Improve webdesign with color linguistics: the effects on e-business behavior.

(1)

Mustafa Avci

Student number: 10899766

Thesis Master Information Studies - Business Information Systems University of Amsterdam Submission date: 23-11-2015

!

Improve webdesign with  

color linguistics:

the effects on e-business behavior.

!

First examiner: Second examiner:

Dick Heinhuis Tom van Engers!

!

(2)

Abstract

This study investigates whether specific colors stimulate web sales for different product categories. In psychology, researchers of color linguistics have shown correlations between color and behavior. In contrast, not much attention has been given to color for e-business websites. This study took the first step in research by creating a

conceptual framework for the relationship between color and web shopping behavior, specifically in the context of e-business categories. An experiment was constructed in order to measure the relationship between webstore preference and webstore color. One hundred and eight participants evaluated websites for seven product categories. Each participant was shown identical websites that differed only on the used

background color. Red, green, and blue colors were tested. The results indicate that when the same webstore is shown using different colors, the webstores are preferred differently based on product category. It appears that this phenomenon is not explained by the feeling of trust since the results indicate that the feeling of trust affects store choice only in certain situations with a limited effect size. This study indicates, with scientific evidence, that color affects store preference based on e-business category, and the results of the experiment in this study led to a model that can be used to create a strategy for web design colors. E-business companies can increase the preference rate of a webstore by using suitable colors.

!

"2

(3)

!

Content

1 --- Introduction 5 1.1 --- Motivation 5 1.2 ---Purpose 5 1.3 ---Research domain 5 1.4 ---Academic and practical relevance 6 2. ---Literature study 7 2.1 ---Present state of conceptual models 7 2.2 ---Building our own conceptual model 10 2.3 ---Investigating criteria about specific colors 12 3. ---Methodology 18 3.1 ---Participants 18 3.2 ---Materials 18 3.3 ---Procedure 19 4. ---Results 21 4.1 ---Color and product category 21 4.2 ---Explanation of trustworthiness 26 4.3 ---Summary of analysis results 28 5. ---Finishing and forming the model 29 6. ---Conclusion and Discussion 31 7. ---References 33 8. ---Appendix I ANOVA-test color & trust 38

(4)

List of figures

Figure 1 ---Modified model of Mehrabian-Russel 7  (Donovan et al. ,1994, p. 284)

Figure 2 ---Proposed model by Bjork (2010) 8 based on Mehrabian & Russel’s model

Figure 3 ---Wang, Hernandez & Minor’s improved version of Mehrabian 8 and Russel’s model which is applied to the webdesign context

Figure 4 ---Proposed model by Koo and Ju (2010) 9 based on Mehrabian and Russel’s model

Figure 5 ---Proposed model of Lee & Rao (2010) 9 Figure 6 ---Color as stimulus and emotion as organism 10 Figure 7 ---First part of conceptual model 11 Figure 8 ---Screenshot example of question survey 20 Figure 9 ---Indication of trust level of colors by participants 26 based on context (1=very low trust, 7=very high trust)

Figure 10 ---Proposed color linguistics strategy model for 30 e-business web design by present author

List of tables

Table 1 ---Overview of indications for the color red 13 Table 2 ---Overview of indications for blue color 14 Table 3 ---Overview of indications for green color 16 Table 4 ---Preference for a specific colored website 21 when buying a specific product

Table 5 ---Chi-Square analysis for the category sports 22 Table 6 ---Chi-Square analysis for the category sexual 22 Table 7 ---Chi-Square analysis for the category relaxation 23 Table 8 ---Chi-Square analysis for the category intellectual 23 Table 9 ---Chi-Square analysis for the category happiness 24 Table 10 ---Chi-Square analysis for the category plants 24 Table 11 ---Chi-Square analysis for the category health 25 Table 12 Summary of ANOVA repeated measures result for trust & colors -- 27 

(5)

1 Introduction

!

1.1 Motivation

As the use of web shopping grows, companies have more need of web presences and embark upon e-commerce. The total internet users who shopped online increased explosively from 6.27 million to 10.3 million from 2005 to 2013 in the Netherlands, according to the Dutch Census Agency (CBS, 2015).

!

Websites providing similar functionalities and identical products and services has also increased, which has created a necessity for the e-business sector to find new ways to improve web design and satisfy these increasing shoppers (Deng & Poole, 2012). Considerable research into web design has been conducted, but this research has not yet found a solution to fulfill this desire to improve web design and satisfy the growing number of customers. Currently, research into web design seems to be saturated. Therefore, the solution could come from research with a new and different

perspective. Some research areas are unexposed and not yet studied. In psychology, researchers of color linguistics have shown correlations between color and behavior. These researchers indicated profound effects of color on humans. In contrast, not much attention has been given to color for e-business web design. Only a few

practical-based studies have investigated this topic, and a conceptual framework is still lacking. This gap in literature forms the motivation to research how color linguistics could be used as design-criteria to improve web design since it could provide a solution for the e-business area to satisfy increasing visitors.

!

1.2 Purpose

The purpose of this thesis is to extend the current status of research in the field of color linguistics toward general insight regarding the use of color linguistics as a strategy for specified e-business environments. It should be possible to explain the use of color linguistics, which is connected with human psychology, in the context of e-business environments. As stated before, in the current situation there is a gap between the e-business and color linguistics areas. By bridging this gap, the e-business area can be helped to make a new jump in website design.

!

1.3 Research domain

Based on this gap in literature, the following research question was formulated:

• In which way can web design for e-business be improved with the use of color linguistics?

(6)

This central question will be further located and demarcated with the following sub-questions:

1. Which theories of color linguistics can be used to explain web shopping behavior?

2. What are indications of the effects of particular colors in research? 3. Could these effects of color be explained by the feeling of trust?

4. Is it possible to create a model based on theories that explain the use of color linguistics?

5. Can this model be confirmed empirically?

!

