The effects of corporate social responsibility and price on consumer responses.

This experiment examined the influence of corporate social responsibility and price on consumer responses. Scenarios were created to manipulate corporate social responsibility and price across two domains (environment and philanthropy). Results from a national sample of adults indicate that

**********

Even before the problems of Enron became headline news, there was evidence of a growing discontent among the American public about the way companies conduct business. A BusinessWeek (Bernstein 2000) cover story describes Americans as "uneasy about Big Business" (145). Their BusinessWeek/Harris Poll found that 72% of Americans say they believe that business has too much power over American life. Furthermore, 66% of those polled agree that companies care more about making large profits than about selling safe, reliable, quality products. At the same time, pressure for companies to take on more responsibilities in their communities seems to be rising. In 1999, a worldwide survey found that two-thirds of consumers surveyed wanted companies to contribute to broader societal goals (Isa 2003). The Cone Corporate Citizenship Study (2002) found 89% of Americans agreeing that, following the recent corporate scandals, "it is more important than ever for companies to be socially responsible."

Many companies appear to have responded, as evidenced by a Cone/ Roper (1999) poll finding that nearly 50% of large corporations have programs associated with a social issue. Based on eight years of research, Cone (2002) concluded that cause marketing has grown to become a widely accepted business practice as it has simultaneously evolved from a short-term sales-enhancement tactic to a means for improving brand equity and corporate image.

Corporations often struggle, however, in deciding how to reconcile these social demands with those of shareholders for profit maximization. Although the majority of studies on the topic have found a significant positive relationship between corporate social responsibility (CSR) and financial performance (see, for example, Margolis and Walsh's [2001] review of 95 studies), the payoff from socially responsible programs is not guaranteed and may take time. For this reason, many managers still view CSR as an expense rather than an investment. If increased CSR were demonstrated to lead to increased sales, however, companies would be encouraged to become more socially responsible. To the extent that increased CSR results in improved communities, consumers who live in those communities would clearly benefit.

Two research questions guided this study: (1) Does a company's social responsibility record matter to consumers? and (2) If socially responsible companies charge higher prices for their products, will consumers still buy them? To investigate these issues, we designed experimental scenarios, which we mailed, along with questionnaires, to a random sample of American adults.

BACKGROUND

Mohr, Webb, and Harris (2001, 47) define CSR as "a company's commitment to minimizing or eliminating any harmful effects and maximizing its long-run beneficial impact on society." Socially responsible behavior, then, includes a broad array of actions such as behaving ethically, supporting the work of nonprofit organizations, treating employees fairly, and minimizing damage to the environment. The definition implies that a socially responsible company considers the effects of its actions on everyone, whether directly related to the company or not. Socially responsible companies, therefore, must be managed according to stakeholder theory.

There are many debates in the academic community over whether companies should be managed using a stakeholder or a shareholder theory. According to Smith (2003), stakeholder theory is based on the principle that companies need to consider the effects of their actions on all constituencies (e.g., shareholders, customers, employees, suppliers, the environment, and the community), even if profitability is reduced, while shareholder theory argues that the only responsibility of a company is to (legally) make profits for its shareholders. Mitchell (2001) argues that the shareholder approach increasingly dominates American companies. Furthermore, it is often interpreted as having one basic goal--maximizing stock prices. This, in turn, leads to a short-term perspective because a growing proportion of stockholders buy and sell stocks based on short-term information. Mitchell goes on to argue that companies are obligated to make profits, but there is nothing in the legal system that requires maximization of stock prices or that restricts the time frame to the short term. Nonetheless, a manager who does not produce sufficient earnings for shareholders risks his/her job and puts the company at risk of a takeover (Martin 2002).

Logically, many financial payoffs from CSR will take time to materialize. For example, it takes time for consumers to learn about a company's social programs, and it takes time for a company to earn the trust of the community. For this reason, a company with primarily a short-term perspective will be less likely to make CSR a focal strategy. When companies have a longer-term perspective, however, they may be more willing to wait for the payoff from CSR.

Martin (2002) argues that when responsibilities to the community are seen as being likely to come at the shareholder's expense, managers usually side with shareholders. For this reason, he developed an analytical tool, the virtue matrix, to help executives decide when social responsibility is likely to pay off financially. He suggests that most socially responsible corporate behaviors are done specifically to enhance shareholder value. These are generally activities undertaken to comply with laws, regulations, or norms. They entail simply meeting the baseline of society's expectations; companies that do not meet basic expectations are likely to lose financially. On the other hand, there are many socially responsible actions taken by companies because managers believe they are the right things to do. Because these actions are outside the norms and may lead to financial losses, they are considered risky. Although research may be helpful in predicting the financial returns on a responsible behavior that is not mandated by society, Martin suggests that it is the lack of vision among executives that is responsible for so few companies being willing to take these risks.

Because most CSR initiatives require up-front costs to companies, one notable area of risk aversion undoubtedly involves the fear that passing along the increased costs to consumers will lead to lower sales. Based on the proportion of advertisements that proclaim sales, bargains, or lower prices, one might conclude that corporate America is obsessed with keeping prices low. Some writers refer to this thinking as the myth that consumers care only about price (e.g., Crawford and Mathews 2001; Graham 2001). Certainly, there are consumers who are price sensitive, and companies such as Wal-Mart have grown dramatically based on a low-price proposition. But this does not mean that consumers care more about getting the lowest price than about anything else. In their surveys and interviews with consumers, Crawford and Mathews (2001, 23) found that consumers care more about getting "fair and honest" prices than about getting the lowest price. In their choice experiments with Amnesty International supporters and students, Auger et al. (2003) found that consumers expressed willingness to pay more for products made ethically (e.g., soap with no animal testing). Berry (1999) argues that companies overuse price as a marketing tool, probably because it can be implemented quickly and it is assumed to be salient to consumers. The cost-cutting that leads to low prices, though, can erode the value customers see in a product and lead sales to decline.

Bhattacharya and Sen (2003) argue that consumers who identify with companies are more likely to be loyal to those companies, promote them to others, and be resilient to negative information about them. They go on to propose that consumers are likely to identify with a company that offers them a positive and meaningful social identity. We suggest that being a leader on a social issue is one powerful way for a company to enhance consumers' and employees' identification with it. For example, since Avon began its Avon Breast Cancer Awareness Campaign in 1993, it has become strongly associated with breast cancer awareness, leading to increased loyalty of both customers and employees (Berger and Drumwright 1998).

Extending Bhattacharya and Sen's (2003) arguments, we suggest that CSR may add value to a product. The broadest definition of value encompasses everything the consumer receives in relation to everything the consumer sacrifices to buy and consume a product (Zeithaml 1988). For consumers who care about CSR, the social responsibility level of the company that makes a product bolsters or diminishes the value of the product. In situations of product parity, it would not take much added value for a consumer to choose a product made by a highly responsible company over one made by a less responsible one. CSR would need to add more value, however, for a consumer to be willing to pay more for a product made by the more responsible company. For this reason, we investigated the impact of CSR on consumer responses in the context of price differences.

