Printer Friendly

Determinants of Successful Brand Extension: Aaker and Keller's Approach.

Byline: Usman Ghani, Muhammad Aamir and Irfan Ullah Shah

Abstract

Brand extension has become a popular alternate approach to introducing new products in today's highly competitive environment. Aaker and Keller (1990) proposed a model identifying factors influencing consumers' evaluations of extensions. Replications of this model resulted in inconsistent findings, putting a question mark on its generalizability. This study attempted to examine the proposed relationships of the Aaker and Keller's model to get further evidence of its generalizability by correcting for the problem of multicollinearity which is argued to be the cause of the findings of the Aaker and Keller study. Results of the study are somewhat different from those of Aaker and Keller, but similar to other replication studies that addressed the multicollinearity issue.

Keywords: Brand extension; Aaker and Keller approach; multicolinearity

Introduction

In today's world of increasing competition, high promotion and product development costs, it has become very difficult for firms to succeed with new products (Aaker and Keller, 1990). It has been estimated that it costs about US$ 80 million to US$ 150 million to develop and introduce a new product. Further, the failure rate of new products is quite high, and some estimate it to range from 30 to 35% (Montoya-Weiss and Calantone, 1994). In these circumstances, a popular alternate approach is the brand extension strategy, which is the use of an existing brand name for launching a product in a new category (Aaker and Keller, 1990). Brand extension has been found to be followed in almost eight out of ten new product launches (Ourusoff, Ozanian, Brown and Starr, 1992) and in the Fast Moving Consumer Goods (FMCGs), 85% of the new products are extensions (Kirmani, Sood and Bridges, 1999). Brand extension uses the equity built up in established brand names for launching new products (Boush and Loken, 1991).

Brand extension is based on the assumption that favourable beliefs and attitudes toward the original brand also facilitate the formation of similar beliefs and attitudes towards the extension (Aaker and Keller, 1990). It, therefore, helps in quick acceptance by customers and reduces the risk of failure. It also helps in reduced marketing research, new product introduction, distribution and promotion costs (Chen and Lue, 2004).

The concept of brand extension has been of interest to researchers since many a decade. However, interests were aroused after a study by Aaker and Keller (1990) which aimed at exploring consumers' evaluations of brand extensions.

Aaker and Keller (1990) proposed a brand extension model in which several factors were posited to influence the success of an extension. These factors (independent variables) were identified as:

1. Quality of the parent brand (labelled as QUALITY)

2. Fit between the parent brand and the extension category. The fit was further considered to have three dimensions. They were:

a. COMPLEMENT, which is the extent to which the parent and the new product are viewed as complements by the customers;

b. SUBSTITUTE, which is the extent to which the parent and the new products are viewed as substitutes by the customers, and

c. TRANSFER, which is how the customers view the relationship between the parent and the new product in terms of their design or manufacturing.

3. The perceived difficulty of making the extension (labelled as DIFFICULT).

The consumers' attitude towards the new extension was taken as the dependent variable.

The following four hypotheses were proposed based on the relationships among the above mentioned factors.

H1: Perceptions of higher quality toward the parent brand leads to a more favourable attitude toward the extension.

H2: The fit between the parent brand and the extension category has a positive association with the attitude towards the extension.

H3: The transfer of a parent product's perceived quality to the extension is enhanced when the two product classes fit together.

H4: The difficulty in making the extension has a direct positive association with the attitudes towards the extension.

Aaker and Keller (1990) found no support for H1, partial support for H2 (only transferability was significant) and H3 (Q C and Q S were significant) and full support for H4.

The Aaker and Keller's (1990) study has been replicated by several authors (Sunde and Brodie, 1993; Bottomley and Doyle, 1996; Holden and Barwise, 1995; Nijssen and Hartman, 1994) but their results have been inconsistent. While responding to the inconsistency in the findings, Aaker and Keller (1993) proposed that it might have been due to the differences in stimuli and cultures. However, Bottomley and Doyle (1996) argued that the differences in the results may be due to the multicollinearity between the main effects and cross-product or interactions. They corrected the problem of multicollinearity in their analysis by using the residual-cantering method proposed by Lance (1988) and found different results. They, therefore, rejected the Aaker and Keller's (1990) proposition that the differences in the findings may have been due to the differences in stimuli and cultures.

