Consumer Perceived Value of Salon Hair Coloring in China

The primary objective of this study is to examine how Chinese consumers perceive the value of salon hair coloring? Based on the findings from a qualitative study and literature review on consumer perceived value (CPV), we developed a survey to measure Chinese consumers’ perceptions of quality value, emotional value, epistemic value, social value, monetary cost, time/effort cost, and health risk cost of salon hair coloring. Quality value and monetary cost were found to be the most important dimensions for Chinese consumers. The results suggest that the model for CPV proposed in this study could be beneficially applied to other areas, such as to other product/service categories, other demographics, and in other cultural settings.


Introduction
In recent years, consumer perceived value (CPV) has gained significance in business research due to its ability to predict buyers' intentions during transaction and the importance of managing customer relationships. There is an extant literature on CPV across different products/industries. Yet, as our literature review documents, there is a paucity of CPV studies on the appearance-related product category, such as beauty and personal care products. In addition, Chinese consumers warrant more attention in consumer research to reveal their preferences: As China continues to increase its economic power in the global market, it is essential for business to shed light on the consumption behavior of its consumers in order to better serve this market. Therefore, the aim of this research is to discover Chinese consumers' CPV components as they associate with an appearance-related service category, namely salon hair coloring.

Consumer Perceived Value (CPV)
Value is a dynamic abstract concept with meanings that vary according to context (Woodruff, 1997). How a consumer perceives value in the market place, i.e. consumer perceived value -attempts to assess the net value a consumer sees he/she reaps when weighing what is obtained against what has to be forfeited in a marketplace exchange. Apart from the obvious importance from a consumer's standpoint, perceived value is of significance to marketing scholars since it may alter the attitudinal direction (satisfied/dissatisfied) of consumers and lead to different behaviors, as well as different behavioral consequences (give-up/purchase/repurchase). This requires identifying the components of a consumer's experience during the purchasing process of a product or service and estimating how these components weigh positively or negatively into the consumption experience. Increasingly, consumption situations are harder to assess, which may install perceived value for consumers as a key tool for firms to compete (Woodruff, 1997). Yet, the literature review on consumer value shows that overall theory building has not been strong in the CPV area (Sanchez-Fernandez & Iniesta-Bonillo, 2009). In addition, methodological hurdles remain high. Sometimes, it can be easily confounded with quality, benefits, and price. As a consequence, there are no generally accepted measurement scales for consumer value (Janawade, 2013). curiosity, offer novelty or satisfy a desire for knowledge. Conditional value refers to the situation in which the value judgment is made. Sheth and colleagues (1991) applied their theoretical approach to three different cigarette consumption cases. The results suggested possible uses of this framework to predict consumption behavior. Moreover, their work provided a strong foundation from which to build a perceived value scale (Sweeney & Soutar, 2001). Tam (2004) applied the concept of consumer perceived value to the restaurant service industry in Hong Kong, proposing an integrative model to examine the relationships among perceived value, consumer satisfaction, and post-purchase behavior. The study hypothesized that perceived value directly influences both consumer satisfaction and post-purchase behavior and indirectly influences post-purchase behavior via consumer satisfaction as well. The model was tested by survey, and the statistical results showed that in terms of influencing post-purchase behavior, perceived value had both a direct and indirect effect, while consumer satisfaction only had a direct effect. Furthermore, perceived value was found to have a greater weight than consumer satisfaction in determining post-purchase behavior. Lai et al. (2009) examined a model that incorporated service quality, perceived value, corporate image, customer satisfaction, and customer loyalty in China at the telecom setting. The statistical result confirmed with Tam (2004) that perceived value did have a both a direct as well as indirect (through customer satisfaction) effect on post-purchase behavior, i.e. customer loyalty in this case.
However, findings in CPV are still limited due to the small number of studies that have been conducted and majority of the research has been done in the United States or other developed nations. The consensus of the extant studies is that CPV has a trade-off structure between benefits and costs, which is more complex than just the difference between a single measure of quality and price. In addition, extant research shows that consumers' perceptions of value vary across the types of products investigated/services provided, which means the measurement of CPV in different settings varies accordingly or there is need to develop a general measure of CPV for all circumstances. Some of the findings suggest that CPV is even more important than consumer satisfaction in deciding behavioral intention/consequences (Tam, 2004;Lai et al., 2009). Thus, it deserves expanded application across a wider range of products and service as well as differing contexts (Swait & Sweeney, 2000), such as in China where the economic condition as well as cultural setting are different from the western world.

