominant role in the purchase of private store brand (PSB).The findings corroborate that involvement influences buying decision and different faces of CIP scales are found to have strong impact on the loyalty behaviour for PSB.
The concept of involvement was theorized by Krugman (1965) and subsequently the concept was refined by various authors.A substantial research work in the field of involvement has been taken to relate the brand loyalty and commitment behaviour of consumers, particularly after two articles were published in the journal of marketing and the journal of marketing research by Laurent andKapferer (1985a, 1985b).However, Traylor (1981) has probably examined initially the relationship between product involvement and brand commitment.Since then a plethora of research articles have been published to relate involvement variables and brand loyalty behavior of consumers for a wide variety of product and services.However, Traylor (1981) has probably examined first the relationship between the product involvement and brand commitment.
In the context of the review of literature presented above, several aspects need to be explained for establishing the justification of the present study.In existing literature the concept of narrow categorizers or broad categorizer has received very little attention from the researchers.Highly involved consumers find fewer brands acceptable.Theory posits that narrow categorizers are likely to be more loyal to the brand they purchase for consumption.On the other hand, consumers who are broad categorizers have a large number of brands in their consideration set and they are very likely to be brand switchers.In view of this, it is perfectly logical to incorporate the number of brands the consumers have in their consideration set.So far our knowledge goes previous studies did not incorporate this important variable for predicting brand loyalty behaviour of consumers.This study is undertaken to predict the involvement and brand loyalty behaviour of Indian teenagers who exert considerable pester power on their parents for the purchase of a brand of their choice.We have made a serious attempt to develop a brand influence score (BIS) scale that is reliable as well as valid to discern the relationship between BIS and brand loyalty.We have then sincerely endeavoured to incorporate this construct which was not considered by previous researchers working in this area.
In this research work, we have introduced both global as well as multi-dimensional measure to capture the construct involvement to probe which measure is more effective in predicting brand loyalty behaviour of teens.

METHODOLOGY
Since the objective of our study is to relate the teenage involvement and brand loyalty behaviour incorporating the brand influence score and the number of brands in the consideration set, we have employed factor analysis to establish scale dimensionality.In addition to this, multiple regression analysis is employed to assess the importance of different variables in predicting the brand loyalty of teenagers considered in our study.Regression analysis is also employed to ascertain the predictive validity of the proposed measure of involvement and brand loyalty.The construct involvement is measured using a twelve item multi-dimensional scale incorporating risk probability, risk importance, pleasure value and the Sign Value.In our study, we have also measured involvement using a five item five-point scale proposed by Zaichkowsky (1995) to compare whether the uni-dimensional or multi-dimensional measure of involvement predicts brand loyalty behaviour of teenagers.Highly involved consumers find fewer brands acceptable (narrow categorizers) and tend to be more loyal.On the other hand, brand switchers are likely to have more brands in their consideration set (broad categorizers) that are likely to be less loyal to their brands.In an attempt to establish this phenomenon we have gathered data from teenagers regarding the number of brand they have in their consideration set.Teenagers play a significant role in deciding the brands they purchase for themselves as well as they shape the brand choice behaviour for other brands purchased for family consumption which are technically known as pester power.In our study, we have developed a seven item five point scale to measure the Brand Influence Score (BIS) of teenagers which is likely to influence the brand loyalty of teenagers.The detailed methodological procedures followed in our study are briefly discussed in the subsequent discussions.

