Introduction

The coronavirus (COVID-19) infectious disease, which struck the world at the end of 2019, has had a considerable global impact (Ministry of Health and Welfare, 2022). Since the first confirmed case in Korea on January 20, 2020, Koreans have been experiencing lifestyle changes due to COVID-19. Consequently, “Post-Corona,” refers to the new era after COVID-19 (Institute of PMG Knowledge Engine Lab, (2020)). However, during the COVID-19 era, predicting when the new era would begin was impossible. Moreover, the global economic downturn and environmental crisis caused by COVID-19 have become catalysts demonstrating sustainability’s importance for multinational companies (Han, 2021). For example, Handfield et al. (2020) emphasized building sustainable supply chains that transcend international crises, highlighting the advantages for companies with highly responsive and dense supply chains during the COVID-19 crisis. The COVID-19 pandemic has also changed human behavior, including consumer behavior and consumption patterns, affecting lifestyles, purchasing intentions and behaviors, desires, and consumption patterns of goods and services (Brzustewicz and Singh, 2021).

Even before the outbreak of COVID-19, the concept of sustainability emerged as a global topic for building the future of society. For example, the United Nations (UN) (2015) General Assembly adopted “responsible consumption and production” (Goal 12 of 17) in its 2030 Agenda for Sustainable Development Goals (SDGs). Further, Gardetti and Muthu (2019) presented a practical plan for applying SDG 12 to the textile and fashion industries, significantly impacting the environment and society. The researchers discussed the fashion industry’s efforts to develop traceability and transparency, a circular economy, functional textiles, and the consumers’ efforts to change their mindset towards sustainable fashion. Thus, the consumer’s pursuit of sustainable fashion has increasingly attracted attention during the current COVID-19 era. The IBM Institute for Business Value’s Survey of more than 14,000 consumers in nine countries conducted in March 2021 found that 93% of respondents said COVID-19 had affected their views on environmental sustainability, 84% of global consumers valued sustainability when choosing a brand, and 62% of consumers said they were willing to change their purchasing behaviors to reduce the negative impact on the environment (Orrell et al., 2021).

After COVID-19, researchers have found an increased need for sustainability. Due to the pandemic’s prolonged global impact, there has been a rise in awareness and demand for eco-friendly and vegan products that have boosted this category of apparel products (Baliyan and Diwan, 2021). The COVID-19 outbreak put this consumption model into a development dilemma. Thus, cultivating consumers’ sustainable behaviors is an effective way to promote the sustainable development of the apparel industry (Xu et al., 2022). In addition, the pandemic has changed people’s views on health, consumption, and the environment. For example, COVID-19 has brought an unprecedented psychological burden to people regarding health protection (Liu et al., 2022). The global pandemic experience has created awareness of sustainable fashion’s importance in rebuilding the fashion industry with high-quality durability. Stakeholders view sustainability as the main product priority, so by adopting a sustainable fashion concept, the expectation is for the fashion industry to work together to create more environmentally friendly innovations, empowering the industry (Kulsum, 2020). Accenture’s (2020) global survey found that shoppers “have dramatically evolved,” as 60% of respondents reported making more environmentally friendly, sustainable, or ethical purchases since the start of the pandemic.

However, researchers often blend the concepts of sustainable fashion and ethical fashion (Choi and Lee, 2020; Joergens, 2006; Kim, 2019; Kim, 2020; Ko, Han (2021); Koh and Lee, 2020; Kwon, 2009; Na and Lee, 2013; Yoon, 2013). Moreover, researchers have almost no consensus definitions or metrics regarding sustainable and ethical fashion consumption concepts.

In recent years, there has been growing recognition of the importance of sustainability and ethics within the fashion industry, especially in light of the COVID-19 pandemic. However, these two concepts are frequently conflated, resulting in theoretical and practical ambiguities (Byun et al., 2019; Koh & Lee, 2020). Recent studies highlight the necessity of distinguishing between these concepts to enhance both consumers understanding and industry strategies. For instance, Perry and Chung (2016) found that while consumers express increasing interest in ethical fashion, many still struggle to differentiate it from sustainable fashion, leading to mixed and inconsistent purchasing behaviors. This conceptual overlap undermines the ability of brands to effectively market their ethical and sustainable initiatives. By employing advanced LDA topic modeling and big data analysis, this study aims to provide empirical evidence that clarifies how consumers perceive these two critical dimensions, filling a significant gap in both academic and industry discourses. Through this approach, the research not only contributes to the academic field by refining the conceptual definitions of sustainability and ethics but also offers actionable insights for fashion brands looking to create clearer, more effective marketing strategies in the post-pandemic era.

For example, Byun et al. (2019) pointed out that if researchers mix similar concepts, the definition of their characteristics will inevitably remain insufficient. The concept of adopting sustainability within the fashion industry is a relatively recent development. While the idea of sustainability had existed for many years, it wasn't until the early 1990s that stylists and design trendsetters began to break away from the established norms in the fashion industry, exploring alternative approaches to apparel construction and consumption beyond the traditional archetype. (Azeem, Adamjee (2023)). Clark (2008) characterizes sustainable fashion as a counterforce to the fast fashion phenomenon. Eco-fashion represents one manifestation of anti-consumption, alongside other forms like eco-consumption, green consumption, ethical consumption, and political consumption, as explored by Joy et al. (2012) and Littler (2009). Furthermore, it is situated within the realm of the cultural economy, as emphasized in Culture Unbound (2014: vol. 6, theme 5).

Sustainable fashion primarily focuses on environmental protection and resource conservation, targeting practices such as recycling, reusing materials, and incorporating eco-friendly resources. For instance, Joergens (2006) highlighted that sustainable fashion aims to minimize the environmental impact of production processes, including waste reduction and recycling efforts. Additionally, Henninger, Alevizou, and Oates (2016) emphasized that sustainable fashion encourages consumers to engage in environmentally protective behaviors, leading to long-term economic and environmental value.

