Consumer Perceptions of Value: a Multi-Item Scale For Its Measurement

ABSTRACT - The paper presents a framework for assessing consumer perceptions of value for electronic product class. The purpose was to develop a most effective and generalizable scale that would allow the researcher to use the same measure across various research studies. Items were developed considering the multidimensional approach to the measurement of the construct. The scale met standards for internal scale reliability, validity by changing wording, reliability over time, criterion-related validity, and convergent validity. Finally, validity was assessed whether the scale is an appropriate operational definition of the construct. The paper concluded with a discussion of the limitations of the study.

Citation:

Humayun Kabir Chowdhury and Shuzo Abe (2002) ,"Consumer Perceptions of Value: a Multi-Item Scale For Its Measurement", in AP - Asia Pacific Advances in Consumer Research Volume 5, eds. Ramizwick and Tu Ping, Valdosta, GA : Association for Consumer Research, Pages: 160-168.

Asia Pacific Advances in Consumer Research Volume 5, 2002      Pages 160-168

CONSUMER PERCEPTIONS OF VALUE: A MULTI-ITEM SCALE FOR ITS MEASUREMENT

Humayun Kabir Chowdhury, Rajshai University, Bangladesh

Shuzo Abe, Yokohama National University, Japan

ABSTRACT -

The paper presents a framework for assessing consumer perceptions of value for electronic product class. The purpose was to develop a most effective and generalizable scale that would allow the researcher to use the same measure across various research studies. Items were developed considering the multidimensional approach to the measurement of the construct. The scale met standards for internal scale reliability, validity by changing wording, reliability over time, criterion-related validity, and convergent validity. Finally, validity was assessed whether the scale is an appropriate operational definition of the construct. The paper concluded with a discussion of the limitations of the study.

The development of effective marketing strategies requires an understanding of the process in which consumers perceive value among the available alternatives. Researchers of consumer behavior have developed a number of theories in the attempt to explain and predict consumers’ perceptions of value (Bearden and Shimp 1982; Dodds and Monroe 1985; Dodds, Monroe, and Grewal 1991; Teas and Agarwal 2000). These theories suggest that consumers’ perceptions of value are generally formed on the basis of "an array of cues" (Cox 1962). The consumer’s task in evaluating any given product is to use cues from this array for making evaluative judgments about that product. These cues mainly can fall into two categories: intrinsic cues or extrinsic cues. Intrinsic cues refer to attributes that cannot be changed without changing the physical characteristics of the product. Extrinsic cues are attributes that are not part of the physical product (Olson and Jacoby 1972; Szylillo and Jacoby 1974). It is still unknown about which of the intrinsic attributes and/or extrinsic cues are chosen by consumers in forming their value in a deal, and why some cues are chosen while others are not. However, it is assumed that when a consumer confronts with a new product, s/he would consider salient attributes of the product and would form an opinion. In this process, the same attribute might have different effects depending on the environment in which it competes. The contrast effect assumes that the effect applies not only to single attributes but also to the tradeoffs between attributes. In addition to this evaluation process, extrinsic cues might influence consumers’ product evaluations in a more global manner by serving as an evaluative context of liking or disliking for the new product. The information processing theory suggests that decision-makers have limitations on their capacity for processing information. It means that consumers cannot always articulate value between or among the alternatives. Such limitations include limited working memory and limited computational capabilities (Olson and Jacoby 1972). The consumer, for example, may lack the motivation to sort out the information that might lead to an objective determination of value in making the decision, may not have the ability to process it, or the information may simply be unavailable. The salience or accessibility of the brand associations depends on their strength in memory, as well as the retrieval cues provided (Keller and Aaker 1992; Lynch and Srull 1982).

In consumer behavior literature, taking altogether, it is argued that the value of a good is not only inherently related with the attributes in the good but also the psychological outcome a person or people have for it (Zeithaml 1988).

THE PROBLEM

To measure value in practice, it is difficult to have a shared understanding of exactly what is value in consumer judgment. While a number of investigators have alluded to perceived value, few have theoretically or empirically examined the nature and measurement procedure of the construct. Until recently, little formal conceptual effort has been directed toward isolating theoretical reasons for the relationships between or among intrinsic quality, extrinsic quality and how such a relationship influences consumers’ perceptions of value judgment (Monroe and Krishnan 1985). Thus, it appears worthwhile at this point to determine measures that may be used to assess the process of consumer value judgement. Since the introduction of a model containing consumer evaluation of price, perceived quality, and perceived value by Dodds and Monroe (1985), much research has been carried out in consumer perceived value studies. In the findings of their models they have shown that consumers willingness to buy is affected by perceived value, and perceived value is affected by perceived quality and perceived sacrifice. In a relatively different research setting, Dodds, Monroe, and Grewal (1991) has extended consumers’ value assessment by adding brand name and store name, and demonstrated the effects of these cues on perceptions of quality, value, and consumers’ willingness to buy. They found persuasive effect of some cues on perception of value in the presence of some cues than when alone.