In addition to the problem of inconsistency of the results, another aspect of these studies is that most of them have been conducted in developed countries, like the USA, New Zealand, the UK, Germany and France; therefore, there is a need to further examine these relationships in the underdeveloped and developing countries to see whether these relationships hold true in these countries or not. This study therefore, attempts to test the relationships proposed by Aaker and Keller (1990) in Pakistan which is a developing country. Further, the study uses the analytical techniques used by Bottomley and Doyle (1996) to control for the multicollinearity problem.

This study, along with replicating the Aaker and Keller's (1990) model, also examines the relationships of the variables at the brand level and compares the findings with some of the later studies conducted on the phenomenon of brand extension to get additional insights.

A summary of the results of the replicated studies including the ones with and without adjustments for the multicollinearity problem is presented in Table 1. As can be seen, none of the seven studies mentioned support the four hypotheses proposed for the present study.

Table 1: Summary of Findings of the Studies on Brand Extension

Table 1: Summary of Findings of the Studies on Brand Extension

Hypothesis###Aaker and###Sunde and###Sunde and###Nijssen and###Bottomley###Barrette et###Bottomley

###Keller (1990).###Brodie###Brodie###Hartman###and Doyle###al. (1999).###and Holden

###(original)###(restated)###(1994).###(1996).###(2001).

###(1993)###(1993)

1. The high###The statistical###Supported###Supported###Supported###Supported###Supported###Supported

quality of###test of this

parent brand###study did not

lead to more###support this

favourable###hypothesis

acceptance of###(Hi).

brand

extension.

2. The higher###Supported to###Supported to###Supported###Supported###Supported###Supported###Supported

fit between###certain###certain###that

parent brand###extent. Only###extent. Only###transfer-

and extension###transferability###substitute is###ability and

is significant in###is significant.###insignificant,###substitute

evaluation of###are

brand###significant.

extension.

3. The###Supported###Not###Supported###Supported###Supported###Supported###Supported

interaction of###that Q C and###supported.###that Q C and###that Q T###that Q T###to certain###that Q T

quality of###Q 5 are###Q T are###and Q 5###and Q C###extent that###and Q C

parent brand###significant.###significant.###are###are###Q T is###are

with fit###significant. significant. significant.###significant.

variables is

helpful in the

evaluation of

extension.

4. The###Supported###Not###Not###N/a###Not###Not###Supported

difficulty of###supported.###supported###supported.###support.

manufacturing

product class

has positive

relationship

with

evaluation of

extension.

Research Method

Three prominent brands including Colgate, Nestle and Lux with two extensions for each brand were used to test the hypotheses of the study.

Primary data were collected through a fully structured questionnaire consisting of five-point Likert scale from a sample of 150 students in the business administration programme of the Institute of Business and Management Studies/Computer Sciences (IBMS/CS), Agricultural University, Peshawar. Convenience sampling procedure was used for selecting the respondents.

Since the study replicates the Aaker and Keller's model, the same variables were used. The independent variables were Quality, Fit and Difficult while the Attitude of consumer was the dependent variable. The Fit between the parent brand and the extended category was conceptualized as consisting of three dimensions, that is, Complement, Substitute and Transfer. The scales for measuring these variables have been adopted from Jahangir et al. (2009).

Data Analysis

Two regression models, namely Main effect and Full effect model were used for data analysis. The Main effect model consists of Quality, Transfer, Complement, Substitute and Difficult. These are the independent variables regressed against the Attitude as a dependent variable while the Full effect include the Main effect model variables in addition to the Quality[?]Complement, Quality[?]Transfer and Quality Substitute. The model which was used for analysis of data is given below.

