29 3. CHAPTER 3 METHODOLOGY 3.1. Research Design This research aims to explore the relationship between the independent variables such as the informational sources of consumer confusion—overload, similarity, and ambiguity confusion, individual decision-making styles—utilitarian and hedonic, decision outcomes—decision quality, decision postponement, and cognitive dissonance, and the mediating effect of previous brand experience. As stated by Yin (2008) research design is “the logical sequence that connects the empirical data to a study's initial research questions and, ultimately, to its conclusions”. In order to collect primary data, there are three distinct methods that can be utilized to obtain the desired input: quantitative, qualitative and mixed methods (Creswell & Creswell, 2018). The research design process is shown below in figure 3.1. Figure 3.1. Research design framework. 3.2. Research Approach The research adopts a mixed method (quantitative and qualitative) approach, use of a quantitative online survey, which is subsequently validated and enhanced through 30 personal interviews to gather qualitative data, meaning the study is conducted in a sequential manner. The choice of a mixed-method approach for this study was based on the findings of Chauhan and Sagar's (2021) systematic review, which highlighted the insufficient attention given to consumer confusion in the marketing literature. The review revealed that a considerable majority of the papers employed empirical or quantitative approaches, there was a notable scarcity of research papers employing mixed-method approaches, accounting for a mere 4% of the total research, underscored the need for a comprehensive investigation that integrates both qualitative and quantitative data. Hence, to acquire an extensive understanding of consumer confusion and its implications, this study opted for a mixed-method approach to leverage the strengths of both quantitative and qualitative methodologies. 3.2.1. Quantitative Approach Quantitatively, this study employed questionnaire surveys to measure key variables and test the model hypothesis based on the collected data. To ensure research reliability and validity, all variable scales were derived from well-established literature and frequently cited in core journals. Bryman and Bell (2015) recommend pilot testing to ensure a questionnaire's adequacy, particularly for self-completion formats, to minimize respondent errors. Consequently, a small-scale pilot test was conducted to assess the questionnaire's reliability and validity, followed by the distribution of large- scale formal questionnaires for data collection. PLS-SEM was utilized for statistical analysis, testing, and verification of research hypotheses and variable relationships. 3.2.1.1. Sample and Data Collection A study by Malhotra (2017) indicated that a minimum sample size of 200 respondents is required in a marketing research study. Therefore, in this research, the data was procured from a total of 505 respondents using the purposive sampling technique. The candidates belonged to the age group of 18–. Targeting respondents aged between 18 and  for the quantitative questionnaire survey is justified due to their significance as the primary consumers in the cosmetics and personal care industry, as well as their significant presence in digital commerce (Statista Consumer Insights 2022). Female 31 respondents were targeted for data collection as they are believed to be more actively involved in the purchase of personal care and cosmetic products and therefore, were presumed to be more knowledgeable on the subject. This study employed an online platform to recruit participants from Java Island, specifically targeting frequent users of cosmetics and personal care products who were well-informed about these products and agreed to experience confusion while shopping for cosmetics and personal care products through all different media sources including social media and the Internet. This study utilizes Google Forms for data collection. To ensure demographic diversity, the survey will be distributed through various channels, including social media platforms and online communities. 3.2.1.2. Measures This study tests the hypotheses developed in Chapter two. All variables involved in this paper were measured by a Likert 5-level scale, where 1 represents “Strongly disagree”, and 5 represents “Strongly agree”. The items of the research instrument were adopted from the literature of consumer confusion. Some of the research items were modified to fit into the context of the current research work. Informational sources of consumer confusion. To measure the three different informational sources of consumer confusion, 10 items were derived from the consumer confusion literature. The scale of overload confusion was designed by referring to the research by Walsh et al. (2007), Matzler et al. (2011). It includes 3 items, which were employed to measure consumers’ perception of product information acquisition and processing during online shopping. The measurement of similarity confusion was designed by referring to the scale by Walsh et al (2007), which measured the degree of confusion experienced by consumers when distinguishing between visually and functionally similar products. There are a total of 3 items in the scale. Ambiguity confusion was measured by referring to the research by Walsh and Mitchell (2010), which involved assessing consumers' tolerance for processing unclear, misleading, or ambiguous products, product-related information, or advertisements. The scale includes 4 items in total. 32 Information confusion. Information confusion was measured with the scale of Sharma et al (2022), which measured consumers' perceptions of confusion during online shopping. The scale consists of 4 items. Impact of decision-making styles. The scale decision-making styles were adapted from the work of Sproles and Kendall (1986) and Kasper et al (2010), which consists of 18 items in total. Mediator of confusion. The scale of brand experience was measured by referring to the research by Brakus et al.