Ethical fashion, in contrast, encompasses social responsibilities such as labor conditions, fair trade, and animal welfare. Kwon (2009) explained that ethical fashion centers on guaranteeing fair trade and protecting workers’ rights, evolving as consumers increasingly recognize corporate social responsibility. Furthermore, Perry and Chung (2016) argued that ethical fashion offers consumers pathways to promote human rights and animal welfare, serving as a catalyst for industry transformation. Hoskins (2022) also supported this view, suggesting that ethical fashion provides a foundation for consumers to contribute to positive social change by choosing products aligned with human rights and animal welfare standards.

Given the present socioeconomic context, there is a growing inclination to associate luxury with sustainable and ethical fashion, as observed in studies (Joy et al. 2012). However, the researchers also emphasized that one can establish an appropriate marketing strategy for the relevant field when there is an arrangement and understanding of the two mixed concepts. Given the potential confusion between these concepts, it is essential to investigate consumer perceptions of sustainable and ethical fashion. Employing big data analysis techniques such as text mining and Latent Dirichlet Allocation (LDA) can provide empirical insights into how consumers differentiate these concepts in practice, thus addressing any ambiguity.

To provide a framework for a sustainable review of judicial dynamics, this paper presents a sustainable approach to three areas of conflict based on Felix Guattari’s Three Ecologies (2000). In Guatari’s model, all three ecologies form a systematic level: (1) environment, (2) society, and (3) spiritual realm. As you can see, the discussion of sustainable fashion suggests a broad change in the fashion system.

In the pursuit of a comprehensive understanding of consumers' perceptions regarding sustainable and ethical fashion, this study aimed to address potential limitations associated with subjective researcher opinions. While existing studies have typically relied on methods such as literature reviews, quantitative surveys, or qualitative interviews, we sought to augment these approaches. Instead, we leveraged the power of search engines to extract real-world data and applied advanced big data analysis techniques. This shift allowed us to empirically assess the nuanced aspects of sustainable and ethical fashion. Specifically, we wanted to discern any disparities between the terminology commonly used interchangeably in previous research and the actual words and phrases prevalent in public discourse. Moreover, we aimed to investigate whether there has been an observable increase in the demand for sustainability in fashion both before and after the onset of the COVID-19 pandemic.

In the context of the fashion industry, sustainability and ethics have gained increasing attention in recent years. This study aims to investigate the impact of the COVID-19 outbreak on discussions related to sustainable and ethical fashion consumption on social media platforms. To achieve this, we will explore What keywords and major topics which are related to sustainable fashion consumption and ethical fashion consumption appeared on social media before and after the outbreak of COVID-19.

Literature review

Analyzing concepts of sustainable and ethical fashions

Previous studies did not clearly define “sustainable fashion” and “ethical fashion,” using the terms in a somewhat blended pattern. From the perspective of environmental protectionism and social responsibility, sustainable fashion is a holistic concept that includes social issues such as material and production energy consumption and recycling, pollution reduction, disposal, consumer health and work environment, fair distribution, etc. Yoon (2013) found that previous studies discussed the term and definition of sustainable fashion regarding environmental protection, correct social ethics, and responsibility.

Researchers consider sustainable fashion “eco-friendly and socially responsible in all parts of the valuechain of the fashion product industry,” confirming the inclusion of these factors. In addition, Na and Lee (2013) define sustainable fashion as “a fashion philosophy that is environmentally friendly and fulfils the social responsibility of the future community” from a broader perspective than eco-friendly fashion and eco-fashion. In addition, sustainable fashion “takes into account cultural values for environmental conservation, economic growth, social contribution, and human welfare and future development” (Kim, 2020; Ko, Han (2021)).

Joergens (2006) defined ethical fashion as “incorporating fair trade principles into Sweatshop-free working conditions that do not harm the environment or workers by using biodegradable organic cotton.” A “fashionable outfit” is ethical fashion. Kwon (2009) defined ethical fashion as “using raw materials that do not negatively impact the global environment, paying fair wages in an environment where the human rights of producers are protected and taking social responsibility not to encourage excessive consumption” (p. xx). Kim (2019) states that ethical fashion fulfils the proper ethics a fashion brand should conduct or follow to create economic profits, maximize benefits to humans and society, and minimize environmental impact. It means fashion design, sourcing, and production.

Another study defines ethical fashion as “fashion considering environmental preservation, economic growth, social contribution, and cultural value” (Ko, Han (2021), p. xx), citing the same definition previously used for sustainable fashion (Choi and Lee, 2020).

Hoskins (2022) pointed out that various brand products from different companies are produced at a single production plant, and the only factor imposing production standards is a fragmented CSR (Corporate Social Responsibility) agreement. CSR programs are often designed for large companies that are sensitive to public pressure. In particular, shoes represent a complex network within the fashion industry, involving numerous value chains and containing potentially harmful chemicals that can affect both the workers producing the shoes and the consumers using them. However, the ability to impact consumer behavior solely based on issues within the production environment, such as adhesive factories, shoelace manufacturers, and outsole production plants, is limited.

CSR is carried out through voluntary participation by companies, and while the process of persuading consumers is important, the government bears the responsibility of safeguarding everyone, including raw materials, the industrial environment, companies, workers, and consumers. Companies that have demonstrated social responsibility should promote government-level incentives such as tax benefits, while those that do not should be subject to continuous regulation and monitoring, with punishments made public. In the fashion industry, sustainable and ethical consumption depends on consumer choices. As consumers increasingly demand more information about products, government agencies should be granted greater authority. This includes promoting legislation tailored to consumers' needs, similar to regulations for products like medicines and processed foods, where valuable information such as raw materials, nutrients, recycling methods, and country of origin is disclosed.