Attitude of consumers =

b0+b1Q+b2C+b3S+b4T+b5Q T+b6Q C+b7Q S+b8 Difficult + error

Where

Q = Quality

C = Complement

S = Substitute

T = Transfer

The procedure of Bottomley and Doyle (1996) was used for analysing the data. They used two levels of analysis: aggregate level analysis and brand extension level analysis. In aggregate level analysis, the data of all individual brand extensions were grouped in SPSS and regressed to get combined result of all brand extension for individual variables. The sample size for the aggregate level analysis was created by multiplying 150 subjects with six brand extensions. The analysis of the main-effect and the full-effect terms create dispute in the results due to the presence of high multicollinearity which Bottomley and Doyle (1996) argued to be the reason for inconsistent results between Aaker and Keller (1990) and their close replication by Sunde and Brodie (1993). In the present study, the multicollinearity problem was identified through Bivariate Correlation. High correlation leads to high multicollinearity. The result of correlation is shown in Table 2.

A rule of thumb is that multicollinearity may be a problem if a correlation is greater than 0.90 or if several are greater than 0.7. As such, Table 2, in accordance with the rule of thumb, presents high multicollinearity between the independent variables of the study.

Table 2: Correlation Result of Main Terms and Interaction Terms

###QUALITY###COMPLEMENT###SUBSTITUTE###TRANSFER###Q C###Q 5###Q T

QUALITY###1.000###.132###.046###.072###.590###.501###.640

COMPLEMENT###1.000###.085###.065###.859###.128###.122

SUBSTITUTE###1.000###.274###.086###.867###.243

TRANSFER###1.000###.080###.273###.794

Q C###1.000###.351###.406

Q 5###1.000###.516

Q T###1.000

To handle the problem of multicollinearity, the Lance (1988) residual cantering method similar to Bottomley and Doyle (1996) was used. It is a two-step process. In the first step for each interaction-term, such as Q C, Q S and Q T, the residual cantering is initiated by an equation in which regression for each interaction-term is created from the independent variables (Quality and Fit). The equation for the first step is written in the following form:

Q Fit = Quality + Fit

To develop residual of the interaction-terms, the above equation is further proceeded.

D1x2 = X1x2 - X1x2 = X1x2 - (C1X1 - C2X2)

Whereas

D1x2 = Q Fit (Residual) X1 = Quality

X2 = Fit

X1x2 = Q Fit

The Q C (residual), Q S (residual) and Q T (residual) were developed with the help of the above equation. In the second step, Q C (residual), Q S (residual) and Q T (residual) were regressed instead of Q C, Q S and Q T against the attitude of the consumers. The equation for the second step is:

Attitude of consumers = b0quality + b1complement + b2Substitute + b3Transfer + b4Q C (resid) + b5Q S (resid) + b6Q T (resid) + b7Difficult

The data were analysed at both the aggregate and the brand extension levels. The results are discussed with reference to the hypotheses of the study that were adopted from the original Aaker and Keller (1990) study on brand extension.

Aggregate Level Analysis Results

In aggregate level analysis, Table 3 presents the results of the main effect model, while Table 4 shows the results of the full effect model after using residual term to reduce multicollinearity.

Table 3: Study Result of Main Effect Model

Table 3: Study Result of Main Effect Model

###Unstandardized Coefficient###Standardized Coefficient

Variables###B###Std. Error###Beta###t###sig.

(Constant)###1.144###0.157###7.281###.000

QUALITY###0.216###0.028###0.230###7.831###.000

COMPLEMENT###0.168###0.021###0.233###7.882###.000

SUBSTITUTE###8.319E-02###0.022###0.117###3.868###.000

TRANSFER###0.229###0.026###0.268###8.855###.000

DIFFICULT###-2.83E-02###0.020###-0.041###-1.396###.163

###.246

The results from Table 3 and Table 4 suggest that quality is significant in the evaluation of brand extension with beta coefficient (0.230) and (0.230) respectively which do not support the Aaker and Keller's (1990) findings for quality. They hypothesized in their study that quality had a direct association with the attitude of the consumer's, but statistically they found inverse relationship between Quality and the attitude of the consumers towards brand extension. For all of the three Fit variables, both Tables 3 and 4 of the present study show significant results. The application of the main effect model yield beta coefficients of 0.233 for Complement, 0.117 for Substitute and 0.268 for Transfer whereas the full effect model shows beta coefficients of 0.233 for Complement, 0.118 for substitute and 0.266 for Transfer.

