The Green-Awakening Customer Attitudes towards Buying Green Products on an Online Platform in Thailand: The Multigroup Moderation Effects of Age, Gender, and Income

Abstract

In order to respond to sustainable consumption and production, customers today are more aware of how their actions affect the environment. As a result, customers are looking for products that use sustainable practices and are environmentally conscious; an online platform is viewed as a channel to approach such products in the digital era. To meet this demand, understanding how green customer attitudes play a critical role in changing consumer behavior is required. This new concept of "green-awakening" customer attitude encapsulates green positive feelings, green intent, green likeness, and green belief to overcome the limitations of measurements used previously. The objectives of this research were (1) to investigate and confirm the existence of green-awakening customer attitudes toward online platforms in the emerging economy and (3) to investigate differences in equivalent measurements across age, gender, and income properties in Thailand. Data on 348 Thai customers were collected from different parts of the country, including the Northeast, North, Central, East, and South. In addition, a multigroup moderation approach was employed to investigate differences among groups segmented by age, gender, and income. SPSS Amos was used on the basis of the composite-based model to analyze multigroup structural invariance among the segmented data. On the one hand, the findings showed that Thai customers’ attitudes toward purchasing green products on an online platform are more likely to be affected by perceived relative advantage, perceived online social norms, and perceived risk. On the other hand, perceived online compatibility did not have a statistically significant effect on how green customers felt at an aggregate level. According to the findings of the multigroup analysis, Thai customers with a mid-high income level perceived green products positively and appeared to be able to afford them on an online platform when compared to other groups, which suggests that an income-based segment could moderate Thai customers’ attitudes toward purchasing green products on an online platform. Marketers could use the concept and idea of “green-awakening” customer attitudes to strengthen how to decide the precise target segmentation of green consumers.

1. Introduction

Green customer growth has become an essential focus for sustainable businesses in recent years [1,2,3], thereby achieving sustainable patterns of consumption and production [4,5]. Green customers have been acknowledged as a crucial driver of sustainable consumption of eco-friendly products [6] and a driving force for the green economy transition [7,8]. Green behavioral changes lie partly in customer attitude as a powerful tool for positive action as it is inherently interwoven into everything customers think, feel, and do [9]. Customer attitudes determine behavioral changes fostered by beliefs, feelings, and behavioral intentions [10]. Green customer attitudes are an imperative part of the online buying process [11,12]. Attitude change is generally complicated, particularly when customers suspect that the marketer has a self-serving manner in bringing about this change (e.g., buying green products online). Marketers are faced with many constraints in terms of slow economic recovery, digital disruption, sustainable lifestyle, and global pandemics, indicating a need for being greener and online complemented along the same path. Understanding the determinants of green customer attitudes to online platforms is required; 76% of Thai customers have become aware of items with eco-friendly packaging or less packaging online [13]. This is a signal of being socially and environmentally conscious as this figure indicates demand for green products somewhat. Green products are products that are less harmful to the environment. [14,15]. These are products made from organic and all-natural ingredients such as recycled toilet tissue, biodegradable cleaning cloths, natural cleaning products, and organic produces. They also come in recyclable, compostable, or biodegradable packaging [14,15]. This study undertakes an evaluation of green customer attitudes to create a new concept, Green-Awakening Customer Attitudes (GACA). Sequentially, it investigates the underlying determinants’ effects on this concept and evaluates the implications for green marketers. Therefore, it comes to the research concept of evaluating the extent of green-awakening Thai customer attitudes toward green products on online platforms. In this context, this study poses four central research questions. First, is there an existence of green-awakening customer attitudes toward online purchasing platforms in Thailand? Second, what are the determinants of green-awakening customer attitudes? Third, do risks, norms, and innovation diffusion matter when customers? Fourth, are there any differences in green-awakening customer attitudes that changes to purchase green products on an online platform across age, gender, and income properties in Thailand?

The problem to be addressed by this study is that the continuous purchase of products online since 2020 has still caused a significant change in age, as this change reflects the generation gap in the reliance on online platforms in Thailand [16]. According to Wunderman Thompson Consultant’s Future Shopper Report 2021 in Thailand, out of 1025 Thai customers, almost 94%were likely to purchase products online. This was higher than the global average of 72% [17]. While almost 45% of Thai customers noted that they were likely to spend up to 8000 THB (or around 220 USD) on online platforms per purchase [18]. The roles of age and income are essential as their differences cause behavioral changes in attitudes and willingness to purchase products online. However in behavioral studies, it is also indispensable to look at how customers buy things in general since the customer journey can interact differently with genders [19,20,21]. Being aware of differences in general buying habits among age, gender, and income groups can help green marketers can help marketers take the first step toward diversifying their marketing campaigns and understand and predict their attitudes, behaviors, and dynamics [22]. Changing customer buying habits on the move by going online for their needs is challenging in the emerging economy, especially in Thailand, due to the digital divide across age, income, and gender.

There is an abundance of existing literature pointing to the importance of green customer behavior as a major determinant of behavioral action change, but relatively little on green customer attitude change. To the best of our knowledge, existing studies have not yet investigated: (1) how green customer attitudes are formed and (2) how green products are rated and perceived as product choices on an online platform. This is especially important for a developing country such as Thailand. Thai citizens have their own set of rules for how to act that they think are very important for comfort, this study then proceeds to investigate the effect of risks, norms, and platform diffusion on the formation of green customer attitudes across age, gender, and income characteristics that are viewed as key characteristics of Thai citizens. Delcea et al. (2019) consider the impact of customer attitudes in online social networks and how their role influences online environments and customers’ positive attitudes toward buying and using green products [23]. They find that the consideration of green products on an online platform requires an environmentally friendly attitude towards the environment, environmental-related action, and customers’ attention to environmental issues [23]. Wang, Gao and Wang (2021) argued that the relative advantage of green products over traditional products in an online platform influences the magnitude of customer decisions and helps determine customers’ differences in attitude toward these products [24]. Meanwhile, Ma and Liu (2022) consider another step of green online community creation when online environmental platform services are provided and launched to grab users’ public welfare participation and try to reduce environmental impact as a result of their consumption [25]. They find that environmental attitude and price sensitivity are key to the success of public green consumption behaviors. Dangelico, Nonino and Pompei (2021) argued that to achieve green purchase behavior, it is required to put green purchase satisfaction, willingness to pay a premium price, and green purchase frequency into effect [26]. In addition, social norms (i.e., green product usage is the right thing to do) need to be relatively associated with natural-materialistic attitudes towards green products. Similarly, Sharma et al. (2020) deem that attributes of green self-concepts and green-self identities can grasp the wide range of green attitudes when predicting the existence of green customer behavior [27]. As a green attitude represents a state of green readiness, Wei et al. (2017) illustrate that green customers would contribute to the demand for green products when they pay more attention to the environment and health, which will result in a green consumption attitude [11]. In general, most recent studies show that people are more likely to buy green products when they have a positive attitude toward them. However, the things that make people think and act in a green way are contingent. However, most of the studies we have talked about so far focused mostly on understanding how green customers act when what we really need is to know how green customers feel. If the green attitude is not successfully formed, green-related action is not taken. The goal of the attitude change is to find out how different types of people feel about being green. This study then views green customer attitudes as green-awakening customer attitudes that match the nature of an emerging market such as Thailand where the awareness of environmental impact is ongoing. To fill the knowledge gap in examining green customer attitudes, this study poses two main research objectives: (1) to investigate and confirm the existence of green-awakening customer attitudes on online platforms in the emerging economy and (2) to investigate differences in equivalent measurement across age, gender, and income properties in Thailand.