We turn now to the development of our hypotheses. We first discuss how external (or stimulus) factors affect consumer responses. We then go on to discuss the impact of personal trait factors that are internal to consumers.

DEVELOPMENT OF HYPOTHESES

Our first issue concerns consumer responsiveness to CSR. If consumer response to CSR was reliable and strong, most all companies would have embraced the concept by now. Research indicates that consumers do vary on their responsiveness to CSR. For example, Webb and Mohr (1998) found a number of interviewees who said they shopped based on price, quality, or convenience rather than choosing retailers because of the social causes they support.

Experimental studies, however, have generally found a positive impact of CSR. In one experiment, Lafferty and Goldsmith (1999) created an ad and newspaper article manipulating both spokesperson and corporate credibility. Positive corporate credibility was created by describing the company's community and environmental contributions, while negative corporate credibility was created by describing the company's poor quality control and alleged Securities and Exchange Commission violations. Results indicated that corporate credibility had a significant impact on attitudes toward the brand and purchase intent, an effect that was stronger than the effect of spokesperson credibility. In another experiment, Brown and Dacin (1997) manipulated the level of a company's philanthropy and employee involvement in the community using a hypothetical company profile and CSR report card. They found that high CSR led to higher evaluation of the company than low CSR, and company evaluation significantly influenced evaluation of the company's product.

In their literature review and conceptual paper, Hoeffler and Keller (2002) argue that corporate societal marketing (i.e., marketing with at least one social welfare objective) can be a major way to build brand equity. Bhattacharya and Sen (2003) make a similar argument for how nonproduct aspects of a company, such as CSR, can lead to customer loyalty and other positive postpurchase outcomes. Based on these results, we offer the following hypotheses:

H1: A high level of CSR will lead to a more positive evaluation of the company and a higher level of purchase intent than a low level of CSR.

Other researchers have attempted to determine whether negative or positive information about a company's social record has the stronger effect. Folkes and Kamins (1999) suggested that negative behavior diminishes evaluations more than positive behavior bolsters them. They reasoned that because people and companies behave ethically much of the time, unethical behavior is thought to be more diagnostic of a person or company's moral character than ethical behavior. To test this, these researchers manipulated product quality and ethics using hypothetical word-of-mouth accounts by a friend describing a brand of telephone. They found that descriptions of unethical behavior (i.e., employing child labor) led to negative attitudes toward the firm, regardless of whether product quality was described as high or low. On the other hand, descriptions of ethical behavior (refraining from hiring children or hiring victims of a natural disaster) led to more varied results depending on product quality. They concluded that their results supported the negativity bias.

Creyer and Ross (1996) created scenarios describing cereal manufacturers generally (e.g., size and emphasis on product improvements) and then as either ethical (i.e., quickly refunding customer's money after discovering incompletely filled cereal boxes) or unethical (i.e., stating that their cereal lowers the risk of heart disease when there was no evidence supporting this claim). They found that, compared to a control group that received only the general information, respondents were not willing to pay more for the cereal made by the ethical firm, but they did demand lower prices on cereal sold by the unethical firm. In an experiment comparing control groups with groups given information about high or low levels of CSR (i.e., being or not being a corporate champion of women and minorities), Sen and Bhattacharya (2001) found consumer evaluations of the company to be more sensitive to negative than to positive CSR information. Based on the above research, we propose the following hypothesis:

H2: A low level of CSR will have a greater impact on evaluation and purchase intent than a high level of CSR.

Several researchers have suggested that important product attributes would interact with CSR in their impact on consumer responses. Specifically, a low level of CSR may overwhelm the importance of other, more traditional, product attributes such as quality and convenience. Folkes and Kamins (1999) manipulated product quality (i.e., the quality of sound of a telephone) in addition to CSR. They found that sound quality had no significant effect on attitude toward the firm when CSR was low, but quality did have a significant positive effect on attitude when CSR was high.

Using scenarios about a new, hypothetical retailer moving into a community, Handelman and Arnold (1999) manipulated traditional store image attributes (i.e., merchandise assortment, prices, and location convenience) and CSR and measured the effects on consumer support (purchase intent, word of mouth, willingness to boycott, and willingness to support a building permit in the city). When the store was described as not being supportive of family, not contributing to local charities, and not buying products made in-country (i.e., low CSR), it received no more consumer support when it was described as strong on store image attributes than when it was described as weak on those attributes. When CSR was high, store image had a positive effect on support. They conclude that community members will only buy from a company that surpasses a minimum tolerable level of CSR. When a company drops below that level, high performance on other dimensions makes no difference.

We propose extending this research by using price as the other traditional product dimension. We suggest that price will be less important to consumers when CSR drops below a minimum acceptable level:

H3: There will be an interaction between CSR and price such that when CSR is low, price will have a weaker effect on purchase intent than when CSR is high.

It seems likely that several personal trait variables would affect whether or how strongly consumers respond to a company's level of social responsibility. One such trait is called socially responsible or socially conscious consumer behavior. Webster (1975, 188) defines the socially conscious consumer as "a consumer who takes into account the public consequences of his or her private consumption or who attempts to use his or her purchasing power to bring about social change." Mohr, Webb, and Harris (2001, 47) define socially responsible consumer behavior (SRCB) as "a person basing his or her acquisition, usage, and disposition of products on a desire to minimize or eliminate any harmful effects and maximize the long-run beneficial impact on society." SRCB can be seen as an enduring personality trait that involves the consumer's self-concept. Persons high on this trait would modify their consumption behaviors in a wide variety of contexts in order to strive toward the ideal of improving society.

This trait has been investigated by a number of researchers (e.g., Antil 1984; Leigh, Murphy, and Enis 1988; Roberts 1995) who have developed measurement scales and explored related demographics and attitudes. Roberts (1995), for example, used cluster analysis to isolate a group of socially responsible consumers, which he estimated to be 32% of the American population. His results indicate that, when compared to most Americans, this group is more liberal and environmentally concerned and has higher levels of perceived consumer effectiveness (perceived ability of individual consumers to influence environmental problems). To our knowledge, however, no one has experimentally tested the impact of this trait on consumer responses to CSR. We suggest the following hypothesis:

H4: When the measured trait of SRCB is high, CSR will have a stronger impact on evaluation and purchase intent than when SRCB is low.