The interaction terms are included only in the full effect model, and Table 4 shows that only Q C is significant in the evaluation of brand extension while Q S and Q T are insignificant, but the negative sign with beta coefficients of 0.063 for Q C and 0.014 for Q T represent inverse relationship with the attitude of the consumers toward brand extension.

Table 4: Study Results of the Full Effect Model

###Unstandardized Coefficient###Standardized Coefficient

Variables###B###Std. Error###Beta###t

(Constant)###0.423###0.501###0.844###.399

QUALITY###0.216###0.028###0.230###7.845###.000

COMPLEMENT###0.168###0.021###0.233###7.891###.000

SUBSTITUTE###8.351E-02###0.022###0.118###3.876###.000

TRANSFER###0.226###0.026###0.266###8.745###.000

Q C###-5.06E-02###0.024###-0.063###-2.069###.039

###8.854E-03###0.024###0.011###0.363###.717

###-1.32E-02###0.028###-0.014###-0.465###.642

DIFFICULT###-2.96E-02###0.020###-0.043###-1.460###.145

###.250

The beta coefficients of both Tables 3 and 4 with values of 0.041 and 0.043 respectively show insignificant results and inverse association between evaluation attitude of consumers and the variable, Difficult.

Tables 5 and 6 show the statistical results of past studies as well as the present study for main effect and full effect models respectively. All studies in both the tables support that the quality of the parent brand is significant in the evaluation of brand extension. The present study provides full support for the fit variables in both the models. As is obvious, only a very few studies supported positive relationship with all the fit variables whereas some supported the fit variables only to a certain extent.

Table 5: Main Effect Model Results Summary of the Conducted Studies on Brand Extension (Beta Coefficient)

###Sunde and###Sunde and###NUssen###Bottomley Barrett, Lye and###Bottomley###Present

Variables###Brodie###Brodie###and###and Doyle###Venkateswalu###and Holden###Study

###(Original)###(restated)###Hartman###(1996)###(1999)###(2001)

###(1993)###(1993)###(1994)

QUALITY###0.38###0.25###0.25###0.22###0.37###N/A###0.230

COMPLEMENT###0.29###0.30###0.01###0.31###0.34###N/A###0.233

SUBSTITUTE###0.13###0.19###0.08###0.18###0.19###N/A###0.117

TRANSFER###0.21###0.26###0.58###0.31###0.25###N/A###0.268

DIFFICULT###0.00###0.03###OFR###0.02###0.00###N/A###-0.041

Adjusted R2###0.48###0.43###0.49###0.47###0.47###N/A###0.246

Sample Size###1413###1558###693###1358###2130###N/A###900

In the full effect model, the main effect terms plus interaction terms were regressed on the attitudes of the consumers. All the studies followed the same model for analysing the main effect along with interaction terms. Table 6 for the full effect model shows the same results for the main terms as presented in Table 5 of the main effect model but two out of eight studies, i.e., Aaker and Keller (1990) and their close replication, Sunde and Brodie (1993), show different results for the main variables between the main effect and the full effect models, and the untreated multicollinearity is the reason for different results. The interactions terms as secondary findings have been identified by previous studies but the results are inconsistent.

Table 6: The Full Effect Model Results: Summary of the Conducted Studies on Brand Extension (Beta Coefficient)

###Aaker###Sunde and###Sunde and###Nijssen###Barrett,

###and###Brodie###Brodie###and###Bottomley###Lye###Bottomley

Variables

###Keller###(Orignal) (restated)###Hartman###and Doyle###and Venkates and Holden###Present