Furthermore, most of the previous research employed a unidimensional single-item measure of perceived value. In different articles unidimensional single-item scale is criticized for their low reliability and inability to capture the latent constructs like perceived value (Churchill 1979). Recent developments strongly suggest a multidimensional approach to the measurement of perceived value (Hunt 1979; Nunnally 1978; Zeithaml 1988). However, some studies used multi-item scales to tap the construct but have not been tested for their stability or validity (Aoki 1994; Dodds, Monroe, and Grewal 1991; Teas and Agarwal 2000). Therefore, we do not know if the discrepancy is due to the methodological problems or to different behaviors. Hence, a reliable, valid and multidimensional measure of perceived value should be useful.

BACKGROUND AND CRITERIA FOR MEASURING PERCEIVED VALUE

Zeithaml (1988) mentioned, "What constitutes value, even in a single product category, appears to be highly personal and idiosyncratic (pp. 13)". Based on the responses from an exploratory study she grouped into four consumer definitions of value. These are: (a) value is low price, (b) value is whatever I want in a product, (c) value is the quality I get for the price I pay, and (d) value is what I get for what I give. These grouping were based on an investigation of consumer value assessment in the product category of beverages. In doing so, Zeithaml noted that the first group of respondents indicated that what they had to give up was most salient in their perceptions of value. The second group of respondents emphasized the benefits they received from the product as the most important components of value. The third group of respondents perceived value as a tradeoff between one "give" component, price, and one "get" component, quality. The final group of respondents conceptualized all relevant "get" components as well as relevant "give" components when describing value. Thus, the same way, this study utilizes Zeithaml’s (1988) definition for the development of perceived value measures. These definitions are summarized with specific examples in Table 1. These 4 definitions served as the basic structure of the perceived value domain from which items were derived.

TABLE 1

CONSUMER DEFINITIONS OF VALUE BY ZEITHAML (1988)

Some scholars may argue that the four definitions are not quite independent to each other. For example, the first dimension of price is included in the dimensions of quality & price and get (want) & give (price). The second dimension of want is included in the dimensions of quality (want) & price and get (want) & give as well. Moreover, the third dimension of quality and price is included in the fourth dimension of get and give. Henceforth, the first and second dimensions of price and want are more specific and less multidimensional and the third and fourth dimensions are more global and multidimensional. Apparently, the concept of perceived value should be composed of two key factors of price and want, instead of four dimensions. Conceptually, this argument makes sense, as there are not many differences between or among the dimensions at the operational level. A significant portion of their domains overlaps. Because of these overlaps, these dimensions seem to have somewhat similar meanings. However, Zeithaml (1988) mentioned, "the diversity in meanings of value depending on the definitions provides a partial explanation for the difficulty in conceptualizing and measuring the value construct in research" (p. 13). A careful evaluation reveals that the third dimension of "give" and "get" can be considered in terms of "price" and "quality" respectively, but the fourth dimension of "get" and "give" are more global where something more than quality is included in the "get" components and non-monetary sacrifice is included in the "give" components. Therefore, to see statistically whether it is reasonable to consider the four definitions provided by Zeithaml (1988) as different factors is one of the important purposes of this study.

Abe (1987) defined construct validity and deliberatively explained its measurement procedures. In explaining the problems of measurement procedures in the international consumer research settings, Abe (1993) mentioned the importance of concept development, concept operationalization, measure equivalence, and data analysis. Despite the explanation is provided in international research settings, certainly the methodological issues explained in the studies can be used in different consumer research topics. Thus, measures of the study were developed considering the theoretical explanations provided by Abe (1987; 1993), and Abe, Bagozzi, and Sadarangani (1996) and were validated in accordance with recommended procedures suggested by Churchill (1979). The steps taken to develop the measure were: Step 1. Conceptualizing the construct; Step 2. Item generation and content validity; Step 3. Internal scale reliability; Step 4. Validity by changing wording; Step 5. Test-Retest reliability; Step 6. Criterion-related validity and convergent validity; Step 7. Empirical analysis of the facets.

CONCEPTUALIZING THE CONSTRUCT

As a hypothetical construct, perceived value cannot be measured directly. Although many previous studies dealt with perceived value, scholars in different disciplines viewed the construct from different quarters due to its highly subjective nature. For instance, Kotler (1998) defined value as, "the consumer’s estimate of the product’s overall capacity to satisfy his or her needs." Anderson et al. (1993) described value from the point of view of business markets. They defined value as, "the worth in monetary terms of the technical, economic, service, and social benefits a customer company receives in exchange for the price it pays for a market offering."

Researchers recently tend to imply a distinction between types of perceived value. For example, Grewal, Monroe and Krishnan (1998) empirically distinguished perceived value between acquisition value and transaction value. They, depending on previous definitions, conceptualized perceived acquisition value as, "a more global and enduring kind of value which takes into account both price and quality". They argued that the perceived acquisition value of the product will be positively influenced by the benefits buyers believe they are getting by acquiring and using the product and negatively influenced by the money given up to acquire the product. Furthermore, consumers examining the financial terms of the price offer might perceive additional value by comparing the selling price to their internal reference prices. Thus, perceived transaction value is defined as, "the perception of psychological satisfaction or pleasure obtained from taking advantage of the financial terms of the price deal" (pp. 48).

As can be seen from the above definitions that abstraction of perceived value have varied from author to author. This melange of definitions has confused the construct of perceived value, and at this time neither its precise nature nor its determinants have been clearly delineated. Although there are a wide variety of arguments in defining the construct, it is equivocally agreed that what a consumer tries to achieve from a purchase is nothing but the value (Zeithaml 1988). It is reasonable to predict that a highly valued product perceived by a consumer would perform better in order to satisfy more needs, or the same needs more effectively, than would a product of lower valued.

Thus, it is critical to characterize a consumer’s value perceptions for a particular task when trying to ascertain why his or her choice processes take a certain form. It may be concluded from the previous definitions that monetary factor in maximizing the accuracy of the deal, cognitive effort required for the justification of the deal, minimizing the experience of negative emotion for te dealBare the determinants of consumer perceived value. This article will adopt the general view of perceived value that focuses on consumer perception of an object. Based on the previous literature, value in this study defined from a consumer’s viewpoint for the purposes of scale development is: consumers’ perceived fulfillment to the satisfaction of their requirements through acquisition and use of the product that results from the tradeoff between the product’s perceived benefits and perceived monetary and nonmonetary sacrifice. This definition recognized past definitions of perceived value, i.e., perceived value is the cognitive tradeoff between perceptions of quality and sacrifice (Dodds, Monroe and Grewal 1991). Additionally, this definition presumes that the cognitive tradeoff among perceptions of quality, perceived monetary and nonmonetary sacrifice results in perceptions of value.

Some scholars may argue that the four dimensions are not quite independent to each other. For example, the first dimension of price is included in the dimensions of quality & price and get (want) & give (price). The second dimension of want is included in the dimensions of quality (want) & price and get (want) & give as well. Moreover, the third dimension of quality and price is included in the fourth dimension of get and give. Henceforth, the first and second dimensions of price and want are more specific and less multidimensional and the third and fourth dimensions are more global and multidimensional. Apparently, the concept of perceived value should be composed of two key factors of price and want, instead of four dimensions. Conceptually, this argument makes sense, as there are not many differences between or among the dimensions at the operational level. A significant portion of their domains overlaps. Because of these overlaps, these dimensions seem to have somewhat similar meanings. However, Zeithaml (1988) mentioned, "the diversity in meanings of value depending on the definitions provides a partial explanation for the difficulty in conceptualizing and measuring the value construct in research" (p. 13). A careful evaluation reveals that the third dimension of "give" and "get" can be considered in terms of "price" and "quality" respectively, but the fourth dimension of "get" and "give" are more global where something more than quality is included in the "get" components and non-monetary sacrifice is included in the "give" components. Therefore, to see statistically whether it is reasonable to consider the four definitions provided by Zeithaml (1988) as different factors is one of the important purposes of this study.

ITEM GENERATION AND CONTENT VALIDITY

A semantic differential scale was to be developed based on the definition mentioned previously. Items representing various facets of the 4 perceived value dimensions were generated to form the initial items for perceived value instrument. In doing so, a list of 48 pair of words was formed to represent the concept of perceived value. The first step was to judge how well the chosen items represent the defined concept. Thus, the main emphasis here was to delete the poor pair of words that do not represent respondents’ perceptions of value. Two Ph.D. candidates were taught the definition of perceived value and were requested to judge the representativeness of the 48 pair of words replacing the last part (a product at the lowest possible cost and sacrifice) of the definition. Name of the product class was replaced in the definition with "a product". For example, "computer" that the respondent is now owning was changed in the place of "a product" in case of computer product class. Orally, each pair of words was rated on either (1) clearly representative of my perception of value, or (2) not representative of my perception of value.

15 out of 48 initially developed pair of words was dropped for their unrepresentativeness of perception of value. Another two juges from the same faculty were again requested to judge the remaining 33 pair of words for different product classes using the same procedure. The reason for this judgment was to determine if the validity of the scale would stand the changes of product classes. For this purpose, only different electronic product classes were considered and their generalizability was judged. Only 21 items were consistently rated as representing the perceived value construct (Computer, TV, Video Camera were considered).

Therefore, a twenty-one item scales emerged from the content validity phase although it seems to be small in a number of items with which to start data collection. Thus, further validation procedures were carried out on these item scales.

INTERNAL SCALE RELIABILITY

Five and six items on each dimension and altogether 21 items existed after content validity stage. Here, the task was to measure the internal consistency of 21 items as a scale. Computer was chosen as a test product as the results indicated that a computer was a personally relevant product for students. The test was performed with a total 63 students from the business faculty of a national university in Japan. A total of 52 usable questionnaires were obtained. 11 Questionnaires which had problems of partially incomplete answers were dropped before data analysis.

The first step to purify the items began with the analysis of Cronbach coefficient alpha, in accordance with Churchill’s (1979) recommendation. Because the perceived value measure was developed based on four dimensions, reliability was assessed for each dimension. Resulting values of coefficient alpha ranged between .6375 and .9133 across the four dimensions and suggested that deletion of certain items with low item-to-total correlations would improve the alpha values. At this point, two items, Q6 and Q9, showing low item-to-total correlations were deleted. Factor analysis, using varimax rotation with squared multiple correlations in the diagnonals for factor extraction, were carried out to check if the items selected for deletion loaded onto one particular dimension or were amorphous across factors. The results of the factor analysis showed that the items selected for deletion did not load together on any unique factor. Coefficient alpha was again assessed for the remaining items (see Table 2).

The high alpha values indicated good internal consistency among items within each dimension (Cronbach 1951). Furthermore, the combined reliability for the 19-item scale was quite high (.9221). Therefore, the 19-item scale was considered to be ready for further testing with new samples.

VALIDITY BY CHANGING WORDING

The 19-items resulted in the internal scale reliability were subject to a second data collection with two changes. First, 6 items were negatively worded from the originally positive statements, in accordance with recommended procedures for scale development (Churchill 1979). And, second, some wording changes to eliminate ambiguities were made to some items. The reason for the negative transformation of items was to determine if the validity and reliability of the scale would stand the change of wording (Churchill and Peter 1984).

TABLE 2

RELIABILITY TEST RESULTS OF PERCEIVED VALUE

Data were gathered from a convenience sample of 147 student respondents in two business classes of a national university in Japan with a prior permission of the Professors. Three products, Computer, TV, and Camera, were used instead of one for the purpose of generalizability. The questionnaire took about seven minutes to answer. A total of 132 usable questionnaires were obtained. First, reliability of each dimension was checked. Values of Cronbach coefficient alpha were .8512, .6776, .7061, and, .7678 respectively. These values are lower than those obtained in the internal scale reliability stage. The combined reliability for the 19-item scale was .8378. There were no significant differences of Cronbach coefficient alpha values over the three products. However, the reason for the poor value is because of transforming positive wording into negative for some items. Low item-to-total correlations were observed in those negatively transformed items. Q2, Q10, Q13, Q15, Q18 and Q19 had a noticeable low item-to-total correlation both in cases of individual dimension and in combined procedure as well. It is worth noting that the item-to-total value for Q10 was low but all the others were quite high in the internal scale reliability stage. Thus, we do not know the exact cause of the inconsistencies of whether these are due to the transformation or changes of wording. Therefore, it deemed unwise to delete those items at this stage and the total 19-items were subjected to further data collection and refinement.

TEST-RETEST RELIABILITY

Coefficient alpha does not adequately estimate, though, errors caused by factors external to the instrument, such as differences in testing situations and respondents over time (Churchill 1979). It is also advised to collect additional data to rule out the possibility that the previous findings are due to chance (Zaichkowsky 1985). Thus, test-retest reliability of the 19-items was performed over the same subjects, with the same products, and in the same procedures as was done in the previous section. About 28 days later of collecting data for section 'validity by changing wording’, the scales were administered again for retesting. Two new items, one of which measured respondents’ overall value perceptions and the other measured respondents’ intention to buy the same brand if s/he were going to buy the product again, were added at the end of the questionnaire. The purpose of these two additional items was to measure the criterion-related validity of the scales (discussed in the next section). A total of 103 usable data were obtained at this data collection stage.