Previous studies about sustainable and ethical fashions using big data analysis

Guo et al. (2021) pointed out that the rapid development of the data analysis method provides opportunities to realize SDGs. Big data could improve the scope of adoption through scientific evidence and innovative technological solutions to achieve SDGs. In previous studies using big data analysis, we could mainly look at the relationships between words or the similarities to form groups. Previous studies examined the relationship between words. For example, in a study using big data analysis related to sustainable fashion, Youn and Jung (2021) analyzed the semantic network using social media big data analysis extracted from Facebook using “sustainable fashion” as a keyword. The keywords “eco-friendly,” “ethical,” and “recycling” had the highest frequency and connection centrality. As a result of keyword classification according to correlation, the researchers could classify the “sustainable fashion supply chain” into four groups: supply chain of sustainable fashion, circular fashion, fashion business concepts for sustainability, and academic importance of sustainable fashion. In addition, Blasi et al. (2020) examined the relationship between fashion and eco-friendliness in consumers’ perceptions of fashion brands. They found that eco-friendliness positively affected the brand image and that this effect was more significant for luxury brands than fast fashion brands. We measured the relationship between brands and the perceptual attributes of fashion and eco-friendliness by applying social network analysis to examine fashion clothing brands with the most followers using data collected from Twitter. Among the SNS collected keywords, the fashion-related keywords were “fashion,” “glamour,” and “style,” while the eco-friendly keywords were “environment” and “ethical business.” We found a correlation between fashion and the eco-friendly recognition attributes of brands, which was much stronger in luxury brands than in fast fashion brands. This difference demonstrates that fashion luxury brands have recently begun considering environmental issues as part of their core business.

A study on ethical fashion (Shin and Koh, 2021) extracted keywords representing the sub-dimensions of ethical consumption defined in previous studies, using “ethical consumption” to understand the changes over time. The authors identified the changes and analyzed the relationship between ethical consumption and consumption values. The results showed that ethical consumption changed over time. Initially, it was price sensitive, but it changed to value-oriented consumption. The researchers also used big data analysis and found that as ethical consumption increased, social perception changed to good consumption, indicating positive emotional value.

Previous studies also focused on grouping. For example, Koh and Lee (2020) analyzed the frequency of data collected through text mining using “sustainable fashion consumption” as a keyword. They conducted a comparative study to examine the frequencies between the two clusters pre-and post-COVID-19 using the convergence of iteration correlation (CONCOR) analysis. The results showed that the main categories of sustainable fashion were “sustainable growth,” “environmental value,” “recycling era,” and “online service” pre-COVID-19 and “sustainable trend” after the outbreak of COVID-19. In addition, the researchers classified the data into “pursuit of eco-friendliness,” “platform consumption,” and “digital expansion” to confirm the prediction direction of the literature research.

Choi and Lee (2020) analyzed domestic ethical fashion research trends through text mining and network analysis. They found that research on ethical fashion has increased since 2009, and the increase in case studies related to the ethical activities of fashion companies could reflect the spread of ethical awareness throughout the fashion industry. Thus, we identified ethical fashion research from the environmental aspect to the social aspect, such as corporate social responsibility and fair trade, and the animal protection aspect, such as vegan fashion.

LDA topic modeling

The topic modeling technique, a big data analysis method, makes it possible to objectively analyze large volumes of textual data (Jang et al., 2020). Topic modeling starts with the idea that there is a specific relationship between a subject and a word, and it defines the meaning according to co-occurring words rather than the word’s meaning. It analyses words that appear simultaneously in a document and recognizes them as a single topic based on the frequency of repetition (Lee et al., 2021). Blei et al. (2003) published the latent Dirichlet allocation (LDA) algorithm, a probabilistic generative model based on Dirichlet distribution. The basic assumption of LDA is that text data are composed of a mixture of topics, and one can infer the underlying topic by estimating the probability that a specific topic in the document exists and that a word exists in a specific topic as the combined probability (Kim, 2020). In other words, the basic logical structure of the LDA algorithm assumes that an individual text has a probability of distribution for a topic, and the author completes the article by selecting an appropriate word according to this distribution. Various research fields currently use the LDA model, making it possible to derive meaning and interpretation equivalent to a qualitative analysis by supplementing the limitations of existing keyword frequency analysis or network-based big data analysis (Lee, 2021).

Previous studies on fashion-related LDA topic modeling mainly focused on research trend analyses and consumer perception studies. In our research trend analysis study, we collected keywords and conducted text mining and topic modeling from abstracts of papers published in the first issue of the Journal of Digital Convergence. For example, to understand research trends in fashion design in Korea (Jang et al., 2020), Lee (2021) conducted a study of 340 papers by setting keywords related to possible fashion, in which the researcher performed topic modeling using the LDA algorithm for the paper’s abstract. On the other hand, An and Park (2017) conducted a consumer perception study related to LDA topic modeling based on text mining analysis on social media for men’s striped shirts. The researchers analyzed the topic and content of the posts using the LDA algorithm and investigated user perception accordingly.

Furthermore, another study conducted an in-depth text analysis of consumers’ online fashion rental experiences using LDA topic modeling and word co-occurrence networks (Lang et al., 2020). Finally, Brzustewicz and Singh (2021) used LDA topic modeling-based Twitter data to derive a topic, confirmed the clustering through the network structure, and analyzed the emotions of Twitter users using the NRC lexicon. Therefore, we propose the following research question based on the above discussion and previous studies.

Method

This study analyzed consumer perceptions of sustainable and ethical fashion consumption pre-and post-COVID-19. We explored texts appearing on social media using keyword analysis and LDA topic modeling among text mining techniques.

Data collection

We used Textom 6.0, a big data analysis software, to collect data for this study. Textom transformed full-text software developed by Professor Loet Leydesdorff to IMC, Inc. for Korean analysis and implemented it on the web. Textom 6.0 allows individuals to collect and refine data from various channels. This study’s data for the pre-COVID-19 period was from January 20, 2018, to January 19, 2020, and the post-COVID-19 data was from January 20, 2019, to October 19, 2021. We collected the texts using Korea’s two largest search engines, Naver and Daum and Google, excluding duplicate content.

