1. Introduction
Online reviews, as one of the manifestations of online word-of-mouth, are not only an important way for people to obtain product information, but also affect consumers’ purchasing behavior. In recent years, the role of online reviews has attracted great attention from the academic circle. Scholars have explored the influence of textual reviews on consumers’ psychology and behavior, mainly focusing on the quality of reviews [1] [2] , the number of reviews [3] [4] , and the reviews’ valence [5] [6] . In addition, scholars’ research had also found different effects of picture reviews and textual reviews in specific situations. Yang and Zhu (2016) found that consumers with high purchase feasibility were more influenced by picture reviews than textual reviews [7] . Zhang and Li (2016) found that picture reviews of search products can effectively enhance the usefulness of online comments [8] . Lin, Lu and Song (2017) found that the picture reviews of sensory products had a greater impact on consumers’ purchase intention than the textual reviews [9] . Previous researches had deeply discussed the influences of textual reviews on consumers, and compared the different functions of picture reviews and textual reviews, but rarely discussed the influence of the characteristics of picture reviews on consumers’ purchase decisions and their psychological mechanism.
Clothing belongs to the type of sensory products, but consumers can’t direct contact with the product itself in the online shopping environment. In addition, textual online reviews cannot clearly describe product information and consumer experience, so picture reviews are used to provide consumers with product information to make up for the deficiency of textual reviews [10] . This research will explore the influence of the physical attractiveness of the people in the online picture reviews on the psychology and the clothing purchase decision of female consumers, so as to provide suggestions for the clothing sellers in the e-commerce platform to manage the reviews’ information.
In the following part of this paper, three hypotheses are proposed through theoretical derivation. Then, two experiments were conducted to verify the hypothesis. Finally, we will discuss the research results and propose future research prospects.
2. Hypothesis
Physical attractiveness is a specific image that enables one to experience pleasure, which contains many contents, such as appearance, body shape, clothing, hairstyle [11] . Research on product aesthetics has proved that visually appealing products can trigger consumers’ positive responses on their emotion, cognitive and even neural level [12] [13] . Scholars found that highly attractive models can bring better results to appearance-related products (e.g., shampoo, deodorant, underwear), and highly attractive models were more professional than normally attractive models, so as to improve the advertising effect of enhancing beauty products (e.g. lipstick, jewelry) [14] [15] .
When shopping online, consumers are more likely to be attracted by pictures because they cannot have direct contact with products, and transfer the positive emotions caused by pictures to products and brands [16] [17] . As a result, picture reviews that include highly attractive person (HAP) will trigger consumers’ feelings of pleasure and positive emotions and ultimately increase their willingness to buy. Therefore,
H1: Compared with the normally attractive person (NAP), the picture reviews that include HAP can arouse women’s higher clothing purchase intention.
Body image refers to the mental image formed by individuals of their own weight, body shape and body [18] . In the online shopping environment, consumers cannot directly appreciate, touch and try on clothing products like the traditional shopping process. They can only imagine their shopping experience in their mind based on the existing information and form a general image. Clothing is called “visible self” and “second skin” of human beings [19] .
When shopping online for clothes, consumers’ brains will form images related to their bodies, namely body image, and judge the clothing products. At the same time, consumers’ mental image of their bodies will directly affect their feelings and emotions, and then they will transfer these feelings and emotions to clothing products. When consumers’ body image is positive enough, it will promote their buying behavior. Conversely, when consumers’ body image is negative, their buying behavior will be reduced. That is to say, as a kind of product information, the physical attractiveness of the person in the picture reviews will affect the consumers’ body image and further influence their buying behavior. Therefore,
H2: Body image was a mediator in the relationship between person’s physical attractiveness in picture reviews and consumers’ purchase intention.
The state body image of an individual is easily influenced by the mass media, for the reason that the individual will compare herself with the model image presented by the media and make a comparative evaluation of her own body [20] [21] , which is related to social comparison in the field of appearance.
Perceived similarity refers to the similarity between the commentator’s characteristics and the reader’s own [22] [23] . According to the selective accessibility (SA) model proposed by Mussweiler (2003), the effect of information comparison is the result of the hypothesis testing process [24] . If the initial evaluation results show that they are similar to the object of comparison, the judge will conduct a similarity test. On the contrary, if the initial evaluation indicates that they are different from the object of comparison, the judge will carry out the dissimilarity test. Similarity test produces assimilation effect, while dissimilarity test produces contrast effect [24] [25] [26] .
Therefore, when consumers are faced with pictures reviews that contain characters similar to themselves, HAP will trigger consumers’ more positive body image (assimilation effect), thus promoting consumers’ purchase intention. When consumers are faced with pictures reviews containing figures that are not similar to them, HAP will cause more negative body image (contrast effect) for consumers, thus reducing consumers’ purchase intention. Therefore,
H3: The influence of person’s physical attractiveness in picture reviews on consumers’ purchase intention was moderated by perceived similarity.
To test our hypothesis, we conducted two experiments in which participants were asked to rate their attitudes. In study 1, striped skirt was used as the experimental material to manipulate the person’s physical attractiveness, and to test the influence of physical attractiveness on consumers’ purchase intention and the mediating effect of body image. In study 2, the striped skirt was also used as the experimental material to manipulate the physical attractiveness and perceived similarity of the persons, and to test the interactive influence of the physical attractiveness and perceived similarity on the purchase intention.
3. Study 1
The study consisted of a 2 (physical attractiveness: NAP vs. HAP) between-subjects experimental design, to test whether HAP in the picture reviews would trigger a higher purchase intention of female consumers, and to explore the mediating role of body image, namely to verify whether hypothesis H1 and H2 are true.
3.1. Pretest
In order to verify the applicability of the manipulated material for physical attractiveness, we conducted a pretest. This study selects the striped skirt as the experimental material. We selected a picture displayed by consumers in the product reviews as the prototype and processed the picture with reference to previous experimental studies [14] [27] , then obtained a pair of pictures with NAP or HAP, as shown in Figure 1.
This study focused on the effect of the physical attractiveness of the people in the picture reviews on female consumers, and the selected experimental materials were also targeted at female consumers, so only female were invited to participate in the study [15] . Sixty-six female students who volunteered to take part in the study for a fee were randomly divided into two groups and asked to look at either picture include NAP or HAP. After viewing the picture material, the
![]()
![]()
(a) (b)
Figure 1. Experimental material. (a) HAP; (b) NAP.
subjects were asked to complete the physical attractiveness scale [28] . The subjects in the HAP group rated the physical attractiveness of the characters (M = 5.06, SD = 0.54) significantly higher than those in the NAP group (M = 4.11, SD = 0.75, t = −5.953, p < 0.001), indicating that it was appropriate to use the pictures as manipulation materials for the physical attractiveness.
3.2. Subjects and Procedure
The main study consisted of 110 female students who volunteered to participate in the study for the experimental reward. Their average age was 20.95 years (SD = 2.27). All of them had the experience of buying clothes online and were paid 5 RMB after the experiment.
The experimental process was completed on the mobile phone screen. We imitate the product page of the e-commerce platform to display pictures. First, show the pictures of the pure clothes (excluding people) of the striped skirt to the participants. Subsequently, the subjects were randomly assigned to any of the 2 (physical attractiveness: NAP vs. HAP) groups, in which the NAP group viewed the picture reviews containing the NAP, and the HAP group watched the picture reviews with HAP. In addition to the people in the picture commentary, the other information (e.g. price, size, picture size, etc.) that the two groups of subjects saw were identical. After viewing the picture reviews, the subjects were asked to rate the physical attractiveness of the figures in the picture reviews [28] , complete the body image state scale and purchase intention scale [29] [30] . All variables were scored by a seven-point Likert scale. Finally, the participants were asked to answer the gender, age, and awareness of the experimental intent, then the experiment was over and rewarded.
3.3. Results
Manipulation Checks. Female subjects in the HAP group rated the physical attractiveness of the figure in the picture higher than those in the NAP group (MNAP = 4.12, MHAP = 5.10, t = −5.649, p < 0.001), indicating that the manipulation of physical attractiveness in this experiment was successful.
Purchase Intention. Women in the HAP group had higher purchase intention than those in the NAP group (MNAP = 2.96, SDNAP = 1.08, MHAP = 4.03, SDHAP = 1.05, t = −5.253, p < 0.001).That is, the main effect of physical attractiveness is significant, and hypothesis H1 is verified.
Mediation Analysis. We used bootstrapping with repeated extraction of 5000 samples to test our process model (model 4). Results revealed that the indirect effect through the body image was significant (effect = 0.63, LLCI = 0.39, ULCI = 0.92), showed that body image had a significant mediating effect. The direct effect for physical attractiveness on women’s purchase intention remained significant (LLCI = 0.06, ULCI = 0.65). Therefore, body image plays a partial mediating role in the influence of characters’ physical attractiveness on consumers’ purchase intention, assuming that H2 is supported by data.