Sen and Bhattacharya (2001) suggest that consumers are more likely to respond to a company's social responsibility record when they identify with the company. Identification is enhanced when consumer perceptions of the company's character are similar (or congruent) to their perceptions of their own character. They go on to argue that consumers form judgments of a company's character based more on its CSR than on its business expertise. When consumers personally support the social issues that the company targets (called support for the CSR domain), they are likely to see greater congruence between themselves and the company. In two experiments, they manipulated CSR (in the contexts of diversity issues and treatment of overseas employees) and measured its effects on evaluation of the company. They found that consumer support for the CSR domain significantly moderated the positive effect of CSR on evaluation. Based on this research we hypothesize the following:

H5: CSR in a specific domain will have a stronger influence on evaluation and purchase intent for consumers who believe that companies have many responsibilities in that domain than for those who believe that companies have few responsibilities in that domain.

METHOD

Study Design

An experiment was constructed to test the hypotheses. Scenarios were created to manipulate CSR (low vs. high) and price (low vs. high) across two CSR domains (environment and philanthropy). A completely randomized full factorial design with eight treatment groups and two control groups (no CSR information with price low vs. high) was used.

The scenario involved shopping for a consumer product, athletic shoes. This product was used because it is one that most everyone, male or female, is likely to be familiar with. The respondent was asked to imagine shopping for athletic shoes or sneakers at a store with a number of brands and finding one that he or she liked made by "Company A" (see Appendix A). Company A was described briefly as a company that has a reputation for making high-quality shoes. To manipulate price, the shoes were described as costing either more or less than the other shoes the respondent has been looking at.

For the experimental treatment groups, a second paragraph described the level of CSR of the company. To provide two contexts for testing the hypotheses, high and low CSR levels were created in two important domains, the environment and philanthropy. Scenarios were developed for more than one domain to create a greater range of settings and provide a more robust test of the hypotheses, which should lead to greater generalizability of the findings across types of CSR.

To manipulate CSR as either high or low, Company A was described as having the best or worst rating in the industry on the environment or corporate giving. In each treatment, three reasons for the rating were provided. The rating was described as given "by a highly respected, impartial organization that evaluates companies every year" to ensure that respondents viewed the information as credible.

Measures

Purchase intent was measured by asking the respondent to rate how likely he or she would be to buy the shoes made by Company A on three 7-point semantic differential items anchored by very unlikely/very likely, impossible/very possible, and no chance/certain. To measure evaluation of the company, they were asked to evaluate Company A on three 7-point semantic differential items anchored by unfavorable/favorable, bad/good, and harmful/beneficial.

The attitudinal variable, support for the domain, was measured using the mean of three 5-point Likert-type items for each domain. Each item tapped the respondent's support for one of the three reasons given in the scenario for the company being rated as high or low on CSR. For example, in the environmental domain, Company A was described as having been given a high (or low) rating because its factories pollute less (or more) than others in the industry. The corresponding support for the domain item is "Companies should make every effort to reduce the pollution from their factories." (See Appendix B for the complete set of items.)

SRCB was measured using the mean of the 26-item Socially Responsible Purchase and Disposal Scale (Webb, Mohr, and Harris 2004). This scale asks respondents to indicate, on a 5-point response scale anchored by "never true" and "always true," how often they engage in a variety of behaviors (see Appendix B). The original 147-item scale was reduced to 26 items through the analysis of expert judging and administrations to two different samples. Factor analysis indicated that the scale is composed of three correlated factors: (1) influence on consumer purchasing of the company's philanthropic activities, hiring practices and treatment of employees; (2) consumer recycling; and (3) consumers' avoidance and reduction in usage of products that harm the environment. For this study, we combined the factors into one measure of SRCB ([alpha] = .95).

The manipulation check for CSR asked respondents about their judgments of Company A's social responsibility in the specific domain described in their scenario. For example, those in the environmental scenarios were asked to rate their agreement to the statement, "Company A has an excellent environmental record." The questionnaire also included a 1-item manipulation check for price. Respondents were asked to rate their level of agreement with the statement "Company A's shoes cost more than shoes made by most other companies."

To determine how credible the scenarios were and whether there were any significant differences in credibility created by the treatments, respondents were asked to evaluate how believable they thought the description of Company A was on a 7-point scale anchored by "completely unbelievable" and "completely believable." The respondents were also asked for demographic information about themselves.

Data Collection

A cover letter and questionnaire were mailed to a sample of 1,997 adults. The questionnaire each person received contained one randomly assigned experimental or control treatment scenario and scales measuring the appropriate variables. The mailing was sent to a random sample of American adults obtained from a marketing research firm.

Using many of Dillman's (1978, 1998) methods, a compelling cover letter was written to engage recipients in the research project, emphasize the credibility of the researchers and their university sponsorship, and guarantee anonymity of responses. The letters were printed on university letterhead, and one of the researchers signed each letter in blue ink. The enclosed questionnaire was kept short (three or four pages), and a stamped, preaddressed return envelope was enclosed. A follow-up reminder postcard was mailed one week later. Of the 1,997 questionnaires mailed, 51 were returned as undeliverable. Of the remaining 1,946 questionnaires, 194 were completed and returned for a 10.0% response rate. Cell sizes were reasonably balanced (ranging from 15 to 23).

RESULTS

The Sample

The sample was 54.1% male. Ages ranged from 19 to 94, with a mean of 53.0. When asked to rate themselves on a 7-point scale from very conservative politically to very liberal politically, responses indicated that they lean slightly to the conservative side, with a mean of 3.6. Almost 20% (19.2%) had never attended college, while 47.7% had completed at least a four-year college degree. The modal household income category was $40,000-$59,999 per year, and 18.0% had yearly incomes of $100,000 or more. Respondents lived in 43 states across the United States. Compared to the overall U.S. population (U.S. Census Bureau 2000), the sample was slightly more male and slightly older. It had a somewhat higher income level and a higher education level.

To address concerns of nonresponse bias, we compared early (the first two-thirds) and late respondents (last one-third) on the demographic factors of age, political conservatism, education, and income and found no significant differences. In an attempt to determine whether the nature of our experiment influenced self-selection (i.e., response bias) into the sample, we conducted an additional analysis. We compared the demographics and SRCB scores of those who received control group scenarios (with no CSR information) with those who received a scenario manipulating CSR. We found no significant differences between the two groups in age, political conservatism, income, or SRCB. There was a significant difference in education (F = 4.03, p < .05), with those responding to the treatments being more highly educated than those responding to the control scenarios. We expect that this difference may be due to the greater length of the treatment scenarios, which may have discouraged potential respondents with less education. The fact that those who responded to the treatment scenarios were not more politically liberal or higher on SRCB than those responding to the control scenarios indicates that the nature of the experimental manipulations did not exert a strong influence on respondent self-selection. In summary, we believe that the relatively high income and education level of our sample is likely due to the fact that highly educated people are more inclined to respond to mail surveys.

Manipulations and Measures

Analyses of variance were conducted to determine the effects of manipulated CSR on perceived CSR. Because CSR was of necessity described differently for the environmental and philanthropic domains, the manipulation checks were conducted separately for each domain. The analysis showed that CSR as manipulated significantly affected perceived CSR in the expected direction for both the environmental (F = 228.67, p < .01) and philanthropic (F = 194.63, p < .01) domains. The analysis also indicated that price as manipulated significantly affected perceived price (F = 141.54, p < .01). Overall, these tests show that the manipulations strongly and significantly influenced the perceived variables as intended.

