Latent Class Analysis (LCA) is commonly used in market research to analyze consumer preferences in the context of product choices. LCA helps identify distinct segments based on the characteristics of observed variables. In this study, we apply LCA to survey data regarding consumer preferences for different types of sports apparel. Our goal is to understand how demographic factors influence purchasing decisions, aiming to gain insights for more targeted marketing strategies in the sportswear industry.

Market research studies often use surveys to examine consumer preferences across various product scenarios (Smith & Johnson, 2021). Our team initially conducted a pilot survey to refine the questions and format before launching the main survey.

As illustrated in the table below, we presented respondents with three different product purchase scenarios, each offering varying levels of feature enhancements: premium feature set 1, premium feature set 2, and standard feature set. These sets had effectiveness ratings of 70%, 50%, and 10% improvement in user experience, respectively. We collected respondents’ willingness to pay (WTP) by asking them to specify their preferred price for each feature set through an open-ended response question.

Based on feedback from our pilot study, we made several adjustments to our main survey. We streamlined the willingness to pay (WTP) options into three distinct amounts: 0, 400, and 600 USD, reflecting the identified clusters and excluding outliers. Additionally, due to frequent responses in the pilot study, we added two new options for those unwilling to pay: "The product should be free" and "I don’t consider this a priority."

The data for our market analysis were collected from the main survey, comprising a total of 460 responses regarding consumer preferences for different types of sports apparel. We identified 8 key factors influencing each respondent's purchasing decision: age, education level, household size, monthly income, frequency of exercising, participation in sports events, proximity to fitness centers, and previous experience with the brand (Smith & Johnson, 2021). For the response options, we converted categorical choices into numerical values (1 to 5). Our analysis classified respondents into three distinct groups based on their spending willingness. The first group, representing 54.1% of respondents, is willing to spend 600 USD for premium features; the second group, comprising 22.2%, is willing to spend 400 USD for mid-tier features; and the third group, with 23.7%, prefers to spend 0 USD for basic features.

Cluster Analysis was used to identify hidden segments among the 8 factors and consumer preferences. We employed R, an open-source statistical software, to fit our data into various segmentations. To determine the most appropriate model, we examined changes in the Silhouette Score and Davies-Bouldin Index, which measure the cohesion and separation of clusters. A higher Silhouette Score or lower Davies-Bouldin Index indicates a more accurate cluster solution. With the optimal number of segments identified, we can then investigate the underlying variables influencing consumer preferences.

Table 1 displays the distribution of each demographic attribute across the three different "preference for product features" options within the overall sample.

We observe a negative correlation between preference for advanced features and the INCOME factor. This is evident from Table 1, where 133 (53.7%) of the 248 respondents willing to invest in premium features are students with no income. Comparing these 133 student respondents to the total 217 student respondents, 61.3% preferred the premium features over the basic options. Additionally, 45.5% of individuals earning between 40,000 and 60,000 USD per year chose the mid-tier features, while 58% of those with a lower annual income of 25,000 USD opted for the top-tier features. The CHARITY factor also significantly influences preference, as among the 7 respondents who regularly support charitable causes, 6 (85.7%) selected the premium feature set.

Breaking down the 3 identified segments, we analyze the quantitative distribution within each segment, as detailed in Figure 2 above. It becomes apparent that certain factors exhibit minimal variance concerning consumer preferences: BRAND, FREQUENCY, LOCATION, and PAST PURCHASES. Additionally, Segment 2 displays the highest level of preference for premium features, followed by Segment 1 and Segment 3, with Segment 3 showing a notable gap in preference.

Between EXPERIENCE and FREQUENCY, Figure 2 reveals a set of roughly parallel lines, indicative of the Halo effect. This suggests that the segmentation might have been influenced by overlapping factors, with the identified segments potentially reflecting varying degrees of EXPERIENCE and FREQUENCY rather than distinct categories.

A correlation matrix is used to quantify the relationship between all 8 factors and consumer preference for product features. The matrix employs the Spearman rank correlation coefficient, which ranges from -1 to 1. Values closer to 1 or -1 indicate a stronger association between the variables.

factors (EXPERIENCE, FREQUENCY, BRAND LOYALTY, and PAST PURCHASES) and consumer preference. The ANOVA test helps determine whether the differences observed among the groups are statistically significant. The null hypothesis posits that there is no significant relationship between the variables. We analyzed the p-value from the ANOVA test, rejecting the null hypothesis when the p-value is less than 0.05.