Appearance Related Consumer Research
One popular way of changing physical appearance has been the application of beauty and personal care products. Research has shown that the wearing of cosmetics by women in Western culture has been under discussion for decades (Tseelon, 1995). It was found that cosmetics can be used not only to help achieve physical attractiveness, but also to make people feel good about themselves, which is especially true for women (Graham & Jouhar 1982). However, a review of the literature reveals that little consideration has been given to appearance-related product categories, such as hair care products, cosmetics, jewelry, and clothing, especially for Chinese consumers. It was found that in modern China, women enjoy their power in making decisions about their appearance and roles. On the other hand, they feel the anxiety and pressure to keep the status quo (Afshar, 1991). Given the importance of hair as a vehicle of adornment and building social advantage, it is important to understand how consumers perceive the value of hair coloring. Furthermore, not enough attention has been paid to male consumers in this consumption area although it has become more obvious that male consumers make up an important potential market for appearance-related products, such as cosmetics (Souiden & Diagne, 2009).
This study therefore tries to address the following question: What are the components of perceived value that influence Chinese consumers' overall perceptions of value for salon hair coloring the most? How important is what they give versus what they receive? In particular, we used in-depth interviews to find out what Chinese consumers perceive as value in salon hair coloring and its consumption. We then developed a conceptual model of CPV for salon hair coloring based off the Zaithaml (1988) framework, which is discussed in the next section of Methodology. In the end, the results of our analysis are presented and followed by a summary with conclusion and some suggestions for future research.

Methodology
In order to explore our research question appropriately, this study consisted of a qualitative stage using in-depth interviews and a quantitative stage using the survey technique (Zhao, 2014). Data collection for the qualitative study was conducted by pre-arranged phone calls with Chinese informants. Overall, eight people agreed to participate in the study and were interviewed. While eight may seem to be a small number for each group, Griffin and Hauser (1993) among others, have found that focus groups or one-on-one interviews including seven www.ccsenet.org/ijms International Journal of Marketing Studies Vol. 6, No. 6;2014 to ten people uncover the vast majority of needs and concerns. Furthermore, a review of the interview records suggested a recycling of ideas, or saturation, indicating the likelihood that little new information would emerge from further interviews.
Overall, eight recurring topical patterns emerged that were common across informants in each group, which could be matched into benefits and sacrifices categories. The results from those interviews are summarized in Table 1. The interviews revealed that Chinese participants seemed not to care about monetary cost. At the same time, they were much more worried about potential health risks. These preliminary results were compared to previous findings in the literature to develop a set of hypotheses that could be tested quantitatively.
To generate the Hypotheses, a conceptual model for this study was developed by adopting Zeithaml's (1988) classification and insights to classify the consumer perceived value for salon hair coloring into benefits and sacrifices (see Figure 1). It was predicted that the dimensions/constructs in the "benefits" category should have a positive effect on CPV, while the dimensions/constructs in the "sacrifices" category should have a negative effect on CPV. Overall, the model utilizes the six dimensions from the literature review that were closely aligned with the recurring topical patterns that emerged from the interview data: quality value, emotional value, epistemic value, social value, monetary costs, and time/effort costs. The model also includes health risk costs, a theme that represents a topical pattern identified in the qualitative study, but not found in the literature.