This study contributes to the body of new knowledge on green-awakening customer attitudes by analyzing the impact of risk, norms, and characteristics of online platforms. Additionally, the investigation is concerned with discovering the facts about green-awakening customer attitudes in Thailand. Firstly, the current study adds to this field by focusing on how green-awakening customers form their attitudes and looking at factors from a theoretical perspective on an online platform. It is surmised that customer attitudes powerfully show how customers express themselves through their psychological tendencies, such as how they feel about buying green products online. The first step to building a critical thoughtful feeling about buying green products online is to be aware of how customers feel. A key factor in customer attitudes is their awareness of going green. Customers learn about green products by judging them (i.e., this is called the green-awakening customer) but not by getting deeply involved (i.e., this is called the green-awakened customer). Secondly, the study contributes to the main features of the proposed model, which are perceived online platform risk, compatibility, digital social norms, and perceived relative advantage. This is based on the theory of risk perception, the theory of social norms, and the theory of innovation diffusion, with age, gender, and income as moderators. The composition of the buying patterns varies according to gender, age, and income. These moderating effects change the basic lens or frame through which customers are aware of buying green products online, whether positive or not. The current understanding of online platforms for green products in a Thailand context is added to generalize the results from the sample of the Thai population of interest as it is supported by the fact that being green can go along with being online for some reasons and levels. In this respect, being online and green is an uphill battle, but Thai customers are encouraged to make eco-conscious online purchases. Later, strategies are proposed based on our empirical results to overcome a struggle. Before that, it is vital to understand the stream of the literature review underlying our research model, as presented in Section 2. To develop an approach that matches the objectives, Section 3 addresses research design and data analysis. Section 4 tests hypotheses and interprets the results, while Section 5 discusses the key findings in comparison with prior studiesconcludes the study, and discusses limitations and future directions.

2. Literature Review

The research topics studied in this paper involve the fields of marketing and innovation. In order to understand the green customer attitude change in purchasing green products on an online platform across specific groups, this paper organizes the streams of literature, including green products on an online platform, green customer attitudes, the essence of income, gender, and age properties, perceived risk, perceived online social norms, perceived online compatibility, and perceived relative advantage as illustrated in Figure 1. However, the theoretical framework of variables and the development of hypotheses made and mixed are reviewed.

2.1. Green Products on an Online Platform

The research context focuses on the purchase of green products on an online platform. The purchase is formed by customer attitudes, representing how customers feel about purchasing green products online. In this study, “green products” are defined as any products whose qualities are designed to be toxin-free, maximize resource efficiency, use water and energy efficiently, and minimize environmental impacts. Their characteristics include being manufactured under hygienic conditions without toxic chemicals, being biodegradable and recyclable, being eco-efficient, using the minimum number of resources, and having a minimal carbon and plastic footprint. Therefore, green products specified in this research are eco-friendly home consumption items such as recycled toilet tissue, biodegradable cleaning cloths, natural cleaning products, and organic produce. In this study, the term “online platforms” refers to the software application or website that an online store uses to handle all the buying needs of its merchants and the selling needs of its own sellers. To sum up, purchasing green products on an online platform involves using e-commerce apps or websites to help people find green products online.

2.2. Green-Awakening Customer Attitude Changes to Purchase Green Products on an Online Platform

The definition of attitude is “the extent to which a person has a favorable or unfavorable view or assessment of the conduct” [28]. From this perspective, attitude can be viewed as an evaluative response to the potential occurrence of a given activity (e.g., the purchase of a product), which could be positive or negative to varying degrees. Furthermore, through learning processes, attitudes are developed over time. Thus, an individual’s action appears to be guided by a previously formed attitude when a decision must be made [29]. This research defines “green awakening customer attitudes” as a set of a customer’s beliefs, feelings, and behavioral intentions toward a sustainable business, which they start to become green-conscious about at a certain level but may not be highly engaged with and committed to. This means this type of green customer is in the awareness stage of green behavioral transformation.

2.3. The Moderation Effect of Age

Age can make a difference in the view of customers when purchasing green products on an online platform. In this research, age is viewed from time and experience perspectives representing career stage age. As this research was carried out in Thailand, which represents an emerging market, the ages of interest in this research are divided by career-starting age into (1) 30 and under [early career age] and (2) over 30 [mid-career age]. Younger people typically have better online experiences and place a greater emphasis on utility and attitude. When it comes to online purchases, older people, on the other hand, see more risks, have trouble placing complicated order instructions, and put more weight on how confident they feel about using an online platform for purchases [30]. Several studies (see Simons et al., 2018 [31]; Al-Somali et al.,2009 [32] and Nguyen et al., 2019 [33]) have incorporated age as a crucial element in explaining online shopping behavior. This is because the nature of online and green customer behavior differs depending on age. Al-Somali et al. (2009) examined the use of e-banking as one of the online platforms by a sample of seasonal customers and discovered that age was not associated with attitude and, thus, did not significantly influence their behavior [34]. However, this finding seems to be opposite to that of Hernández et al. (2011) [35], when a higher rate of online purchases depends on the variety of age groups and their online experience.

The online or internet experiences of older people influenced their assessment of the risks associated with Internet use. Previous experiences with online buying would influence elder customers’ perceptions of risk and benefit. However, age and internet experience were not found to increase online purchase intentions significantly but customer choices and availability were. Nguyen et al. (2019) [33] demonstrated that age affects customers’ choices when purchasing “green commodities” [33]. Prior studies by Watanabe et al. (2021) [34] and Yin et al., 2016 [35] found that customers are more likely to be able to afford green products such as organic food when their early career ages started to gain mid-high income at a certain level. Thus, age changes the strength of the relationship between green customer attitudes and makes a difference in buying green products online; the following hypothesis was proposed:

Hypothesis 1 (H1).

Age has a moderation effect on green customer attitudes and their determinants, causing differences in green-awakening customer attitudes toward buying green products on an online platform.

2.4. The Moderation Effect of Gender

The effect of gender differences on customer behavior has been of particular interest in the marketing field. Quoquab et al. (2020) suggested that gender is a significant personal trait that influences the decisions and actions of individuals [36]. Suki (2013) demonstrated that gender is a significant demographic determinant of customers’ green purchasing behavior [37]. Pickett-Baker and Ozaki (2008) also found that most green customers are females since they tend to be more confident in online capabilities [38]. Similarly, Felix et al. (2022) found that female customers are more likely to engage in environmentally conscious behavior than male customers [39]. When attitudes cross gender lines, the roles of customers’ gender in green attitude transformation are important. For example, online customers based on gender differences in choices and outcomes may like green food and products when they see a lower risk, a style of consumption that fits with their own, online references and recommendations, and attractiveness. [40,41,42]. However, this is independent of their own behavior [42,43]. Men and women often have distinct preferences, which may help to account for gender variations in attitudes and decisions. It is impossible to deny that the relationship between green customer attitudes and their determinants differs as a function of the customers’ gender, which leads to the following hypothesis:

Hypothesis 2 (H2).

Gender has a moderation effect on green customer attitudes and their determinants, causing differences in green-awakening customer attitudes toward buying green products on an online platform.

2.5. The Moderation Effect of Income

Online customers perceive lesser implicit risks associated with online transactions when they have a higher income, influencing their demand for internet-based goods and services [44]. Low income makes people less likely to buy things online, and when incomes go up, self-efficacy, usability, and usefulness should improve because people can handle possible financial losses. He et al. (2019) suggested that customers’ purchasing attitudes and behavior could be influenced by their economic level [45]. Since eco-friendly products typically involve a higher price premium than conventional products, customers’ income may impact their behavioral attitudes and purchasing decisions. With higher salaries, customers will have greater affordability and freedom to make environmentally conscious choices. Additionally, they will be better able to put their optimistic outlook and feeling of responsibility into action. Moreover, Alrawad et al. (2023) indicated that all experienced e-shoppers exhibit identical online purchase patterns, regardless of their incomes [46]. Income acts as a categorical moderator to change how customers feel, believe, and behave about buying green products online, which consistently reflects green customer attitudes. Thus, the following hypothesis was proposed:

Hypothesis 3 (H3).

