If China Makes a Better Product, Will Consumers Respond?

ABSTRACT - Previous research has found that people hold detailed country-of-origin impressions, but those impressions are not very robust. This research was designed to investigate how product trial would affect those expectations. Most important, the expectations were divided into three components-characteristics that were observable during trial, characteristics that were not observable, and price. Ratings of both observable and unobservable characteristics responded to trial, but price expectations did not.


Jan Hack Katz (1994) ,"If China Makes a Better Product, Will Consumers Respond?", in AP - Asia Pacific Advances in Consumer Research Volume 1, eds. Joseph A. Cote and Siew Meng Leong, Provo, UT : Association for Consumer Research, Pages: 43-46.

Asia Pacific Advances in Consumer Research Volume 1, 1994      Pages 43-46


Jan Hack Katz, Cornell University


Previous research has found that people hold detailed country-of-origin impressions, but those impressions are not very robust. This research was designed to investigate how product trial would affect those expectations. Most important, the expectations were divided into three components-characteristics that were observable during trial, characteristics that were not observable, and price. Ratings of both observable and unobservable characteristics responded to trial, but price expectations did not.

Each year, Americans buy billions of dollars of goods from the less industrialized countries of Asia (U.S. Dept. of Commerce, 1986), yet country-of-origin studies show that those countries induce low expectations in consumers minds. Lumpkin, Crawford and Kim (1985), for example, found that consumers believed that buying a product from China, Singapore, Korea, Taiwan, and Hong Kong was more risky than buying one from an industrialized country. Thorelli, Lim, and Ye (1989) selected Taiwan to signal moderate to low quality merchandise in their research.

The country-of-origin research has shown the construct to be remarkably detailed. Bannister and Saunders (1977) and Han and Terpstra (1988), among others, found that country stereotypes are specific at the product attribute level, with consumers distinguishing the technical advancement, workmanship, and appearance of source countries. Consumers also distinguish across product categories. Gaedeke (1973) found differences in country ranking between electronic items, textiles, and food products.

As Bilkey and Nes (1982) point out however, country-of-origin is often presented in studies as a lone cue, with consumers asked to judge abstract product scenarios (eg. How well made are Japanese cars?). Schooler (1971) and Schooler and Wildt (1968) used actual products, but the plain cloth, unmarked pen, and unadorned goblet that were used were chosen to permit "respondents to see in the products whatever characteristics they are predisposed to see," (Schooler, 1971; 75). The products were not really intended as ancillary cues for the consumers.

Under those conditions, it is no wonder that country-of-origin cues produce a significant effect. While the effect is often called a bias, it is really more a stereotype. Subjects given an abstract product scenario imagine products that they expect to come from the cued country-of-origin (Bilkey and Nes, 1982). They are therefore rating stereotypic products rather than objectively similar ones. In this research, subjects are asked to rate a real product, which a pre-test showed is not a neutral example of its class. The results here do measure country-of-origin bias and not the stereotype.

Studies that have provided additional information to country-of-origin found that the country-of-origin stereotypes are relatively weak constructs. They provide little explanatory power (Thorelli et al., 1989) and are moderated by other factors such as branding (Han and Terpstra, 1988), warranty, and retail channel (Thorelli, et al., 1989). Since a real product is rated in this study, we will be able to see whether a product experience that provides real product cues, also moderates the country-of-origin effect. If consumers correctly perceive the product and are not biased by the county-of-origin, then producers from countries with low country-of-origin stereotypes should work hard to induce trial as they improve their products. If country-of-origin continues to bias product evaluations, other marketing techniques that shift consumers' country-of-origin attitudes might be more effective in upgrading a national image.

Still, we must consider that product experience provides information on some product attributes, such as stylishness, but does not provide information on a host of other attributes, such as dependability. Country-of-origin studies show that consumer expectations are formed for both the observable and the unobservable product attributes. In the study, the impact of actual product experience on country-of-origin expectations of both types of product attributes is considered.