Survey Constructs and Measurement Items
We derived our constructs and item measurements from the extant literature in combination with the results obtained from our qualitative study. In total, we examined eight constructs in this study, seven of which are exogenous latent constructs (quality value, emotional value, epistemic value, social value, monetary cost, time/effort cost, health risk); and one is an endogenous latent construct, namely consumer perceived value (CPV).
Quality Value: Quality value is a product or service's overall excellence or superiority as judged by the consumer (Zeithaml, 1988). We measure quality value using Petrick's (2002) Quality Scale. It was developed as a multi-dimensional scale for gauging consumer perceived value of a service. Survey respondents were asked to rate statements such as "salon hair coloring is outstanding quality." This quality scale is a four-item scale and has a composite reliability score of 0.79, which indicates that the scale can reliably measure the construct of interest (Fornell & Larcker, 1981).
Emotional Value: Emotional value is a product or service's ability to generate utility through feelings or affective states (Sweeney & Soutar, 2001). We quantify the emotional value using Sweeney and Soutar's (2001) five-item scale called the Emotional Value Scale. Survey respondents were asked to rate statements such as "salon hair coloring is one that I would enjoy." They originally developed this multiple item scale for measuring consumer perceived value of a durable good at the brand level. Despite the fact that the scale was originally designed for a product, not a service, it also performs suitably to capture the essence of the emotional value of a service. At a value of 0.94 in the study on durable goods, the scale exhibited a high level of composite reliability.
Epistemic Value: Epistemic value is a consumer's experienced curiosity, novelty, or knowledge obtained from the consumption of a product or service (Sheth et al., 1991). We determine epistemic value using Pura's (2005) Epistemic Value Scale, which was originally designed to study consumer perceived value of mobile phone services. Survey respondents were asked to rate statements such as "I use salon hair coloring to test the new looks". This phrase frequently emerged in the in depth interviews of the qualitative study, as informants wanted to try "new looks" using hair coloring. Cronbach's alpha for this three-item scale is 0.80, which indicates good reliability.
Social Value: Social value is the utility a consumer experiences if the product or service enhances the consumer's social self-concept (Sweeney & Soutar, 2001). We gauge social value by using Sweeney and Soutar's (2001) Social Value Scale. This four-item scale was also developed for their previously mentioned study assessing consumer perceived value of a durable good at the brand level. Survey respondents were asked to rate statements such as "salon hair coloring would help me to feel acceptable." As the focus of this study in on adornment, this scale appears to be appropriate for hair coloring. At a value of 0.82 in the original study on durable goods, the scale exhibited an acceptable level of composite reliability. Monetary Cost: Monetary cost is the total financial price of a service a consumer pays (Petrick, 2002). We measure monetary cost with Sweeney and Soutar's (2001) Price/Value Scale. This four-item scale was also developed for their previously mentioned study assessing CPV of a durable good at the brand level. Survey respondents were asked to rate statements such as "salon hair coloring is reasonably priced." At a value of 0.80 in the original study on durable goods, the scale exhibited a satisfactory level of composite reliability.

Time/Effort Cost:
Time/effort cost is a consumer's perceived cost of the time required to experience hair coloring (Rindfleisch & Crockett, 1999). Informants pointed out in the preliminary research that the time/effort cost of salon hair coloring included the trip to the salon, possible waiting time as well as the time spent on the actual hair-coloring process. We measure time/effort cost using Rindfleisch and Crockett's (1999) Time Risk Scale, which they developed for their study of perceived risk of smoking. Survey respondents were asked to rate statements such as "I would waste a lot of time by having salon hair coloring." Cronbach's alpha for this five-item scale is 0.81, which indicates good reliability.

Health Risk Cost:
Health risk cost is a consumer's perceived risk that hair coloring may negatively impact the consumer's physiological health (Rindfleisch & Crockett, 1999). Previous research (Trueb, 2005) and informants in the preliminary qualitative study indicate that the health risk of hair coloring consumers perceive includes unhealthy hair, allergy, headache, and cancer. We measure health risk using Rindfleisch and Crockett's (1999) Health Risk Scale developed for the previously mentioned study on the perceived risk of smoking. Survey respondents were asked to rate statements such as "salon hair coloring would get me cancer." Cronbach's alpha for this five-item scale is 0.93, which indicates high reliability. We exchanged "cancer" with "unhealthy hair," "headache," and "allergy" in variations of the original question.

Consumer Perceived Value (CPV): CPV is a consumer's overall judgment of the net value of what is received
and what is given up for the consumption of a product or a service (Zeithaml, 1988). We measure CPV using Lin, Sher, and Shih's (2005) Sacrifice Scale. This three-item scale was developed to gauge the value of eTail services in Taiwan. Survey respondents were asked to rate statements such as "Compared with the price you paid, salon hair coloring provides good service value." In the original study, the authors only reported that this scale had a composite reliability score greater than the acceptable level of 0.70. This indicates that the scale is reliable measuring CPV (Fornell & Larcker, 1981).