###(1990)###(1993)###(1993)###(1994)###(1996)###walu###(2001)###Study

###(1999)###

QUALITY###-0.01###0.67###0.25###0.24###0.22###0.33###N/A###0.23

COMPLEMENT###-0.02###0.24###0.30###-0.00###0.31###0.28###N/A###0.33

SUBSTITUTE###-0.08###0.99###0.18###0.06###0.18###0.15###N/A###0.118

TRANSFER###0.15###0.16###0.26###0.60###0.31###0.21###N/A###0.266

Q C###0.25###0.05###0.05###-0.02###0.05###-0.01###N/A###-0.63

Q 5###0.18###-0.99###-0.01###-0.07###0.03###0.01###N/A###0.011

Q T###0.12###0.07###0.08###0.08###0.08###0.06###N/A###-0.014

DIFFICULT###0.12###0.00###0.03###OFR###0.01###0.00###N/A###-0.043

###0.26###0.48###0.43###0.49###0.48###0.50###N/A###0.250

Sample Size###2140###1413###1558###693###1358###2130###N/A###900

The proposed relationship between the Difficult variable and consumers' attitude towards the extension is only supported by Aaker and Keller (1990) and Bottomley and Holden (2001) while all other studies including the present study do not support this relationship.

Brand Extension Level Analysis

In the brand level analysis, the respondent data of each brand extension was separately regressed. The benefit of this type of analysis over the aggregate level analysis is to run regression model separately for each brand extension to avoid the combined effect of brand extension on each another. In the brand level analysis, both the main and the interaction terms were regressed separately against the attitude of consumers. The results for the main terms at brand extension level analysis are shown in Table 7.

Table 7: Present Study Results of Main Effect Terms at Brand Level Analysis

Brand Extension###R2###QUALITY###COMPLEMENT###SUBSTITUTE###TRANSFER###DIFFICULT

Colgate Tooth

###0.207###0.137###0.197###0.197###0.249###0.066

Brush

Colgate Mouth

###0.128###0.282###0.158###0.017###0.064###0.126

Wash

Lux Lotion###0.205###0.194###0.238###0.092###0.224###0.011

Lux Hand Wash###0.284###0.153###0.305###0.189###0.300###0.015

Nestle Corn Flacks###0.319###0.265###0.233###0.089###0.311###0.037

Nestle Milo###0.407###0.324###0.284###-0.095###0.418###0.052

Results from Table 7 support the Fit variable to some extent. The beta coefficients 6/6 for Complement and 5/6 for Transfer demonstrate that both of the fit variables are significant in the formation of consumers' attitude towards brand extension, while the beta coefficients 2/6 for the substitute show its weak positive relationship with evaluation of the brand extension. The values of some beta coefficients are not significant. The negative beta coefficient shows inverse relationship between substitute and attitude. No significant beta coefficient for the difficult variable is listed in Table 7.

To determine the effect of interaction terms in the first step, the residual terms, such as, Q C (Resid), Q S (Resid) and Q T (Resid) were created by using the Residual Centring Technique of Lance (1988), and then were analysed against consumer attitude through regression. The purpose of this analysis was to determine the effect of interaction terms on the attitude of consumers towards the brand extension. The results of the brand level analysis moderator terms are shown in Table 8 (below).

In Table 8, the beta coefficients 5/6 of Q C (Resid) and 4/6 of Q T (Resid) support the hypothesis of interaction terms, while the beta coefficient 1/6 (negative) shows insignificant result for Q S (Resid). The negative sign is an indicator of inverse relationship of Q S (Resid) with the attitude variable.

Table 8: Present Study: Results of Interaction Terms at Brand Level Analysis

Brand Extension###R2###Q C (Resid)###Q S (Resid)###Q T (Resid)

Colgate Tooth Brush###0.009###0.037###0.077###-0.063

Colgate Mouth Wash###0.095###0.187###0.030###0.169

Lux Lotion###0.199###0.215###0.076###0.252

Lux Hand Wash###0.221###0.194###0.155###0.237

Nestle Corn Flacks###0.291###0.250###0.074###0.317

Nestle Milo###0.335###0.304###-0.188###0.494

Table 9 (below) shows a summary of the results related to the interaction terms at brand level analysis. All of these studies have followed the same statement for interaction terms. Although all the past studies listed show less support for all of the interaction terms and describe insignificant role of interaction terms in brand extension evaluation process, the present study supports the hypothesis with two out of three significant interaction terms. The interaction of quality with complement (Q C) and the interaction of quality with transfer (Q T) are important for the attitude of consumers towards brand extension because only 5 out of 6 beta coefficients for Q C and 4 out of 6 beta coefficients for Q T are respectively significant. The interaction of quality with substitute (Q S) was not found essential in the evaluation of brand extension since only 1 out of 6 beta coefficients is significant.