Another objective of this research project is to test the impact of product experience on expectations about price. While price can be considered a product attribute, economists would argue that it is determined only in part by factors intrinsic to the product. Product attributes that determine customer valuation generate the demand side of the price determination function. Other factors, including country-of-origin characteristics such as wage rates, affect the supply side of that function. If consumers do intuitively understand that price is determined by both demand and supply, then price expectations based on country-of-origin cues should be less affected by actual product experience than are other product attributes.


Research on expectations and their impact on later evaluations lead us to believe that country-of-origin expectations should affect ratings of actual products. Both the stereotype literature and the assimilation/contrast literature suggest that individuals manipulate information that they receive to align it more closely with their prior beliefs. This can be done through selective processing-ignoring or undervaluing stereotype inconsistent information-or through interpretation-reinterpreting information to make it more consistent (Bodenhausen, 1988).

In the stereotype literature, Weissbach and Zagon (1975) asked subjects to evaluate an individual who presented himself via videotape. In that study, evaluations were affected by subjects' prior induced belief about his sexual orientation. Darley and Gross (1983) found that subjects' assessments of a young girls academic skills were significantly lower if they were led to belief she was from a poor background than if they believed she was from a wealthy background.

In that research, subjects were presented with a real stimulus, such as a videotape of a person or a written dossier about the person. Subjects, then, had expectations, based on their own prior beliefs, and some cues about the individual being evaluated. Under those conditions, the expectations affected evaluations. Unfortunately, the stimuli were often designed to be neutral, and so, we are actually left with the same question as to whether additional cues were really given as was raised regarding the research by Schooler (1971) mentioned above.

The assimilation/contrast literature, while dealing with the impact of personal opinions rather than stereotypes, provides some suggestion as to what might happen with a non-neutral stimulus. In that work, subjects were presented with speeches or essays regarding a topic of social or political significance. If the material falls within the subjects' latitude of acceptance (general area with which they agree), they may perceive it as more similar and are more likely to respond to the material and update their attitudes (Hovland, et al., 1957; Sherif, et al., 1965). If the material falls within their latitude of rejection, attitude change will not occur (Eagly, 1981)

Since the country-of-origin literature has shown that the construct is relatively weak, probably including high variance in expectations and significant individual uncertainty, it seems that the country-of-origin construct should have a large latitude of agreement. In fact, a pre-test questionnaire that is described in the methods section shows that indeed individuals believe that there is high variance in the products of most countries and they have limited confidence in their own ratings of most countries' products. As a result, we expect subjects to perceive the product somewhat accurately, though perhaps assimilate product characteristics more toward a prior country-of-origin stereotype. [The assimilation/contrast literature also suggests that country-of-origin stereotypes might shift with product experience if products are within an expected range. If they are outside the expected range, however, the Weber and Crocker (1983) research suggests that consumers might create a subtype within the stereotype rather than shift the overall stereotype. This topic must be left to another research project as multiple product exposures were not used here.] This leads to the following hypothesis:

H1: Ex poste ratings of products should differ from ex ante expectations based on country-of-origin and be more similar to unbiased ratings (no c-o-o cue given) of the product.


Consumer research on category match suggests that, with most products, if consumers' ratings of observable characteristics respond to trial, so too should their ratings of unobservable characteristics. Sujan (1985) concludes that when new product information matches expectations based on category knowledge, cognitive processing occurs at a category level and judgement reflects a more global effect. In this condition, the consumer may not even be aware of missing data (Gardial and Biehal, 1990). As will be mentioned in the methods section, pre-test indicated that the product used in this study matched both the countries and the product category signalled. Therefore, the category match literature suggests that if the first hypothesis is upheld (ie. subjects' product evaluations reflect the objective product) then we also expect:

H2: Ratings of product characteristics that are not observable during limited product trial will move with ratings of observable characteristics.


Basic neo-classical economics suggests that in the presence of country-of-origin signals, price should not be fully linked to quality. In the neo-classical model, price is a function of both demand and supply (Samuelson, 1992). While demand, at a given price, is a function of consumer valuation, supply is determined by both production and producer characteristics. The production characteristics generally support a quality/price linkage, but the producer characteristics are quite independent. Countries, for example, vary in their wage rates, tax structures, and transportation costs. If consumers implicitly understand both the demand and the supply side of price determination, those country factors would make price expectations relatively stable in the presence of a country-of-origin cue. This suggests a third hypothesis:

H3: Price expectations will not be revised with quality ratings.


Pre-Test Survey on Countries-of-Origin

To select the countries and products to be used in this study, some information was needed. First, the assimilation/contrast literature suggested that country-of-origin bias would be moderated by actual product taste if the product was within the latitude of acceptance, in this case, the range of expectations that people would expect from a country. Earlier research had indicated that the country-of-origin construct was relatively weak, which suggests that the range is broad, but we wanted to test this. Second, for the category match literature, it was necessary to know that the product used was consistent with peoples' expectations about the countries.

To gather this information, 27 college students participated in a survey of country impressions. Each person was asked to indicate what they would expect, in terms of price and quality, from products of a range of countries. For each country, they were asked to rate not only the most likely quality and price, but also the range of expected quality and price. The two countries chosen, France and the Peoples' Republic of China, had different "most likely" scores, but both had a range for both quality and price of 3.2 points on a seven point scale (other countries, such as Japan did have significantly lower ranges of expectations). In the absence of extreme products (ie. extremely low or high quality), then, the product should be within the subjects' latitude of acceptance. This is further supported by ratings that respondents were asked to give about their confidence in their ratings. In the case of France, the average level of confidence was 4.0 on a seven point scale (1 = extremely confident, 7 = not at all confident). The average level of confidence for China was 4.5. Given the absence of complete confidence in their assessments, the latitude of acceptance is likely very wide.

In the same survey, subjects were asked to indicate products that they associated with those countries. Some respondents were unable to spontaneously provide any examples of products from the countries, but of those who did respond, food and food-related items were the most common response for France and China. The use of a food product then, will likely match subjects ideas of the country categories.

Stimulus Selection

Given the results of the pre-test, a food product that could conceivably be from either France or China had to be selected. Further, the product needed to be non-neutral within its general category. Jam was selected as the product category because it is actually imported into the United States from many countries, and so, was expected to be believable as an import from both France and China. [Test of the cover story following the experiment suggested that subjects did indeed believe the stated country-of-origin of the product.] Apricot jam specifically was chosen because a pre-test showed that these college students viewed strawberry jam as the typical, relatively neutral product, but apricot jam, with its distinctive flavor, color, and chunkier texture generated stronger sentiments. In this experiment, subjects were explicitly asked to give their opinion of the apricot jam that they were given; so, product trial did give them cues that were distinct from the country-of-origin cue.




A total of 130 university undergraduates participated in this study. Due to a food allergy, one subject was unable to complete the study and so, was dropped from the analysis. Subjects were randomly divided into five groups. Two of the groups were not given a product sample, and were asked to indicate what they expected of a jam from China(France). Ratings were made on six scales indicating three observable features (taste, color, and texture) and three unobservable features (hygiene, company responsiveness to complaint, and expected attractiveness of packaging). Subjects were also asked to report what they would expect to pay for such a product. They were given an anchor price of a jar of Smuckers apricot jam (a U.S. national brand of jam). Data from those two groups was used to generate the country-of-origin stereotype for France and China in this product category.

The other three (France, China, control) groups actually tried the product. Each person was given a sample of the apricot jam on an unsalted cracker. The samples were presented on a plate-no product packaging was shown. After trying the product, subjects were asked to rate it and write what they would expect to pay for the product, as above.


In the initial stage of analysis, two summary variables were generated. Data on the three observable dimensions (taste, flavor, texture) were summed to generate a variable called "observable" and data on three other dimensions (hygiene, company responsiveness, packaging) were summed to generate the variable "unobservable". The cost data was left as it was gathered. There are therefore two variables that indicate ratings of the product (observable, unobservable) and one variable describing subjects cost expectations. Analyses were done using MANOVA, including those experimental groups relevant to each analysis (for example, only those in the expectation condition). For reference during the discussion of the results, the rating means are presented above in Table 1.

Consistent with the idea that people hold country-of-origin stereotypes, differences in the ratings for France and China in the expectation condition were significant for the observable characteristics, the unobservable characteristics, and the price expectation (.05, .01, .0001 levels respectively). Not surprisingly given prior research on French country-of-origin effects, subjects expected the products from France to be better and to cost more.

The difference between the expectation and trial groups (France and China only) was statistically significant, but more detailed analysis shows that this is due primarily to the difference in observable and unobservable ratings between the expectation and trial groups for China (both significant at the .01 level). There was also a difference in the rating of observable characteristics for the French groups (.06 level of confidence). [While we cannot rule out the possibility that the French country-of-origin stereotype is stronger than that of China, it appears that the absence of more significant differences between the expectations and trial groups for France can be explained by the unfortunate coincidence that the expectation for France was very similar to the control group (i.e., no country-of-origin signal) rating of the product.] In the trial condition, there were no statistically significant differences among the three groups (France, China, control) in regard to the observable and unobservable characteristic ratings.

Those findings support Hypothesis 1 which states that ratings of actual products would differ from the country-of-origin stereotype and reflect the unbiased (ie. control) rating. The findings also support Hypothesis 2, which stated that the unobservable ratings would follow the observable ratings. Subjects appear to have accurately perceived the product and not to have been biased by country-of-origin expectations.

Unlike the ratings for observable and unobservable characteristics, cost expectations did not differ significantly between the expectation and trial conditions. Across the three trial groups (France, China, control), the cost variable was significant at the .10 level. Further investigation of that cost effect shows that it is driven by a significant (<.05) difference between the price for the French and control groups. This indicates that while subjects responded to the objective taste of the apricot jam, they still expect certain price levels from certain countries. This provides support for hypothesis three, which was based on the neo-classical economics assumptions that individuals implicitly understand both the demand and the supply side of the price equation.


Support for all three hypotheses developed in this paper were found. Country-of-origin stereotypes for observable characteristics appear to be moderated by actual product trial. The category match literature suggested that the impact of product trial would spill over to unobservable characteristics and in this study, with a category match product, that was found to be true. These findings suggests that companies from countries with a poor image should try to induce trial to cause consumers to form an accurate perception of their product.

Schooler and Wildt (1968) found that price reductions did cause people to try products from countries with a poor image. In that study, the price reductions were seen to compensate either for poorer quality or greater risk associated with products from poor image countries. In this study, the trial condition rating suggests that even when consumers perceive quality and risk to be comparable across products, they still expect price differences. A strategy of using price reduction to increase trial would need to respond to expected price from the country-of-origin rather than a price based on the inherent quality of the good.

This study's results beg the question as to whether producers from poor image countries can command prices comparable to those of equal quality products from better image countries. A neo-classical economics explanation would say that they could-expectations should not affect consumer satisfaction, given a correct appraisal of product quality. Behavioral explanations are more complicated, depending on the timing of price and country-of-origin information, and so, depending on the expectation formed. Future research could certainly investigate these points.

Another interesting question suggested by this study is whether product trial leads to an updating of the country-of-origin stereotype itself. Crocker and Weber (1983) investigated the updating of social group stereotypes and found that updating does occur if the new information differs from the stereotype along a limited number of dimensions. If there is a radical difference from the stereotype (a tea-totalling fraternity member who works hard on classes) then subtyping occurs, whereby that individual is placed in their own category (eg. members of XYZ fraternity), and the overall stereotype (fraternity members) remains intact. Once again, future research with repeated product exposures and stereotype measurements could investigate those questions.

This study, then, answers some questions that have been raised by prior work in the country-of-origin literature. Product trial does moderate consumers country-of-origin bias. Under some conditions, at least, that moderation will extend to unobservable characteristics. Price expectations based on country-of-origin, however, are more stable and are not moderated by product trial. In addition to these answers, however, the work also generates several questions, showing that this area of research still has substantial room for growth.


Bannister, J.P. & Saunders, J.A. (1977). UK Consumers' Attitudes Towards Imports: The Measurement of National Stereotype Image, European Journal of Marketing, 12(8): 562-570.

Bilkey, Warren J. & Nes, Erik (1982). Country-of-Origin Effects on Product Evaluations, Journal of International Business Studies, Spring/Summer: 89-99.

Bodenhausen, G.V. (1988). Stereotypic Biases in Social Decision Making and Memory: Testing Process Models of Stereotype Use. Journal of Personality and Social Psychology, 55:726-737.

Darley, J.M. & Gross, P.H. (1983). A Hypothesis-Confirming Bias in Labeling Effects, Journal of Personality and Social Psychology, 44:20-33.

Eagley, A. H. (1981). Recipient Characteristics as Determinants of Responses to Persuasion, in Cognitive Responses to Persuasion, R.E. Petty, T.M. Ostrom, and T.C. Brock (eds.) (Erlbaum; Hillsdale, NJ)

Gaedeke, Ralph (1973). Consumer Attitudes Toward Products "Made In" Developing Countries, Journal of Retailing, 49(2):13-25.

Gardial, S.F. and Biehal, G. (1987). Measuring Consumers' Inferential Processing in Choice, Advances in Consumer Research XIV, M. Wallendorf and P.F. Anderson (eds.) (ACR: Provo, UT), 101-105.

Han, C. Min & Terpstra, Vern (1988). Country-of-Origin Effects for Uni-National and Bi-National Products, Journal of International Business Studies, Summer:235-255.

Hovland, C.I., Harvey, D.J., & Sherif, M. (1957). Assimilation and Contrast Effects in Reactions to Communication and Attitude Change, Journal of Abnormal and Social Psychology, 55: 244-252.

Lumpkin, J.R., Crawford, J.C., & Kim, G. (1985). Perceived Risk as a Factor in Buying Foreign Clothes, International Journal of Advertising, 4:157-171.

Samuelson, Paul (1992). Economics (McGraw-Hill: New York).

Schooler, Robert (1971). Bias Phenomena Attendant to the Marketing of Foreign Goods in the U.S., Journal of International Business Studies, Spring: 71-80.

Schooler, Robert & Wildt, Albert (1968). Elasticity of Product Bias, Journal of Marketing Research, 5:78-81.

Sherif, C.W., Sherif, M., & Nebergall, R.E. (1965). Attitude and Attitude Change: The Social Judgment-Involvement Approach (W.B. Saunders; Philadelphia).

Sujan, Mita (1985). Consumer Knowledge: Effects on Evaluation Strategies Mediating Consumer Judgements, Journal of Consumer Research, 12:31-46.

Thorelli, Hans B., Lim, Jeen-Su, & Ye, Jongsuk (1989). Relative Importance of Country-Of-Origin, Warranty and Retail Store Image on Product Evaluations, International Marketing Review, 6(1):35-46.

United States Department of Commerce, Bureau of the Census (1986). Highlights of U.S. Export and Import Trade, (U.S. Government Printing Office: Washington, D.C.).

Weber, Renee & Crocker, Jennifer (1983). Cognitive Processes in the Revision of Stereotypic Beliefs, Journal of Personality and Social Psychology, 45(5):961-977.

Weissbach, T.A. & Zagon, G. (1975). The Effect of Deviant Group Membership Upon Impressions of Personality, The Journal of Social Psychology, 95:263-266.



Jan Hack Katz, Cornell University


AP - Asia Pacific Advances in Consumer Research Volume 1 | 1994

Share Proceeding

Featured papers

See More


When Prominent Logos Make You Feel Competent

Sara Caprioli, Technical University of Munich
Christopher Fuchs, Technical University of Munich
Nailya Ordabayeva, Boston College, USA

Read More


L3. Categorizing Engagement Behaviors from the Perspective of Customer Resources

Xianfang Zeng, University of Calgary, Canada
James Agarwal, University of Calgary, Canada
Mehdi Mourali, University of Calgary, Canada

Read More


Consumer Response to Innovations: The Differential Effects of Focused and Defocused Attention on Perceived Novelty, Usefulness and Symbolism

Katarina Hellén, Univeristy of Vaasa
Maria Sääksjärvi, Delft University of Technology, The Netherlands

Read More

Engage with Us

Becoming an Association for Consumer Research member is simple. Membership in ACR is relatively inexpensive, but brings significant benefits to its members.