1.4 Academic and practical relevance

The academic relevance of this study for the ICT field is twofold. Firstly, this thesis provides insight into the theories relevant for describing the use of color linguistics, especially from the academic field of psychology. Many indications have been given in psychologic research about the link of color with different emotional states, but an explanation of how this link can be used as a strategy by organizations to improve their e-business environments is still lacking. Secondly, this thesis leads to the addition of new constructs and is intended to create a model to explain the link between

consumer behavior and color linguistics, which leads to insight into how organizations can make strategies for their e-business environments. Step-by-step a model, which can be applied to the information system field specifically for e-business environments, will be generated. This model will provide practical value for e-business companies

because they can apply color linguistic strategies to their websites in order to improve effectivity and sales. The model will give information about the appropriateness of different color schemes based on context. With this model, companies can determine and choose colors that stimulate sales for their products.

!

In this study, the main research question will be answered with the help of

sub-questions. First, in Chapter 2, literature will be reviewed in order to form a conceptual model. Afterward, in Chapter 3, a methodology will be presented to validate this model. Based on this methodology, an experiment was conducted and will be

discussed in Chapter 4. According to these results, a model will be formed in Chapter 5. Finally, Chapter 6 will present a conclusion and discussion.

!

"6

(7)

2. Literature study

To research in which way website design can be improved with the use of color linguistics, an investigation was conducted into how customers are affected by color. First, the present state of models that describe the behavior of visitors on e-business websites will be discussed in Section 2.1. Afterwards, in Section 2.2, theories from color linguistics, used to create a conceptual framework for the specific context of this study, will be presented. Finally, empirical evidence about the effects of particular colors were examined to gather criteria, which was used to form a model, and will be explained in Section 2.3.

!

2.1 Present state of conceptual models

To expound on the effect of web design elements on users, research used the model of Mehrabian and Russel (1974) as a basis. (Bjork, 2010; Donovan et al., 1994; Koo et Al., 2010; Wang et Al., 2010) This model describes a relationship between Stimulus, Organism, and Response and indicates that changes within an environment will lead to change in consumer behavior. Previous studies used Mehrabian and Russel’s (1974) model to create new conceptual models for the e-business context. It appears that this model is the basis of almost all conceptual models of relevant studies, except one. The conceptual models discussed in this section came into prominence by investigating studies from peer reviewed journals focused on the relationship between the e-business website environment and customer behavior. These models will be discussed now one by one.

!

Donovan et al. (1994) were the first to apply Mehrabian and Russel’s (1974) model to the retail context. Donovan et al. (1994) applied Mehrabian and Russel’s (1974) model to the retail context by describing the relationship between the shopping environment and emotions and behavior of customers. The shopping environment plays a role as stimuli, which induces the emotional states pleasure or arousal. In turn, these

emotional states induce approach or avoidance responses.

!

"

Figure 1. Modified model of Mehrabian-Russell (Donovan et al., 1994, p. 284)

!

Bjork (2010) applied Mehrabian and Russel’s (1974) model to the website shopping environment. The stimulus, organism, and response sections were reflecting and representing the shoppers situations and states. The shopping environment affects the shoppers’ emotional state, which in turn affects the shopping behavior outcome.

(8)

"

Figure 2 . Proposed model by Bjork (2010) based on Mehrabian & Russel’s (1974) model

!

Wang, Hernandez, and Minor (2010) also improved and applied Mehrabian and Russel’s (1974) model to the web design context. This conceptual models includes behavior outcomes from previous marketing research.

"

Figure 3. Wang, Hernandez, and Minor’s (2010) improved version of Mehrabian and Russel’s model (1974), which is applied to the web design context

!

Aesthetics affect the organism because the perceived online service quality and

satisfaction is affected by aesthetics, which in turn affects the behavior outcomes of the visitor. Purchasing is one these behavior outcomes.

!

Koo and Ju (2010) also built upon Mehrabian and Russel’s (1974) model and determined the model on different levels for web design context. Graphics, colors, links, and menus are elements that correspondent with Stimulus in Mehrabian and Russel’s (1974) model. Stimulus can arouse pleasure or arousal. These emotions

correspond with Organism in Mehrabian and Russel’s (1974) model and can influence the behavior of visitors and induce approach or avoidance. Thus, the intention of visitors are influenced and dependent on emotions.

(9)

"

Figure 4. Proposed model by Koo and Ju (2010) based on Mehrabian and Russel’s (1974) model

!

Lee and Rao (2010) were the first to investigate color for e-business websites. They argued that website colors affect store choice because color causes a difference in trust. However, the validity of this model is questionable since, in the experimental study, only one website (bookstore) was used as material when measuring. Also, the results slightly exceeded the maximum significance level, but the authors argued that the results could still be accepted as significant because the excess was small.

Figure 5. Proposed model of Lee and Rao (2010)

!

The relationship of store choice with trust will also be investigated in this study since it is one of the sub-questions. Lee and Rao s (2010) study is the only study that argued that color provokes different trust levels instead of emotions. However, the argument was weak because of limitations in measurement and results.

(10)

Mehrabian and Russel (1974) formed the root for current theoretical frameworks in web design research with their model. Bjork (2010); Wang, Hernandez, and Minor (2010); Koo and Ju (2010); and Lee and Rao (2010) described, in their proposed models, how shoppers are affected by the environment. They showed that Mehrabian and Russel s (1974) model is highly suitable to apply to web shopping behavior.

However, they only investigated the e-business context in general. Lee and Rao (2010) focused on color in the e-business context, but the validity of their model is

questionable because of limitations in terms of measurement and results. As stated, this study is specifically focused on color for e-business context. Therefore, in the next section, we will discuss customer behavior in relation to color, as well as a new model. We applied Mehrabian and Russel’s (1974) model to the e-business web design context in relation to color.

!

2.2 Building our own conceptual model

Previous studies using Mehrabian and Russel’s (1974) model were discussed in Section 2.1. This model proved to be suitable to describe behavior of customers on shopping websites. However, our study is slightly different; it is specifically focused on color in relation to customer behavior on e-business websites. Therefore, this study built a new model from the S-O-R model of Mehrabian and Russel (1974) by investigating

different e-business contexts.

!