Table 9: Comparison of Significant Interaction Terms of Present Study with Past Research Studies at Brand Level.

Variables###Sunde and Brodie###Bottomley and Doyle###Barrett, Lye and###Present Study

###(Restated) 1993###(1996)###Venkateswarlu (1999)

Q C (Resid)###3/18###1/18###1/16###5/6

Q S (Resid)###1/18###1/18###1/16###1/6

Q T (Resid)###0/18###0/18###2/16###4/6

Table 10: Comparison of Significant Main effect Terms of Present Study with Past Research Studies at Brand Level

Variables###Sunde and Brodie###Bottomley and Doyle###Barrett, Lye and###Present Study

###(Restated) 1993###(1996)###Venkateswarlu (1999)

QUALITY###13/18###09/18###14/16###5/6

COMPLEMENT###12/18###12/18###11/16###6/6

SUBSTITUTE###12/18###15/18###12/16###2/6

TRANSFER###05/18###07/18###11/16###5/6

DIFFICULT###01/18###01/18###02/16###0/6

It is clear from Table 10 that all the previous studies as well as the present study show a strong support for Quality and Fit variables at brand level analysis; however, the present study shows insignificant result for the Substitute variable. None of the studies provides support for the Diifficult variable at the brand level, similar to the aggregate level.

Discussion and Conclusion

The original study of Aaker and Keller (1990) on brand extension has been replicated in numerous studies for generalization of the determinants of brand extension. All of these studies including the present study provide strong support for Quality. For Fit variables some studies have full support, while other studies support two out of three Fit variables. The full effect model of the original study of Aaker and Keller (1990) did not find support for Quality, while only partial support was found for the Fit variables with only Transfer as significant. The hypothesis for the interaction term states that transfer of quality is enhanced when there is Fit between the original brand and the extension category. The result of the present study for the aggregate level analysis found no support for the interaction terms whereas previous studies provide conflicting results.

Bottomley and Doyle (1996) concluded from their results that consumers' attitude towards brand extension was driven primarily by the main effects and moderated by the interaction effects. Some of the studies have actually used brand level analysis instead of aggregate level analysis. At brand level analysis, previous studies did not find support for interaction term while the present study found favourable support for interaction terms except Q S. The results of brand level analysis of all studies provide support for Quality and Fit Variables but show lack of agreement for the Difficult variable.

The exclusion of the Difficult variable by Nijssen and Hartman (1994) is an indication of an unrelated matter. In previous studies, only Aaker and Keller (1990) and Bottomley and Holden (2001) found a support for Difficult.

Previous research and the present study on brand extension provide evidence to drive us towards the conclusion that Quality and Fit variables play a significant role in the judgment of brand extensions, though the original study of Aaker and Keller (1990) on brand extension gives contrary results. They do not support the Quality variable and provide nominal support for the Fit Variables. It is concluded that this lack of agreement for Quality and Fit is due to the untreated multicollinearity since the interaction terms were analysed as secondary effects. Even the present study at the aggregate level analysis does not find support for the interaction terms whereas at the brand level analysis provide only partial support for these terms with only the substitute as insignificant. It is concluded that unsupportive result for these terms at the aggregate level analysis may be the reason of the combined effect of brands on each other.

The role of Difficult in the evaluation of brand extension remains an issue. Some studies state that it does not influence the consumer attitude towards extension category, while only two studies found it to be a significant factor. The supportive studies of Difficult are in minority against unsupportive studies. The omission of this factor from Nijssen and Hartman (1990) suggests that it is an unrelated factor and on the basis of the results obtained from the present study, it is also concluded that Difficult has no influence on consumer attitude towards band extension.