The postings collected through social media included voluntary feedback on ethical consumption reflecting people’s perceptions, attitudes, and preferences. Using social media as collection channels has the advantage of converging in real-time a constantly growing large amount of user opinions at no cost (Ahn, Jang (2020)). Furthermore, given that social media postings are not limited in scope as opposed to the conventional user opinion convergence methods of receiving responses to fixed questions, it enables exploratory analysis of new information that escaped researchers’ attention and are thus of great help for academic research (Lee et al, 2021). The collected keywords were “sustainable fashion consumption” and “ethical fashion consumption” to examine consumer perception of sustainable and ethical fashion.

During the pre-COVID-19 period, we collected 5,338 documents on sustainable fashion consumption using the collection keyword. During the post-COVID-19 period, we collected 5,467 documents. In addition, we collected 4,610 documents on the keyword “ethical fashion consumption” during the pre-COVID-19 period and 5,420 documents during the post-COVID-19 period. Table 1 presents the data collection results.

Table 1 Results of Data Collection.
Full size table

Data analysis

We used Textom 6.0 to analyse this study’s data and divided the procedure into text frequency analysis and LDA topic modeling. After collecting the data with Textom 6.0, we fulfilled data refinement. For instance, we looked at adverbs, deleted stop words, and merged synonyms. Specifically, we conducted data cleansing and preprocessing unstructured text data to eliminate irrelevant or generic words. Consequently, we split the entire text document into individual words, known as word tokenization. Then, we removed stop-words and unified synonyms and applied word lemmatization to filter meaningful words in natural language data. Next, we performed a keyword frequency analysis to extract the top 30 keywords.

We also performed cross-validation with three clothing science doctoral graduates to ensure the reliability of the data purification process. We then conducted LDA topic modeling to determine the 10 major topics. We analyzed the topic, and its interpretation based on the meaning of the extracted keywords, the relationship between keywords within the topic, and the original data containing the keyword. In consideration of the contents, we selected four researchers through consensus. Based on these research results, we divided the terms “sustainable fashion consumption” and “ethical fashion consumption” by period, pre- and post-COVID- 19, respectively. Finally, we compared and empirically analyzed consumer perceptions.

Results

Results of research questions 1 and 2: keywords relating to consumption of sustainable and ethical fashions pre- and post-COVID-19

After the data purification process to compare extracted keywords on sustainable and ethical fashion consumption before and after the COVID-19 outbreak, we arranged the top 30 extracted keywords by their frequency, as shown in Table 2.

Table 2 Results of keyword frequency analysis of sustainable fashion consumption and ethical fashion consumption pre-and post-COVID-19.
Full size table

Results of Keyword Frequency Analysis Based on “Sustainable Fashion Consumption”

Among the keywords extracted for sustainable fashion consumption pre-COVID-19, the high-frequency keywords were “sustainable” (2,219), “fashion” (1,790), “brand” (832), “consumption” (722), and “product” (622), and the keywords with the highest frequency post-COVID-19 were “consumption” (3,177), “sustainable” (2,465), “fashion” (2,457), “eco-friendly” (1,683), and “brand” (985). Comparing the data collected for sustainable fashion consumption keywords, we found no significant difference between the periods – 5,338 cases pre-COVID-19 and 5,647 cases post-COVID-19. However, when comparing the sum of the frequencies of the top five keywords, the frequency increased by approximately 1.74 times, from 6,185 cases pre-COVID-19 to 10,767 cases post-COVID-19 outbreak. In addition, in the extracted keywords related to sustainable fashion consumption, we found keywords such as “eco-friendly” and “material” mentioned with a higher frequency post-COVID-19 outbreak than pre-COVID-19. On the other hand, “fast fashion,” mentioned in sustainable fashion consumption pre-COVID-19 and presented as a task to overcome, was not in the top 30 frequency order post-COVID-19 outbreak, and “generation” and “earth” were not mentioned pre-COVID-19. The keyword “online” also emerged after post-COVID-19.

Results of keyword frequency analysis based on “Ethical fashion consumption”

Among the extracted keywords for ethical fashion consumption pre-COVID-19, the keywords with high frequency were “fashion” (1,452), “eco-friendly” (899), “consumption” (859), “ethics” (815), and “brand” (757). During the post-COVID-19 outbreak, the most frequent keywords were “fashion” (3,497), “eco-friendly” (1,236), “consumption” (1,044), “ethics” (962), and “brand” (869). The number of keywords for ethical fashion consumption collected post-COVID-19 was 5,420, and 4,610 pre-COVID-19, showing a larger difference (810) than the respective number of keywords for sustainable fashion consumption (309). As with sustainable fashion consumption, the sum of the frequency of the top five keywords by period for ethical fashion consumption increased by about 1.59 times, from 4,782 cases pre-COVID-19 to 7,608 cases post-COVID-19 outbreak. When comparing the frequencies of keywords related to ethical consumption, we extracted more post-COVID-19 (856) than before (553). We separately extracted the keyword “vegan fashion,” the top keyword pre-COVID-19, as “vegan” post-COVID-19. Similar to sustainable fashion consumption, in ethical fashion consumption, the keyword “material” emerged as an extracted keyword post-COVID-19.

Results of research questions 3 and 4: major topics relating to consumption of sustainable and ethical fashions in pre- and post-COVID-19

We conducted LDA topic modeling using data collected as keywords for “sustainable fashion consumption” and “ethical fashion consumption” by period: pre-and post-COVID-19.

Results of LDA topic modeling for keyword “Sustainable fashion consumption”

We derived the top 10 topics using LDA topic modeling on “sustainable fashion consumption” from January 20, 2018, to January 19, 2020, before the COVID-19 outbreak. Table 3 summarizes the topics emerging from the related keywords.

Table 3 Results of LDA topic modeling pre-COVID-19 on the keyword “Sustainable fashion consumption”.
Full size table

The title of Topic 1 was “fashion technology,” with related keywords derived from “fashion,” “sustainability,” “technology,” “product,” “Japan,” “Korea,” and “consumption.” The title of Topic 2 was “sustainable exercise environment,” with related keywords derived from “sustainable,” “self,” “exercise,” “environment,” “clothing,” “benefit,” and “service.” The title of Topic 3 was “local culture,” with related keywords derived from “culture,” “country,” “region,” “China,” “power,” “product advertisement,” and “soft.” The title of Topic 4 was “generational consumption trend,” with related keywords derived from “fashion,” “sustainability,” “consumer,” “product,” “generation,” “ethics,” and “society.” The keywords “fast fashion” and “generation” were only in Topic 4. As a result of checking the original text of the keyword “generation” in the collected data, we found many references to the millennial-Z (MZ) generation. We named Topic 5 “wholesale market trend,” with related keywords derived from “China,” “trend,” “product,” “fashion,” “wholesale,” “cosmetics,” and “brand.” The title of Topic 6 was “eco-friendly business brand,” with related keywords derived from “sustainable,” “fashion,” “brand,” “eco-friendly,” “consumption,” “industry,” and “environment.” The title of Topic 7 was “fashion industry trend,” with related keywords derived from “fashion,” “trend,” “clothing,” “sustainable,” “industry,” “growth,” and “brand.” The keywords “sustainable,” “product,” “consumption,” “fashion,” “increase,” “online,” and “trend” were the derived related keywords for Topic 8, “online fashion consumption.” The title of Topic 9 was “sustainable fashion production,” with related keywords derived from “fashion,” “brand,” “sustainable,” “product,” “meaning,” “production,” and “growth.” Finally, Topic 10 was “sustainable product growth,” with related keywords derived from “sustainable,” “product,” “growth,” “market,” “China,” “industry,” and “inspection.” We derived country-related keywords such as “China,” “Korea,” “USA,” “Japan,” and “Thailand” together from related keywords in the topics on sustainable fashion consumption pre-COVID-19. In addition, we identified keywords related to real estate and stocks concerning social issues.

The intertopic distance map (IDM) shows the weight and distance between topics. In addition, it shows the degree of relevance between topics. For example, the IDM for Topic 6 of the 10 topics extracted from LDA topic modeling in pre-COVID-19 sustainable fashion consumption is in Fig. 1.

Fig. 1
figure 1

IDM for topic 6 derived from LDA topic modeling on pre-COVID-19 sustainable fashion consumption.

Full size image

Next, we used LDA topic modeling and analyzed “sustainable fashion consumption” from January 20, 2020, to October 19, 2021, the period post-COVID-19 outbreak, deriving ten topics, as summarized in Table 4.

Table 4 Results of LDA topic modeling post-COVID-19 on the keyword “Sustainable fashion consumption”.
Full size table

The title of Topic 1 was “eco-friendly fashion consumption,” with related keywords derived from “eco-friendly,” “consumption,” “fashion,” “sustainable,” “product,” “brand,” and “clothing.” We classified the keywords “fashion,” “sustainable,” “consumption,” “eco-friendly,” “brand,” “industry,” and “trend” under the title “fashion industry trend” (Topic 2). The title of Topic 3 was “eco-friendly brand value,” with related keywords derived from “consumption,” “fashion,” “sustainable,” “eco-friendly,” “brand,” “value,” and “clothing.” We classified the keywords “brand,” “consumption,” “eco-friendly,” “sustainable,” “gender,” “fashion,” and “global” as “eco-friendly brand consumption (Topic 4). “Real estate,” “consumption,” “sustainability,” “fashion,” “global,” “growth,” and “potential” were the keywords for the title “real estate growth” (Topic 5). Topic 5 combines the potential for real estate growth and sustainable fashion consumption. Topic 6 was “brand evaluation,” with related keywords derived from “consumption,” “sustainable,” “fashion,” “brand,” “city,” “evaluation,” and “clothing.” The title of Topic 7 was “fashion technology,” with related keywords derived from “happiness,” “technology,” “wholesale,” “fashion,” “consumption,” “growth,” and “Korea Securities Dealers Automated Quotation (KOSDAQ).” The keywords “happiness” and “technology” had the highest frequencies under Topic 7. Topic 8 was “increasing clothing consumption” because related keywords derived from “consumption,” US, “rising,” clothing, “related stocks,” and “market conditions.” The keywords “consumption,” “sustainable,” “fashion,” “industry,” “clothing,” “eco-friendly,” and “production” fell under “fashion industry production” (Topic 9). Finally, the title of Topic 10 was “online fashion consumption” as related keywords derived from “consumption,” “fashion,” “sustainable,” “brand,” “product,” “dollar,” and “online.” If the country name derived as a keyword related to the period pre-COVID-19, we found that “global” derived from several topics as a keyword related to the period post-COVID-19.

The IDM for Topic 1 of the 10 topics extracted from LDA topic modeling in post-COVID-19 sustainable fashion consumption is in Fig. 2.

Fig. 2
figure 2

IDM for topic 1 derived from LDA topic modeling on post-COVID-19 sustainable fashion consumption.

Full size image

Results of LDA topic modeling on the keyword “Ethical fashion Consumption”

We derived 10 titles using LDA topic modeling for “ethical fashion consumption” from January 20, 2018, to January 19, 2020, pre-COVID-19. Table 5 summarizes the topics and related keywords.

Table 5 Results of LDA topic modeling pre-COVID-19 on the keyword “Ethical fashion consumption”.
Full size table

The title of Topic 1 was “eco-friendly fashion brand,” with related keywords derived from “eco-friendly,” “fashion,” “brand,” “ethics,” “consumer,” “animal,” and “product.” The title for Topic 2 was “eco-friendly product culture,” with related keywords derived from “eco-friendly,” “society,” “culture,” “product,” “beauty,” “fashion,” and “brand.” We classified the keywords “fashion,” “brand,” “sustainable,” “trend,” “eco-friendly,” “ethics,” and “clothing” under “fashion brand trend” (Topic 3). This topic differs from other topics in that it includes the keyword “BBYB,” a bag brand pursuing vegan fashion, as a brand name. The title of Topic 4 was “fashion industry ethics,” with related keywords derived from “fashion,” “consumption,” “sustainability,” “ethics,” “industry,” “value,” and “eco-friendly.” The keywords “bag,” “culture,” “society,” “product,” “use,” “brand,” and “good consumption” were under the title “good consumption culture” (Topic 5). The title for Topic 6 was “fashion consumption trend,” with related keywords derived from “fashion,” “clothing,” “product,” “trend,” “consumption,” “ethical consumption,” and “eco-friendly.” Only Topic 6 included the keywords “fast fashion” and “upcycling.” Furthermore, the title of Topic 7 was “logistics development,” with related keywords derived from “development,” “logistics,” “trend,” “brand,” “fashion,” “consumption,” and “generation.” The keywords “fashion,” “clothing,” “consumer,” “brand,” “global warming,” “water,” and “ethics” fell under the title “fashion consumer ethics” (Topic 8). In addition, we included specific keywords, such as “warming” and “water. The keywords for Topic 9 were “eco-friendly,” “ethics,” “problem,” “trend,” “thought,” “consumer,” and “product,” and the title was “eco-friendly ethical issues.” The title of Topic 10 was “fur fashion,” derived from “eco-friendly,” “fur,” “animal,” “fashion,” “clothing,” “thought,” and “brand.” The IDM for Topic 3 of the 10 topics extracted from LDA topic modeling in pre-COVID-19 ethical fashion consumption Fig. 3.

Fig. 3
figure 3

IDM for Topic 3 derived from LDA topic modeling on pre-COVID-19 ethical fashion consumption.

Full size image

We also conducted LDA topic modeling of “ethical fashion consumption” from January 20, 2020, to October 19, 2021, post-COVID-19. As a result, we derived 10 topics from these keywords, as summarized in Table 6.

Table 6 Results of LDA topic modeling post-COVID-19 on the keyword “Ethical fashion consumption”.
Full size table

The title of Topic 1 was “investment in distribution development,” with related keywords derived from “investment,” “development,” “distribution,” “fashion,” “generation,” “programming,” and “brand.” We regarded the topic as an extension of “logistics development trend,” which was Topic 7 under “ethical fashion consumption,” and both titles included “generation” as a related keyword for these two topics. The title of Topic 2 was “vegan fashion trend,” with related keywords derived from “fashion,” “vegan,” “ethical consumption,” “trend,” “animal,” “consumption,” and “clothing.” The title of Topic 3, “ethical consumption brand,” derived from the keywords “fashion,” “clothing,” “brand,” “consumption,” “ethical consumption,” “use,” and “bag.” The keyword “conscious,” included in the other topics, was also related. The title of Topic 4 was “ethical consumption industry,” with related keywords derived from “fashion,” “sustainability,” “clothing,” “ethical consumption,” “industry,” “ethics,” and “product.” The title of Topic 5 was “culture,” which was the only keyword for Topic 5 after refining the data. Topic 6 was “eco-friendly fashion brand,” with related keywords derived from “fashion,” “eco-friendly,” “ethical consumption,” “sustainable,” “ethics,” “vegan,” and “brand.” The keywords for Topic 7 were “fashion,” “logistics,” “brand,” “consumption,” “eco-friendly,” “generation,” and “ethical consumption,” and the title was “fashion brand logistics.” As in Topic 1, “distribution and development investment,” the keyword “generation” was also in Topic 7. Further, the title of Topic 8 was “online ethical consumption,” with related keywords derived from “fashion,” “development,” “consumption,” “online,” “ethical consumption,” “brand,” and “market.” The keywords “fashion,” “society,” “ethics,” “culture,” “brand,” “use,” and “problem” (Topic 9) fell under the title “fashion social problem.” The keywords “fashion,” “brand,” “good consumption,” “ethical consumption,” “bag,” “leather,” and “clothing” were under the title “good consumption brand” (Topic 10).

Even pre-COVID-19, “vegan fashion” was a keyword related to ethical fashion consumption. However, we did not select it as the title name because the probability of appearing within the topic was low. However, post-COVID-19 confirmed that keywords related to “vegan,” such as those under Topic 2 (“vegan fashion trend”), were more likely pre-COVID-19. As such, we found that under “ethical fashion consumption” post-COVID-19, there were more titles with keywords related to “animal” and those related to “vegan” than pre- COVID-19.

The IDM for Topic 7 of the 10 topics extracted from LDA topic modeling in post-COVID-19 ethical fashion consumption is in Fig. 4.

Fig. 4
figure 4

IDM for Topic 7 derived from LDA Topic Modeling on Post-COVID-19 ethical fashion consumption.

Full size image

Conclusion

As the need for sustainable fashion consumption and ethical fashion consumption increases due to COVID-19, this study investigated the difference in consumers’ perceptions of two concepts: sustainable fashion and ethical fashion. To this end, we set “sustainable fashion consumption” and “ethical fashion consumption” as collection keywords. We examined consumer perceptions using keyword frequency analysis and LDA topic modeling using social media data collected before and after the COVID-19 outbreak.

The results of this study are as follows. First, the keyword frequency analysis to find out the main keywords used by consumers on social media pre-and post-COVID-19 regarding “sustainable fashion consumption” showed that the main keywords pre-COVID-19 were “sustainable,” “fashion,” “brand,” “consumption,” and “product,” and the main keywords post-COVID-19 outbreak were “consumption,” “sustainable,” “fashion,” “eco-friendly,” and “brand.” The difference in the volume of data collected pre-and post-COVID-19 was insignificant. However, when comparing the sum of the frequencies of the top five keywords, it was about 1.74 times more during the period post-COVID-19 than before the outbreak. We found that starting with the COVID- 19 outbreak, consumers’ attention focused on specific keywords for sustainable fashion consumption, considering they mentioned keywords such as “eco-friendly” and “material” with higher frequency post-COVID-19 than pre-COVID-19. Thus, the interest in eco-friendliness provides a helpful direction for sustainable fashion consumption post-COVID-19. We found an expanding consumer interest in materials. This finding aligns with a preliminary study (Koh and Lee, 2020) in which consumers recognized eco-friendly value as a growth stage before COVID-19 but gradually shifted to identifying product value and pursuing eco-friendliness for conscious consumption after COVID-19. A key theme that emerged from the analysis is the increased focus on eco-friendly materials and sustainable fashion practices post-COVID-19. Prior studies, such as Henninger et al. (2016) and Joy et al. (2012), have emphasized the growing consumer interest in environmental sustainability, particularly in how products are produced and disposed of. Our findings reinforce this, showing a significant rise in the use of keywords like “eco-friendly“ and “sustainable“ post-pandemic, mirroring earlier observations that eco-consciousness is a driver for consumer engagement with sustainable fashion. As Koh and Lee (2020) noted, eco-friendliness was already a rising concern before the pandemic, but the current findings show a broader acceptance and prioritization of sustainability among consumers.

Second, the keyword frequency analysis showed that the main keywords used by consumers on social media pre- and post-COVID-19 outbreak regarding ethical fashion consumption were “eco-friendly,” “consumption,” “ethics,” and “brand,” which ranked equally. As with sustainable fashion consumption, we confirmed that the sum of the frequencies of the top five keywords for ethical fashion consumption by period increased more post-COVID-19 than pre-COVID-19. In addition, we can infer that veganism is expanding from the fashion domain to other domains. For instance, “vegan fashion,” the top keyword pre-COVID-19, was extracted as a separate keyword, “vegan,” post-COVID-19 outbreak. These results show that, as in the previous study (Guo et al., 2021), due to COVID-19, there is an emphasis on health, environment, and sustainability importance. Further, vegan materials are growing in the fashion industry, which includes animal welfare and environmental protection, and the demand for veganism, such as vegan cosmetics, vegan interiors, and vegan packaging, is expanding into all areas of life. The findings on ethical fashion consumption align with prior research that highlight growing consumer concern over labor practices and animal welfare (e.g., Stringer et al., 2020). Post-COVID-19, there was an evident shift in consumer discourse toward ethical considerations like veganism and animal welfare, reflected in keywords such as “vegan“ and “animal.“ This parallels the findings of Phua et al. (2020), who observed the increasing role of veganism in shaping ethical consumption choices. Furthermore, our analysis supports Guo et al. (2021), who pointed out that post-pandemic consumers are more concerned with health, environment, and sustainability, which may explain the surge in interest in vegan products across not only fashion but also other industries like cosmetics and packaging.

Third, as a result of examining the keywords used by consumers on social media before and after the outbreak of COVID-19 regarding “sustainable fashion consumption” using LDA topic modeling, we derived 10 topics for each period. We changed the title of each topic based on keywords. For example, the keyword “eco-friendly,” the title of Topic 1, “eco-friendly industrial brand,” occurred pre-COVID-19, developing into “eco-friendly fashion consumption,” “eco-friendly brand value,” and “eco-friendly brand consumption” during post-COVID-19. Likewise, in Topics 3, 5, 8, and 10, “local culture,” “wholesale market trend,” “online fashion consumption,” and “sustainable product growth” focused on specific countries in the pre-COVID-19 period. On the other hand, post-COVID-19, the titles of the topics “fashion industry trend,” “eco-friendly brand value,” “eco-friendly brand consumption,” “real estate growth,” and “brand evaluation” expanded with a focus on the keyword “global” to increase sustainable fashion consumption. In terms of the keyword “generation,” the title of Topic 4, “generational consumption trend,” developed into “eco-friendly brand value” after the COVID-19 outbreak. Previous research (Nikolić et al., 2022) also confirms this, focusing on Generation Z, which pursues sustainable fashion consumption. In this study, Generation Z explored the trend of actively choosing brands that value health, well-being, and sustainability. This brand has a competitive advantage for the new generation of consumers during the pandemic era. Fourth, in terms of “ethical fashion consumption,” we used LDA topic modeling and derived 10 titles from the keywords used by consumers on social media for each period, pre-and post-COVID-19. Focusing on the keyword “eco-friendly,” pre-COVID-19, we derived the titles “eco-friendly fashion brand,” “eco-friendly product culture,” and “eco-friendly ethical issues.” However, post-COVID-19, “eco-friendly fashion brand” was the only title. On the other hand, the title “ethical fashion consumption” was impressive, focusing on the keyword “animal.” Pre-COVID-19, the derived titles were “eco-friendly fashion brand,” “fur fashion,” and “fashion social issues,” with related keywords, indicating an expanding consumer focus on animal welfare and ethical issues within the fashion industry. These findings align with previous studies, such as Stringer et al. (2020), which explored the growing ethical concerns related to animal welfare, labor practices, and environmental impact on the fast fashion industry. Stringer et al. (2020) found that consumers increasingly reject fast fashion brands over concerns like animal cruelty and unsustainable production practices. Similarly, our study demonstrates that post-pandemic, there is heightened consumer awareness of animal-related issues, reinforcing the idea that consumer perceptions of ethics in fashion have broadened.

In addition, as Phua et al. (2020) suggest, this can support understanding the importance of animal welfare issues given the growth of the vegan market. Phua et al. (2020) highlighted the rise of veganism not just as a dietary preference but as a broader lifestyle choice that encompasses fashion, cosmetics, and more. The shift in our study's keyword focus, from “fur fashion“ to broader concerns around “animal“ and “vegan“ post-pandemic, mirrors this trend, emphasizing that animal welfare is becoming a core concern in ethical fashion consumption.