Income has a moderation effect on green customer attitudes and their determinants, causing differences in green-awakening customer attitudes toward buying green products on an online platform.

2.6. Perceived Risk of Green Products on an Online Platform

Perceived risk encompasses the type and degree of uncertainty customers evaluate when purchasing goods or services [47]. Risk perception has been identified as one of the primary motivators of customer behavior. Due to the inability of customers to physically examine things before purchase, internet purchasing is associated with a substantially higher risk perception than traditional shopping channels [48,49,50,51]. Additionally, there is a growing knowledge asymmetry between buyers and sellers in the e-commerce market [52]. When converting customers who merely browse for green products or services on the Internet into actual buyers, research indicates that it is essential to comprehend and lower their risk perception [53].

Perceived risk has been discovered to have a significant impact as one of the key predictors of online customer behaviors [54,55,56]. Perceived risk has also been found to have a considerable impact on other e-commerce factors, including customers’ e-commerce attitudes, trust, customer satisfaction, and repurchase intent [49,50,57,58]. Thus, perceived risk has been demonstrated to be one of the most significant obstacles to online purchases [50].

Previous studies have identified four risks associated with online purchase behavior: financial, product, time, and psychological risks [59]. Moreover, it is believed that the greater the uncertainty surrounding a product innovation, the greater the perceived risk, which acts as a barrier. This demonstrates that the risk is primarily caused by a lack of trust in environmentally friendly products. The significance of perceived risk in determining online impulse purchases has not been established despite the popularity of online purchasing behavior [60,61]. Users’ perceptions of online services are influenced by their risk perception [62]. Thus, the following hypothesis was developed:

Hypothesis 4 (H4).

Perceived risk of green products on an online platform has a negative effect on green customer attitudes toward buying green products on an online platform.

2.7. Perceived Online Social Norms

Social norms are intrinsically linked to members of a social network and can impact their conduct. The social norm indicates the strength of normative beliefs and the motivation of the individual to comply with these ideas [63]. Social norms reflect the social pressure that an individual perceives regarding topics; they substantially influence their behavioral intentions. The individual’s need for approval primarily determines social norms [64].

Individuals perceive social pressure while contemplating a specific activity to satisfy their need for social acceptability [65]. This condition is shown by adhering to societal norms and exhibiting accepted behavior. The social norms comprise descriptive and injunctive motivators for individuals [66]. Descriptive norms are associated with the beliefs of what most people in the reference group do. Individuals are motivated to imitate the actions of others due to their reflecting nature. The injunctive norms, on the other hand, are associated with moral endorsement or disapproval and which action should be performed or avoided within the reference group. Previous research has established the influence of social norms on the pro-social conduct of others [67]. When it comes to online, this research coined the term "online social norms," which represents how customers should act in certain situations online and is set by shared standards and practices of acceptable online behavior by groups. Therefore, social norms are viewed as one determinant of being influenced by other collective groups of people. This affects customer attitudes.

Kim et al. (2012) [68] indicate that socially descriptive and injunctive norms have a higher influence on buying behavior, in which attitudes act as intermediaries to intervene in customer behavior. Additionally, recent studies (see Wan and Choi (2018) [69]; Sadiq et al. (2021) [70] and Suk et al. (2021) [71]) indicate that social norms affect how customers interact in terms of the alignment of purchasing feelings and actions [71,72,73]. (Online) social norms were found to be key factors that encourage the adoption of an online platform to buy green products; in turn, norms also determine users’ attitudes [62]. Based on the findings, this research hypothesized:

Hypothesis 5 (H5).

Perceived online social norms have a positive effect on green awakening customer attitudes toward buying green products on an online platform.

2.8. Perceived Online Compatibility

Compatibility is defined as “the extent to which potential adopters perceive an innovation to be consistent with their existing values, past experiences, and needs” [72]. Several studies in the field of technology adoption have demonstrated that compatibility significantly influences users’ attitudes and intentions to use new technology [29,73]. It was argued that when technology is compatible with users’ current requirements, uncertainty decreases, and adoption rates improve. In this research, online compatibility is defined as how well the use of an online platform matches the values, experiences, and needs of potential customers when it comes to the purchase of green products online.

The predicted association between personal traits and the amount of time spent shopping is a significant part of marketing research on shopping and personal characteristics [62]. The lifestyle analysis is an effective way for segmenting the target market. Lifestyle is a well-known notion in market research and customer behavior; nonetheless, it is occasionally conflated with subcultures, social movements, and status groups [74]. Consequently, lifestyle compatibility takes hold of customer behavior and product, brand, and service selections. Previous research indicated that the influence of compatibility on customer sentiments regarding online services is substantial [62]. This research assumed:

Hypothesis 6 (H6).

Perceived online compatibility has a positive effect on green awakening customer attitudes toward buying green products on an online platform.

2.9. Perceived Relative Advantage of Green Products on an Online Platform

Relative advantage is the degree to which an innovation outperforms existing concepts or procedures [75]. Prior research has highlighted relative advantage, compatibility and outcome expectations as significant factors of a behavior’s functional performance. Particularly, these variables were identified as the most influential predictors of behaviors such as the adoption of online channels [74]. The customers must be satisfied with the added benefits of online purchasing, or they would opt to continue using conventional shopping methods. Comparing online buying with traditional purchasing from the customer’s perspective, internet shopping offers distinct advantages and benefits. Firstly, online shopping enables customers to acquire goods and services anytime and from any location. Second, online purchases allow customers to save money, time, and effort when purchasing goods. Thirdly, online shopping provides customers with the ability to search and gather more information, as well as a high level of convenience and transparency. Therefore, these benefits would have a considerable and favorable impact on customers’ perceptions of online buying [76]. The previous study demonstrated that relative advantage had a positive effect on the utilization of cloud computing [76]. It was found that the relative advantages of online platforms are crucial variables that stimulate the adoption of such platforms [76]. This research assumed:

Hypothesis 7 (H7).

Perceived relative advantage has a positive effect on green awakening customer attitudes toward buying green products on an online platform.

3. Methods

3.1. Sampling and Data Collection

Electronic commerce was chosen as a research context because it is essential to all online business and customer behavior settings [77]. It was also the interactions of customers in an online platform and ecosystem. Purposeful random sampling was used to find a possible representative sample for the research [78]. During 2020, data were collected from customers who would probably have experience with a possible existing e-commerce platform. When looking at products online in Thailand, it is important to know how Thai customers feel about them. It was asked what makes customers feel differently about green products online and if their age, income, and gender affect how they feel about green products such as:

• recycled toilet tissue,
• biodegradable cleaning cloths,
• natural cleaning products, and
• organic produces

This study used these products as the basis for a survey to teach customers about green products and figure out how they felt about them. These green products appeared to influence customer attitudes so that customers may or may not feel positive about recycled toilet tissue after hearing about it. The category of green products in this research was focused on eco-friendly home consumption items, which customers might consume commonly.

To choose an appropriate sample out of the Thai population to participate in the surveys, the inclusion criteria were set up during the sampling process as follows:

• The respondents must be over 20 years old. This practice is defined by research ethics and compliance.
• The respondents must identify their geographical region.
• The respondents must make online purchases on any online platform.

The respondents will be excluded from the study if one of these criteria is missing.

To adequately test multigroup moderation hypotheses, 100 observations per group are the minimum sample size [79]. However, the sample size for this study was based on the number of observed variables. As a rule of thumb, 10 observations per indicator variable were used to determine the sample size. An adequate sample size of 160 was suggested. A total of 348 Thai citizens were approached in different parts representing the Northeastern, Northern, Central, Eastern, and Southern regions of Thailand. Because there was no complete list of sample names for random selection, this study approached the targets by posting permission on a specific Facebook group (i.e., a group relevant to online shopping). The survey was completed via an online channel using a Google Form.

3.2. Measures

All the question items were measured on a 7-point Likert scale and came from previous research to make sure that the constructs were valid and reliable. Two sets of 7 points on a Likert scale were applied: (1) the level of agreement, and (2) the level of risk concern. Conventional translation and back-translation, following Sperber’s (2004) procedures [80], were performed to ensure equivalence of meaning and translation accuracy. The pilot study was conducted on 30 Thai customers to check the relevance of the question content. The wording of the questionnaires was made clearer and easier so that customers with different levels of education could understand them. However, these 30 pretest samples were removed, and the survey was re-distributed. Appendix A contains the list of survey instruments and the question patterns that were used to measure each key latent variable. The questionnaire included respondents’ demographic background, such as gender, geographical regions, age, income, types of green product purchases, and online purchase frequency, and key questions relating to the perceived relative advantage of green products on an online platform; perceived online compatibility; perceived online social norms; perceived risk of green products on an online platform and customer attitude changes to purchase green products on an online platform.

Perceived relative advantage of green product on an online platform: Respondents’ relative advantage was evaluated from their perception with the three-item scale adapted from Huang (2018) [74]; Khayer et al. (2020) [75] and Mombeuil and Uhde ( 2021) [76]. They were asked to rate their agreement level with each key statement to reflect their perceived relative advantage of green products on an online platform. Key elements measuring relative advantage included (PRA3) money and time saving; (PRA2) green certified product and (PRA1) green product information.

Perceived online compatibility: The three-item scale developed by prior research, such as Elnadi and Gheith (2022) [62], was used for this study. The scale contains three dimensions of perceived online compatibility, such as (POC3) content needed; (POC2) online lifestyle fit and (POC1) quick response.

Perceived online social norms: Respondents’ online social norms were a composite of three items adapted from Çelik (2011) [63]; Elnadi and Gheith (2022) [62] and Mou and Lin (2015) [64]. Each item was designed to assess respondents’ perceptions on a different aspect, such as (POS3) online reference group; (POS2) most people involved and (POS1) collective opinion and recommendation.

Perceived risk of green products on an online platform: This research adapted a mix of a three-item scale developed by Amirtha et al. (2021) [47]; Chiu et al. (2014) [51]; Kumar and Bajaj (2019) [48] and Mou et al. (2017) [49] that measures activities that are designed to capture (PR3) an unmatched product portfolio, (PR2) information privacy, and (PR1) delivery risk. Unlike other constructs, the construct of perceived risk was applied with reverse scoring runs going in the opposite direction.

Green awakening customer attitude changes to purchase green products on an online platform: The four-item scale was based on previous studies by Barrutia and Echebarria (2021) [81]; Lee and Chow (2020) [28]; Sharma and Foropon (2019) [14]; Srisathan and Naruetharadhol (2022) [82]. The respondents were asked to rate how they felt about the purchase of green products on an online platform, such as (GACA 1) positive feeling; (GACA 2) likeness; (GACA 3) intent and (GACA 4) belief.

4. Data Analysis and Results

The data for this research were multi-sample in nature; therefore, all hypotheses were analyzed using the multigroup moderation approach. To examine moderation using multigroup analysis, when the moderator is a categorical variable (i.e., segmented into different groups of interest), predictive models can be used to investigate interactions. For example, when age-based segmentation is categorical where age groups with a group equal to 30 and under 30 = 0 and with a group above 30 = 1, a multigroup SEM can be tested estimating separate slopes ${\beta }_{\le 30}$ and ${\beta }_{>30}$ in one model. In addition, using latent variables usually presupposes the measurement invariance tests briefly described above before proceeding to the predictive analyses of group differences. Therefore, the multigroup moderation approach was justified as a useful tool for seeing more precise insights into segment characteristics than the data on an aggregate level.

To test moderation using multigroup analysis, this research carried out several steps of data analysis. The first step of the data analysis was to investigate confirmatory factor analysis (CFA) with AMOS 28 and then the measurement model. The second step was to estimate the aggregate structural model as the baseline model. To ensure that scores from the variable operationalization have the same meaning under different conditions, the third step was to perform measurement invariance in CFA. The fourth step was to test structural invariance. This was conducted by looking at the differences between groups based on age, gender, and income.

To describe the characteristics of the sample, descriptive analysis was conducted.

Table 1. Profile of respondents.

Demographics	Sub-Variables	Gender (n =348)			Age (n = 348)			Income (n = 348)
		Male (n = 157)	Female (n = 191)	Pearson Chi-Square Test	≤30 (n = 167)	>30 (n = 181)	Pearson Chi-Square Test	<20K (n = 109)	20K–30K (n = 108)	>30K (n = 131)	Pearson Chi-Square Test
Geographical Regions	The North	12	15	Chi-square = 2.798 (p-value = 0.592)	23	4	Chi-square = 52.676 (p-value < 0.001 ***)	7	16	4	Chi-square = 83.771 (p-value < 0.001)
	The North-east	76	86		70	92		67	27	68
	The Central	34	44		20	58		10	19	49
	The East	14	11		23	2		6	17	2
	The South	21	35		31	25		19	29	8
Status	Single	66	109	Chi-square = 7.786 (p-value = 0.005 **)	119	56	Chi-square = 7.786 (p-value < 0.001 ***)	83	60	32	Chi-square = 65.399 (p-value < 0.001 ***)
	Married	91	82		48	125		26	48	99
Green Product Types	Beverage	30	35	Chi-square = 0.035 (p-value = 0.852)	52	13	Chi-square = 32.816 (p-value < 0.001 ***)	23	33	9	Chi-square = 22.477 (p-value < 0.001 ***)
	Food	127	156		115	168		86	75	122
Online Purchase Frequency	Less than 3 times a week	102	127	Chi-square = 2.022 (p-value = 0.568)	72	157	Chi-square = 80.966 (p-value < 0.001 ***)	65	48	116	Chi-square = 64.251 (p-value < 0.001 ***)
	4–6 times a week	24	36		41	19		27	23	10
	7–9 times a week	13	11		20	4		7	13	4
	More than 10 times a week	18	17		34	1		10	24	1

Note: *** p < 0.001; ** p < 0.01. Source. Data from this study (2022).

shows the summary of the respondents’ profiles. Most of the responding customers were in the northeastern region of Thailand, which accounted for 48.41% of the males and 45.03% of the females, while 57.96% of the males were married and 57.07% of the females were single. Both men and women tended to buy green food more than beverages, which accounted for over 80%. The number of times that most customers buy green products was less than three times a week for both male and female groups, which accounted for 64.97% and 66.49%, respectively. When segmented by a range of ages, most customers with ages below 30 and above 30 lived in the northeastern part of Thailand. As for those below 30 years old, they were single, which accounted for 71.26%, while those above 30 years old were married, accounting for 69.06%. Furthermore, the majority of green products they bought were food items, which accounted for 68.86% of those under 30 years old and 92.82% of those over 92.82%. Like gender-based segmentation, most customers in a range of below 30 years old and above 30 years old tended to purchase green products online up to three times a week; the group above 30 years old was considered for 86.74%, followed by those below 30 years old for 43.11%. Looking at the income features, over half of the customers tended to have incomes below 20,000 THB (61.47%) and more than 30,000 THB (51.91%) in the northeastern region, while those with incomes between 20,000 and 30,000 THB were in the southern region. Married individuals had incomes over 30,000 THB, while single individuals had income levels of less than 20,000 THB (76.15%) and in a range of 20,000–30,000 THB (55.56%). From all ranges of income perspectives, over 60% of customers went for green food. The majority of Thai customers still bought those green products up to three times a week, no matter how much they earned, accounting for over 40% of each income level.