Direction for Further Research

The purpose of this study was to add knowledge to the field of marketing. Previous studies stated that the difference in results between the Aaker and Keller (1990) and their replication by Sunde and Brodie (1993) may be due to such factors as dissimilarity in brand, sample nature and cultural difference. Majority of the studies that were conducted in different cultures provide similar results for the main terms. The present study suggests that the nature of the sample and culture differences are uncontrollable factors but the brand should be changed according to availability and knowledge of consumers. The possibility of using the same brands is nominal because consumers have knowledge only about the brands available in the local market. Evidence from previous research indicates that Difficult is not an influencing factor. It is suggested that in future studies, other factors such as the design of the packaging should be analysed instead of Difficult.

References

Aaker, D.A. and Keller, K.L. (1990). Consumer Evaluation of Brand Extension. Journal of Marketing, 54(1), 27-41.

Barrette, J., Lye, A., and Venkateswarlu, P. (1999). Consumer Perceptions of Brand Extension: Generalizing Aaker and Keller's Model. Journal of EmpiricalGeneralizations in Marketing Science, 4, 1-21.

Bottomley, P.A. and Doyle, J.R. (1996) The Formation of Attitudes towards Brand Extension: Testing and Generalizing Aaker and Keller's Model. InternationalJournal of Research in Marketing, 13, (4) 365-377.

Bottomley, P.A. and Holden, S.J.S. (2001). Do We Really Know How Consumers Evaluate Brand Extension? Empirical Generalizations Based on Secondary Analysis of Eight Studies. Journal of Marketing Research, 38(4), 494-500.

Boush, D.M. and Loken, B. (1991). A Process Tracing Study of Brand Extension Evaluation, Journal of Marketing Research, 28 (1), 16-28.

Chen, K.F. and Lue, C.M. (2004). Positive Brand Extension Trial and Choice of Parent Brand, Journal of Product and Brand Management, 13(1), 25-36.

Holden, S.J.S and Barwise, P. (1995). An empirical investigation of what it means to generalize. Proceedings of the European Marketing Academy, Cergy-Pontoise, France.

Jahangir, N., Parvez, N., Bhattacharjee, D. and Ahmed, K.K.B. (2009). The Relationship Between Brand Affect, Brand Quality, and Customers' Brand Extension Attitude: Exploring the Mediating Role of Customer Loyalty. The Cambodian Management Journal, 1(1), 20-34

Kirmani, A., Sood, S. and Bridges, S. (1999). The Ownership Effect in Consumer Responses to Brand Line Stretches. Journal of Marketing, 63(1), 88-101.

Lance, C.E. (1998). Residual Centering, Exploratory and Confirmatory Moderator Analysis, and Decomposition of Effects in Path Models Containing Interactions. Applied Psychology Measurement, 12(2), 163-175.

Montoya-Weiss, M.M. and Calantone, R. (1994). Determinants of New Product Performance: A Review and Meta-Analysis. Journal of Product Innovation Management, 11 (5), 397-417.

Nijssen, E.J. and Hartman, D. (1994). Consumer Evaluations of Brand Extensions: An Integration of Previous Research. European Marketing Academy Conference, 673-683

Ourusoff, A., Ozanian, M., Brown, P.B. and Starr, J. (1992). What's in a Name? What the World's Top Brands are Worth. Financial World, (September), 32-49.

Sunde, L and Brodie, R.J. (1993). Consumer Evaluations of Brand Extensions: Further Empirical Results. International Journal of Research in Marketing, 10(1), 47-53.

a Institute of Management Sciences, Peshawar, Pakistan, b IBMS, Agricultural University, Peshawar, Pakistan
COPYRIGHT 2010 Asianet-Pakistan
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2010 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Publication:The Journal of Humanities and Social Sciences
Article Type:Report
Geographic Code:9PAKI
Date:Dec 31, 2010
Words:4968
Previous Article:A corpus-based study of Urdu borrowings in Pakistani English Newspapers.
Next Article:A Study of Naguib Mahfouz's Midaq Alley (Zuqaqq-al-Midaqq) and the Beginning and the End (Bidaya-wa-Nihaya).
Topics:

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters