Measuring Values’ Importance: the Use of Different Scales to Measure the Lov

Aviv Shoham, University of Haifa, Israel
Moshe Davidow, University of Haifa, Israel
Maja Makovec Brencic, University of Ljubljana, Slovenia
ABSTRACT - The relationship between values and consumer behavior has been a central theme in many studies in recent years. Kahle (1983) sees values as guidelines for desirable and enduring end-states, which affect attitudes and behaviors. Values are stable, abstract cognitions; they provide people with structure, organization, and purpose for interactions with their environments (Kahle 1996). Furthermore, values provide purpose and motivation for one’s actions. Numerous studies have documented the link between values and attitudes and behaviors. For example, they influence leadership, sports participation, gift giving, and role relaxation tendencies, among others. Previous operationalizations of the List of Values (LOV) have resulted in highly skewed distributions, limiting their potential explanatory power. Consequently, further development is required to provide more balanced measures of the LOV This paper reports the results of two Israeli studies, designed to test two additional approaches to LOV measurement. In all cases, age and gender’s relationships with the LOV, commonly used to predict the importance of values, were examined for each of the three operationalizations. Additionally, we tested the relationships between the LOV and role relaxation tendencies, an outcome construct used in previous research. In summary, we find that one of the two new operationalizations of LOV outperforms the standard measurement approach, while the second approach did not. Our findings and their implications are discussed and areas of future research are indicated.
[ to cite ]:
Aviv Shoham, Moshe Davidow, and Maja Makovec Brencic (2003) ,"Measuring Values’ Importance: the Use of Different Scales to Measure the Lov", in E - European Advances in Consumer Research Volume 6, eds. Darach Turley and Stephen Brown, Provo, UT : Association for Consumer Research, Pages: 154-161.

European Advances in Consumer Research Volume 6, 2003      Pages 154-161

MEASURING VALUES’ IMPORTANCE: THE USE OF DIFFERENT SCALES TO MEASURE THE LOV

Aviv Shoham, University of Haifa, Israel

Moshe Davidow, University of Haifa, Israel

Maja Makovec Brencic, University of Ljubljana, Slovenia

ABSTRACT -

The relationship between values and consumer behavior has been a central theme in many studies in recent years. Kahle (1983) sees values as guidelines for desirable and enduring end-states, which affect attitudes and behaviors. Values are stable, abstract cognitions; they provide people with structure, organization, and purpose for interactions with their environments (Kahle 1996). Furthermore, values provide purpose and motivation for one’s actions. Numerous studies have documented the link between values and attitudes and behaviors. For example, they influence leadership, sports participation, gift giving, and role relaxation tendencies, among others. Previous operationalizations of the List of Values (LOV) have resulted in highly skewed distributions, limiting their potential explanatory power. Consequently, further development is required to provide more balanced measures of the LOV This paper reports the results of two Israeli studies, designed to test two additional approaches to LOV measurement. In all cases, age and gender’s relationships with the LOV, commonly used to predict the importance of values, were examined for each of the three operationalizations. Additionally, we tested the relationships between the LOV and role relaxation tendencies, an outcome construct used in previous research. In summary, we find that one of the two new operationalizations of LOV outperforms the standard measurement approach, while the second approach did not. Our findings and their implications are discussed and areas of future research are indicated.

INTRODUCTION

Personal values have been studied for many years. Four conceptualizations of values have dominated the literature: the List of Values (LOV; Kahle 1983), the Rokeach Value Survey (RVS; Rokeach 1968, 1973), Values and Life Style (VALS), and Activities, Interests, and Opinions (AIO; Plummer 1974). The LOV seems to have been the most popular in marketing research. It was developed as an alternative to Rokeach’s value inventory because the terminal values identified by Rokeach were too abstract and difficult to apply to marketing situations (Assael 1998). It includes nine values with more direct marketing applications: self respect, security, warm relationship with others, sense of accomplishment, self fulfillment, sense of belonging, being well respected, fun and enjoyment in life, and excitement. Respondents are asked to rate the importance of these values in their daily life. The scale is introduced by: "The following is a list of things that some people look for or want out of life. Sometimes you find that you have to give up a little of something important because something else is most important to you. Please rate the following on how important it is to you in your daily life." In such research, single values have been studied in some cases, whereas in other cases, they have been combined into dimensions (Keng and Yang 1993; Razzaque 1995; Tai and Tam 1996). While values and demographics influence consumption (McCarty and Shrum 1993), most previous work viewed demographics as values’ antecedents (Kahle 1996; Keng and Yang 1993; Lascu, Manrai, and Manrai 1996; Razzaque 1995).

Recently, studies have utilized the LOV to examine values from a global perspective and have explored inter- and intra-national differences. Kahle (1996) compared value distributions in seven countries. Rose and Shoham (2000) compared value differences of mothers in the US and Japan. Kropp, Jones, Rose, Shoham, Florenthal, and Cho (2000) compared value importance ratings in Australia, the US, Canada, and Israel.