It was important for the information about the company described in the scenario to be seen as credible by respondents. On a 7-point scale (where 1 = completely unbelievable and 7 = completely believable), the mean across the entire sample was above the midpoint at 4.51. To determine whether any of the treatments affected credibility, a three-way analysis of variance was conducted using price, CSR, and domain as the independent variables and credibility as the dependent variable. No main effects or interactions were statistically significant, indicating that all treatment scenarios were perceived as similarly credible.

Cronbach's alpha and interitem correlations were examined to determine if any of the multiitem measures had unacceptable reliability levels (see Appendix B). All had high alphas (ranging from .84 to .95). Because of this, for our analyses, we were able to combine the items of each measure into one using the mean. For example, we used the mean of the three purchase intent items to measure purchase intent.

Hypothesis Tests

H1 predicted that CSR would positively impact the evaluation of the company and purchase intent. To test this, we conducted a multivariate analysis of variance (MANOVA) with CSR and domain as independent variables and evaluation and purchase intent as dependent variables. (Because domain was a replication variable in this study, it was included in the analyses as an independent variable along with the variables of interest only to check for interactions.) As can be seen in Table 1, the multivariate effects of CSR, domain, and the interaction between CSR and domain were significant. Univariate analyses indicated that the significant interaction is driven by its effect on evaluation. Examination of the means on the 7-point evaluation scale showed that when CSR was high, there was little difference between respondents in the environmental and philanthropic domains. When CSR was low, however, the company described in the environmental domain was evaluated lower than the company described in the philanthropic domain.

Because of the significant interaction, all remaining analyses were run separately for each domain. Further testing H1, the multivariate effect of CSR on evaluation and purchase intent was found to be significant for both the environmental and the philanthropy scenarios. Univariate analyses showed that CSR significantly affected both dependent variables in both domains in the predicted direction. H1 is, therefore, supported.

H2 suggests that provision of information about low social responsibility will have a stronger impact on evaluation of the company and purchase intent than provision of information about high social responsibility. To test this, four MANOVAs were conducted with evaluation and purchase intent as dependent variables. Separately for each domain, the low-CSR groups were compared to the control groups and the high-CSR groups were compared to the control groups. Compared to the control groups who were given no information about the company's social responsibility, information on the company's low level of CSR was found to significantly reduce evaluation and purchase intent for both the environmental and philanthropic domains (see Table 2). In neither domain did information on the company's high level of CSR significantly increase evaluation or purchase intent over the control groups. H2 is, therefore, supported.

H3 suggests that there will be an interaction between CSR and price such that the relationship between price and purchase intent will be weaker when CSR is low than when it is high. To test this hypothesis separately for each domain, we conducted a two-way analysis of variance with CSR and price as independent variables and purchase intent as the dependent variable. As can be seen in Table 3, the results were consistent across both domains: price had a significant inverse effect on purchase intent, CSR had a significant positive effect on purchase intent, and there were no significant interactions between CSR and price (failing to support H3). Although the prediction that price would have less of an impact when CSR is low was not supported, an examination of the F levels suggests that CSR had a stronger effect on purchase intent than price.

To further examine the relative effects of CSR and price, we conducted post hoc analyses comparing the purchase intent means of each treatment group (i.e., each CSR-price combination) within each of the two domains (see Figure 1). For subjects given information on the company's environmental behavior, every group was significantly different from every other group. Most interestingly, purchase intent was higher for those rating the high-price shoes made by the responsible company than for those rating the low-price shoes made by the irresponsible company (t = -5.36, p < .01). For subjects given information on the company's philanthropy record, the difference between these two treatment groups was in the same direction but did not reach significance.

H4 proposes that CSR will have a stronger effect on evaluation and purchase intent for respondents scoring high on the trait of SRCB than for those scoring low on this trait. To conduct this analysis the sample was first split at the median on the SRCB score (3.27 on a 5-point scale). Then, separately for each domain, multiple analysis of variance was conducted using CSR and dichotomized SRCB as independent variables and evaluation and purchase intent as dependent variables. As can be seen in Table 4, the results showed that when the scenarios described CSR in the environmental context, the interaction between CSR and SRCB was significant. In univariate tests, the interaction was significant for evaluation but not for purchase intent. The direction of the interaction (see Figure 2) is consistent with the hypothesis: CSR affected evaluation of the company more strongly for those high in SRCB than for those low on this trait. In the philanthropy domain, the interaction between CSR and SRCB was in the hypothesized direction but not significant. H4 is, therefore, partially supported.

H5 suggests that CSR will have a stronger impact on evaluation and purchase intent for consumers who more strongly support the domain of the company's social responsibility programs than for those who exhibit weaker support. Because the measure of support for the domain is specific to each domain, a median split of the sample was conducted separately for those exposed to the environmental and philanthropy scenarios. On 5-point scales, the median support for CSR in the environmental domain was 4.33 and in the philanthropic domain was 4.00.

Multiple analysis of variance was then conducted separately for each domain using CSR and dichotomized support for the domain as independent variables and evaluation and purchase intent as dependent variables. For the subjects given the environmental scenarios, significant multivariate effects were found for the interaction between support for the domain and CSR (see Table 5). In univariate tests, the interaction was found to have a significant effect on both evaluation and purchase intent. As can be seen in Figure 3, group means indicate that the impact of CSR on evaluation of the company was stronger when support for the domain was high than when it was low. The results were similar for purchase intent: when support for the domain was high, the impact of CSR was stronger than when support was low (see Figure 4). For subjects responding to the philanthropy scenarios, multivariate tests also indicated a significant interaction between CSR and support for the domain. Univariate tests showed that the significant interaction is driven by its effect on evaluation. As can be seen in Figure 3, the effect of CSR on evaluation of the company was stronger when support for the domain was high than when it was low. Although this interaction was not significant when purchase intent was the dependent variable, Table 5 shows that the effect was in the hypothesized direction. Overall, three of the four analyses provide support for H5.

DISCUSSION

The experimental results indicate that CSR in both the environmental and philanthropy domains had a significant positive effect on evaluation of the company and purchase intent. These results support the idea that aspects of a company beyond its products, such as its reputation (e.g., Brown and Dacin 1997), influence evaluation of the company and buying choices.