Hypotheses
Based on the in-depth interviews from the qualitative research, the Chinese informants appeared to have a strong preference for salon hair coloring instead of hair coloring done at home. The reasons expressed for this preference included salons being better equipped, salon personnel creating better outcomes, personnel providing expert suggestions as needed, avoiding damage to the hair, and avoiding the mess of doing it at home. The study informants expressed clearly that they felt more secure about the results from salon hair coloring than doing it on their own. Therefore, it appeared that the quality value dimension of CPV is the dominant value dimension. Based on the above, it is hypothesized that: H1: Quality value will be the most influential value for Chinese consumers relative to their perception of salon hair coloring.
On the other hand, the number one concern for them was the health risks associated with hair coloring. Informants' concerns included worry about the potential of skin cancer, brain cancer, headaches, allergies, and harm to their hair. They expressed clearly that should any validated harm take place, they would stop hair coloring immediately. Based on the preliminary research finding, it is postulated that: H2: Health risk cost will be the most influential cost for Chinese consumers relative to their perception of salon hair coloring.

Data Collection and Analysis
College students were selected as a convenience sample. A review of the literature shows that a number of studies on consumer issues have involved using college students as subjects in the research (Peterson, 2001;Valentine & Powers, 2013;Shim & Maggs, 2005;Lachance & Choquette-Bernier, 2004). 700 surveys were distributed in three universities in a provincial capital city in western China, out of which 249 were completed and used for the final data analysis. The two hypotheses were tested by Confirmatory Factor Analysis (CFA) in Structural Equation Modeling (SEM). The Software package used to obtain our results was LISREL 8.
Since many statistical procedures used in this study assume that data are normally distributed, we reviewed all variables to examine possible departures from normality and to check for univariate outliers. We used SPSS to calculate descriptive statistics of each variable as well as correlations between variables to describe the profile of www.ccsenet.org/ijms International Journal of Marketing Studies Vol. 6, No. 6;2014 our sample. If the absolute value of the skewness coefficient or (excess) kurtosis is larger than one, this indicates that a variable may not be normally distributed and deserves further analysis. For these variables, the Kolmogorov-Smirnov test was performed, which compares the observed cumulative distribution function for a variable with a specified distribution, i.e. a normal distribution. If the value of the Kolmogorov-Smirnov's statistic is significant, this implies that the variable in question is not approximately normally distributed and could not be used in further statistical analyses. All statistical tests were considered significant at an alpha level ≤ 0.05.
We first collapsed the original variables used in the survey into composite variables as the hypothesis related to composite constructs rather than individual measurement items. We assessed all eight composite constructs in the model by first testing their validity and reliability. All those tests delivered satisfactory results. In particular, validity is the extent to which a scale measures the concept under discussion (Hair, Anderson, Tatham, & Black, 1998). There are multiple options to assess the validity of each construct, such as face validity, convergent validity, and divergent validity. We established the face validity of constructs in this study with the results from the survey pre-test. Face validity is confirmed when the survey constructs appear to the researcher logically and measure what they are supposed to measure. Reliability on the other hand is defined by how consistently different measurements of a variable reflect the construct of interest (Hair et al., 1998). We examined reliability by calculating Cronbach's alpha in SPSS for each construct. The threshold point set for the study's eight constructs was a Cronbach's alpha criterion value of 0.7, as suggested by Nunnally (1978). Once requirements for validity and reliability are met, we calculated the mean value of each construct as the composite measure (Hair et al., 1998). Finally, we calculated descriptive statistics including frequency distributions, means, standard deviation, skewness, and kurtosis for each construct. Except for the construct of epistemic value (EPV) in the sample having a lower reliability of 0.63, all the other constructs showed high internal reliability, with an alpha coefficient greater than 0.80. The measurement scale for the epistemic value construct was adapted from Pura's (2004) study conducted in Finland, where the scale demonstrated a good reliability of 0.80. Although Cronbach's alpha for the EPV construct fell below Nunnally's 0.7 criterion, taking into account the exploratory nature of this research, a Cronbach's alpha of 0.63 is considered to be acceptable (Hair et al. 1998, p.118).

Confirmatory Factor Analysis (CFA)
A Confirmatory Factor Analysis in SEM was conducted to investigate the overall fit of the proposed conceptual model for the sample.  Figure 2 reports the fit of the proposed model for the sample. The diagram links the effect of the seven exogenous latent variables to the endogenous latent variable, where the numbers along the arrows represent the path coefficients. The diagram shows that all exogenous latent variables exhibited the predicted influence on CPV except for time/effort cost (TC). Specifically, the effects of quality value (QV), emotional value (EV), epistemic value (EPV), and social value (SV) on CPV were all positive as predicted and reflect the "benefit" constructs that would be anticipated for CPV. Conversely, the effects of monetary cost (MC) and health risk cost (HC) were negative as predicated and reflected the "sacrifice" constructs that would be anticipated for CPV. Note that time/effort cost (TC) had a positive effect on CPV for this sample, which was not consistent with the prediction and deserves to be further investigated.