Unfortunately, the LOV suffer from a methodological drawbackCthe distribution of responses has tended to be highly skewed. In other words, mean importance ratings for the nine values in the LOV tended to average between 6.0 and 8.0 on 9-point scales. For example, mean importance ratings were 6.0-7.9, 6.0-7.6, 5.9-7.6, and 6.3-8.0 for Australians, Americans, Canadians, and Israelis, respectively (Kropp et. al. 2000). Consequently, the potential of predictors of values’ importance (e.g., demographics) to explain the low variance in importance is low. Conversely, the potential explanatory role of values as predictors of numerous outcomes is reduced. This problem of variance restriction has been noted in other research areas in marketing, such as satisfaction (Peterson and Wilson 1992) and service quality (Parasuraman, Berry, and Zeithaml 1993).

Our two studies were designed to address this shortcoming in previous research. Specifically, one study examined the commonly used approach to measure the importance of values, using a scale of 1 to 9; as well as a second approach, using the same nine items, but introduced by a different introductory question and utilizing a scale from 0 to 10. Both approaches were used within one sample. Respondents provided item ratings twice, each time in a different context. In the second study, we used two smaller sub-samples. One provided importance ratings in the standard approach, whereas the second provided importance ratings on differently anchored scales (details provided in a later section).

The studies were conducted in Israel, a Western-oriented country, which has undergone many changes over the last decade. First, the local economy has shifted from a more centrally planned economy to a more competitive one. Second, there has been an ongoing peace process with neighboring countries and the Palestinians (though not in the last two years). Both have made Israel a desirable market for global firmsCMcDonald’s, Office Depot, Toys-R-Us and Burger King exemplify recent entrants to Israel. Thus, research on values in Israel is important as a basis around which multinational firms can design and improve their marketing strategies, including positioning, segmentation, and advertising. Better measures of values’ importance can help global firms to identify differences between Israeli consumers and consumers in other markets. This would lead to more accurate adjustments in global marketing strategies to Israeli consumers. Additionally, accurate measures should allow for a better value-based segmentation of the Israeli population.

In sum, our study contributes to values’ research in two ways. Substantively, it adds to the limited body of research on antecedents and consequences of values in Israel, utilizing three different measurement approaches to advance our knowledge of the cross cultural impact of LOV. Methodologically, we test these three approaches in an effort to identify which of the three provides the most psychometrically sound results, thus advancing the general measurement of the LOV construct.

THEORY

Personal Values

The importance of personal values has long been recognized. As early as 1979, Williams theorized that values affect consumers’ judgments, preferences, and choices. We define a value as "a centrally held, enduring belief which guides actions and judgments across specific situations and beyond immediate goals to more ultimate end-states of existence" (Rokeach 1968, p. 161). As consumers internalize values, "they become unconscious and conscious criteria in guiding behavior" (Lascu et al. 1996, p. 148).

Values’ impact on behavior has been established in many studies (for a review, see Kahle, Rose, and Shoham 2000). Becker and Conner (1981) reported that values affect consumers’ tendency to be exposed to mass media. Specifically, Beatty et al. (1985) linked values with television programs’ preferences. In a different domain, Rose et al. (1994) documented that values influence buying preferences for fashion products (see also Shoham and Florenthal 2001). Given their impact on consumer behavior, values have been popular as bases for segmentation, positioning, and promotion (Lascu et al. 1996).

Cumulatively, previous studies have suggested a causal pattern from Values to Attitudes to Behavior (Homer & Kahle 1988). We do not argue that global values alone can predict behavior; rather, similar to Kahle (1996, p. 136), we view values as "anchors or cognitive sources." In short, values are the basis upon which attitudes are built, leading to behavioral and consumption outcomes (Kahle 1996).

Measuring Values

Previous literature used four values’ operationalizations. Rokeach (1973) developed the Rokeach Value Survey (RVS) with 18 instrumental and 18 terminal values. Instrumental values determine desirable behavior; terminal values define desirable end-states. In practice, the administration of RVS requires respondents to rank all values’ pairs, a daunting task for respondents. Substantively, some RVS values, such as "a world at peace" are not directly relevant to the study of consumer behavior (Kahle 1996). These shortcomings explain why the RVS has not been used often in consumer behavior research.

AIO (Activities, Interests, and Opinions) was designed to operationalize consumers’ lifestyles. It concentrates on activitiesChow people spend time and money; interestsCwhat they see as important in their surrounding environment; and opinionsCwhat they feel about themselves and their world (Tai and Tam 1996). AIO suffers from several weaknesses. First, its scope is narrow; it excludes some values and subsequent attitudes. Additionally, since the AIO includes many statements, it is difficult to administer (Tai and Tam 1996). Moreover, no standard instrument exists to operationalize the AIOCusers select AIO statements from an inventory of about 300 statements, thus exacerbating the methodological issue.