Items Q2, Q10, Q13, Q15, Q18 and Q19 were dropped at this stage based on two criteria; (a) significantly low average test-retest correlation values and (b) low item-to-total correlations. The average Pearson correlation value on the 13 items Time 1 and Time 2 together with the Cronbach coefficient alpha value for each product are shown in Table 3. A reasonably high test-retest correlations and alpha values indicated good consistency among items between time lags. Moreover, the reliability of linear combinations of the remaining 16 items was also quite high as can be seen in the table.

Therefore, although the thirteen-items can be presumed to grasp perceived value adequately but still further analysis is deemed necessary to ensure high level of reliability.

CRITERION-RELATED VALIDITY AND CONVERGENT VALIDITY

According to Churchill (1979),"A fundamental principle in science is that any particular construct or trait should be measurable by at least two, and preferably more, different methods. Otherwise the researcher has no way of knowing whether the trait is anything but an artifact of the measurement procedure (pp. 70)." Criterion-related validity investigates the empirical relationship between the scores on a test instrument (predictor) and an objective outcome (the criterion). The most commonly used measure of criterion-related validity is a validity coefficient, which is the correlation between predictor and criterion scores.

TABLE 3

TEST-RETEST RESULTS OF PERCEIVED VALUE

TABLE 4

MEAN VALUE BASED ON THE ITEM >OVERALL VALUE PERCEPTION=

As noted in the previous section that two new items were constructed one of which measured overall perceived value and the other willingness to buy served as criterion variables. The two items were used differently to ceck criterion-related validity by correlating average scores of all the perceived value measure items with scores for the items assessing overall value and willingness to buy (Parasuraman et al. 1988). Correlation between average scores of all the items and perceived overall value was .87 and correlation between average scores of all the items and willingness to buy was .73. A reasonably high value of the correlation between average scores of all the items and perceived overall value indicates the possession of criterion-related validity of the scale items.

Finally, a one-way ANOVA was performed with ratings of the overall value measure as the independent variable and the average values of the 13-items as the dependent variable. The purpose was to check convergent validity of the final scales as advocated by Parasuraman, Zeithaml, and Berry (1988). Duncan’s Multiple Range Test revealed significant differences between groups. Each mean was different from those of the others (see Table 4). This confirms the convergent validity of the items was distinguished between different levels of perceived value.

EMPIRICAL ANALYSIS OF THE FACETS

Construct validity measures whether a scale is an appropriate operational definition of an abstract variable, or a construct. Factor analysis can be useful in establishing construct validity. Conducting a factor analysis on a single summated scale will show whether all items within the summated scale load on the same construct, or whether the summated scale actually measures more than one construct. The researcher should specify both the number of dimensions in the construct and the specific items or scales which are hypothesized to load on those dimensions a priori.

Factor analysis was performed on these 13 items. A principal components factor analysis using a varimax rotation technique generated 4 dimensions with clear factor patterns. All the items for the same dimension loaded high on the respective factor and low on other factors (see Table 5). As can be seen in Table 5 the loadings on item 8 and 17 are not very salient compared to other items, and thus, an iterative sequence deleting these two items was performed. Factoring did not remain distinct in each iterative analysis dropping these two items. Therefore, the final iteration was accepted that yielded the 13 question items given in the appendix.

Confirmatory Factor Analysis

In fact, the results of any single analysis are always less than perfectly dependable. The problem is especially pernicious because the results of a single factor analysis usually look plausible. But plausibility is no guarantee of validity or even stability (Wells and Sheth 1971). A confirmatory factor analysis can be conducted when a researcher has theory and guidance regarding the expected factor structure (Stewart 1981). A confirmatory factor analysis specifies the expected number of factors that should suffice in describing the data. In confirmatory factor analysis, researchers would have theoretical reasons or past empirical evidence to believe they could predict the number of factors, the pattern of which variables should load, on which factors, and the actual values of those loading (Stewart 2001). Since the factor patterns have been achieved by performing Principal Component Analysis, a Confirmatory Factor Analysis using Amos 4.0 was performed to determine the construct validity (see Figure 1). The main premise of performing Confirmatory Factor Analysis is that if the 13-items included in the instrument measure the four distinct dimensions identified in the previous sections, then the survey data should produce results that conform to the model.

TABLE 5

PRINCIPAL COMPONENT FACTOR ANALYSIS RESULTS OF THE FINAL 13 PERCEIVED VALUE MEASURE ITEMS

FIGURE 1

STANDARDIZED ESTIMATES OF THE CONFIRMATORY FACTOR ANALYSIS

Results of the Analysis

The Chi-square value of 73.1 (df=59; p=0.13) shows reasonably good fit of the model. Supplemental results of absolute fit (GFI=0.949; AGFI=0.922; RMR=0.047) and indices of comparative fit (NFI=0.964; IFI=0.995; CFI=0.995) also support the fit of the data to the model. The squared multiple correlations can be interpreted based on the variance explained with regard to the specific items. Bagozzi and Yi (1988) suggest that variance extracted should be greater than or equal to .50. A reasonable percentage of the variance of the items respectively are accounted for by the variance in the common factors. The remaining percentage of the variance cannot be explained by this model and are attributed to the unique factors. In this model, factor loadings of all the items to explain the latent variables are quite satisfactory.

SUMMARY AND LIMITATIONS

In fact, perceived value is an elusive construct on that a comprehensive understanding in a particular customer setting may appear monumentally difficult. The purpose of this study was to develop a most effective and generalizable scale that would allow the researcher to use the same measure across various research studies of consumer value judgment. If valid, the instrument could be used for a variety of purposes such as tracking consumers’ perceptions of value of a product offered by a manufacturer or measuring consumers’ perceptions of the differences in value among competing products. This information would be useful for designing marketing strategies. It can also help in pinpointing areas requiring managerial attention and action to improve consumer perceptions of value. Depending on the definition developed by Zeithaml (1988) based on an exploratory study, items were generated and content validity was performed by expert judges at the first phase of the scale development. Internal scale reliability and validity by changing wording were checked over two different samples. Test retest reliability was performed to check the stability of the scale over time. Criterion-related validity of the scale were checked to investigate the empirical relationship between the scores on the predictor and the criterion and convergent validity to see the extent to which the score converged with other methods designed to measure the same construct. Finally, validity was assessed whether the scale is an appropriate operational definition of the construct.

The main limitation of this study is the sample size used to measure the internal scale reliability. 52 is very low a number of subjects for testing reliability although the number was increased in the following reliability and validity test stages. The second limitation of this study lies with the initial items. Only 21-items existed after the item generation and content validity stage and served as the base for testing different types of validity. Indeed, this is too low a number with which to start data collection. After the judgment by the experts and depending on the content validity it seemed unreasonable in this study to consider more than 21-items. Furthermore, considering the easiness to use the instrument, we refrained from creating new items. However, despite the small number of items per dimension, the Cronbach alpha values are proved to be satisfactory. The third limitation of this study is with Churchill’s (1994) suggestion that we may feel quite comfortable with the items included in a measure, while a critic may argue that we have failed to sample from some relevant domain of the characteristic. In fact, content validity cannot be determined statistically, it can only be determined by experts. It is a judgment, by experts, of the extent to which a scale truly measures the concept that it intended to measure, based on the content of the items. For this study, Ph.D. students were selected to determine the content validity, although students to fulfill the requirements in pinpointing the domains of the content might be questioned of being insufficient. However, due to the lack of empirical research, one of the purposes of this study was to see the construct empirically that could lead to evolving knowledge and a sophisticated understanding of the content. Henceforth, the content validity of the construct might be improved over time by further theory building and theory verification by experts. The forth limitation lies on the student sample that has been used in this study. There are many arguments in favor and against the convenience samples containing students. Several authors have enumerated the dangers of using student samples in research (Beltramini 1983; Oakes 1972). These authors have generally cited threats to external validity as their primary concern, arguing that students are atypical of the "general population", and that any findings based on student samples may therefore not be generalizable to other populations. However, some scholars disagree on this issue. Oakes (1972) contends that such arguments are specious because, regardless of what population is sampled, generalization can be made only with caution to other populations. In fact, a student sample, with its homogeneous characteristics, is often advocated because its use can increase internal validity (Calder, Philips, and Tybout 1981) and statistical conclusion validity (Judd and Kenny 1981) through a reduction in error variance. The end result is that statistical conclusion validity is improved, or conversely, the probability of a Type II error is reduced. This situation is particularly desirable when researchers are engaged in theory testing, or are testing specific theoretical predictions. All that is required is that the sample be chosen to allow a test of the theoretical predictions under consideration. Because the primary focus of this study was a theory test and not effects generalization, considerations of internal validity were paramount and a student sample was appropriate (Calder, Philips, and Tybout 1982; Cook and Campbell 1975). Concerns about external validity were secondary. Finally, the study cannot satisfy the convenience criterion that the items would not make sense for every product besides electronic product class. The findings are based on a limited set of brands and hence generalization beyond that set should be made with caution. However, in order to advance the theory as well as to develop reliable and valid measures of perceived value further, considerable empirical work is surely needed.