In this study, both Harman’s single factor and the common latent factor (CLF) were used to find the common variance. Harman’s single factor test yielded the total variance extracted by a single factor at 38.828%, which was not over 50% of the recommended variance. Then, the method of the common latent factor was used to confirm the reasonable common variance among all observed variables in the model. After common method bias is evaluated, it is essential to evaluate measurement models in terms of reliability and validity. In order to do this,

Table 2. Construct validity and reliability.

Constructs	Items	λ	p-Value	AVE	CR	${\mathsf{\rho }}_{\mathit{T}}$	μ	${\mathit{R}}^2$	VIF
Perceived online social norms	POS1	0.84	***				3.825	0.706	3.401
	POS2	0.781	***				3.716	0.61	2.564
	POS3	0.668	***	0.587	0.809	0.805	3.739	0.446	1.805
Perceived online compatibility	POC1	0.844	***				3.305	0.712	3.472
	POC2	0.829	***				3.023	0.688	3.205
	POC3	0.868	***	0.718	0.884	0.883	3.135	0.754	4.065
Perceived risk of green products on an online platform	PR1	0.849	***				3.672	0.721	3.584
	PR2	0.856	***				3.615	0.733	3.745
	PR3	0.799	***	0.697	0.873	0.872	3.693	0.638	2.762
Perceived relative advantage of green product on an online platform	PRA1	0.848	***				3.882	0.72	3.571
	PRA2	0.891	***				3.894	0.793	4.831
	PRA3	0.742	***	0.688	0.868	0.864	3.963	0.55	2.222
Green awakening customer attitude changes to purchase green products on an online platform	GACA 1	0.785	***				3.733	0.617	2.611
	GACA 2	0.823	***				3.822	0.677	3.096
	GACA 3	0.771	***				3.882	0.595	2.469
	GACA 4	0.771	***	0.621	0.867	0.867	3.851	0.594	2.463

Source. Data from this study (2022). Note: *** p < 0.001; λ = Standardized factor loading; ${\rho }_T$ = Tau-equivalent reliability (Cronbach’s alpha); μ = mean; $R^2$ = Squared multiple correlations; AVE = Average variance extracted; CR = Composite reliability; VIF = Variance inflation factor. The test of construct validity and reliability is a result of aggregate data, which is data input.

shows that the confirmatory factor analysis was first looked at; the standardized regression weights yielded values of more than 0.60 and ranged from 0.668 to 0.891, which was in line with the value of between 0.6 and 0.7 recommended by Hair et al. (2017) [83]. The CFA model with the CLF was then performed; the standardized regression weights with the common latent factor yielded values above 0.6 and ranged from 0.669 to 0.888. When comparing the standard regression weights of the CFA model to the standard regression weights of a model with the CLF, the large differences were found to be less than 0.2 or 20%, as recommended by Kock et al. (2021) [84]. Both methods demonstrate that the common method bias was not present.

Since all models could be wrong, academic research using statistics cannot obtain a correct one through excessive elaboration; the estimation of model fit is required.

Table 3. Model fit indices.

Fit Indices	CMIN/DF	SRMR	RMSEA	IFI	CFI
CFA (Aggregate)	2.962	0.0431	0.075	0.949	0.948
CLF	2.314	0.038	0.062	0.972	0.971
SEM (Aggregate)	2.962	0.0431	0.075	0.949	0.948
Grouped by age	2.437	0.0425	0.064	0.922	0.921
Grouped by gender	2.179	0.0737	0.058	0.939	0.939
Grouped by income	2.218	0.0488	0.059	0.905	0.903
Threshold	<3	<0.08	<0.08	>0.9	>0.9
Assessment	✓	✓	✓	✓	✓

Source. Data from this study (2022).

shows the overall measurement model to estimate model fit indices. At the aggregate level, a confirmatory factor analysis (CFA) test was then used to look at the fit indices of the measurement model. The CFA results demonstrated that the proposed factor structure fits the data well: the normed chi-square = 2.962; the standardized root mean square residual (SRMR) = 0.0431; the root mean square of approximation (RMSEA) = 0.075; the incremental fit index (IFI) = 0.949 and the comparative fit index = 0.948. The measurement model for the common latent factor was satisfactory: the normed chi-square = 2.314; the root mean square residual (RMR) = 0.038; the root mean square of approximation (RMSEA) = 0.062; the incremental fit index (IFI) = 0.972 and the comparative fit index = 0.971. As for the structural model, the overall fit indices suggest a good fit for the measurement model: the normed chi-square = 2.962; the standardized root mean square residual (SRMR) = 0.0431; the root mean square of approximation (RMSEA) = 0.075; the incremental fit index (IFI) = 0.949 and the comparative fit index = 0.948. The standardized factor loadings were statistically significant at the 0.001 level, and the regression weights were higher than the recommended threshold of 0.6. This means that indicator variables have enough variation to show how important they are in explaining a construct. All of the extracted average variances (AVEs) were greater than the cutoff of 0.5, and the composite reliability (CR) was greater than the threshold of 0.7. This means that the construct is convergent.

After convergent validity was confirmed for indicator variables belonging to each construct, discriminant validity was required to confirm whether a test measures the concept it was designed to measure.

Table 4. HTMT Discriminant validity.

Heterotrait-Monotrait Ratio of Correlations (HTMT)
	GACA	PRA	PR	POC
PRA	0.892
PR	0.021	0.104
POC	−0.147	−0.094	0.625
POS	0.850	0.853	0.125	−0.207

Note: PR = Perceived risk; POS = Perceived online social norm; POC = Perceived online compatibility; PRA = Perceived relative advantage; GACA = Green awakening customer attitudes. Source. Data from this study (2022).

shows the result for assessing discriminant validity using the heterotrait-monotrait ratio of correlations (HTMT). The HTMT criterion demonstrated an average correlation between two reflective constructs that was not over 0.90, as recommended by Henseler et al. (2015) [85]. This indicates that the constructs are distinct from one another. Purposely, the cross-loading approach can be traced back to exploratory factor analysis, where this research aimed to examine indicator loading patterns to identify indicators that have high loadings on the same factor and those that load highly on multiple factors. In

Table 5. Item-level discriminant validity.

Construct	Items	PRA	PR	POC	POS	GACA
Perceived relative advantage of green product on an online platform	PRA3	0.742	0.077	−0.07	0.63	0.661			1
	PRA2	0.891	0.093	−0.084	0.756	0.794			0.9
	PRA1	0.848	0.088	−0.08	0.72	0.756			0.8
Perceived risk of green products on an online platform	PR3	0.083	0.799	0.499	0.1	0.017			0.7
	PR2	0.089	0.856	0.535	0.107	0.019			0.6
	PR1	0.088	0.849	0.53	0.106	0.018			0.5
Perceived online compatibility	POC3	−0.082	0.542	0.868	−0.179	−0.128			0.4
	POC2	−0.078	0.518	0.829	−0.171	−0.122			0.3
	POC1	−0.079	0.527	0.844	−0.174	−0.124			0.2
Perceived online social norms	POS3	0.567	0.083	−0.138	0.668	0.566			0.1
	POS2	0.663	0.098	−0.161	0.781	0.662			0
	POS1	0.714	0.105	−0.173	0.84	0.713			−0.1
Green awakening customer attitude changes to purchase green products on an online platform	GACA 4	0.687	0.017	−0.114	0.654	0.771			−0.2
	GACA 3	0.687	0.017	−0.114	0.654	0.771			−0.3
	GACA 2	0.733	0.018	−0.121	0.698	0.823			−0.4
	GACA 1	0.7	0.017	−0.116	0.666	0.785			−0.5
									−0.6
									−0.7
									−0.8
									−0.9
									−1

Source. Data from this study (2022).

below, the bolded items are the factor loadings for each construct, while the non-bolded items for the same construct are the cross-loadings. Despite no theoretical justifications or empirical proof of item-level or cross-loading discriminant validity, this research applied a similar criterion in which the correlation between items of a different construct should not be over 0.9 or −0.9 [85]. The result yielded a range from −0.179 to 0.891. The cross-loading for each construct is slightly moderate, indicating good discriminant validity.

To assess if an instrument is interpreted in the same way across different groups of customers, measurement invariance is conducted. The results of the hierarchical invariance levels were satisfactory, which allowed us to claim full measurement invariance as illustrated in

Table 6. Measurement Invariance.

Gender
Levels of Invariance	CMIN/DF	RMSEA	IFI	CFI	Assessments
Configural equivalence (Unconstrained)	2.179	0.058	0.939	0.939	✓
Metric equivalence (Measurement weights)	2.115	0.057	0.939	0.938	✓
Scalar equivalence (Measurement intercepts)	2.113	0.057	0.934	0.934	✓
Thresholds	<3	<0.08	>0.90	>0.90
Age
Levels of Invariance	CMIN/DF	RMSEA	IFI	CFI	Assessments
Configural equivalence (Unconstrained)	2.437	0.064	0.922	0.921	✓
Metric equivalence (Measurement weights)	2.454	0.065	0.916	0.915	✓
Scalar equivalence (Measurement intercepts)	2.739	0.071	0.891	0.89	✓ (Partial)
Thresholds	<3	<0.08	0.8 ≤ IFI ≤ 0.9	0.8 ≤ CFI ≤ 0.9
Income
Levels of Invariance	CMIN/DF	RMSEA	IFI	CFI	Assessments
Configural equivalence (Unconstrained)	2.218	0.059	0.905	0.903	✓
Metric equivalence (Measurement weights)	2.322	0.062	0.888	0.886	✓ (Partial)
Scalar equivalence (Measurement intercepts)	2.55	0.067	0.854	0.853	✓ (Partial)
Thresholds	<3	<0.08	0.8 ≤ IFI ≤ 0.9	0.8 ≤ CFI ≤ 0.9

Source. Data from this study (2022).

. The fit indices of configural invariance, where the same factor structure in each group is unconstrainedly evaluated, were reasonably fit. The normed chi-square = 2.179; the root mean square of approximation (RMSEA) = 0.058; the incremental fit index (IFI) = 0.939 and the comparative fit index (CFI) = 0.939. Moving to constrain measurement weights equally across groups, metric invariance is considered. The fit indices of metric invariance were reasonably fit: the normed chi-square = 2.115; the root mean square of approximation (RMSEA) = 0.057; the incremental fit index (IFI) = 0.939 and the comparative fit index = 0.938. To answer the question of whether means are also equivalent across groups, the scalar invariance model is proposed. The fit indices of metric invariance, where a constraining item intercepts or means equal across groups are evaluated, were reasonably fit: the normed chi-square = 2.113; the root mean square of approximation (RMSEA) = 0.057; the incremental fit index (IFI) = 0.934 and the comparative fit index = 0.934. After the confirmation of measurement invariance in the CFA model, the next step is to assess multigroup moderation analysis.

To see if defined data groups are different in their standardized path coefficients, multigroup structural invariance is examined. The multigroup measurement models were taken into account for three individual tests. The measurement model for an age-based sample was an acceptable fit: the normed chi-square = 2.437; the standardized root mean square residual (SRMR) = 0.0425; the root mean square of approximation (RMSEA) = 0.064; the incremental fit index (IFI) = 0.922 and the comparative fit index (CFI) = 0.921. The fit indices for the gender-based sample suggest an acceptable fit: the normed chi-square = 2.179; the standardized root mean square residual (SRMR) = 0.0737; the root mean square of approximation (RMSEA) = 0.058; the incremental fit index (IFI) = 0.939 and the comparative fit index (CFI) = 0.939. The measurement model for the income-based sample proposes acceptable fit indices: the normed chi-square = 2.218; the standardized root mean square residual (SRMR) = 0.0488; the root mean square of approximation (RMSEA) = 0.059; the incremental fit index (IFI) = 0.905 and the comparative fit index (CFI) = 0.903.

5. Discussion and Conclusions

The main purpose of this study was to investigate and confirm the existence of green-awakening customer attitudes to buying green products on online platforms in the emerging economy across equivalent measurements across age, gender, and income properties in Thailand. The findings supported all the proposed hypotheses. This research enriches previous theoretical work by providing the determinants of green-awakening customer attitudes. This section responds to the research findings that causally confirm the existence of green-awakening customer attitudes on online platforms in the emerging economy, especially Thailand.

Once this research has made sure that the way to measure constructs is accurate and reliable, it is important to move on to the next step, which is evaluating the results of the structural model. Figure 2 and Appendix B, which illustrates standardized path coefficients (β) for the aggregate structural model, show that Thai customer attitudes toward purchasing green products on an online platform are more likely to be positively affected by perceived relative advantage (β = 594; p < 0.001; critical ratio = 5.58) and perceived online social norms (β = 0.373; p < 0.001; critical ratio = 3.375) but negatively affected by perceived risk (β = −0.127; p < 0.05; critical ratio = −2.231), whereas perceived online compatibility (β = 0.065; p > 0.05; critical ratio = 1.092) was not statistically significant to determine the extent of green customer attitudes. So, the baseline model supported Hypotheses 3, 4, 5, and 7, but not Hypothesis 6. To see the proportion of the variance for a dependent variable explained by an independent variable, according to

Table 7. Structural model (Aggregate data).

Hypothesis	Path Relationship	β	p-Value	Critical Ratio	${\mathit{R}}^2$	Results
H4	PR → GACA	−0.127	0.026 *	−2.231	0.833	Supported
H5	POS → GACA	0.373	***	3.375		Supported
H6	POC → GACA	0.065	0.275	1.092		Not Supported
H7	PRA → GACA	0.594	***	5.58		Supported

Note: *** p < 0.001; * p < 0.05. PR = Perceived risk; POS = Perceived online social norm; POC = Perceived online compatibility; PRA = Perceived relative advantage; GACA = Green awakening customer attitudes. Source. Data from this study (2022).

, 83.3% of the squared multiple correlations ($R^2$) in Thai customers’ attitudes toward purchasing green products on an online platform can be explained by taking into account relative advantage, social norms, and risk.

We found that the perceived relative advantage of green products on an online platform positively affects customer attitudes toward green products. This finding is consistent with Mombeuil and Uhde (2021), [76], that the relative advantage of the online platform can favorably change customers’ or users’ attitudes when they perceive some advantages over the traditional (purchase) methods. Our current research finding, Hypothesis 7, demonstrates the significant relative advantages of searching for green products online over other traditional purchase methods. In the case of the purchase of green products on an online platform, the length of time that customers spend on the platform may affect how they judge how easy and useful it is compared to other traditional or physical purchases [74,75]. This is determined by their attitudes, which means that the more they perceive relative advantages in money, time, and confirmation of green certification, the more positively they tend to view purchasing green products online as a viable option.

Furthermore, we found that perceived digital social norms had a significant positive effect on customer attitudes toward purchasing green products on an online platform. This means that the more other groups of people are involved, the more positive or favorable their tendencies become. The finding is consistent with previous research by Wan and Choi (2018) [69]; Sadiq et al. (2021) [70] and Suk et al. (2021) [71], which found that social norms in an online preference group, as well as the large number of people involved, may influence how potential customers feel and act. The finding also showed that markets should pay attention to group opinions and recommendations; this is what the research offers and makes it different from the previous studies. The better the group’s opinions and recommendations, the more positive customers feel, act, and behave when purchasing green products online. However, as Kim et al. (2012) [68] suggest, marketers should consider descriptive and injunctive norms as two relevant types of social norm approaches, by which customer attitudes or behaviors differ from others.

In addition, we found that perceived online compatibility directly affects green awakening customers’ attitudes toward purchasing green products on an online platform. This result indicated that green product online platforms that fit well into customers’ lifestyles and daily habits would be more easily adopted for green consumption. This finding is in support of previous studies by Groß (2018) [86] and Elnadi and Gheith (2022) [62], which found that a deep understanding of the conditions of products is positively compatible with green and online lifestyles. To replace traditional purchases with online purchases, marketers must consider two additional factors: quick response and appealing content. These two points are brought up to talk about how buying green products online will fit with how potential green adopters feel about the problem that green products will help solve for the better. Most Thai customer behaviors do not involve consuming green products online all at once. Instead, they do it over time as they learn more and integrate it. It is all about their attitude.

However, this study also found that customers are less likely to buy green products online if they think they are risky. This is similar to what was found in earlier studies of online behavior [60,61] which found that risk decreases as customers’ favorable perception increase when they shop online. Risk has an inverse relationship with what and how customers feel, act, and behave. Taking control of risks regarding both green products and online platforms is the marketer’s job. According to the current evidence from this research, most Thai online food purchasing platforms did a good job of managing risks such as mismatched product portfolios, information privacy, and delivery risks.

To investigate differences in equivalent measurements across age, gender, and income properties in Thailand, the results of the multigroup structural invariance analysis are presented in

Table 8. Multigroup moderation analysis of structural invariance.

Hypothesis	Path relationship	Age (H1)
		30 and under (${\mathit{R}}^2=0.836$)				Above 30 (${\mathit{R}}^2=0.766$)
		β		p-value		β		p-value
H4	PR → GACA	−0.091		0.148		−0.367		0.033 *
H5	POS → GACA	0.389		0.057		0.302		0.026 *
H6	POC → GACA	−0.009		0.895		0.373		0.034 *
H7	PRA → GACA	0.545		0.004 **		0.689		***
Hypothesis	Path relationship	Gender (H2)
		Male ($R^2= 0.858$)				Female ($R^2=0.816$)
		β		p-value		β		p-value
H4	PR → GACA	−0.224		0.147		−0.138		0.08
H5	POS → GACA	0.421		0.159		0.407		***
H6	POC → GACA	0.114		0.491		0.106		0.216
H7	PRA → GACA	0.539		0.07		0.591		***
Hypothesis	Path relationship	Income (H3)
		Below 20,000 ($R^2=0.917$)			20,000–30,000 ($R^2=0.72$)		Above 30,000 ($R^2=0.891$)
		β	p-value		β	p-value	Β		p-value
H4	PR → GACA	−0.101	0.454		−0.122	0.123	−0.233		0.318
H5	POS → GACA	0.647	0.16		0.352	0.03	0.129		0.422
H6	POC → GACA	0.008	0.971		0.065	0.387	0.31		0.157
H7	PRA → GACA	0.306	0.375		0.557	***	0.869		***

Note: *** p < 0.001; ** p < 0.01; * p < 0.05. PR = Perceived risk; POS = Perceived online social norm; POC = Perceived online compatibility; PRA = Perceived relative advantage; GACA = Green awakening customer attitudes. Source. Data from this study (2022).

. When the multigroup analysis was evaluated, the structural model appeared to fit the response data for the above-30 group. The findings of Hypothesis 4 proposed that the perceived risk of green products on an online platform had a negative impact on green customer attitude change in the over-30 age group while finding no group-moderation effects in any other groups at the significance level of 0.05 (β = -0.367; p < 0.05). Although Hypothesis 4 was found to have significance in the data for the respondents over 30 years old, it did not show statistical significance among or between groups, as illustrated in Figure 3. Still, this supported Hypothesis 1, there was a difference in the age segmentation.

The findings of Hypothesis 5 revealed that perceived online social norms had a positive effect on green customer attitude change in the over-30 age group at a significance level of 0.05 (β = 0.302; p < 0.05) and in the female group at a significance level of 0.001 (β = 0.407; p < 0.001). Figure 4 illustrates the result of a multigroup structural model where gender properties are analyzed. The findings supported both Hypotheses 1 and 2. Differences in age- and gender-based attitudes toward considering the purchase of green products on an online platform were found among groups.

The findings suggested that female customers are more likely than male customers to engage in environmentally conscious and online behaviors as a result of peer pressure. The results were consistent with the previous study by Felix et al. (2022) [39], which found that green product-based effects may be varied by gender roles and color. When people buy green products online, their decisions are based on how they feel and see the products and what they value about them, such as helping to solve social and environmental problems [39].

The findings of Hypothesis 6 showed that perceived online compatibility had a positive influence on green customer attitude change in the over-30 age group at a significance level of 0.05 (β = 0.373; p < 0.05), supporting Hypothesis 1.

The presence of age segmentation supported the significant findings of Hypothesis 7, which revealed that the perceived relative advantage of green products on an online platform had a positive influence on green customer attitude change among the over-30 age group at a significance level of 0.01 (β = 0.545; p < 0.01) and the 30-and-under age group at a significance level of 0.001 (β = 0.689; p < 0.001), fostering Hypothesis 1.

In support of Hypothesis 2, this Hypothesis 7 relationship was observed in females with a difference at a significance level of 0.001 (β = 0.591; p < 0.001). Figure 5 illustrates the result of a multigroup structural model where income properties are analyzed. Looking at income-based gaps, it appeared that there was a 0.1% probability that the results were due to random chance and that Thai customers with 20,000–30,000 THB (β = 0.557; p < 0.001) and with more than 30,000 THB (β = 0.869; p < 0.001) perceived the relative advantage of green products on an online platform and had more positive attitudes toward purchasing green products on an online platform, supporting Hypothesis 3.

To further determine whether the groups were different, the critical ratio difference was calculated as shown in

Table 9. Critical Ratios for Differences between Parameters.

Hypothesis	Path Relationship	Critical Ratio Difference
		≤30 vs. >30	Male vs. Female	<20K vs. 20K–30K	20K–30K vs. >30K	<20K vs. 30K	Threshold
H4	POS → GACA	| −0.533 |	| −0.368 |	| −0.446 |	| −1.05 |	| −1.016 |	| ± 1.96 | *
H5	POC → GACA	| 1.789 |	| −0.297 |	| 0.286 |	| 0.824 |	| 0.801 |	| ± 1.96 | *
H6	PR → GACA	| −0.516 |	| 0.935 |	| −0.518 |	| 0.066 |	| −0.359 |	| ± 1.96 | *
H7	PRA → GACA	| 0.907 |	| −0.279 |	| 1.159 |	| 2.3 | *	| 2.959 | *	| ± 1.96 | *

Note: The critical ratio difference in absolute value corresponds to |1.96| at 5% significance level (p < 0.05 *). PR = Perceived risk; POS = Perceived online social norm; POC = Perceived online compatibility; PRA = Perceived relative advantage; GACA = Green awakening customer attitudes. Source. Data from this study (2022).

. The multigroup analysis showed that, to compare those with incomes higher than 30,000 THB (Δ critical ratio = |2.3| > z-test = |1.96|), Thai customers who held a range of incomes between 20,000 and 30,000 seemed to have a statistically significant difference where the effect could be observed. When comparing Thai customers’ earnings below 20,000 THB (Δ critical ratio = |2.959| > z-test = |1.96|), the same thing happened. The findings suggest that green awakening customer attitudes may be more valuable to Thai customers, particularly those with mid-to-high income levels because purchasing green products on an online platform is relatively expensive. Thus, income-based segmentation is most critical when Thai customers need to purchase green products on an online platform. In the context of Thailand, positive values and attitudes that are most strongly linked to green customer behaviors are socially altruistic; decisions to act green are based on what people think it will cost them to act in a green way. The critical ratio difference analysis answers the questions about whether there are differences in the green-awakening customer attitude changes to purchasing green products on an online platform across age, gender, and income properties in Thailand. This research finding revealed that the green awakening customer attitude model depends on some customer characteristics, such as gender, age, and income.

Furthermore, according to the test result of loading differences, current researchers found that income moderates the perceived relative advantage of online platforms and customer attitude changes. People who earn middle incomes are concerned about saving money and time when purchasing green products online. They are also concerned about green-certificated products and green product information when they purchase green products from online platforms.

5.1. Theoretical Implications

The primary goal of this study is to provide a thorough understanding of the construct of customer attitude change, which represents green-awakening customer attitudes in Thailand, and to investigate its determinants from customer perspectives in green product dynamics. By examining an integrated model of green customer attitude change with demographic segmentation, the study has provided a few theoretical and methodological contributions to the literature pertaining to customer attitudes in business, green innovation, and marketing management.

To begin with, the findings have extended the literature on green-awakening customer attitudes, which represents an increased awareness of collective green actions. This is how customers realize and feel supported about how much they rely on nature. Previous research (e.g., Sharma and Foropon, 2019 [14]; Sharma et al., 2020 [27]; Mehta and Chahal, 2021 [87]; Zaremohzzabieh et al., 2021 [88]) on green customer behavior assessed the impact of planned/cognitive determinants on behavioral intention in various cultural situations or assumed that customers shared collective conceptions of customer attitudes towards green products, but this study contends differently. In particular, the current study advances the knowledge of green customers beyond the cluster differences proposed by Mehta and Chahal (2021) [87] and emphasizes a deeper understanding of the purchase of green products on an online platform in dynamic consumption. This is the first study to divide green customer attitudes into two categories: green-awakening and green-awakened. As customers become more positive about green products, there is evidence of a green awakening. To measure green awakening, this research conceptualizes green customer attitudes through the lenses of green likeness, positive feeling, green belief, and green intent. These lenses are composite of the awareness of the green awakening.

In addition, this research is the first attempt to combine multiple determinants from three different theories into the proposed research framework. The purpose of this approach was not to discredit the results of previous work but to provide an alternative perspective on attitude testing in green products on an online platform. Using ideas from risk perception theory, social norms theory, and innovation diffusion theory, key variables show how people feel about green products on an online platform. In this research context, risk, social norms, and platform adoption are all used as predictors of customer attitudes.

Furthermore, the research framework included not only the impacts of perceived risk, perceived online social norms, perceived compatibility, and perceived relative advantage on green-online customer attitudes, but it also examined customer segmentation representing demographic differences. Demographic segmentation divides the market into segments based on age, gender, and income. Our findings suggest that green products might not be of interest to certain age groups, e.g., 30 and under. This study uses age, gender, and income as moderators. Our multigroup moderation analysis responds to the call of Hernández, Jiménez and Martín (2011) [30] that a multigroup moderator should be shown in relation to the penetration of attitudes as well as interests and opinions, giving clear direction for future studies on online (purchase) behavior. Unlike previous studies that have used aggregate analysis (see Lee and Chow, 2020 [28]; Dangelico, Nonino and Pompei, 2021 [26]; Sadiq, Adil and Paul, 2021 [89]) and cluster analysis (see Mehta and Chahal, 2021 [87]) of green customer behavior, this study measured a direct group-level score of green customer behavior focusing on green customer attitudes, instead of aggregating customers’ perceptions of green products on an online platform. In addition to demographics, attitudes are another major factor in categorizing customers. This study also expands on the findings that demographic profiles can identify what factors contribute to different customer attitudes, especially income.

5.2. Practical Implications

The findings can be applied as a reference for marketers, given the positive variability of customer attitude changes toward purchasing green products on an online platform. The results of the aggregate structural model show that characteristics of online commerce, including perceived relative advantage and compatibility, significantly affect green customer attitudes. Therefore, the perception of these two attributes of an online platform should be fulfilled.

Based on our empirical results, the following tips would help marketers prepare key strategies:

Clarifying the relative advantage of green products on an online platform. Marketers need to figure out how a green product-based advantage could be visualized over other traditional products online. Thus, this paper consolidated that the relative advantage of green products should be clearly demonstrated on an online platform. Marketers could leverage the key points of the relative advantage of green products online by using money and time, green-certified products, and green product information. As a part of the marketing mix, pricing may be a relative advantage when customers are price-conscious owing to limited purchasing power; this point is derived from PRA3, in which money and time matter.

Being consistent with compatibility. That is what matters, not just the purchase of green products. The comparability of both green products and the use of online platforms may influence and change purchasing attitudes. Marketers need to consider that it might take much discussion and time for customers to warm up to some green-innovative products on an online platform because green products on an online platform will be bought if those customers use online platforms as usual. This means that online purchases may be appropriate for customers who live an online lifestyle [POC2]. Marketers must ensure that the attractive content of green products is posted on an online platform for customers who do not shop online [POC3]. Furthermore, when customers need assistance or have questions, marketers must respond quickly on an online platform [POC1].

Ensuring positive presence of social norm messages. Marketers need to ensure the positive cognitive representations of online reference groups. Behavior and attitude patterns that are consistent within a social group’s message and set it apart from other groups. Key review messages and testimonials should be carefully delivered. For example, customers would not endorse the online purchase of green products because they could not try them.

Minimizing risk. Online platforms, merchants, and marketers should carefully act on the issues of online purchase-related risks such as unmatched product portfolio, information privacy, and delivery risk. For example, customers may buy from online merchants they have heard of, especially those with a good reputation.

5.3. Limitations and Future Research

This study is not without limitations. Because the data were collected in Thailand, the generalizability of the findings may be limited. In addition, the context of this study is limited to the online commerce industry available in Thailand. Although the online commerce industry is fundamental to all marketing environments, the findings should be cautiously interpreted. Because the data were collected before the COVID-19 pandemic, the effects of perceived risk, relative advantage, compatibility, and social norm on green customer attitudes toward purchasing green products on an online platform should be understood under normal circumstances. This is because the COVID-19 pandemic has shifted more customers to online shopping than before [82]. Findings may differ in the post-pandemic environment due to changes in online commerce patterns balanced with green products commercialized in physical stores. Therefore, future research could replicate this study in different contexts to examine the effects of perceived risk, relative advantage, compatibility, and social norms of green products on an online platform.

Future research is also encouraged to examine the measurement of green-aware customer attitudes in relation to purchasing green products on an online platform. As this current research was limited to the measure of "green-awakening" customer attitudes, the addition of "green engagement" and "green commitment" could help shape the cognitive methods of measuring these characteristics.