Figure 2. Path diagram of the proposed model
In a final step, we need to discuss whether our initial hypotheses were supported or rejected. Recall that Hypothesis 1 states that "Quality value will be the most influential value for Chinese consumers relative to their perceptions of salon hair coloring," which means we would expect that quality value (QV) would have the largest absolute path coefficient value among the "benefits" category. As shown in Figure 2, the absolute value for the path coefficient of QV was 0.22, which was larger than the absolute values for the path coefficient of EV, EPV, and SV in the "benefits" category, although not by much. Therefore Hypothesis 1 was supported by the data analysis: quality value appeared to be the most influential value for Chinese consumers.
Hypothesis 2 predicts that "Health risk cost will be the most influential cost for Chinese consumers relative to their perceptions of salon hair coloring. In other words, health risk cost (HC) would have the largest absolute path coefficient value in the "sacrifices" category. As shown in Figure 2, the absolute value for the path coefficient of HC was 0.08, which was actually smaller than the absolute values for the path coefficient of MC and TC in the "sacrifices" category. Therefore Hypothesis 2 was not supported by the data analysis: monetary cost (MC) appeared to be the most influential cost for Chinese consumers.

Conclusion and Implication
The purpose of this study was to gauge consumers' perceived value of salon hair coloring in China. Pursuing this research is important for several reasons. From a consumer perspective, the hair coloring category is a product category growing in importance both for older and younger consumers: Given that more women and men today are turning to hair coloring to cover their gray and to remain youthful-looking, this is an important and growing market segment. Moreover, manufacturers provide fashionable, strong, and bright colors to tempt the consumers into experimenting with hair color (Moore, 2005). However, little research has focused on hair coloring and how consumers value it.
This study is also important from an academic perspective. Theoretically, the findings of this study have added to the body of knowledge in CPV by demonstrating that Zeithaml's (1988) classification of consumer perceived value into benefits and sacrifices seemed to be applicable to the case of salon hair coloring. Results from both the qualitative and the quantitative research suggested that consumers perceive hair coloring as a tradeoff www.ccsenet.org/ijms International Journal of Marketing Studies Vol. 6, No. 6;2014 between what they "receive" (i.e. quality, emotional value, epistemic value, and social value) and what they have to "give up" (i.e. health risk, monetary cost, and time/effort). Furthermore, this study suggested that the model which was originally developed in a Western cultural setting, i.e., the United States, could also be applied to an Eastern cultural setting, i.e., China.
A methodological contribution of this study is the adaptation of three scales, including the scales for epistemic value (EPV), time/effort costs (TC), and health risk costs (HC), which were not originally developed for consumer perceived value studies. The three scales proved to have satisfactory internal reliability (Cronbach's alpha between 0.81 and 0.92) except for the application of the EPV scale in this case (Cronbach's alpha being 0.63). It could be that the EPV scale's low reliability is due to a ceiling effect, an effect whereby data cannot take on a value higher than some "ceiling," that is respondents may have given consistently higher responses on this construct resulting in reduced variation. A possible solution would be to add more items to this scale to improve its reliability. Overall, the high performance of theses adapted scales not only showed the credibility of the translation procedure used in this study, but also added possible tools to examine the role of EPV, TC, and HC further in subsequent studies. In addition, due to the establishing of partial invariance and the strong Cronbach's alpha levels for the constructs, the set of scales used in this study may prove helpful for future cross-cultural studies that are concerned with consumer perceived value.
For this particular sample, six out of the seven dimensions of CPV seemed to have similar levels of influence on CPV given that their path coefficients were close to one another in absolute value (i.e. between 0.22 and 0.15), with the only exception being the path coefficient for health risk cost at 0.08. From a research perspective, this finding from the SEM results is novel and thus important as no research comparing the relative impact between individual CPV dimensions has been available in the literature. By the same token, the path coefficient for the time/effort cost (TC) turned out to be positive, which contradicts the expected negative impact that a "sacrifice" construct should have on CPV. One possible explanation for this discrepancy was that for the Chinese consumers, the higher they thought the time/effort costs to be, the better service they thought they were getting, and possibly the higher value they perceived relative to salon hair coloring. As a result, the time/effort costs may have appeared to them as having a positive influence on CPV, which deserves further investigation in future research.