When compared to control groups given no CSR information, our respondents reacted more strongly to negative than to positive CSR. This supports the negativity bias found by other researchers (e.g., Creyer and Ross 1996; Folkes and Kamins 1999; Sen and Bhattacharya 2001). The fact that companies with no information about CSR were rated as positively as companies with high levels of CSR does not, however, support the idea that respondents were as positive toward companies that did nothing as toward companies who worked to be socially responsible. This is because the low-CSR treatments were created by describing companies as making no effort to be socially responsible (e.g., as not doing something to reduce pollution, as not using recycled materials in manufacturing, as not making substantial donations to charity). The relatively positive response to the control groups seems to indicate that, when no information on CSR was provided, respondents either did not think about CSR or they assumed (using the adage "no news is good news") that the company was behaving responsibly.

The effect of environmental CSR on evaluation of the company was found to be stronger for those who scored high on a measure of SRCB, a personality trait that includes responses to companies and products across a variety of social responsibility issues. SRCB did not significantly affect the relationship between philanthropic CSR and evaluation or between CSR in either domain and purchase intent, but the results were in the predicted direction. Overall, these results provide some support for the idea that consumers who see themselves as trying to improve the world through a variety of purchase, consumption, and disposal behaviors (vs. those who do not) respond more strongly to information about the level of a company's social responsibility. Perhaps, stronger support would have been found for the relationship between SRCB and consumers' responses to CSR if the measure for SRCB had been more balanced between environmental and other socially responsible behaviors. This is particularly true for philanthropic CSR.

Not surprisingly, stronger interaction effects were found with the more specific attitudinal variable, support for the domain. Environmental CSR had a stronger effect on both evaluation and purchase intent for those with a high than with a low level of support for company responsibilities to the environment. Similarly, philanthropic CSR had a stronger effect on evaluation of the company for those respondents most supportive of corporate philanthropy, and the effects on purchase intent were in the same direction although not significant. These results provide support for Sen and Bhattacharya's (2001) suggestion that identification with a company is enhanced when consumers believe in that company's social agenda, and identification strengthens positive responses to the company.

The only hypothesis that did not receive some support in this study was the one predicting that price would have less impact on purchase intent when CSR was low as opposed to high. For our respondents, price significantly inversely affected purchase intent even when CSR was low. In fact, contrary to the hypothesis, the effect of price was slightly stronger when CSR was low than when it was high. This contradicts the results found by Folkes and Kamins (1999), where product quality lost its effect for the low-CSR company, and by Handelman and Arnold (1999), where store image attributes had no impact for the low-CSR company.

Although negative CSR information was not sufficient to eliminate a price effect, price did not have as strong an effect on purchase intent in this experiment as CSR did, especially in the environmental domain. Because CSR was manipulated using a paragraph and price was manipulated using only a sentence, the manipulation of CSR may be stronger than that of price, and this could account for this effect. We deliberately created a strong manipulation of CSR, so we could test whether CSR can overpower price. Furthermore, we believe that communicating a construct as complex as CSR requires information on more than one behavior. With price, on the other hand, we believe that a simple statement is sufficient and realistic. To our knowledge, this is the only experiment with a broad-based national sample that has demonstrated that information on CSR can affect purchase intentions more strongly than price.

Because the two domains of environment and philanthropy were used as a replication variable, testing the relative strength of CSR across domains was not a focus of this study. Nonetheless, we note that CSR had a stronger effect on evaluation of the company when it was in the environmental than in the philanthropic domain. The difference results primarily from different responses to low levels of CSR. This is likely due to the differences inherent in low responsibility across these domains. That is, not contributing philanthropically is seen as less negative for the community than harming the environment (e.g., by polluting).

CONCLUSION AND IMPLICATIONS

We begin by discussing limitations of this study and suggestions for future research. We then suggest implications of our results for consumers, policy makers, and consumer educators.

We believe that the major limitation of this research is inherent in scenario-based experiments. That is, subjects were asked to role-play by imagining themselves in a hypothetical shoe-buying situation. It is likely that the respondents paid more attention to the CSR descriptions than they would if they had read them in a newspaper or magazine. Furthermore, they read these descriptions immediately before evaluating the company and rating their likelihood of buying the company's shoes. Both these factors make it likely that the effect of CSR was stronger than it would be in an actual store. For this reason, replicating this study in a field experiment would provide a more conservative test of the effects of CSR. While our study demonstrates that CSR information can affect consumer responses, field research should be conducted to determine when it will have this effect. We also acknowledge that our sample was more highly educated and had a higher income level than the general population.

Survey research cited earlier shows that many consumers want companies to behave more responsibly. The more the consumers use CSR as a consideration when making purchase decisions, the more strongly CSR will affect sales and the more likely companies will be to adopt CSR practices. On the negative side, however, the potential to realize increased sales in response to CSR is also tempting to less scrupulous marketers. Consumers need information about companies' levels of social responsibility to differentiate sound CSR programs from those that simply pay lip service to social responsibility.

This study found that, when consumers are given information that they trust about a company's level of social responsibility, it affects how they evaluate the company and their purchase intentions. Furthermore, a low price did not appear to compensate for a low level of social responsibility. These results indicate that many American consumers value CSR and may use it as a purchasing criterion even when there is not a product parity situation. Unfortunately, this leaves consumers in a vulnerable position. Our finding that CSR can affect purchase intentions more strongly than price emphasizes the importance of consumers' abilities to accurately evaluate CSR programs.

Responses to CSR are, therefore, dependent on consumers having valid information. While many companies communicate the good things they are doing, consumer trust of corporate communications is low. Nonprofit organizations can and do play a role here by publicizing the support they receive from companies. Research shows that consumers are less skeptical when companies demonstrate a long-term commitment to an issue (e.g., reducing environmental harm) or to a nonprofit organization (Webb and Mohr 1998). However, getting trustworthy information on corporate commitments to CSR may be difficult and take time.

Disinterested third parties may be a source of credible information on how socially responsible companies are. For example, Fortune publishes an annual ranking of companies based on their treatment of employees (Levering et al. 2004). Other examples include the Business Ethics awards published by Business Ethics magazine (Asmus 2003) and corporate social performance ratings developed for investors by firms such as Kinder, Lydenberg, and Domini (Waddock and Graves 1997). Dow Jones provides sustainability indices based on economic, environmental, and social criteria (Dow Jones 2004). Nonprofit organizations such as Co-Op America provide consumers with ratings of companies based on their level of social responsibility. At their Web site, consumers can find company profiles listed by company, brand, or product category; compare companies within a product category on social and environmental issues; or learn about current boycotts (Co-Op America 2004).

The government also plays a role in disseminating information about the CSR levels of companies. One good example is the ENERGY STAR Web site of the Environmental Protection Agency and the Department of Energy (ENERGY STAR 2003). This Web site rates the energy efficiency of products such as home appliances and companies such as home builders. It also selects ENERGY STAR Partner of the Year Awards, publicizing the contributions selected companies have made to reducing greenhouse gas emissions through energy efficiency. Governmental agencies could publicize their corporate ratings more widely so that more consumers could make use of them in their purchasing decisions.