VALS (Values and Life Style) has been used as an alternative to RVS. Theoretically, it is based on Maslow’s (1954) hierarchy of needs and on social character. Operationally, it uses 34 demographic and attitudinal measures to classify respondents into one of nine lifestyle categories, which provide a distinct profile of people in each category (Kahle 1996; Kahle, Beatty, and Homer 1986). VALS and LOV (discussed below) evolved from Maslow’s theory. Thus, the two approaches overlap. For example, an "achiever" in VALS would probably assign a high level of importance to LOV’s "sense of accomplishment".

Kahle developed a fourth operationalization, the LOV (Beatty et al. 1985; Kahle 1983). Based on Maslow’s needs’ hierarchy (1954), the LOV incorporates elements of social adaptation theory (Kahle 1983). Kahle saw values (1996, p. 135) as the "most abstract type of social cognition that people use to store and guide general responses to classes of stimuli." The LOV includes nine items: self respect, security, warm relationship with others, sense of accomplishment, self fulfillment, sense of belonging, being well respected, fun and enjoyment in life, and excitement, operationalized by Likert scales. As noted, distributions of responses to the LOV tend to be skewed (Kahle 1996), which makes the determination of the values’ hierarchy difficult (Razzaque 1995). However, the LOV has several advantages over VALS. First, it enables a separation of the impact of demographics and values on behavior (Kahle et al. 1986). In contrast, demographics are part of the measurement method in VALS (Novak and MacEvoy 1990). Second, administration of the LOV is easier (Kahle et al. 1986). Third, whereas VALS is culturally biased toward the USA, the LOV is not (Kahle 1996), making the LOV less sensitive to cross-cultural variations. Kahle et al. (1986) and Novak and MacEvoy (1990) compared the two methods and found the LOV to be superior to VALS.

Therefore, we chose to use the LOV in our study for three reasons. First, the LOV is standardized and can be compared cross-culturally. Second, it has demonstrated validity and reliability in cross-cultural studies (Beatty et al. 1985; Grunert, Grunert, and Beatty 1989; Kahle 1996). Third, the LOV is short, making it easy to administer (Razzaque 1995).

Values and Demographics

Values and demographics are useful in predicting consumer behavior; therefore, both are useful segmentation bases (Kahle 1996; Lascu et al. 1996; McCarty and Shrum 1993). Since values and demographics are instrumental in developing marketing strategies, both should be studied simultaneously. A number of studies have examined the relationships between age, gender, and values. Lascu et al. (1996) reported that values’ importance varies with age. Similarly, Keng and Yang (1993) showed that age and gender affected values of Taiwanese consumers. For example, consumers over 30 years old valued security more, whereas consumers between the ages 19 and 29 valued self-respect and being well respected more. Similar findings have been reported in Bangladesh (Razzaque 1995). Vale differences were observed across numerous demographics (e.g., age and gender). To illustrate, highly educated, older male managers with a high income, valued sense of accomplishment, self-fulfillment, and self respect more than other respondents did. Finally, Polish consumers’ age was correlated significantly with instrumental values (Lascu et al. 1996). Accordingly:

H1: Values’ importance differs for males and females.

H2: Values’ importance differs by age.

Values and Role-Relaxation Tendencies

The term role-relaxation was coined by Chris Riley, a director of Wieden & Kennedy Advertising (Kahle 1995). Role-relaxed consumers decide how to act and what to buy while remaining intentionally oblivious to social demands (Kahle and Shoham 1995). It is closely related to consumer susceptibility to interpersonal influence (CSII)Crole-relaxed consumers should be less susceptible to such influence. First introduced by Kahle (1995), role-relaxation is a relatively new concept. Kahle and Shoham (1995) examined the relationship between role-relaxation, values, and consumption in the US. Role-relaxed consumers, who are affluent and self-confident, tend to emphasize functional over social product attributes (Kahle 1995).

Kahle (1995) argued that the more role-relaxed the consumer, the higher the assigned importance of self-respect compared to that of being well respected because role-relaxed consumers are less dependent on others for approval or guidance compared to low role-relaxed consumers. Kahle and Shoham (1995) reported negative relationships between the importance of four social values (being well respected, beauty, competitiveness, and excitement) and role-relaxation tendencies.

Thus, the three social LOV values (being well respected, warm relationship with others, and sense of belonging) should explain role-relaxed behavior and CSII. The more important these values are to individuals, the less role-relaxed and the more CSII they should be. Conversely, the three internal values (self respect, sense of accomplishment, and self fulfillment) should be related positively with role-relaxation and negatively with CSII. In sum:

H3: The higher the importance of social valuesCbeing well respected, warm relationship with others, and sense of belongingCthe lower the role-relaxation and the higher the CSII. The higher the importance of internal valuesCself respect, sense of accomplishment, and self fulfillmentCthe higher the role-relaxation and the lower the CSII.