In addition, as Phua et al. (2020) suggest, this can support understanding the importance of animal welfare issues given the growth of the vegan market. The conclusions of this study are as follows. Post-COVID-19, consumer preferences in sustainable and ethical fashion consumption exhibit distinct trends. Keywords such as “fashion,“ “consumption,“ “eco-friendliness,“ and “brand“ have gained increasing interest. Notably, the MZ generation has emerged as a significant consumer group, with brands shifting their focus accordingly. Sustainable fashion consumption, which previously centered on products, now places greater emphasis on the “eco-friendly“ aspect in the post-COVID-19 era. In contrast, ethical fashion consumption maintains steady interest in eco-friendliness, with a particular focus on animal-related aspects. LDA topic modeling highlights these differences and suggests the emergence of veganism as a cross-cutting concern beyond fashion, aligning with global environmental awareness during the pandemic.

The practical implications of this study are significant and can be applied in the following ways. First, targeted Marketing Strategies for Fashion Brands: Fashion brands can use the insights from this study to design more focused and effective marketing campaigns. Understanding the distinction between sustainable and ethical fashion allows brands to tailor their campaigns accordingly. For example, brands emphasizing sustainable fashion can highlight eco-friendly practices such as the use of recycled materials, while ethical fashion brands might focus on fair labor practices and animal welfare.

Second, Product Development and Innovation: The results of this study provide guidance for fashion companies in developing products that align with specific consumer preferences. Sustainable fashion lines can focus on resource conservation and environmentally friendly materials, while ethical fashion products can emphasize fair trade and humane practices, catering to distinct market segments.

Third, Corporate Social Responsibility (CSR) Policies: This research helps fashion companies refine their CSR strategies by providing insights into consumer expectations around sustainability and ethics. Companies can implement more focused CSR initiatives, such as environmental programs aimed at reducing carbon footprints or ethical initiatives targeting improved labor conditions and animal welfare, thereby aligning their actions with consumer values and ultimately improving brand reputation and customer loyalty.

The study's results offer actionable insights for the fashion industry, emphasizing the need for strategies that align with evolving consumer needs in the post-COVID-19 era. Understanding the shifting consumer preferences and the importance of eco-friendliness can guide brands in developing effective marketing and product strategies. Theoretical frameworks can benefit from the insights generated by this research. These frameworks have the potential to enhance our comprehension of consumer choices within the fashion industry, shedding light on the complex interplay of factors influencing sustainable and ethical fashion consumption.

Future studies should consider employing quantitative analyses, such as t-tests, to compare consumer perceptions and behaviors across pre- and post-pandemic periods. This approach would enable a more robust examination of how consumer trends and values related to sustainability and ethics have shifted over time. Furthermore, qualitative research methods can complement these quantitative analyses by delving deeper into the associations, motivations, and attitudes underlying sustainable and ethical fashion consumption. Such qualitative insights could help us better understand the complex emotional and psychological drivers behind consumer behavior, building on empirical findings from this study and others (Perry and Chung, 2016; Joy et al., 2012).

In addition to these mixed methods approaches, future research could focus on expanding the scope of topics derived from the LDA topic modeling. For example, “Sustainable Fashion Consumption“ is connected to major topics such as “Fashion Technology,“ “Eco-Friendly Industrial Brand,“ “Sustainable Fashion Production,“ and “Sustainable Product Growth.“ These areas can be further explored by fashion industry brands or companies looking to align their strategies with evolving consumer preferences. Developing sustainable corporate and brand operations through comprehensive value chain management will not only secure consumer confidence but may also contribute to building a virtuous cycle economy, impacting both local and global markets. Future research could also explore cross-cultural comparisons of sustainable and ethical fashion consumption, given that consumer priorities may vary across different regions. By analyzing consumer behavior in diverse cultural contexts, scholars can explore how fashion brands might tailor their sustainability and ethical initiatives to resonate more deeply with international audiences. Finally, longitudinal studies could track changes in consumer behavior over time, helping to determine whether the shifts observed post-COVID-19 are temporary or indicative of lasting changes in consumer values.

Methodological limitations of our studies are as follows. While the use of text mining and LDA topic modeling in this study provides powerful insights into consumer perceptions of sustainable and ethical fashion by analyzing large amounts of data, there are inherent limitations to these methods. One significant limitation is the selection of keywords used in the LDA model, which can influence the comprehensiveness and direction of the analysis. To mitigate this, we carefully selected keywords that reflect core themes in sustainable and ethical fashion based on both industry terminology and prior academic research (e.g., Koh and Lee, 2020; Guo et al., 2021). However, we acknowledge that keyword selection may still introduce bias into the analysis, particularly by underrepresenting certain nuanced consumer behaviors or perceptions. Additionally, the scope of databases and social media platforms analyzed also presents a limitation. While we selected widely used platforms, the analysis may not fully capture all consumer sentiment, especially from niche or less popular platforms. Future research could address this by expanding the scope of data collection to include a broader array of platforms and international data sources, thus enhancing the generalizability of the findings. Moreover, while text mining allows for efficient analysis of large data sets, it does not capture the depth and context of consumer behavior as effectively as qualitative methods, such as interviews or focus groups. We recognize that this limitation could affect the nuanced understanding of the motivations behind consumer choices in sustainable and ethical fashion. Future studies should consider complementing text mining with qualitative methods to gain deeper insights into the drivers of consumer behavior, addressing this gap and providing a more holistic understanding of the phenomena.

Kotler and Armstrong (2020) saw that sustainable pathways not only benefit the world but also benefit businesses, making it reasonable for businesses to pursue greening the supply chain for lower costs and greater profitability.

In addition, when focusing on the reason why “Eco-friendly“ was extracted as a related keyword as “Eco-friendly Ethical Issues“ Topic, it is possible to think about what areas of value consumers value in consumption by expanding to nature protection, ethical rights and values. Accordingly, it will be possible to expand consumer research topics related to animal protection, worker human rights, and consumer rights protection. For example, in the case of shoes, made from plant-based raw materials, animals will be free from the pain and death of intensive livestock and shoe supply chains, and changes in consumer perception and public and public pressure to improve workers' human rights can exert power on companies that rely on social contracts (Hoskins, 2022).

By verifying whether the relationship is appropriate for the definition of sustainable fashion and ethical fashion concepts, changes in corporate and consumer perceptions and the resulting positive changes in society can be expected.