Scale development
While developing the scale to measure the involvement construct, we have followed the recommended scaling procedures which are very commonly found in psychometric literature (Nunnally, 1978).Following Churchill's (1979) suggestion, we generated a pool of items for each facet from different involvement scales developed by Laurent and Kapferer (1995b), Jain and Srinivasan (1990), Lastovicka andGardner (1979), andZaichkowsky (1985).In addition, a preliminary in-depth discussion with a sample of Bhattacharya and Saha 35 respondents (n=21) pursuing management programme was also an important source from which we generated a few other items (Bhattacharya, 2000).Altogether, 28 five point semantic differential items were initially developed to reflect the four facets of involvement.These items were then judged for content validity by a small panel of experts (n=3) resulting in 17 semantic differential statements.The panel comprised both academicians and marketing professional having adequate knowledge in this field.These 17 items were then administered to an initial sample of post-graduate University (n=42) students over two products categories per student.
Following suggestions of Zaichkowsky (1985) and Gaski and Etzel (1986), statements with items to total correlation (within each component) of r = 0.50 or more were retained.In this process five more items were dropped and finally 12 items were retained to measure 4 facets of involvement.It was required to establish scale dimensionality since the Interest and Pleasure items of CIP scale continued to fuse into a single factor.
Data for the survey are obtained from a convenience sample of 447 teens drawn from the four major metros in India.In addition to meeting the socio-demographic criteria, the choice of the convenience sample is made so that the teenagers have to be a user of the product on which their responses are sought.Due to financial constraint, it was not feasible for us to adopt a probability sampling technique Convenience sample, though not very scientific, helps in getting over this limitation.Moreover, since our objective is to determine the degree and direction of relationship between various facets of involvement and their influence on the teenagers' loyalty behavioural aspect and no generalizations about the sample teenagers were envisaged, a convenience sample was considered adequate for this study.The sample size was not very large but previous research in this area also conducted similar type of studies covering a sample size ranging from 150 to 450 in most of the cases.The data for the study were collected from different coaching centres by personally administering the questionnaire.The respondents were given a complementary gel pen as a token gift for participating in the study.

Selection of stimulus products
In our present study, a good deal of exploratory work is needed to select the products to be included in the study.While selecting the stimulus products for the study we have to resolve some important issues.First, the individual considered for the interview as a user of the products for which his response is sought.Secondly, products are deliberately chosen to represent contrasting profiles on various dimensions of involvement viz.risk, pleasure and sign (self expression factor) associated with the product.
The final list of products retained for this study is done through a series of qualitative in-depth interviews with the teenagers.

Psychometric performance of the scale
The twelve-item involvement scale was initially administered to a sample of students enrolled in the Department of Commerce of North Bengal University to assess the reliability and validity of the proposed measure where each student had to give response on two product categories.We computed internal consistency reliability by Cronbach's alpha as well as by test-retest reliability.It is quite evident from the table that the reliability coefficients are reasonably high and it can be concluded that the scale which we intend to use in our study possesses sufficient degree of internal consistency despite a small number of items in each scale.It has to be remembered that consistency is a necessary but not sufficient condition for validity (Nunnally, 1978).Therefore, in the subsequent discussion we address this important issue in detail.
The assessment here will begin with construct validity, which refers to the extent to which the hypothetical, unobservable construct of interest corresponds to its purported measure (Peter, 1981).In order for a measure to have construct validity, each of the measurement items must relate to the characteristics of the construct, and each item must be free from contamination by elements of other constructs.These two requirements are operationalised by two validity tests, viz.(a) Content Validity and (b) Scale Dimensionality.These two issues are briefly addressed below.

Content validity
When a test is constructed so that its content of term measures what the whole test claims to measure, the test is said to have content or circular validity.It was done essentially by a systematic examination of the items included by researchers while capturing the domain of the construct.In addition to this, initial scale items (17 pairs) were judged by a small sample of experts who expressed that these items could be used to capture the domain of the construct.Moreover, statistical tests were applied to ensure content validity.In our study, the level of internal consistency measured by Cronbach's alpha provided sufficient evidence for the content validity.