If informed consumers are the goal, consumer educators can assist by developing and disseminating lists of sources that consumers can use for evaluating companies on their social responsibility records. They can also work toward increasing the proportion of consumers who are motivated enough to seek out such information by discussing issues such as socially responsible consumption and the results of corporate behaviors within specific CSR domains.

APPENDIX A: EXPERIMENTAL SCENARIOS

Environmental Domain

Imagine that you are shopping for athletic shoes or sneakers for yourself or a family member. You go to a store that has a number of brands of shoes and begin comparing them. You find a shoe that you like that is made by Company A. Company A is an American company that has its headquarters in the United States. This company has a reputation for making high-quality shoes. The shoes you like, made by Company A, cost more (less) than the other shoes you have been looking at.

You recently learned that Company A was rated as having the best (worst) environmental record in the industry. This rating was given by a highly respected, impartial organization that evaluates companies every year. Some of the reasons for the high (low) rating are that Company A's factories pollute less (more) than others in the industry, it uses a high percentage of (does not use any) recycled materials in manufacturing its shoes, and its factories have good (have never had any) programs to conserve water and energy.

Philanthropy Domain

Imagine that you are shopping for athletic shoes or sneakers for yourself or a family member. You go to a store that has a number of brands of shoes and begin comparing them. You find a shoe that you like that is made by Company A. Company A is an American company that has its headquarters in the United States. This company has a reputation for making high-quality shoes. The shoes you like, made by Company A, cost more (less) than the other shoes you have been looking at.

You recently heard that Company A was rated best (worst) in the industry on corporate giving. This rating was given by a highly respected, impartial organization that evaluates companies every year. Some of the reasons for the high (low) rating are that Company A's donations to charity are larger (smaller) than others in the industry, it has (does not have) programs to recognize employees for their volunteer work in the community, and it donates (has never donated any of the) products it makes, such as shoes and clothing, to disaster victims or others in need.

Note: Respondents in the control groups received only the first paragraph of the scenario.

APPENDIX B: MULTIITEM MEASURES

Support for the Domain: Environment ([alpha] = .84)

1. Companies should make every effort to reduce the pollution from their factories.

2. Companies should use recycled materials in manufacturing new products.

3. Companies should have factory programs to conserve water and energy.

Support for the Domain: Philanthropy ([alpha] = .84)

1. Companies should regularly make donations to charity.

2. Companies should have programs to recognize employees for their volunteer work in the community.

3. Companies should donate some of their products to people in need.

Socially Responsible Purchase and Disposal ([alpha] = .95)

1. I recycle aluminum cans.

2. I recycle steel/tin cans.

3. I limit my use of energy such as electricity or natural gas to reduce my impact on the environment.

4. I try to buy from companies that support victims of natural disasters.

5. When I am shopping, I try to buy from companies that are working to improve conditions for employees in their factories.

6. I recycle paper.

7. I make an effort to buy products and services from companies that pay all their employees a living wage.

8. I try to buy from companies that help the needy.

9. I try to buy from companies that hire people with disabilities.

10. I recycle cardboard.

11. I avoid buying products or services from companies that discriminate against minorities.

12. I recycle plastic containers.

13. I recycle magazines.

14. I avoid buying from companies that harm endangered plants or animals.

15. Whenever possible, I walk, ride a bike, car pool, or use public transportation to help reduce air pollution.

16. I avoid using products that pollute the air.

17. When given a chance to switch to a retailer that supports local schools, I take it.

18. I try to buy from companies that make donations to medical research.

19. I make an effort to buy from companies that sponsor food drives.

20. I avoid buying products that pollute the water.

21. I make an effort to avoid products or services that cause environmental damage.

22. I avoid buying products that are made from endangered animals.

23. When given a chance to switch to a brand that gives back to the community, I take it.

24. I avoid buying products made using child labor.

25. When given a chance, I switch to brands where a portion of the price is donated to charity.

26. I avoid buying products or services from companies that discriminate against women.

REFERENCES

Antil, John H. 1984. Socially Responsible Consumers: Profile and Implications for Public Policy. Journal of Macromarketing, 4 (Fall): 18-30.

Asmus, Peter. 2003. 15th Annual Business Ethics Awards. Business Ethics, 17 (4): 8-16.

Auger, Pat, Paul Burke, Timothy M. Devinney, and Jordan J. Louviere. 2003. What Will Consumers Pay for Social Product Features? Journal of Business Ethics, 42 (February): 281-304.

Berger, Ida E. and Minette Drumwright. 1998. Social Alliances: Collaborative Efforts Between Business, Government, and Non-profits. Marketing and Public Policy Conference Proceedings, 8: 1-2.

Bernstein, Aaron. 2000. Too Much Corporate Power? BusinessWeek, September 11, 144-158.

Berry, Leonard L. 1999. Discovering the Soul of Service. New York: The Free Press.

Bhattacharya, C.B. and Sankar Sen. 2003. Consumer-Company Identification: A Framework for Understanding Consumers' Relationships with Companies. Journal of Marketing, 67 (April): 76-88.

Brown, Tom J. and Peter A. Dacin. 1997. The Company and the Product: Corporate Associations and Consumer Product Responses. Journal of Marketing, 61 (January): 68-84.

Cone, Carol. 2002. Best Practices: The Evolution of Cause Branding. Holmes Report. http://www. holmesreport.com.

Cone Corporate Citizenship Study. 2002. http://www.coneinc.com.

Cone/Roper Cause Related Trends Report. 1999. The Evolution of Cause Branding. Boston: Cone, Inc. Co-Op America. 2004. http://www.coopamerica.org.

Crawford, Fred and Ryan Mathews. 2001. The Myth of Excellence: Why Great Companies Never Try to Be the Best at Everything. New York: Crown Business.

Creyer, Elizabeth and William T. Ross. 1996. The Impact of Corporate Behavior on Perceived Product Value. Marketing Letters, 7 (2): 173-185.

Dillman, Don A. 1978. Mail and Telephone Surveys: The Total Design Method. New York: Wiley and Sons.

--. 1998. Mail and Other Self-Administered Surveys in the 21st Century: The Beginning of a New Era. Working Paper. Pullman, WA: Washington State University.

Dow Jones. 2004. http://www.dowjones.com.

ENERGY STAR. 2003. http://www.energystar.gov.

Folkes, Valerie S. and Michael A. Kamins. 1999. Effects of Information about Firms' Ethical and Unethical Actions on Consumers' Attitudes. Journal of Consumer Psychology, 8 (3): 243-259.

Graham, John R. 2001. Myths That Menace Selling Success. The American Salesman, March, 11-16.

Handelman, Jay M. and Stephen J. Arnold. 1999. The Role of Marketing Actions with a Social Dimen-sion: Appeals to the Institutional Environment. Journal of Marketing, 63 (July): 3348.

Hoeffler, Steve and Kevin Lane Keller. 2002. Building Brand Equity through Corporate Societal Marketing. Journal of Public Policy and Marketing, 21 (1): 78-89.