Discussion. The results of study 1 verified the hypothesis H1, that person’s physical attractiveness in picture reviews would influence consumers’ purchase intention. Viewing picture reviews of people with high physical attractiveness tended to boost consumer’ buying behavior, whereas viewing picture reviews of people with average physical attractiveness had the opposite effect. At the same time, this study also confirmed the mediating effect of body image, supporting hypothesis H2. Picture reviews with HAP promote consumers’ purchase intention by improving their body image, while picture reviews with NAP do the opposite.
4. Study 2
The second study was a 2 (physical attractiveness: NAP vs. HAP) × 2(perceived similarity: dissimilar vs. similar) between-subjects experimental design, to test the moderating effect of perceived similarity on the effect of person’s physical attractiveness in picture reviews on consumers’ purchase intention, namely to verify whether hypothesis H3 is true.
4.1. Pretest
In order to verify whether the manipulation method of perceived similarity is applicable, we conducted another pretest. This experiment uses the pictures of study 1as the experimental material. In the pretest, 100 female college students who volunteered to participate in the study were randomly assigned to one of 2 (physical attractiveness: NAP vs. HAP) × 2 (perceived similarity: dissimilar vs. similar) groups. They were asked to look at pictures of people who were highly or normally attractive. Regarding the manipulation of perceived similarity, in the similar group, we told subjects that the clothing size of the figure in the picture was the same as that of the subject, while the dissimilar group did not provide relevant information [23] . After viewing the picture materials, the subjects were required to complete the physical attractiveness scale [28] and perceived similarity scale [31] .
A 2 (physical attractiveness: NAP vs. HAP) × 2 (perceived similarity: dissimilar vs. similar) ANOVA with physical attractiveness as the dependent variable indicated the main effect of physical attractiveness (MNAP = 4.19, SDNAP = 0.67, MHAP = 4.98, SDHAP = 0.59, F(1, 96) = 39.158, p < 0.001). Meanwhile, the main effect of perceived similarity (F (1, 96) = 1.458, p > 0.05) and the interaction effect of physical attractiveness and perceived similarity (F(1, 96) = 0.173, p > 0.05) were not significant. These results suggest that it is appropriate to use these pictures as manipulation material for physical attractiveness.
A 2 (physical attractiveness: NAP vs. HAP) × 2 (perceived similarity: dissimilar vs. similar) ANOVA with perceived similarity as the dependent variable indicated the main effect of perceived similarity (Mdissimilar = 3.63, SDdissimilar = 1.03, Msimilar = 4.29, SDsimilar = 1.18, F(1, 96) = 8.750, p < 0.01). Meanwhile, the main effect of physical attractiveness (F(1, 96) = 0.188, p > 0.05) and the interaction effect of physical attractiveness and perceived similarity (F(1, 96) = 0.371, p > 0.05) were not significant. These results suggest that it is appropriate to use these pictures as manipulation material for physical attractiveness. To sum up, it is appropriate to manipulate the perceived similarity of subjects through clothing size.
4.2. Subjects and Procedure
148 female undergrads were invited to participate in the experiment and were paid after the experiment. Their average age was 21.85 years (SD = 3.15), and their average BMI was 21.31 (SD = 2.65). All of them had the experience of buying clothes online, and they were paid 5 RMB after the experiment.
We simulated the product page of taobao.com to display the pictures, and the subjects completed the questions set in this experiment through the link of the questionnaire. First, the subjects were asked about their height, weight and clothing size, and were presented with the pictures of the pure clothes (excluding people) of the striped skirt. Afterwards, the subjects were randomly assigned to either 2 (physical attractiveness: NAP vs. HAP) × 2 (perceived similarity: dissimilar vs. similar) groups to read the picture reviews. The manipulation of physical attractiveness and perceived similarity was the same as that in the pretest. Except for the person and clothing size information in the picture reviews, the other information of the four groups was exactly the same. After viewing the picture reviews, the subjects were asked to rate the physical attractiveness of the characters in the picture reviews [28] , and completed the body image state scale, purchase intention scale and perceived similarity scale [29] [30] [31] . In addition, the usefulness and credibility of reviews were measured as control variables [32] [33] . All variables were scored using a seven-point Likert scale. Finally, the participants were asked to answer the gender, age, and awareness of the experimental intent, and then the experiment was over and rewarded.
4.3. Results
Manipulation Checks. A 2 (physical attractiveness: NAP vs. HAP) × 2 (perceived similarity: dissimilar vs. similar) ANOVA showed that the manipulation of physical attractiveness in this experiment was effective (MNAP = 4.26, MHAP = 4.93, F(1, 144) = 11.430, p < 0.001). Similarly, the manipulation of perceived similarity in this experiment was also successful (Mdissimilar = 3.54, Msimilar = 4.37, F(1, 144) = 14.075, p < 0.001). In addition, the results of two-factor ANOVA with the usefulness and the credibility as dependent variables showed that the main effect of the physical attractiveness, the main effect of the perceived similarity and the interaction effect between them were not significant, excluding the influence of experimental materials on the usefulness and the credibility of the reviews.
Purchase Intention. A 2 (physical attractiveness: NAP vs. HAP) × 2 (perceived similarity: dissimilar vs. similar) ANOVA showed that the main effect of physical attractiveness was significant (F(1, 144) = 4.154, p < 0.05), the main effect of perceived similarity was also significant (F(1, 144) = 14.098, p < 0.001). Most importantly, the interaction effect between physical attractiveness and perceived similarity was significant (F(1, 144) = 27.301, p < 0.001).
Specifically, when subjects perceived that the characters in the pictures were similar with themselves, subjects in the group with HAP had a higher purchase intention than those in the NAP group (MNAP = 3.29, SDNAP = 1.48, MHAP = 4.95, SDHAP = 1.43, F(1, 144) = 22.32, p < 0.001). On the contrary, when the subjects perceived that the persons in the pictures were not similar with themselves, the subjects in the group with HAP had lower purchase intention than those in the group with NAP (MNAP = 3.62, SDNAP = 1.31, MHAP = 2.89, SDHAP = 1.33, F(1, 144) = 3.97, p < 0.05). In summary, the main effect of physical attractiveness is significant, that is, hypothesis H1 is verified again. The interaction effect between physical attractiveness and perceived similarity is also significant, so hypothesis H3 is verified.
Mediation Analysis. We used bootstrapping with repeated extraction of 5000 samples to test our process model (model 8). The results showed that body image had significant mediating effect (effect = 2.68, LLCI = 1.95, ULCI = 3.53). Specifically, when subjects perceived that the figure in the picture was similar to themselves, the mediating effect of body image was significant (effect = 1.97, LLCI = 1.45, ULCI = 2.51). Meanwhile, the mediating effect of body image was also significant when subjects perceived that the figure in the picture was not similar to themselves (effect = −0.71, LLCI = −1.27, ULCI = −0.21). In addition, after controlling for the mediating variable body image, the effect of the independent variable physical attractiveness on the purchase intention of the dependent variable was no longer significant. Therefore, body image plays a completely mediating role in the influence of persons’ physical attractiveness and perceived similarity on consumers’ purchase intention.
Discussion. The results of study 2 reaffirmed the hypothesis H1that person’s physical attractiveness in picture reviews would affect the purchase intention of consumers, and the picture reviews with HAP would tend to promote the purchase behavior of consumers. Furthermore, when the female perceived that the figure in the picture was similar to her, the picture reviews containing HAP would increase her purchase intention by improving her body image. However, when female consumers perceived that the figure in the picture was not similar to them, the picture reviews containing HAP would reduce the consumers’ body image and thus lead to a decrease in their purchase intention.
5. Discussion
This study explores the influence of the physical attractiveness of the characters in the picture reviews on e-commerce platforms on consumers’ purchase intention. The first experiment proved that the picture reviews with high attractive person can promote the purchase intention of consumers, and the body image plays a mediating role. Study 2 proved that perceived similarity plays a moderating role in the influence of person’s physical attractiveness on consumers’ purchase intention. To be specific, when people in the picture are perceived to be similar, the picture reviews with HAP can promote consumers’ purchase intention. When people in the picture are perceived to be dissimilar, the picture reviews with HAP would reduce consumers’ purchase intention. In addition, the clothing size of the figure in the picture reviews can affect the perceived similarity of consumers. Therefore, when consumers take pictures with similar characters as reference, pictures with HAP are beneficial to clothing sales. However, when consumers refer to pictures containing dissimilar figures, the pictures with HAP are not conducive to clothing sales. Based on the above conclusions, we can provide suggestions to marketers. In general, beautiful people lead to better sales results. However, when consumers perceive that the person from picture reviews is not similar to them, beautiful persons will bring worse results to the product. Therefore, blindly pursuing beautiful picture reviews is not a good marketing method.
In addition, this study also has some deficiencies. First, the phenomenon verified in this study only exists among female consumers, and whether the same phenomenon exists in the process and decision-making of male consumers in the purchase of clothing has not been discussed. Second, this study only considers clothing size as a factor affecting consumers’ perceived similarity, and does not discuss other influencing factors, such as height and weight. Third, this study has not yet explored the behavior of consumers after purchase, such as return. All these shortcomings need to be further studied in the future.
Acknowledgements
This research was funded by the National Natural Science Foundation of China (71302151), the Ministry of Education Humanities Social Science Foundation (12YJC630208), and Jinan University Key Discipline Construction Foundation (GY14011).