We note here that how values’ importance is measured should affect the strength (but not the direction) of the relationships discussed in the three research hypotheses. We explore this issue in subsequent sections of the paper, which describe our two studies.

METHODOLOGYCSTUDIES 1 AND 2

Sample CharacteristicsCStudy 1

Data for the first study were collected from a convenience sample of 82 students in an Israeli university. Given the methodological goals of the study, as well as their roles as consumers, students were seen as acceptable respondents. All 82 agreed to participate and provided complete questionnaires with the original and a revised version of the LOV (discussed below).

There were more male (59.8%) than female respondents (40.2%) in the sample. This was not surprising given the engineering orientation of the university, and closely matched the overall gender distribution of the university. The average age of the respondents was 25.1 (s. d.=2.1), with the range being 21-31.

Sample CharacteristicsCStudy 2

Data for the second study were collected from a convenience sample of 80 Israeli consumers, recruited in community and shopping centers. All agreed to participate and provided complete questionnaires. About half (36) responded to the original LOV version while the second half (44) responded to a different version of the LOV (discussed below).

The sample included 40 males and 40 females. Average age of the respondents was 31.2 (s. d.=11.2), ranging between 19 and 63. Education-wise, 30% completed at least some high school and 70% completed at least some college education. Since the two forms of the questionnaires were distributed randomly, we did not expect the sub-groups to differ on age and education. However, we tested for possible cross-group differences. Neither education-based chi-square (ceducation=3.96; 2 degrees of freedom; p>0.10) nor age-based F (F=0.01; 1 degree of freedom; p>0.10) were significant, thus suggesting sub-group equivalency.

Measures for the Two Studies

The role-relaxation scale was designed in English (Kahle and Shoham 1995). In developing an equivalent Hebrew version, we used the back-translation method. One bilingual individual translated the scale to Hebrew. Another individual, blind to the original, back-translated the scale. Then, a third bilingual individual compared the original and back-translated English versions. Adjustments were made by consultation of the three individuals. The same procedure was used for the consumer susceptibility to interpersonal influence (CSII) scale (Bearden, Netemeyer, and Teel 1989). In an effort to enhance the response rate we did not include the CSII scale in this study.

The LOV was translated using a similar procedure. The Hebrew version has been used in a number of Israeli studies (e.g., Shoham, Florenthal, Kropp, and Rose 1997). We used the original Hebrew version for the actual LOV items in both studies. However, in each of the three versions of the LOV, either the introductory passage and anchor words or the scale end-points differed. In study 1, we used the standard version of the LOV, introduced by: "People tell us that different things are important to them in their daily lives. We’d like to know how important each of the items listed below is to you in your daily life." A scale of 1 (important to me) and 9 (extremely important to me) anchored each value. In a later section of the questionnaire, we introduced a different version of the LOV, utilizing the same items, but introduced by: "People tell us that different things are important to them in their daily lives. We’d like to know how important each of the items listed below is to you in your daily life. Unlike the first question on this issue, please rate how often you think about each of the following." Thus, the second version captures a "share-of-thoughts" facet of the importance of values as opposed to the original version. Additionally, we attempted to increase variance by using values of 0 (almost never) and 10 (constantly) to anchor each item value.

In Study 2, about half of the consumers responded to the original version of the LOV, anchored by 1 (important to me) to 9 (extremely important to me). The other half responded to the same introductory passage but rated the importance of each value on scales ranging from B5 (important to me) to +5 (extremely important to me). In Study 1 the scales were changed slightly thus altering the mindset of the respondent and the frame of reference for the scales, thus allowing us to utilize the entire sample to test both versions. In Study 2, the frame of reference remains constant, with only the actual scaling changed, thus necessitating the use of a split sample to test both versions.

Tables 1-2 include descriptive statistics for the scales and LOV terms in Studies 1 and 2. Notably, the CSII scale includes two sub-dimensions (normative and informational influences), which are listed separately in both Tables. The Tables also include reliability coefficients (a) for multi-item scales and skew statistics for the LOV.

RESULTS AND ANALYSISCSTUDY 1

Skew Statistics

As can be seen in Table 1, the share-of-thoughts-based LOV had better skew statistics. Whereas five of the six LOV skew statistics in the traditional LOV exceeded the cut-off rate of " 1, suggesting departure from normality, only one exceeded this rate in the new version. This provides an initial confirmation for the superiority of the revised approach over the standard approach. Below, we compare the nomological validity of the two approaches by examining the impact of demographic variables on values’ importance and the impact of values’ importance on role relaxation and CSII.