Isa, Mohd Khairi Mohd. 2003. Applying the Triple Bottom Line Approach. Business Times, May 12.

Lafferty, Barbara A. and Ronald E. Goldsmith. 1999. Corporate Credibility's Role in Consumers' Attitudes and Purchase Intentions When a High Versus a Low Credibility Endorser Is Used in the Ad. Journal of Business Research, 44 (2): 109-116.

Leigh, James H., Patrick E. Murphy, and Ben M. Enis. 1988. A New Approach to Measuring Socially Responsible Consumption Tendencies. Journal of Macromarketing, 8 (Spring): 5-20.

Levering, Robert, Milton Moskowitz, Julia Boorstin, Christopher Tkaczyk et al. 2004. The 100 Best Companies to Work For. Fortune, 149 (1): 56.

Margolis, Joshua Daniel and James Patrick Walsh. 2001. People and Profits ? The Search for a Link between a Company's Social and Financial Performance. Mahwah, NJ: Erlbaum.

Martin, Roger L. 2002. The Virtue Matrix: Calculating the Return on Corporate Responsibility. Harvard Business Review, 80 (March): 68-75.

Mitchell, Lawrence E. 2001. Corporate Irresponsibility: America's Newest Export. New Haven, CT: Yale University Press.

Mohr, Lois A., Deborah J. Webb, and Katherine E. Harris. 2001. Do Consumers Expect Companies to Be Socially Responsible? The Impact of Corporate Social Responsibility on Buying Behavior. Journal of Consumer Affairs, 35 (1): 45-72.

Roberts, James A. 1995. Profiling Levels of Socially Responsible Consumer Behavior: A Cluster Analytic Approach and Its Implications for Marketing. Journal of Marketing--Theory and Practice, 3 (4): 97-117.

Sen, Sankar and C.B. Bhattacharya. 2001. Does Doing Good Always Lead to Doing Better? Consumer Reactions to Corporate Social Responsibility. Journal of Marketing Research, 38 (May): 225-243.

Smith, H. Jeff. 2003. The Shareholders vs. Stakeholders Debate. MIT Sloan Management Review, 44 (4): 85-90.

U.S. Census Bureau. 2000. American FactFinder. http://www.factfmder.census.gov.

Waddock, Sandra A. and Samuel B. Graves. 1997. The Corporate Social Performance--Financial Performance Link. Strategic Management Journal, 18 (4): 303-319.

Webb, Deborah J. and Lois A. Mohr. 1998. A Typology of Consumer Responses to Cause-Related Marketing: From Skeptics to Socially Concerned. Journal of Public Policy and Marketing, 17 (2): 226-238.

Webb, Deborah J., Lois A. Mohr, and Katherine E. Harris. 2004. Defining and Measuring Socially Responsible Consumption. Working Paper.

Webster, Frederick E., Jr. 1975. Determining the Characteristics of the Socially Conscious Consumer. Journal of Consumer Research, 2 (December): 188-196.

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

Lois A. Mohr is an associate professor (retired) at Georgia State University. Deborah J. Webb is an assistant professor at University of West Georgia (dwebb@westga.edu).

This research was supported by grants from the Georgia State University Marketing Department and the Robinson College of Business. The authors also wish to thank Beverly Wright and Anita Whiting for their research assistance.

TABLE 1
Effects of CSR and Domain on Evaluation and Purchase Intent (H1)

                                 MANOVA Results

Independent variables        Wilks' [lambda]        F

CSR                              0.35          137.00 **
  Low
  High
Domain                           0.95            4.19 *
  Environment
  Philanthropy
CSR x domain                     0.95            3.99 *
  CSR low, environment
  CSR high, environment
  CSR low, philanthropy
  CSR high, philanthropy
CSR (environmental domain)       0.23          122.70 **
  Low
  High
CSR (philanthropic domain)       0.49           38.23 **
  Low
  High
                              Univariate Results

                                 Evaluation

Independent variables        Mean   n        F

CSR                                      272.89 **
  Low                        2.89   76
  High                       6.42   76
Domain                                   6.77 **
  Environment                4.40   75
  Philanthropy               4.91   77
CSR x domain                             7.65 **
  CSR low, environment       2.30   37
  CSR high, environment      6.44   38
  CSR low, philanthropy      3.45   39
  CSR high, philanthropy     6.40   38
CSR (environmental domain)               244.77 **
  Low                        2.30   37
  High                       6.44   38
CSR (philanthropic domain)               77.08 **
  Low                        3.45   39
  High                       6.40   38

                             Univariate Results

                              Purchase intent

Independent variables        Mean   n        F

CSR                                      104.46 **
  Low                        3.43   76
  High                       6.11   76
Domain                                    0.77
  Environment                4.68   75
  Philanthropy               4.87   77
CSR x domain                              2.18
  CSR low, environment       3.12   37
  CSR high, environment      6.19   38
  CSR low, philanthropy      3.74   39
  CSR high, philanthropy     6.04   38
CSR (environmental domain)               76.48 **
  Low                        3.12   37
  High                       6.19   38
CSR (philanthropic domain)               34.78 **
  Low                        3.74   39
  High                       6.04   38

* p<.05; ** p<.01.

TABLE 2
Effects of Low, High, and No CSR Information on Evaluation
and Purchase Intent (H2) Univariate Results

                              MANOVA Results

Independent variables   Wilks' [lambda]       F
Environmental domain
  CSR                        0.32         79.69 **
    Low
    Control
  CSR                        0.96          1.49
    High
    Control
Philanthropic domain
  CSR                        0.62         23.91 **
    Low
    Control
  CSR                        0.96          1.54
    High
    Control

                            Univariate Results

                             Evaluation

Independent variables   Mean   n       F
Environmental domain
  CSR                               151.98 **
    Low                 2.30   37
    Control             5.99   41
  CSR                                 2.99
    High                6.44   38
    Control             5.99   41
Philanthropic domain
  CSR                                48.37 **
    Low                 3.45   39
    Control             5.99   41
  CSR                                 2.76
    High                6.40   38
    Control             5.99   41

                         Univariate Results

                          Purchase intent

Independent variables   Mean   n       F
Environmental domain
  CSR                               48.72 **
    Low                 3.12   37
    Control             5.81   41
  CSR                                1.70
    High                6.19   38
    Control             5.81   41
Philanthropic domain
  CSR                               29.15 **
    Low                 3.74   39
    Control             5.81   41
  CSR                                0.44
    High                6.04   38
    Control             5.81   41

* p<.05; ** p<.01.