First, the link between Stimulus and Organism from Mehrabian and Russel s (1974) was investigated for color in e-business context.by examining theories and empirical studies from the color linguistics area. Color affects humans emotions because certain colors and emotions have a connection. This discovery was first made by psychologists. They initially started studying emotions, and later, they found that emotions have a conformity with colors. The German mathematician and poet Johann Wolfgang von Goethe described this phenomenon in 1810 (Goethe, Eastlake, & Lock, 1840) In the beginning, psychologists developed models about emotions, which were applied when analyzing personality (Block, 1957; Leary, 1952). Over time, it discovered that those models have something in common with color chart models created by color scientists; it appeared they could be merged (Plutchik, 2001). Empirical studies focused on the relationship between color and mood states came to the same conclusion that humans emotions are affected by website colors (Elliot & Maier, 2014; Karlsson, 2007; Pelet & Papadopoulou, 2012; Porat & Noam, 2012;

Westerman et al., 2012; ).

!

Compared to the S-O-R model of Mehrabian and Russel (1975), website color correspondents with Stimulus because as it induces emotions. Emotions correspond with Organism because influenced and dependent on color.

"10

Figure 6. Color as stimulus and emotion as organism

(11)

After determining the first link between color and emotion as Stimulus and Organism, we examined the connection to the third step: Response. Colors affect customers because their emotions are influenced by colors, and this in turn affects their purchasing behaviors. The interface color influences aesthetics by increasing or decreasing weight attached to a judgment and can influence preference for a product choice or alternative (Westerman et al., 2012), which means that customers purchase intentions can be influenced by colors (Eroglu & Machleit, 2002). This influence is present without conscious ideas about the visual appearance (Basso, Ave, Park,

Greenspan, & Weimer, 2001), which mean that the customers purchasing behavior is positively related to colors and emotions, and color does not only influence the

evaluation by the visitor, but also influences the probability to recommend it to others (Gorn, Chattopadhyay, Sengupta, & Tripathi, 2004).

!

Compared to Mehrabian and Russel’s (1974) S-O-R model, it appears that product choice judgement corresponds to Response because the emotions induced by color influence product choice judgment.

The first part of the conceptual model was formed by performing literature review and embraces theories which can be used to explain decisions for different e-business websites. We then applied Mehrabian and Russel's (1974) model to the situation, as the

relationship between color and behavior of e-business visitors is described. The findings from the color linguistics area indicate a three way process. They can be compared to Mehrabian and Russel's (1974) model, where three parts correspond with each other. Color acts as stimulus and triggers emotions, since emotions are dependent on color, which trigger a certain

purchasing behavior in response.

Thus, sub-question 1 can be answered: Which theories of color linguistics can be used to explain the web shopping behavior? As this section shows, there are several useful theories of color linguistics that can be used to explain web shopping behavior, and when combined, they form the basis of our model. The theoretical portion of our model has been explained. In

the following section, we will discuss the practical portion through explaining criteria related to the effects of individual colors.

!

Figure 7. First part of model formed by applying Mehrabian and Russel (1974) to color for e-business context

(12)

2.3 Investigating criteria about specific colors

In this section, we will discuss our investigation into influences of specific colors. Sub-question 2 can be answered by investigating empirical indications about effects of particular colors in research. Based on these indications, hypotheses were formed and describe relationships between colors and sales for specific product categories. The hypotheses were statistically tested with survey data and results will be presented in Chapter 5. Afterward, the conceptual framework from Section 2.2 will be expanded with criteria based on the results of this test in order to form a model.

!

Scholars conducted research on the meaning of colors and gave several indications about links between color types and emotional moods. A library investigation was conducted to collect as much empirical evidence as possible from scientific journals. As a result, it appears that red, green, and blue are important because many studies indicate these colors. We summarized study results about these colors in separate tables. Because there were not many empirical indications about colors other than red, green, and blue, other colors were not included in this investigation. The results about red, green, and blue were used to form hypotheses. The hypotheses describe

connections between these colors and stimulation of sales for specific product categories.

!

Red triggers an outward focus and induces physical action. Results show that sport teams that were randomly assigned red clothes had relatively higher winning rates than those teams assigned other colors. Red functions as a jolt for dominance in human competition and enhances performance (Allen & Jones, 2013; Attrill, Gresty, Hill, & Barton, 2008). Based on these findings, Hypothesis 1 was formed:

Red-colored websites stimulate sales for products related to sports.

!

Red is seen as a color that that provokes emotions like arousal, warmness, and joy, as well as higher levels of anxiety (Goethe, Eastlake, & Lock, 1840; Goldstein, 1942; Soldat & Sinclair, 1997). The color is linked to anger or annoyance and triggers

testosterone (Fetterman, Robinson, Gordon, & Elliot, 2011; Hill & Barton, 2005). Red also induces sexual emotions. For example, research found that wearing red-colored dresses makes sexual attraction higher (Elliot, Kayser, Greitemeyer, Lichtenfeld, & Gramzow 2010; Elliot & Niesta, 2008; Elliot &Pazda, 2012). Based on these findings, Hypothese 2 was formed:

Red-colored websites stimulate sales for sex-related products.

!

"12

(13)

(14)

!

Blue is seen as a relaxing and calming, as well as triggering an inward focus (Goldstein, 1942). Blue-colored websites are seen as relaxing, calming, and reliable (Gorn et al. 2004; Lee & Rao, 2010). Based on these findings, Hypothese 3 was formed:

Blue-colored websites stimulate sales for products related to relaxation.

!

Blue is linked with openness. If continually exposed to blue, attention and exploratory behavior is awakened. Mehta and Zu (2009) found that people exposed to blue have an improved intellectual task-solving capacity. Based on these findings, Hypothese 4 was formed:

Blue-colored websites stimulate sales for intellectual material.

!

Research findings suggest that green is a positive color. A study illustrated that people were in a more positive mood when they were exposed to a screen showing a green environment (Gil, Bigot, 2014). People were asked to give their associations with green, and it appeared that green is linked with nature, ease-fulness, rest, and peace (Adams & Osgood, 1973; Clarke & Costall, 2008; Grieve, 1991). Research indicated that even to view green nature outside from a living room can put one in a better mood and help relieve (Kaplan, 2001). Kliger and Gilad (2012) indicated, in their statistical study, that green gives a feeling of safety, growth, and financial trust. Based on these findings, Hypothese 5 was formed:

"14

(15)

Green-colored websites stimulate sales for products related to happiness.

!

Green is also associated with fertility, growth, and natural environments (Gil & Bigot, 2014). Until the Middle Ages, green was seen as a symbol for fertility and hope and, for this reason, was often used for wedding clothes (Lichtenfeld, Elliot, Maier, & Pekrun, 2012). The word green may originate from plant growth and vegetation (Wierzbicka, 1990). Based on these findings, Hypothesis 6 was formed:

Green-colored websites stimulate sales for products related to nature.

!

Green has a positive relationship with healthiness. In fact, several empirical studies indicated a positive relationship between the amount of green space in the habitat and health (De Vries et al., 2003; Groenewegen, van den Berg, Maas, and Verheij, & de Vries, 2012; Takano, Nakamura, and Watanabe, 2002). Based on these findings, Hypothesis 7 was formed:

Green-colored websites stimulate sales for products related to health.

!

(16)

!

Table 3 overview of indications for green color

!

P.S. see next page for sequel

!

"16

(17)

!

In this section, tables summarizing empirical indications about color were presented, and seven hypotheses were formed based on this. These findings indicate the colors red, green, and blue have specific connections with moods and activities. Thus, the hypotheses state that these colors stimulate sales for specific product categories. In this section, sub-question 2 was answered by giving indications about the effects of particular colors in research. In the next chapter, the methodology used to validate these hypotheses will be presented.

(18)

3. Methodology

Our study was concerned with the research question in which way e-business

webdesign could improved with the use of color linguistics. In the literature review, we discussed results from color linguistics research. Research indicated three colors are important: Red, green, and blue. In Section 2.3, we formed hypotheses based on these indications. The hypotheses address relationships between these three colors and sales for different product categories. An experiment was performed to investigate those relationships.

!

3.1 Participants

A total of 108 people participated in the experiment and completed the survey. The survey was available online via author s Facebook and LinkedIn channel. The survey was completed by people who randomly saw the post on Facebook or LinkedIn. All of the respondents, except four, were Dutch. Respondents ages varied between 18 and 58 years old, with a mean of 25.24 and standard deviation of 7.702. In order to maintain a differential selection and avoid statistical regression, no specific participant selection was made for the survey. For example, we did not perform the experiment only with university students, like Gorn et al. (2004). However, there were still possible drawbacks. For example, this sample could only represents the Dutch population since almost all of the respondents were Dutch. Because of the limited timespan of this study, a convenience sample with mostly Dutch respondents was used. The hypotheses in this study were based on empirical findings about color from all over the world. They were tested with data from almost entirely Dutch respondents, and results were significant. Therefore, it seems that the meaning of color is something substantial despite nationality. However, the effect of nationality needs to be studied separately in order to be able to make conclusion about it. The used sample in this study could, therefore, only be considered reliable when considering it is a convenience sample for the Dutch population, limited to the author s social media channels.

!

3.2 Materials

We intended to measure store preference based on color, and its relationship with the feeling of trust, with the use of webstore page screenshots. In the survey, we showed each webstore page as a picture of 400 x 300 pixels. In previous studies, pictures of websites that were subject of analysis proved to be reliable and convenient as material (Gorn et al., 2004; Lee et al., 2010) In these studies, websites were shown as pictures that could not be modified and only differed in background color. The observed differences in the feeling of trust could only be explained by the difference in website colors. In this study, we developed hypotheses about seven product categories. The following categories were questioned separately in the survey: Sports, sex, relaxation, intellect, happiness, nature, and health. We choose to use Likert scales to collect data

(19)

about the feeling of trust in the same manner as previous studies on similar topics (Gorn, Chattopadhyay, Sengupta, & Tripathi, 2004; Lee & Rao, 2010).

!

The survey was designed for internal validity based on Blumberg s (2011) research methods book. First, a prototype of the survey was created and then was given to potential respondents. The respondents who gave feedback about the prototype were excluded from participation in the real experiment. Experiment time length appeared to be an important threat for internal validity because the potential respondents expressed continued interest in serious participation only up to five minutes, on

average. Based on this feedback, we decided to keep the number of questions low, only seven products were questioned, in order to fit this short timespan. In this way, we intended to prevent our participants from becoming bored, hungry, or tired because these factors can influence response results, according to Blumberg (2011). Material for measuring appeared to be another important threat for internal validity since websites can differ in view based on used software or screen size. Therefore, webstores were shown as pictures in the survey in order to prevent manipulation or view

distortion. No buttons or options could be clicked on the shown webstores, and fixed dimensions were used for pictures in order to make sure that every participant rated exactly the same content. As stated before, we used the same methods as previous studies (Gorn et al., 2004; Lee et al., 2010).

!

3.3 Procedure

In order to answer the research question, an online survey was conducted. In the survey, respondents were first asked to provide personal information, like age and gender. The collection of demographic data was for classifying purposes. Next, the participants were asked to give webstore preference based on product. Three nearly identical website pictures were shown next to each other, which only differed in

background color (see Figure 7). Respondents were asked to indicate their preferences by choosing one of these websites. After they gave their preferences, they were asked to rate the trustworthiness for each colored website variant on a Likert scale (low trust to high trust). We used a 7-point Likert scale, like in previous studies (Gorn et al., 2004; Lee et al., 2010). Respondents gave their webstore preferences and indication of trustworthiness repeatedly for different products. On each page, a new product was shown. After finishing all questions, respondents were shown a thank you message for their responses.

!

The questioned products were chosen based on the hypotheses (see Section 2.3). The next products were used for categories: Football (sports), playboy magazine (sex), hammock (relaxation), books (intellectual), birthday card (happiness), plant (nature), and vitamins (health). Only one product was questioned for each category because of limited available time. As explained, the available timespan was an important

(20)

have been questioned for each category. Therefore, future studies could elaborate on this study by investigating more products and categories.

!

Scientific research is divided about whether Likert scales should be used as ordinal or interval scales. Previous studies investigating web design showed that Likert scales can be effectively analyzed as interval scales (Gorn, Chattopadhyay, Sengupta, & Tripathi, 2004; Lee & Rao, 2010). Therefore, this study also used a similar method and

analyzed the Likert scales as interval scales.

"20

(21)

4. Results

Data from the survey response was analyzed statistically and will be presented in this chapter. First, the relationship between store choice and product categories was

analyzed. As explained in Chapter 3, on each page a different product was questioned in the survey. Each time, three nearly identical websites were shown next to each other, which only differed in background color. Respondents gave their preference for one of these websites regarding desire to buy the shown product. In the second section of this chapter, the analysis of these preferences in relation to the feeling of trust will be discussed. Finally, in Chapter 5, our model will be further developed and finished based on the results of these analyses.

!

4.1 Color and product category

Our main question was about how web design for e-business could be improved with the use of color linguistics. In order to investigate whether visitors have a preference to buy on websites with specific colors, respondents were asked to give their store

preference as if they were buying specific products. Table 4 shows a summary of the total number of preferences for each color based on product. The table shows that different colors stand out depending on the type of product.

!

Each category was statistically tested for color preference. Results show that

participants had a different store preference based on product category (for further detail, see analysis). The results show a significant relation for all categories, except happiness.

!

Table 4. Preference for a specific colored website when buying a specific product

!

Color Sports Sexual Relaxation Intellectual Happiness Plants Health

Red 14 64 10 19 27 4 16

Green 58 15 57 31 36 88 57

(22)

Category Sports: Hypothesis 1

For the category Sports, a football was used as product in the example website. The green store was most preferred with 58 votes, while 14 participants chose red and 36 blue. The Chi-Square statistic was 14.83 (p < 0.01), indicating this difference is significant. Hypothesis 1 was not confirmed because it stated the color red would be preferred. However, the color green was significantly preferred.

!

Table 5.Chi-Square analysis for the category sports

!

Category Sexual: Hypothesis 2

For the category Sexual, a playboy magazine was used as the product in the example website. The red store was most preferred with 64 votes, while 15 participants chose green, and 29 blue. The

Chi-Square statistic was 17.24 (p < 0.01), supporting that this difference is significant. Hypothesis 2 was thus confirmed.

!

Table 6. Chi-Square analysis for the category sexual

!

"22 Chi-Square 14.83 deg of freedom 2 Asymp. Sig 0.000602 Frequencies

Store preference for Sports

Observed N Expected N Residual

Red 14 36 -22

Green 58 36 22

Blue 36 36

Total 108

Frequencies Store preference for Sexual

Red 64 36 28 Green 15 36 -21 Blue 29 36 -7 Total 108 Chi-Square 17.24 deg of freedom 2 Asymp. Sig 0.00018

(23)

Category Relaxation: Hypothesis 3

For the category Relaxation, a hammock was used as the product in the example website. The green store was most preferred with 57 votes, while 10 participants chose red and 41 blue. The Chi-Square statistic was 19.76 (p < 0.01), supporting that this difference is significant. Hypothesis 3 was not confirmed because it stated the color blue would be preferred. However, green was significantly preferred.

!

Table 7. Chi-Square analysis for the

category relaxation

!

Category Intellect: Hypothesis 4

For the category Intellect, books were used as the product in the example website. The blue store was most preferred with 58 votes, while 31 participants chose green and 19 red. The Chi-Square statistic was 10.78 (p < 0.01), indicating that this difference is significant. Hypothese 4 was thus being confirmed.

!

Table 8. Chi-Square analysis for the category intellect

!

Frequencies

Store preference for Relaxation

Red 10 36 -26 Green 57 36 21 Blue 41 36 5 Total 108 Chi-Square 19.76 deg of freedom 2 Asymp. Sig 0.000051 Chi-Square 10.78 deg of freedom 2 Asymp. Sig 0.004562 Frequencies

Store preference for Intellect

Red 19 36 -17

Green 31 36 -5

Blue 58 36 22

(24)

!

Category Happiness: Hypothesis 5

For the category Happiness, a birthday card was used as the product in the example website. The blue store was most preferred with 45 votes, while 36 participants chose green and 27 red. The Chi-Square statistic was 2.29 (p > 0.01), so there was no indication that this difference is significant. Hypothese 5 was not confirmed, and the results do not provide other clues about color preference for Happiness.

!

Table 9. Chi-Square analysis

for the category happiness

!

Category Plants: Hypothesis 6

For the category Plants, a bush was used as the product in the example website. The green store was most preferred with 88 votes, while 16 participants chose blue and four red. The Chi-Square statistic was 55.1 (p < 0.01), indicating that this difference is significant. Hypothese 6 was thus confirmed.

!

Table 10. Chi-Square analysis for the category plants

!

"24

Frequencies Store preference for Plants

Red 4 36 -32 Green 88 36 52 Blue 16 36 -20 Total 108 Chi-Square 55.1 deg of freedom 2 Asymp. Sig 0 Frequencies

Store preference for Hapiness

Red 27 36 -9 Green 36 36 0 Blue 45 36 9 Total 108 Chi-Square 2.29 deg of freedom 2 Asymp. Sig 0.318224

(25)

Category Health: Hypothesis 7

For the category Health, vitamins were used as the product in the example website. The green store was most preferred with 57 votes, while 35 participants chose blue and 16 red. The Chi-Square statistic was 12.45 (p < 0.01), indicating that this difference is significant. Hypothese 7 was thus confirmed.

!

Table 11. Chi-Square analysis for category health

!

Frequencies Store preference for Health

Red 16 36 -20 Green 57 36 21 Blue 35 36 -1 Total 108 Chi-Square 12.45 deg of freedom 2 Asymp. Sig 0.001979

(26)

4.2 Explanation of trustworthiness

Sub-question 3 was about whether these webstore preferences could be explained by the feeling of trust. Figure 9 shows the average trust level indication by respondents for colors red, green, and blue. These differences were analyzed now for significance, separately for all product categories.

!

Figure 9. Indication of trust level of colors by participants based on context (1=very low trust, 7=very high trust)

!

"26 0 1.5 3 4.5 6

Sports Sexual Relaxation Intellectual Hapiness Plants Health

4.35 4.34 4.63 4.81 4.68 3.9 4.46 4.83 5.09 4.61 4.44 4.6 3.6 4.66 3.47 3.34 3.73 3.76 3.56 4.34 3.24 Red Green Blue

(27)

The ANOVA-test (see Appendix I) gives information about the relationship between trust and different colors. Table 12, below, provides a short summary of the results.

!

Table 12 summary of ANOVA repeated measures result for trust based on colors

!

The results show that only for Hypotheses 6 and 7 color choice had a complete connection with trust. Hypotheses 1, 2, 3, 4, and 5 show significant differences in connection with trust, but not between all colors. For example, for hypothesis, 1, red was significantly different than blue and green, but blue was not significantly different than green.

!

In short, Section 4.1 showed there are significant preferences for buying from stores with specific colors based on product. The results in this section indicate that these preferences are explained by the feeling of trust only for hypotheses 6 and 7 (plants and health), with an effect size of 0.265 and 0.462, respectively.

!

Hypothese Product

category F-statistic & p-value Statistically sign. different based on trust:

1 Sports F=60.531

p=0.000 Red vs. Blue, Red vs. Green 2 Sexual F=22.028 p=0.000 Green vs. Blue 3 Relaxation F=34.015 p=0.000 Red vs. Blue 4 Intellectual F=40.391 p=0.000 Red vs Blue 5 Happiness F=36.000 p=0.000 Red vs. Blue 6 Nature F=18.760

p=0.000 Red vs. Blue vs. Green 7 Health F=44.620

(28)

4.3 Summary of analysis results

!

Sport-related products (Hypothesis 1)

Participants preferred buying the football on a green-colored website (See Section 4.1) Hypothesis 1 was not confirmed because it assumed red would be preferred for sport products. However, green was significantly preferred. This choice cannot be completely explained by the feeling of trust since trust was only significant with the color blue.

!

Sex-related products (Hypothesis 2)

Participants preferred buying the sexual magazine on a red-colored website (See Section 4.1). Hypothesis 2 was confirmed, this choice cannot be explained by a feeling of trust.

!

Relaxation-related products (Hypothesis 3)

Participants preferred buying the hammock on a green-colored website (See Section 4.1). Hypothesis 3 was not confirmed because it assumed blue would be preferred. However, green was significantly preferred. This choice cannot be explained by a feeling of trust.

!

Intellect-products (Hypothesis 4)

Participants preferred buying books on a blue-colored website (See Section 4.1). Hypothesis 4 was confirmed. This choice cannot be explained by a feeling of trust since trust corresponded to the color red.

!

Happiness-related products (Hypothesis 5)

There was no significant difference in store preference for this category (See Section 4.1). Hypothesis 5 was not confirmed, and the results did not provide other clues about color preference. Additionally, there was not a significance difference in feeling of trust. Only red and blue showed a significant difference compared to each other.

!

Nature-related products (Hypothesis 6)

Participants preferred buying plants on a green-colored website (See Section 4.1). Hypothesis 6 was confirmed. This choice can be explained by a feeling of trust by 27%. As the results of the ANOVA test show, there is a significant difference between all colors.

!

Health-related products (Hypothesis 7)

Participants preferred buying vitamins on a green-colored website (See Section 4.1).

Hypothesis 7 was confirmed. This choice can be explained by a feeling of trust by 46%. As the results of the ANOVA test show, there is a significant difference between all colors.

!

(29)

5. Finishing and forming the model

Sub-question 4 concerned whether it was possible to create a model based on theories that explain the use of color linguistics. In Chapter 2, theories from color linguistics were investigated in order to build a conceptual framework, and based on this

framework, an experiment was created. The results of the experiment will be used in this chapter to expand on the conceptual framework in order to form a model. The indications in the literature review were a basis for hypotheses, which in turn formed the basis to create a model. Figure 10 shows the resulting model.

!

At the top, a representation of the analyzed colors red, green, and blue can be found. These colors correspond with Stimulus from the model of Mehrabian and Russel (1974). The arrows between the first and second rows represent the connection between the colors and emotions.

!

In the second row, emotions provoked by colors from the first row, and which were discussed in the literature review in Section 2.3, are described. The second row corresponds with Organism from the model of Mehrabian and Russel (1974).

Emotions related to red are described as arousing in the model, based on the following indications:

•! Red is seen as a jolt for dominance (Attrill et al., 2008; Allen et al., 2013) •! Red provokes emotions like arousal, warmness, and joy, as well as higher levels

of anxiety (Goethe et al., 1840; Goldstein, 1942; Soldat et al., 1997)

•! Red is linked with anger or annoyance and triggers testosterone(Fetterman et al., 2011; Hill et al., 2005)

•! Red induces sexual emotions (Elliot et al., 2008; Elliot et al., 2010; Elliot et al., 2012)

!

Emotions related to green are described as pleasing in the model, based on the following indications:

•! Green gives people more positive moods (Gil & Bigot, 2014)

•! Green is linked with ease-fulness, rest, and peace (Adams et al., 1973; Clarke et al., 2008; Grieve, 1991)

•! Viewing green can give help relieve stress (Kaplan, 2001).

•! Green gives a feeling of safety, growth, and trust. (Kliger et al. 2012)

!

Emotions related to blue are described as calming in the model, based on the following indications:

•! Blue is seen as relaxing and calming, as well as triggering an inward focus (Goldstein, 1942)

(30)

•! Blue-colored websites are seen as relaxing, calming, and reliable (Gorn et al., 2004)

•! Blue colors are relaxing (Lee et al., 2010).

!

In the third row, product categories are represented for each color. These product categories were analyzed in this study s experiment, and results showed that participants had a preference for buying on websites with specific colors, based on product category. These product categories correspond with Response from the model of Mehrabian and Russel (1974). Finally, it appears to be possible to create a model based on theories that explain the use of color linguistics, and our model shown in Figure 10 was created from these results.

!

Figure 10. Proposed color linguistics strategy model for e-business web design by present author

(31)

!

6. Conclusion and Discussion

Finally, with the results of this study a model is formed which gives indications about the appropriateness of colors related to different e-business website categories. The model came as a result of research about how webdesign for e-business could be improved with the use of color linguistics, as this was our main research question. Color linguistics is an area which is not entered by webdesign studies yet, but it is definitely worth to be investigated by future research. As this study shows that it is possible for webdesign research to evolve by entering this area, and open up a new way to innovate in the e-business industry. This study took the first step in research by creating a conceptual framework for e-business webdesign based on theories and indications from color linguistics, and can serve as a starting point for future research to expand on. The color linguistics area proofed to have theories and indications which can be used to improve webdesign, as significant results were found in the experiment. Beside the scientific value, the e-business sector can make use of the model of this study to improve their webstores. As the model can help to create a webdesign strategy based on color linguistics.

!

The theoretical framework in this study describes a relation between color and moods. Based on this framework an experiment was set up to investigate the relation between webstore color and sales for different product categories, and the results of this

experiment indicate that such a relation is existent. Secondly, we investigated if this relation could be addressed to the feeling of trust. The feeling of trust appeared not to be enough to explain this relation. As the results show that this relation is addressable to the feeling of trust only for green colored webstores in the category health (with 26.5% effect size), and nature (with 46.2% effect size). Perhaps the reason for this could be a specific quality of green color for giving trustful feelings. This subject could be a research object for future research, as they could further investigate the role of trust to expand on this study. Future research could also investigate variables different than trust, which can explain store choice when purchasing, to find an explanation.

!

The results of the statistical test in this study were very strong with a significance at alpha level less than 0.01. Beside this, the observed power was greater than 0.80 for all tests in our study, according to the statistics book of Cohen (1988), this means that disapproved hypotheses were rejected correctly. With a small alpha level and high observed power the results of this study can be considered statistically valid, but there are some limitations regarding reliability. The reliability of a sample value is the closeness with which it can be expected to approximate the relevant population value according to Cohen (1998). A potential drawback for reliability could be that the sample in this study is representing only the Dutch population. As almost all of the respondents were Dutch. Because of a limited timespan, a convenience sample with mostly Dutch respondents, limited to the social media channels of the author of this study was used. The hypotheses in this study were based on empirical findings about

(32)

color which came from different countries than the Netherlands. But, the hypotheses were tested in the experiment with data from almost only Dutch respondents, and still significant results were found. According to this situation, it seems that the effect of color is something substantial notwithstanding nationality. The effect of nationality could be a good subject for future research to expand on this study.

The model gives criteria about blue, green, red and six product categories. The investigated colors and products in this study were limited due short timespan. Future research could therefore investigate relations between more colors and product

categories, in order to expand on this study.

Another limitation of the used sample was age of respondents, as the convenience sample in this study consisted mostly of younger respondents, which had an average of 25 years old.

The used sample in this study could therefore only be considered reliable if we consider it is a convenience sample for the Dutch population with young age, and is limited to the social media channels of the author of this study.

!

Finally, with the results of this study a model is formed which gives indications about color criteria based on different e-business categories. The model gives answer to the main research question by providing criteria for webdesign, as the main research question was about how webdesign for e-business could be improved with the use of color linguistics. As a practical value, companies can use this model to make decisions for color strategies on their e-business website. This study can be a very good starting point for future research, as it comes with a conceptual framework for an area which is not studied yet. As stated before, this study has some limitations, therefore there are many subjects which can be investigated in order to expand on this study.

!

"32

(33)

7. References

!

Adams, F. M., & Osgood, C. E. (1973). A Cross-Cultural Study of the Affective Meanings of Color. Journal of Cross-Cultural Psychology.

!

Allen, M., & Jones, M. (2014). The home advantage over the first 20 seasons of the English Premier League: effects of shirt colour, team ability and time trends.

International Journal of Sport and Exercise Psychology, 12(1), 10–18.

!

Attrill, M. J., Gresty, K. A., Hill, R. A., & Barton, R. A. (2008). Red shirt colour is associated with long-term team success in English football. Journal of Sports Sciences, 26(6), 577–582.

!

Basso, A., S, L. A., Ave, P., Park, F., Greenspan, S., & Weimer, D. (2001). First Impressions : Emotional and Cognitive Factors Underlying Judgments of Trust E-Commerce. 3rd ACM Conference on Electronic Commerce, 137–143.

!

Bjork, P. (2010). Atmospherics on tour operators’ websites: Website features that stimulate emotional response. Journal of Vacation Marketing, 16(4), 283–296.

!

Block, J. (1957). Studies in the phenomenology of emotions. Journal of Abnormal and Social Psychology, 54, 358–363.

!

Blumberg, B, Cooper, R, Schindler, P. (2011). Business research methods. McGraw-Hill.

!

CBS. (2015). Statistical information about internet usage in The Netherlands. Retrieved from: http://www.cbs.nl/nl-NL/menu/themas/vrije-tijd-cultuur/cijfers/ computer-en-internet/2013-internet-sl.htm at: 11-04-2015.

!

Chaldecott, J. a., & Earles, M. P. (1973). The British Society for the History of Science. The British Journal for the History of Science, 6(04), 464.

!

Chen, M.-Y., & Teng, C.-I. (2013). A comprehensive model of the effects of online store image on purchase intention in an e-commerce environment. Electronic Commerce Research, 13(1), 1–23.

!

Clarke, T., & Costall, A. (2008). The emotional connotations of color: A qualitative investigation. Color Research and Application, 33(5), 406–410.

(34)

!

Cohen, J. (1988). Statistical power analysis for the behavioral sciences. Statistical Power Analysis for the Behavioral Sciences.

!

De Vries, S., Verheij, R. A., Groenewegen, P. P., & Spreeuwenberg, P. (2003). Natural environments...healthy environments? An exploratory analysis of the relationship between greenspace and health. Environment & Planning A, 35, 1717–1731.

!

Deng, L., & Poole, M. S. (2012). Aesthetic design of e-commerce web pages - Webpage Complexity, Order and preference. Electronic Commerce Research and Applications, 11(4), 420–440.

!

Donovan, R. (1994). Store atmosphere and purchasing behavior. Journal of Retailing, 70(3), 283–294.

!

Elliot, A. J., Kayser, D. N., Greitemeyer, T., Lichtenfeld, S., Gramzow, R. H., Maier, M. A., & Liu, H. (2010). Red, rank, and romance in women viewing men. Journal of Experimental Psychology. General, 139(3), 399–417.

!

Elliot, A. J., & Maier, M. a. (2014). Color psychology: effects of perceiving color on psychological functioning in humans. Annual Review of Psychology, 65, 95–120.

!

Elliot, A. J., & Niesta, D. (2008). Romantic red: red enhances men’s attraction to women. Journal of Personality and Social Psychology, 95(5), 1150–1164.

!

Elliot, A. J., & Pazda, A. D. (2012). Dressed for sex: Red as a female sexual signal in humans. PLoS ONE, 7(4).

!

Eriksen, B. A., & Eriksen, C. W. (1974). Effects of noise letters upon identification of a target letter in a non- search task. Perception And Psychophysics, 16(1), 143–149.

!

Fetterman, A. K., Robinson, M. D., Gordon, R. D., & Elliot, A. J. (2011). Anger as Seeing Red: Perceptual Sources of Evidence. Social Psychological and Personality Science.

!

Gil, S., & Le Bigot, L. (2014). Seeing life through positive-tinted glasses: Color-meaning associations. PLoS ONE, 9(8).

!

Goethe’s theory of colours, tr. from the G. J. M. (1840). Goethe’s theory of colours, tr. from the German J. Murray., 1840.

(35)

!

Goldstein, K. (1942). Some experimental observations concerning the influence of colors on the function of the organism. Occup. Ther. Rehabil., 21:147(51).

!

Gorn, G. J., Chattopadhyay, A., Sengupta, J., & Tripathi, S. (2004). Waiting for the Web: How Screen Color Affects Time Perception. Journal of Marketing Research, 41(2), 215–225.

!

Grieve, K. W. (1991). Traditional beliefs and colour perception. Perceptual and Motor Skills, 72(3 Pt 2), 1319–1323.

!

Groenewegen, P. P., van den Berg, A. E., Maas, J., Verheij, R. a., & de Vries, S. (2012). Is a Green Residential Environment Better for Health? If So, Why? Annals of the Association of American Geographers, 102(5), 996–1003.

!

Guttman, L. (1954). A New Approach to Factor Analysis: the Radex. In Mathematical Thinking in the Social Sciences (pp. 258–348).

!

Hill, R. A., & Barton, R. A. (2005). Sporting contests: seeing red? Putting sportswear in context. Nature, 437(7063).

!

Kaplan, R. (2001). The Nature of the View from Home: Psychological Benefits. Environment and Behavior.

!

Karlsson, M. (2007). Expressions, emotions, and website design. CoDesign, 3(sup1), 75–89.

!

Kliger, D., & Gilad, D. (2012). Red light , green light : Color priming in financial decisions . Journal of Socio-Economics, 41(5), 738–745.

!

Koo, D. M., & Ju, S. H. (2010). The interactional effects of atmospherics and perceptual curiosity on emotions and online shopping intention. Computers in Human Behavior, 26(3), 377–388.

!

Leary, T. (1957). Interpersonal diagnosis of personality: A functional theory and methodology for personality evaluation. New York: Ronald Press.

!

Lee, S., & Rao, V. S. (Chino). (2010). Color and Store Choice in Electronic Commerce: The Explanatory Role of Trust. Journal of Electronic Commerce Research, 11(2), 110–126.

(36)

!

Lichtenfeld, S., Elliot, a. J., Maier, M. a., & Pekrun, R. (2012). Fertile Green: Green Facilitates Creative Performance. Personality and Social Psychology Bulletin, 38(6), 784–797.

!

Machleit, K. a, & Eroglu, S. a. (2000). Describing and Measuring Emotional Response to Shopping Experience. Elsevier, 2963(99).

!

Manganari, E. E., Siomkos, G. J., & Vrechopoulos, A. P. (2009). Store atmosphere in web retailing. European Journal of Marketing, 43(9/10), 1140–1153.

!

McDougall. (1992). An introduction to social psychology. Boston: Luce.

!

Mehrabian, A., & Russell, J. . (1974). An approach to environmental psychology. MIT Press.

Mehta, R. Zhu, R. (2009). Blue or Red? Exploring the Effect of Color on Cognitive Task Performances. Science, (February), 1226–1229.

!

Nitse, P. S., Parker, K. R., Krumwiede, D., & Ottaway, T. (2004). The impact of color in the e-commerce marketing of fashions: an exploratory study. European Journal of Marketing, 38(7), 898–915.

!

Pelet, J.-É., & Papadopoulou, P. (2012). The effect of colors of e-commerce websites on consumer mood, memorization and buying intention. European Journal of Information Systems, 21(4), 438–467.

!

Plutchik, R. (2001). The Nature of Emotions: Human emotions have deep evolutionary roots, a fact that mayexplain their complexity and provide tools for clinical practice. American Scientist, 89(4), 344–350.

!

Porat, T., & Noam, T. (2012). No It’s a Pleasure Buying Here: The Effects of Web-Store Design on Consumers' Emotions and Attitudes. Human – Computer

Interaction, 27:3, 235–276.

!

Rafal, R., Gershberg, F., Egly, R., Ivry, R., Kingstone, A., & Ro, T. (1996). Response channel activation and the lateral prefrontal cortex. Neuropsychologia, 34(12), 1197– 1202.

!

Schlosberg, H. (1952). The description of facial expressions in terms of two dimensions. Journal of Experimental Psychology, 44(4), 229–237.

(37)

!

Soldat AS, Sinclair RC, M. M. (1997). Color as an environmental processing cue: External affective cues can directly affect processing strategy without affecting mood. Soc. Cogn., (15:55), 71.

!

Stroop, J. R. (1935). Stroop color word test. J. Exp. Physiol., (18), 643–662.

!

Takano T, Nakamura K, W. M. (2002). Urban residential environments and senior citizens’ longevity in megacity areas: the importance of walkable green spaces. Journal of Epidemiology and Community Health, 18(56:913).

!

Wang, Y. J., Hernandez, M. D., & Minor, M. S. (2010). Web aesthetics effects on perceived online service quality and satisfaction in an e-tail environment: The moderating role of purchase task. Journal of Business Research, 63(9-10), 935–942.

!

Westerman, S. J., Sutherland, E. J., Gardner, P. H., Metcalfe, R., Nash, J., Palframan, S., & Woodburn, N. (2012). Ecommerce interface colour and consumer decision making: Two routes of influence. Color Research and Application, 37(4), 292–301.

!

Wierzbicka, A. (1990). The meaning of color terms: semantics, culture, and cognition. Cognitive Linguistics.