APPENDIX

FINAL 13 ITEMS OBTAINED IN THE PERCEIVED VALUE SCALE AFTER PURIFICATION

REFERENCES

Abe, Shuzo (1987), "Construct Validity and LISREL," in Okuda Kazuhiro and Abe Shuzo (Ed). Marketing Theory and Measurement (pp. 27-46), Chuo Keizai Sha, 27-46 (in Japanese).

Abe, Shuzo (1993), "Methodological Problems in Cross-cultural Consumer Research," Yokohama Keiei Kenkyu. 13, 31-44 (in Japanese).

Abe, Shuzo, Bagozzi, Richard P., & Sadarangani P. (1996), "An Investigation of Construct Validity and Generalizability of the Self Concept: Self-Consciousness in Japan and the United States," Journal of International Consumer Marketing, 8, 97-123.

Anderson, J. C., Jain, D. C., & Chintagunta, P. K. (1993), "Customer Value Assessment in Business Markets: A State-of-Practice Study," Journal of Business-to-Business Marketing, 1, 3-28.

Anderson, J. C. & Narus, J. A. (1998), "Business Marketing: Understand What Customers Value," Harvard Business Review, November-December, 53-65.

Aoki, Michiyo (1994), "Evaluation of Green Products-The Process of Recognizing Information and Forming Qualitative Judgments," In Asia Pacific Advances in Consumer Research, 1, pp. 36-42

Bagozzi, Richard P. (1881), "Attitudes, Intentions, and Behavior: A Test of Some Key Hypotheses," Journal of Personality and Social Psychology, 41, 607-627.

Bearden, W. O. & Shmp, T. A. (1982), "The Use of Extrinsic Cues to Facilitate Product Adoption," Journal of Marketing Research, 19, 229-239.

Beltramini, R. F. (1983), "Student Surrogates in Consumer Research," Journal of the Academy of Marketing Science, 11 (Fall), 438-443

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

Calder, J. Bobby, Lynn W. Philips and Alice M. Tybout (1982), "The Concept of External Validity," Journal of Consumer Research, 9 (December), 240-243

Chattopadhyay, A. and Basu, K. (1990), "Humor in Advertising: The Moderating Role of Prior Brand Evaluation," Journal of Marketing Research, 27, 466-76.

Churchill, G. A. Jr. (1979), "A Paradigm for Developing Better Measures of Marketing Constructs," Journal of Marketing Research, 16, 64-73.

Churchill, G. A. Jr. (1994), Marketing Research: Methodological Foundations, (6^th ed.). The Dryden Press, Harcourt Brace College Publishers.

Churchill, G. A. Jr. and J. Paul Peter (1984), "Research Design Effects on the Reliability of Rating Scales: A Meta Analysis," Journal of Marketing Research, 21 (November), 360-375.

Cook, T. and D. Campbell (1975), "The Design and Conduct of Experiments and Quasi-Experiments in Field Settings," in Handbook of Industrial and Organizational Research, Martin Dunette, ed., Chicago, Rand McNally

Cox, D. F. (1962). "The Measurement of Information Value: A Study in Consumer Decision-Making" (pp. 413-421). In W.S. Decker (Ed.), Emerging Concepts in Marketing., Chicago: American Marketing Association.

Cronbach, L. J. (1951), "Coefficient Alpha and the Internal Structure of Tests," Psychometrika, 16, 297-334.

Dodds, W. B. & Monroe, K. B. (1985), "The Effect of Brand and Price Information on Subjective Product Evaluation," In Elizabeth Hirschman and Morris Holbrook, (eds.), Advances in Consumer Research (Vol. 12, pp. 85-90). UT: Association for Consumer Research.

Dodds, W. B. & Monroe, K. B. & Grewal, D. (1991), "Effects of Price, Brand, and Store Information on Buyers’ Product Evaluation," Journal of Marketing Research, 28, 307-319.

Grewal, D., Monroe, K. B., & Krishnan, R. (1998), "The Effects of Price-Comparison Advertising on Buyers’ Perceptions of Acquisition Value, Transaction Value, and Behavioral Intentions," Journal of Marketing, 62, 46-59.