Demographics and the LOV

To identify significant differences in value importance across age and gender, we used 12 ANOVA models. In each, the standardized LOV measure served as the dependent variable and age and gender as independent variable (Table 3). Two models in the standard approach (for sense of belonging and warm relationships with others) and one in the share-of-thoughts approach (for warm relationships with others) had significant main effects. Whereas females valued sense of belonging more than males did (under the standard approach), older consumers valued warm relationships with others more than younger consumers (in both approaches). These findings provide modest support for H1-2 and for the superiority of the traditional LOV over the share-of-thoughts LOV based on the main effects only. However, this superiority disappears when we take into account the interaction between age and gender. When including the interaction between age and gender in the models (not reported to conserve space), the results for sense of belonging and warm relationships with others were significant (p<0.05) under both approaches. The substantive results remained the sameColder respondents and females rated these values as more important than younger respondents and males did. Additionally, the impact of age was stronger for females than for males.

In sum, the data in Study 1 provide modest support to H1-2. They leave the issue of nomological superiority of the LOV as an outcome of demographics unresolved.

TABLE 1

DESCRIPTIVE STATISTICS FOR STUDY 1: LOV, ROLE-RELAXATION, AND CSII

TABLE 2

DESCRIPTIVE STATISTICS FOR STUDY 2: LOV AND ROLE-RELAXATION

LOV, Role-Relaxation, and CSII

We tested H3 by examining the correlation coefficients between the importance of the six values and between role-relaxation and CSII. Given the directional hypothesis, we used one-tailed tests. Under the traditional approach to measuring the LOV, all correlation coefficients were lower than |0.17|. None of the 18 correlation coefficients (between the six values, role relaxation, and the two components of CSII) was significant; only one was marginally significant (sense of belonging and informational CSII). In contrast, under the share-of-thoughts approach, two coefficients (normative CSII and being well respectedC-0.21Cand informational CSII and warm relationships with othersC0.35) were significant and two were marginally significant (informational CSII with self-fulfillmentC-0.14Cand role-relaxation with self respectC-0.14).

Given the coding used, these findings provide some support to H3. In line with H3, people who value being well respected tend to be more normatively susceptible to interpersonal influence. People who value warm relationships with others and self-fulfillment tend to be more information-based susceptible to interpersonal influence. The former is as hypothesized; the latter is contrary to expectations. Finally, supporting H3, those who value self-respect tend to be more role-relaxed than those who value it less.

In sum, the data support H3. They suggest that the share-of-thought approach to measure the LOV is operationally and nomologically superior to the traditional LOV.

TABLE 3

ANOVA MODELS FOR STUDY 1

TABLE 4

ANOVA MODELS FOR STUDY 2

RESULTS AND ANALYSISCSTUDY 2

Skew Statistics

The traditional LOV had better skew statistics than the "5-anchored LOV (Table 2). Whereas two LOV skew statistics in the traditional LOV exceeded the cut-off rate of "1, suggesting departure from normality, five of the six exceeded it in the "5-anchored LOV. This provides an initial confirmation for the superiority of the traditional approach over the " 5-anchored LOV approach. Below, we compare the nomological validity of the two approaches by examining the impact of demographics on values’ importance, as well as the impact of values’ importance on role-relaxation.

Demographics and the LOV

We used 12 ANOVA models to identify significant differences in values’ importance across age and gender. The standardized LOV measures served as the dependent variables and age and gender served as the independent variables (Table 4).

The main effects of two models in the standard approach (for self-fulfillment and self-respect) were significant (p<0.05). Females and older consumers valued self-fulfillment and self-respect more than males and younger consumers did. None of the main effects in the six models utilizing the "5-anchored LOV approach were significant. These findings provide further support for H1-2. They also provide support for the superiority of the traditional LOV over the "5-anchored LOV approach. When we tested 12 models with the interaction of age by gender, there was no difference in the conclusions and this advantage remained substantively identical.

In sum, the data in Study 2 provide further support to H1-2. They suggest that the traditional LOV approach might be nomologically superior to the "5-anchored approach in terms of values as an outcome of demographics.

LOV and Role-Relaxation

We tested H3 by examining the correlation coefficients between the importance of the six values in each approach and between role-relaxation. Under both approaches to measuring the LOV, not one of the twelve correlation coefficients (between the six valuesCmeasured twice in the two sub-samplesCand the role relaxation scale) was significant.

In sum, the data in Study 2 provide no support to H3. Neither the +/- 5-anchored LOV nor the traditional LOV appear to be nomologically superior as predictors of role-relaxation.

DISCUSSION

Substantive Findings

Our results about higher values’ importance for females parallel the results reported in Taiwan (Keng and Yang 1993) and the US (Kahle 1996). Sense of belonging, warm relationships with others, self-fulfillment, and self-respect were more important to Israeli females than to Israeli males. As noted by Florenthal, Treister, and Shoham (1999), warm relationships with others might well be the global female value, the most important value to women all over the world. However, there also appears to be a gender bias in value importanceCfemales tend to assign a higher level of importance to all nine values in the LOV compared to males.

With respect to the impact of age, our studies reveal that as consumers mature, they assign a higher level of importance to values under all operationalizations. Specifically, we find that older consumers in Israel see sense of belonging, warm relationships with others, self-fulfillment, and self-respect as more important than younger consumers do.

We examined the impact of values on two outcome constructs (role-relaxation and a two-dimensional CSII). Recall that people who value being well respected tend to be more normatively susceptible to interpersonal influence. Additionally, those who value self-respect tend to be more role-relaxed than those who value it lessCin line with H3. Moreover, as hypothesized, people who value warm relationships with others tend to be more information-based susceptible to interpersonal influence. These findings suggest that as social/external values gain importance, they become manifest in and have an impact on the social end of the continuum for both role-relaxation and susceptibility to interpersonal influence. These findings substantiate Kahle (1995), who theorized that role-relaxed and low CSII consumers are less dependent on others for approval or guidance compared to low role-relaxed and high CSII consumers. This line of reasoning implies that social values, which, by definition, depend on significant others to be realized, would result in low levels of role-relaxation and high values of SCII, as was found here.

In contrast, we find that, contrary to expectations, high importance for self-fulfillment tends to be associated with high information-based susceptibility to interpersonal influence. In explaining this finding, we note that it only holds for the informational, but not the normative component of CSII. It might be that the information provided by others is necessary actualization of self-fulfillment. In other words, to the extent that self-fulfillment is based, in part, on purchase and consumption of certain products and services, high self-fulfillment importance might force consumers to seek information from knowledgeable others. This natural reliance on word of mouth would help explain the unexpected result.

Methodological Findings

Why did the relationships materialize for only a subset of the values in the LOV both as dependent and independent variables in our models? Notably, Florenthal, Treister, and Shoham (1999) and Kahle and Shoham (1995) have observed similar patterns in their Israeli and US studies, respectively. In these previous studies, as in ours, only some of the values in the LOV were significantly predicted by demographics and significantly affected role-relaxation. Below, we discuss these findings in the context of our methodological aims.

Recall that we measured the LOV using three different operationalizations. In study 1, we used the standard version of the LOV, measured on nine-point scales (1=important to me to 9=extremely important to me). We used a different introductory paragraph and different scale anchors, which were designed to capture a "share-of-thoughts" facet of the importance of values. We attempted to space out responses by using 11-point scales (0=almost never to 10=constantly) for each value, instead of the original 9-point scale. In Study 2, half responded to the standard version of the LOV, identical to that used in Study 1. The other half responded to the same introductory passage but rated the importance of each of the nine values on scales ranging from B5 (important to me) to +5 (extremely important to me).

These new approaches were designed to provide better measures for the importance of values compared to the traditional approach. Notably, using the traditional introductory paragraph with different anchors (-5 to +5) resulted in a less satisfactory set of distributions for the nine values. Skew statistics were higher than for the traditional LOV. Furthermore, values in this new measurement approach were not as good nomologicallyCfewer values were significantly predicted by age and gender compared to the traditional approach. Moreover, fewer values were significant predictors of role-relaxation and CSII relative to the traditional approach. One possible explanation for the ineffectiveness of the new scale may be the difference between the word anchors and the scale numbers. While it is possible to rate importance, from important to very important, on a scale from 1 to 9 (original scale), there might be a perceptual problem rating important as a B5, since the negative number does not imply importance. Thus, this new approach should not be used in future values’ research.

In contrast, the "share-of-thoughts" introductory paragraph, followed by the traditional, 9-point scales, appears to outperform the "daily life" introductory paragraph. Its skew statistics were lower. In fact, only one of the six distributions for items of the LOV departed significantly from normality, a marked improvement over the traditional LOV. Additionally, for the most part, this new approach resulted in stronger demographics-based predictions for values and in values being a stronger determinant of role-relaxation. Thus, the "share-of-thoughts" approach shows promise and deserves consideration in future research.

Yet, even with this improvement, not all values were predicted by demographics and not all values affected role relaxation and CSII. This was the basis for the question posed in the first paragraph of this section of the paper. We now discuss this by examining the "share-of-thoughts" findings, which, as noted, were the most promising.

As seen in Table 1, the means for the values under the "share-of-thoughts" approach were similar or slightly lower than for the traditional approach. When tested for mean differences, only one value (sense of belonging) was significantly different (and lower). Thus, the new approach did not change the means. However, the standard deviations for the six values under the new approach were consistently higher than under the traditional approach. The new approach resulted in more spaced out responses, but the differences are still modest. Thus, not all hypothesized relationships materialized because the new approach, while a step in the right direction, still does not go far enough. Additional research is needed to improve and refine the "share-of-thoughts" approach to arrive at more balanced LOV distributions.

MANAGERIAL IMPLICATIONS

We begin by examining the Israel-specific implications of our findings. First, older (versus younger) Israeli consumers place more importance on sense of belonging, warm relationships with others, self-fulfillment, and self respect. The marketing mix should reflect these emphases when targeting the older Israeli consumers. For example, advertising to older consumers should emphasize elements of group activity, because warm relationships with others and sense of belonging are more important to such consumers. More generally, advertising targeted to different age groups should account for these differences as well.

Second, females (versus males) place more importance on sense of belonging, warm relationships with others, self-respect, and self-fulfillment. These should be reflected in the marketing mix. When designing an image for feminine products, marketers should stress these values. Even for family products, managers should take into account the values that are important to females. Numerous studies have examined family decision-making (e.g., Davis and Rigaux 1974). Specifically, Shoham, Rose, Botstein, and Pinzi (1996) studied family decision-making in Israel and reported tat most decisions, for most products and decision phases, were female- or equal-dominant. Thus, since the four values are more important for females then for males, they can be used for targeting and positioning. The value warm relationships with others, which we refer to as "the global" female value, is probably the most useful in designing promotion and advertising for female-dominant products.

Combining these insights with the reported impact of social values on role-relaxation and CSII suggests that both outcomes can be stressed in message design. Executions depicting group settings may be especially effective for advertisements that emphasize social/external values such as being well respected.

Regarding cross-cultural implications, marketers should consider differences in value importance in the design of strategies and the decision whether to standardize or adapt them. Importantly, warm relationships with other are more important to females than to males regardless of culture. It was found to be very important to females in Israel, Taiwan, and the US. Thus, standardizing advertising messages to females by emphasizing warm relationships with others appears to be robust to the cultural context of the campaign.

We set two goals for this study, namely to add to the limited body of research on antecedents and consequences of values in Israel and to improve on the psychometric properties of the LOV. Both of these goals have been attained within the limits of this study.

Limitations and Directions for Future Research

This study has several limitations. Respondents came from convenience samples. Additionally, the samples did not account for Israeli sub-cultures. For example, Arabs were not fully represented in Studies 1 and 2, although they are an important segment in Israel, accounting for about 10% of the population. Consequently, future research is needed with more representative samples of Israel’s diverse sub-cultures.

Moreover, future research should compare different sub-cultures in Israel to identify differences and similarities in value preferences. These should be studied in the context of differential effects on consumption behavior. Another interesting direction for future research may involve a comparison of values’ differences between Israeli consumers and consumers in neighboring Arab countries. Such research should be valuable to global firms as the Middle-East becomes more attractive as a market because of increasing standards of living and the reduction in political instability. Further research is also needed to improve on the psychometric properties of the LOV scale. While this study showed one possible solution, additional efforts are needed.

REFERENCES

Assael, Henry (1998), "Consumer Behavior and Marketing Action," Cincinnati, Ohio: South-Western College Publishing.

Bearden, William O., Richard, D. Netemeyer, and Jesse E. Teel (1990), "Measurement of Consumer Susceptibility to Interpersonal Influence," Journal of Consumer Research, 15(March), 473-81

Beatty, Sharon E., Lynn R. Kahle, Pamela M. Homer, and S. Misra (1985), "Alternative Measurement Approaches to Consumer Values: The List of Values and the Rokeach Value Survey," Psychology and Marketing, 3, 181-200.

Becker, Boris W. and Patrick E. Conner (1981), "Personal Values of Heavy Users of Mass Media," Journal of Advertising Research, 21(October), 37-43.

Davis, Harry L. and Benny P. Rigaux (1974), "Perceptions of Marital Roles in Decision Processes," Journal of Consumer Research, 1(June), 51-62.

Florenthal, Bella, Miriam Treister, and Aviv Shoham (1999), "The Influence of Social Values and Demographics on Role-Relaxed Behavior," in Proceedings of the Society for Consumer Psychology, Madhu Viswanathan, Larry Compeau, and Manoj Hastak (eds.), St. Petersburg, FL: Society for Consumer Psychology, 162-5.

Grunert, K., S. Grunert, and S. Beatty (1989), "Cross-Cultural Research in Consumer Values," Marketing and Research Today, 17(1), 30-9.

Homer, Pamela M. and Lynn R. Kahle (1988), "A Structural Equation Test of the Value-Attitude-Behavior Hierarchy," Journal of Personality and Social Psychology, 54(4), 638-46.

Kahle, Lynn R. (1983), Social Values and Social Change: Adaptation to Life in America, NY: Praeger.

Kahle, Lynn R. (1995), "Role Relaxed Consumers: A Trend for the Nineties," Journal of Advertising Research, 35(2), 66-71.

Kahle, Lynn R. (1996), "Social Values and Consumer Behavior: Research from the List of Values," In Clive Seligman, James M. Olson, and Mark P. Zanna (Eds.), The Psychology of Values: The Ontario Symposium, 8, Hillsdale, NJ: Lawrence Erlbaum, 135-51.

Kahle, Lynn R., Sharon E. Beatty, and Pamela M. Homer (1986), "Alternative Measurement Approaches to Consumer Values: The List of Values (LOV) and the Values and Life Style (VALS)," Journal of Consumer Research, 13(4), 405-9.

Kahle, Lynn R., Gregory M. Rose, and Aviv Shoham (2000), "Findings of LOV throughout the World and Other Evidence of Cross-National Consumer Psychographics: Introduction," in Cross-National Consumer Psychographics, Lynn Kahle (ed.), NY: IBP, 1-14.

Kahle, Lynn R. and Aviv Shoham (1995), "Role Relaxed Consumers: Empirical Evidence," Journal of Advertising Research, 35(3), 59-62.

Keng, Kau Ah and Charles Yang (1993), "Personal Values, Demographics and Consumption Behavior: A Study of Taiwanese Consumers," Journal of International Consumer Marketing, 6(1), 27-48.

Kropp, Fredric, Marilyn Jones, Gregory M. Rose, Aviv Shoham, Bella Florenthal, and Bongjin

Cho (2000), "Group Identities: A Cross-Cultural Comparison of Values and Group Influences," in Cross-National Consumer Psychographics, Lynn Kahle (ed.), NY: IBP, 117-31.

Lascu, Dana-Nicoleta, Lalita A. Manrai, and Ajay K. Manrai (1996), "Value Differences between Polish and Romanian Consumers: A Caution Against Using a Regiocentric Marketing Orientation in Eastern Europe," Journal of International Consumer Marketing, 8, 145-67.

Maslow, Abraham H. (1954), Motivation and Personality, New York, Harper.

McCarty, John A. and L. J. Shrum (1993), "The Role of Personal Values and Demographics in Predicting Television Viewing Behavior: Implications for Theory and Application," Journal of Advertising, 22(4), 77-101.

Novak, Thomas P. and Bruce MacEvoy (1990), "On Comparing Alternative Segmentation Schemes: The List of Values (LOV) and Values and Life Styles (VALS)," Journal of Consumer Research, 17(June), 105-9.

Parasuraman, A., Leonard L. Berry, and Valarie A. Zeithaml (1993), "More on Improving Service Quality Measurement," Journal of Retailing, 69(1), 140-7.

Peterson, Robert A., and William R. Wilson (1992), "Measuring Customer Satisfaction: Fact and Artifact," Journal of the Academy of Marketing Science, 20(1), 61-71.

Plummer, Joseph T. (1974), "The Concepts and Application of Life Style Segmentation," Journal of Marketing, 38(1), 33-7.

Razzaque, M. A. (1995), "Demographics, Psychographics and Consumer Value Dimensions: A Study of Consumers in a Traditional Asian Society," in European Advances in Consumer Research, Flaming Hansen (Ed.), 2, 183-91.

Rokeach, M. (168), Beliefs, Attitudes, and Values, San Francisco: Jossey Bass.

Rokeach, M. (1973), The Nature of Human Values, NY: Harper.

Rose, Gregory, M., Aviv Shoham, Lynn R. Kahle, and Rajeev Batra (1994), "Social Values, Conformity, and Dress," Journal of Applied Social Psychology, 24, 1501-19.

Rose, Gregory and M., Aviv Shoham (2000), "The Values of American and Japanese Mothers: An Application of LOV in Japan and the US," in Cross-National Consumer Psychographics, Lynn Kahle (ed.), NY: IBP, 45-62.

Shoham, Aviv, and Bella Florenthal (2001), "Values, Conformity, and Demographics as Determinants of the Importance of Fashion Attributes," presented at the conference of the Society for Consumer Psychology.

Shoham, Aviv, Bella Florenthal, Fredric Kropp, and Gregory M. Rose (1997), "Value Differences in Religious Subcultures: A Case Study of Druze and Jews in Israel," in Society for Consumer Psychology 1997, Cornelia Pechmann and S. Ratneshwar (eds.), St. Petersburg, FA: Society for Consumer Psychology, 5-10.

Shoham, Aviv, Gregory M. Rose, Ann Botstein, and Yael Pinzi (1996), "Marital Roles in Young Israeli Families," presented at the Multicultural Marketing Conference, Virginia Beach.

Tai, Susan H. C. and Jackie L. M. Tam (1996), "A Comparative Study of Chinese Consumers in Asian MarketsBA Lifestyle Analysis," Journal of International Consumer Marketing, 9(1), 25-42.

Williams, Robin M. (1979), "Change and Stability in Values and Value Systems: A Sociological Perspective," in M. Rokeach, (Ed.), Understanding Human Values: Individual and Societal, NY: Free Press, 15-46.

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