Scale dimensionality
The scale dimensionality may be reviewed via factor analysis which is a collection of mathematical procedures for determining which variables belong to which factor or underlying construct.Through factor analysis, specific expectations concerning the number of factors and their loadings are tested on sample data.Campbell (1960) and Nunnally (1978) suggest that each scale should measure a single facet if it is considered to have construct validity.Discriminant validity, on the other hand, represents the distinctiveness of each scale vis-à-vis others.To test simultaneously construct and discriminant validity, we conducted a factor analysis of the items using student samples for two different product categories.
With a few notable exceptions, the scale items were loaded on the factors they were supposed to measure.Apart from this, for other applications, factor analysis led to the results we expected: one factor per item, all items from an antecedent on the same factor and one factor per antecedent.
The results of factor analysis presented in Tables 1 and 3 amply demonstrate that the proposed measure is not contaminated with elements from the domain of other constructs or error.The systematic extraction of four factors can be interpreted as supportive evidence of construct validity.

RESULTS AND DISCUSSION
There is no doubt from past literature that involvement with product plays a dominant role in explaining the loyalty behaviour of consumers belonging to different socio-demographic strata from which the samples are drawn.The unique approach of this paper is that we wanted to apply two very important product involvement scales frequently cited in marketing literature.The global measure suggested by Zaichkowsky (1985) which is found to be highly reliable because of number of items is included in the scale.During the same year, Laurent and Kapfarer (1985a) developed a multi-dimensional measure to capture various constructs of involvement using a scale which included five distinct dimensions.We have not come across any study incorporating both these scales to measure the relationship that exists between brand loyalty and product related involvement.The multiple regression results using multidimensional measure for the brands viz.Laptop; Apparel; Cell Phone; Toothpaste; Health Drink and Toys (modified Laurent and Kafperer scale, 1985).The idea behind employing these scales for six brands considered in our study was to discern the predictive ability of the measures included in the explanatory variable set.Major findings of the survey relating brand loyalty and product involvement are presented in Tables 4-9 where six products have been considered.
Although the tables are self explanatory, a few comments are necessary to focus on the weightage of variables co-efficient in predicting brand loyalty.For cell phone, the risk probability factors as well as risk importance factors have been found to be insignificant though there are numerous brands in the market which may require lot of information processing.However, one possible reason behind this result may be attributed to consumer reliance to Nokia brand of cell phone which enjoys strong brand popularity in India.Similar findings have been reported by Quester and Lim (2003, pp 33-33).
However, for toys, we find that risk importance facet is highly significant.The buyers probably are not sure about how long the product will last and whether it would be socially acceptable.Surprisingly for toys, a lot of spurious brands are trafficked in India and many buyers have expressed their concern whether they are really getting the original brand produced by a particular company or not.
As hypothesized, it is expected that brand influence score would exert a positive influence on the brand loyalty construct.For toys the coefficient is positive and significant whereas for cell phone the same is not significant.
The size of the consideration set negatively influences   the brand loyalty variable signifying the fact that the buyers considering higher number of brands are likely to be brand switchers.However, the results amply demonstrate that the consideration set and brand loyalty behavior are inversely related.For both the brands, the adjusted R square values are significant beyond p< 0.000.In case of laptop, the behavior of coefficients did not vary significantly but the value of R square drops significantly.For cell phone, the consideration set is significant beyond p<0.05 but the same is highly significant for a brand of laptop where all variables are found to be significant.Brand Influence Score is the most important determinant variable influencing the brand loyalty behavior of teens.
For a product like toothpaste the involvement level is found to be significantly influencing the brand loyalty behavior of consumers and it can be concluded that though buyers develop a habitual buying behavior and are not reluctant to switch over to other brands, they search different flavor within their brand choice.A remarkable variation has been observed between the   results of brand loyalty measure when deodorant purchase is concerned with only notable exception that prediction of brand loyalty gives a better result when all the facets of involvement are retained.Product like health drink, the size of the consideration set positively influences the brand loyalty variable signifying the fact that the buyers considering less number of brands are likely to be more brands loyal.However, the results amply demonstrate that the consideration set and brand loyalty behavior are inversely related.For this brand, the adjusted R square values are significant beyond p< 0.000.
It was found that there is positive significant impact of these independent variables on the dependent variable having p<.000.The value of Beta coefficients for all the independent variables shows a positive association within the model.The value of adjusted R-Square predicts a goodness of fit between the set of independent variables and the dependent variable.