Hill, T. (1994), Manufacturing Strategy, Burr Ridge, ILL: Irwin

Houston, M. J. & Rothchild, M. L. (1978), "Conceptual and Methodological Perspective on Involvement," In S. J. Jain (Ed.), AMA 1978 Educators’ Proceedings (pp. 184-187), Chicago: American Marketing Association

Hunt, Keith (1979). Conceptualization and Measurement of Consumer Satisfaction and Dissatisfaction, Cambridge, Mass: Marketing Science Institute

Judd, Charles M. and D. A. Kenny (1981), Estimating the Effects of Social Interventions, Cambridge, MA: Cambridge University Press

Keller, K. L. & Aaker, D. A. (1992), "The Effects of Sequential Introduction of Brand Extensions," Journal of Marketing Research, 29, 35-50.

Kotler, P. (1998), Marketing Management: Analysis, Planning, Implementation, and Control (9^th ed.), New Jersey 07458, Prentice-Hall.

Lee, D. (1999), "Effects of Partitioned Country of Origin Information on Buyer Assessment of Binational Products," In Advances in Consumer Research, 26, 344-51

Lynch, J. G. Jr., & Srull, T. K. (1982), "Memory and Attentional Factors in Consumer Choice: Concepts and Research Methods," Journal of Consumer Research, 9, 18-36.

Min, H. C., & Terstra, V. (1988), "Country-of-Origin Effects for Uni-National and i-National Products," Journal of International Business Studies, 16, 235-56.

Monroe, K. B. & Krishnan, R. (1985), "The Effect of Price on Subjective Product Evaluations," In J. Jacoby and J. C. Olson (eds.), Perceived Quality- How Consumers View Stores and Merchandize (pp. 209-232), MA: Lexington Books.

Nunnally, J. C. (1978), Psychometric Theory (2^nd ed.), New York: McGraw-Hill.

Oakes, W. (1972), "External Validity and the Use of Real People as Subjects," Americal Psychologist, 27, 959-962

Olson, J. C. & Jacoby, J. (1972), "Cue Utilization in the Quality Perception Process," in M. Venkatesan (ed.), Advances in Consumer Research, 3, Iowa: Association for Consumer Research, 77-91.

Parasuraman, A., Zeithaml, V. A., & Berry, L. L. (1988), "SERVQUAL: A Multiple-Item Scale for Measuring Consumer Perceptions of Service Quality," Journal of Retailing, 64, 12-40.

Payne, J. W., Bettman, J. R. & Johnson, E. J. (1993), The Adaptive Decision Maker. Cambridge University Press

Richardson, P. S., Dick, A. S. & Jain, A. K. (1994), "Extrinsic and Intrinsic Cue Effects on Perceptions of Store Brand Quality," Journal of Marketing, 58, 28-36.

Shimp, T. A. & Bearden, W. O. (1982), "Warranty and Other Extrinsic Cue Effects on Consumers’ Risk Perceptions," Journal of Consumer Research, 9, 38-46.

Stewart, David W. (1981), "Application and Misapplication of Factor Analysis in Marketing Research," Journal of Marketing Research, 18, 51B62

Stewart, David W. (2001), "Exploratory versus Confirmatory Factor Analysis," Journal of Consumer Psychology, 10 (1 and 2), 75- 82

Stokes, R. C. (1985), "The Effects of Price, Package Design, and Brand Familiarity on Perceived Quality," in J. Jacoby and J. C. Olson (eds.), Perceived Quality- How Consumers View Stores and Merchandize (pp. 233-246), MA: Lexington Books.

Szylillo, G. J. & Jacoby, J. (1974), "Intrinsic versus Extrinsic Cues as Determinants of Perceived Product Quality," Journal of Applied Psychology, 59, 74-78.

Teas, R. K. and Agarwal, S. (2000), "The Effects of Extrinsic Product Cues on Consumers’ Perceptions of Quality, Sacrifice, and Value," Journal of the Academy of Marketing Science, 28, 278-290.

Wells, William D. and Jagdish N. Sheth (1971), "Factor Analysis in Marketing Research," in Multivariate Analysis in Marketing: Theory and Application, David A. Aaker, eds., Wodsworth Publishing Company, Inc. Belmont, California 94002.

Wheatley, J. J. and Chiu, J. S. (1977), "The Effects of Price, Store Image, and Product and Respondent Characteristics on Perceptions of Quality," Journal of Marketing Research, 14, 181-186.

Zaichkowsky, J. L. (1985), "Measuring the Involvement Construct," Journal of Consumer Research, 12, 341-352.

Zeithaml V. A. (1988), "Consumer Perceptions of Price, Quality, and Value: A Means-End Model and Synthesis of Evidence," Journal of Marketing, 52, 2-22.

----------------------------------------