TABLE 3
Effects of CSR and Price on Purchase Intent (H3)

                         Environmental Domain   Philanthropic Domain

Independent Variable     Mean   n       F       Mean   n       F

CSR                                  86.05 **               36.62 **
  Low                    3.06   38              3.74   39
  High                   6.19   38              6.04   38
Price                                 4.47 *                13.01 **
  Low                    4.80   42              5.68   33
  High                   4.41   34              4.27   44
CSR x price                           0.50                   0.25
  CSR low, price low     3.45   23              4.63   16
  CSR low, price high    2.47   15              3.12   23
  CSR high, price low    6.44   19              6.67   17
  CSR high, price high   5.95   19              5.52   21

* p <.05; ** p <.01.

TABLE 4
Effects of CSR and SRCB on Evaluation and Purchase Intent (H4)

                                MANOVA Results

Independent variables     Wilks' [lambda]       F
Environmental domain
  CSR                          0.21         130.03 **
    Low
    High
  SRCB                         0.94           2.35
    Low
    High
  SRCB x CSR                   0.88           4.58 **
    SRCB low, CSR low
    SRCB low, CSR high
    SRCB high, CSR low
    SRCB high, CSR high
Philanthropic domain
  CSR                          0.51          35.02 **
    Low
    High
  SRCB                         0.94           2.21
    Low
    High
  SRCB x CSR                   0.96           1.34
    SRCB low, CSR low
    SRCB low, CSR high
    SRCB high, CSR low
    SRCB high, CSR high

                          Univariate Results

                              Evaluation

Independent variables     Mean   n        F
Environmental domain
  CSR                                 261.88 **
    Low                   2.30   37
    High                  6.44   38
  SRCB                                  1.11
    Low                   4.27   34
    High                  4.50   41
  SRCB x CSR                            9.28 **
    SRCB low, CSR low     2.81   19
    SRCB low, CSR high    6.13   15
    SRCB high, CSR low    1.77   18
    SRCB high, CSR high   6.64   23
Philanthropic domain
  CSR                                  70.75 **
    Low                   3.45   39
    High                  6.40   38
  SRCB                                  4.00 *
    Low                   5.52   36
    High                  4.37   41
  SRCB x CSR                            1.19
    SRCB low, CSR low     4.09   15
    SRCB low, CSR high    6.54   21
    SRCB high, CSR low    3.06   24
    SRCB high, CSR high   6.24   17

                           Univariate Results

                             Purchase intent

Independent variables     Mean   n       F
Environmental domain
  CSR                                 82.65 **
    Low                   3.12   37
    High                  6.19   38
  SRCB                                 4.62 *
    Low                   4.87   34
    High                  4.51   41
  SRCB x CSR                           3.61
    SRCB low, CSR low     3.79   19
    SRCB low, CSR high    6.24   15
    SRCB high, CSR low    2.41   18
    SRCB high, CSR high   6.16   23
Philanthropic domain
  CSR                                 30.71 **
    Low                   3.74   39
    High                  6.04   38
  SRCB                                 3.63
    Low                   5.45   36
    High                  4.36   41
  SRCB x CSR                           2.71
    SRCB low, CSR low     4.58   15
    SRCB low, CSR high    6.08   21
    SRCB high, CSR low    3.21   24
    SRCB high, CSR high   5.98   17

* p<.05; ** p<.01.

TABLE 5
Effects of CSR and Support for the Domain on Evaluation and Purchase
Intent (H5)

                                      MANOVA Results

Independent variables        Wilks' [lambda]       F
Environmental domain
  CSR                             0.18         160.20 **
    Low
    High
  Support for domain              0.96           1.32
    Low
    High
  Support x CSR                   0.78           9.63 **
    Support low, CSR low
    Support low, CSR high
    Support high, CSR low
    Support high, CSR high
Philanthropic domain
  CSR                             0.44          45.71 **
    Low
    High
  Support for domain              0.94           2.22
    Low
    High
  Support x CSR                   0.87           5.27 **
    Support low, CSR low
    Support low, CSR high
    Support high, CSR low
    Support high, CSR high

                             Univariate Results

                                 Evaluation

Independent variables        Mean   n        F
Environmental domain
  CSR                                    320.66 **
    Low                      2.30   37
    High                     6.44   38
  Support for domain                       2.25
    Low                      4.31   40
    High                     4.50   35
  Support x CSR                           19.51 **
    Support low, CSR low     2.87   22
    Support low, CSR high    6.07   18
    Support high, CSR low    1.47   15
    Support high, CSR high   6.77   20
Philanthropic domain
  CSR                                     91.03 **
    Low                      3.45   39
    High                     6.40   38
  Support for domain                       4.19 *
    Low                      5.14   39
    High                     4.68   38
  Support x CSR                            8.52 **
    Support low, CSR low     4.17   21
    Support low, CSR high    6.26   18
    Support high, CSR low    2.61   18
    Support high, CSR high   6.53   20

                              Univariate Results

                               Purchase intent

Independent variables        Mean   n       F
Environmental domain
  CSR                                    85.48 **
    Low                      3.12   37
    High                     6.19   38
  Support for domain                      0.16
    Low                      4.55   40
    High                     4.82   35
  Support x CSR                           7.88 **
    Support low, CSR low     3.56   22
    Support low, CSR high    5.76   18
    Support high, CSR low    2.47   15
    Support high, CSR high   6.58   20
Philanthropic domain
  CSR                                    35.82 **
    Low                      3.74   39
    High                     6.04   38
  Support for domain                      1.20
    Low                      5.01   39
    High                     4.73   38
  Support x CSR                           1.18
    Support low, CSR low     4.13   21
    Support low, CSR high    6.04   18
    Support high, CSR low    3.28   18
    Support high, CSR high   6.03   20

* p < .05; ** p < .01.

FIGURE 1
Analysis of Variance Results: Effects of CSR and Price on
Purchase Intent

                      Price

Purchase Intent    Low    High

Low CSR (Env.)     3.45   2.47
High CSR (Env.)    6.44   5.95

                      Price

Purchase Intent    Low    High

Low CSR (Phil.)    4.63   3.12
High CSR (Phil.)   6.67   5.52

Note: Table made from line graph.

FIGURE 2
Analysis of Variance Results: Effects of CSR and SRCB
on Evaluation of the Company

                                CSR

Evaluation of the Company   Low    High

Low SRCB (Env.)             2.81   6.13
High SRCB (Env.)            1.77   6.64

Note: Table made from line graph.

FIGURE 3
Analysis of Variance Results: Effects of CSR and Support for the Domain
on Evaluation of the Company

                                CSR

Evaluation of the Company   Low    High

Low Support (Env.)          2.87   6.07
High Support (Env.)         1.47   6.77

                                CSR

Evaluation of the Company   Low    High

Low Support (Phil.)          4.17   6.26
High Support (Phil.)         2.61   6.53

Note: Table made from line graph.

FIGURE 4
Analysis of Variance Results: Effects of CSR and Support for the
Domain on Purchase Intent

                                CSR

Purchase Intent             Low    High

Low Support (Env.)          3.56   5.76
High Supprt (Env.)          2.47   6.58

Note: Table made from line graph.