CONCLUSION AND MANAGERIAL IMPLICATION
As a basis for an assessment of the psychometric performance of the scale administered in our study, in this conclusive section, we begin our discussion with the findings of the reliability analyses discussed in this paper.
There are two basic dimensions of reliability: repeatability and internal consistency.Assessing the repeatability property of measure is the first aspect of reliability.The test-retest correlation coefficients amply demonstrated the repeatability property adequately.
The second underlying dimension of reliability is concerned with the homogeneity of the measure.To ensure homogeneity property internal consistency of multiple item measure has to be established.The coefficient of Alpha and the Split-half reliability estimates are within acceptable limit in spite of a very small number of items in each scale.
Consistency is a necessary condition for validity but it is not a sufficient condition for establishing scale validity.Keeping this aspect in mind several estimates of validity have been provided in our study.We begin our discussion with face or content validity.The relevance of the scale items was judged by a short panel of experts.It appeared evident to the experts that the measure provided adequate coverage of the construct.Moreover, the item to total correlation coefficients and the internal consistency measures also provided sufficient evidence for content validity.
In an attempt to establish discriminant and construct validity, factor analysis was conducted.Factor analysis of the items confirmed the multidimensional nature of the consumer involvement profile.
These research findings are of significance to marketing practitioners and reveal the teenagers influence of involvement on brand loyalty.Results show that teenagers attach more importance to 'interest and pleasure' dimension followed by 'risk importance'.From managerial point of view, these results imply that teenagers can be persuaded to buy a particular brand of toys by consistently adding new features that offer unique benefits.Precisely, the concept of 'innovation through technology' needs to be focused on.It is suggested that marketing professionals should conduct surveys to identify the expectation of teenage users, which changes frequently.
Specifically the present study offers brand executives a meaningful and valuable insight to guide them in winning competition.'Pleasure' has emerged as another important factor in the involvement scale.The respondents feel that pleasure facet is a driving force selecting a particular brand of toys and laptop this provides a clue to the corporate that the store ambience and behaviour of the store personnel should be accentuated in a manner so as to highlight the pleasure aspect.Executives can perform Multi Dimension Scaling Technique to identify the positions of competitive brands in the market and select unique positioning for their brand.This can be achieved by creating specific association (Aaker, 1991) for their brand.
'Sign' dimension has been extracted as an important factor in the analysis.It confirms that mobile brands do reflect the personality of teenage users.This result is highly relevant to managers involved in developing an identity for their brands.They can explore the possibility of launching special models exclusively for high-end teenage consumers and help establish a sense of pride by owning that brand.
This outcome of the results of multiple regression analysis suggests that 'pleasure' and 'sign value' influences brand loyalty significantly.This is the testimony that 'innovative features' of the product is the key determinant of brand selection.In today's teen world, innovation seems to be the key to ensuring continuous patronage and the products must be regularly upgraded in terms of new features that offer a fresh experience with regard to product usage.This fact points towards the overriding importance of new product launches, either as an upgraded version of an already existing product, or a totally new product itself.

LIMITATIONS AND DIRECTION FOR FUTURE RESEARCH
Cautions should be made while generalizing the findings of this study, considering sample size and area of study.The research conducted among the Indian consumers may be subject to cultural influence and the similar study of brand loyalty in other countries is recommendable.This study focused only on limited variety of products and hence, the results are not applicable to other products.Further research is required for other products and services and comparisons could be made across different product classes.It is suggested that an interesting avenue to pursue research would be to investigate whether loyal consumers and switchers differ in their information search, promotional sensitivity, and the extent to which brand loyalty is affected by sales promotion offers.