INTRODUCTION

The global fashion industry’s sales are immense. The apparel market is diverse and competitive. In 2019, Peru exported US$274,976,022 (^{Asociación de Exportadores [ADEX], 2020}) in fashion apparel. According to the Sociedad Nacional de Industrias (^{SNI, 2021}), Peru’s textile and apparel sector generates more than 398,000 formal jobs, meeting domestic and foreign market requirements. This sector represents 6.4% of the country's manufacturing GDP. There has been a decline in this sector in recent years due to the COVID-19 pandemic. To increase sales in domestic and international markets, it is necessary to determine consumers’ aesthetic, physical, and emotional needs and interpret them correctly. The low competitiveness of the Peruvian fashion sector stems from several factors: weak development of productive forces, technological backwardness, and inadequate interpretation of customer needs, among others. Product development cycles are variable and difficult to control.

If product development takes a long time, the competition anticipates new products, and thus losses are incurred. This study contributes to the advancement of knowledge because it allows for an understanding of the perception of Lima consumers of a fashion product. The results provide guidance to the producer to focus on the most important qualities of the products to stimulate purchase intention. It is of interest to companies that manufacture garments. The objective of the research study is to determine the importance of the physical attributes of fashion apparel that influence the purchase intention of 20 to 24-year-old women in Lima. The theory of planned behavior (TPB) is used to determine the influences of the three physical attributes on purchase intention. No similar studies have been conducted using this method. Knowing the importance of the attributes allows an understanding of consumer preferences. The results are valuable because they allow the allocation of resources to strengthen the most important qualities of the product, thus favoring producers, who are more likely to invest in a new product line.

METHODOLOGY

General Hypothesis

Physical attributes of fashion apparel have a significant effect (SE) on the purchase intention (PI) of 20 to 24-year-old university women in Lima.

Specific Hypotheses

H₁ Shape of fashion apparel has a SE on the PI of 20 to 24-year-old university women in Lima.

H₂ The mechanical properties of fabric of fashion apparel have a SE on the PI of 20 to 24-year-old university women in Lima.

H₃ The color of fashion apparel has a SE on the PI of 20 to 24-year-old university women in Lima.

Identification of Variables

Independent variable (IV): Physical attributes (X)

Dependent variable (DV): Purchase intention (Y)

The variables were determined based on the classification by ^{Abraham-Murali and Littrell (1995}). Characteristics were grouped until the following were obtained: shape (X₁), mechanical properties of fabric (X₂), and color (X₃).

This is an applied study with a non-experimental design, following a quantitative, cross-sectional, and correlational-causal approach. A representative sample of women in Peruvian society was deemed necessary to understand the consumer. We decided to conduct this study at Universidad Nacional Mayor de San Marcos, administering a questionnaire to female students. The questions were related to the three physical attributes identified in the theoretical framework, which were related to purchase intention. The influence of the attributes was measured based on the three dimensions defined in the TPB, thus determining their relationship with the dependent variable. The relationship degree was measured using the multiple regression statistical method.

The sample size was determined considering an infinite or unknown population of 20 to 24-year-old female university students in Lima. Based on this data, we can infer a sample of 196 respondents using the formula for infinite population sample determination (Z = 1.96 at 95%, margin of error = 0.07, p = 0.5, q = 0.5). A margin of error of 7% was chosen, as the complexity involved in measuring perceptions can lead to slightly larger errors. Respondents were 20 to 24-year-old female undergraduate students from Universidad Nacional Mayor de San Marcos (UNMSM), who were contacted via e-mail and WhatsApp, and voluntarily participated in the study.

A questionnaire was created and distributed via the web to analyze the variables. Before conducting the main study, the questionnaire was tested on 11 participants to ensure completeness, correct wording, clarity, and appropriate structure. After a few modifications, the final questionnaire had five core sections: Block 1 that collected important general information from the respondents; Blocks 2, 3, and 4 that measured attitude, subjective norm, and perceived behavioral control towards shape, mechanical properties of fabric, and color; and Block 5 that measured purchase intention for a product. Data were collected using two techniques: documentary analysis and surveys, which made it possible to understand the variables and draw the conclusions described.

The instrument was designed using three variables and 25 specific questions (Table 3). Cronbach’s alpha was used to measure reliability, obtaining a value of 0.892 (Table 4), which supports the design of this questionnaire. It was developed based on other relevant research with high reliability and validity in its variables. The references of the attitude were taken from ^{Zhang et al. (2017}) and ^{Singh and Verma (2017}); the instrument to measure subjective norm was taken from Singh and Verma (2017) and ^{Al-Swidi et al. (2014}); the index of perceived behavioral control were taken from ^{Yadav and Pathak (2016}) and Al-Swidi et al. (2014); and, finally, the tools used to determine acquisition intention were taken from Zhang et al. (2017). The questions were adapted to the fashion industry; a Likert-type scale was used.

RESULTS

Multiple regression was used to analyze the data; therefore, hypothesis testing was performed to demonstrate that the proposed variables influence the purchase intention of fashion apparel.

General Hypothesis

H₀: Physical attributes of fashion apparel do not have a SE on the PI of 20 to 24-year-old university women in Lima.

𝛽 1 = 0

H₁: Physical attributes of fashion apparel have a SE on the PI of 20 to 24-year-old university women in Lima.

One or more 𝛽 𝑖 ≠ 0

Specific Hypothesis 1

H₀: Shape of fashion apparel does not have a SE on the PI of 20 to 24-year-old university women in Lima.

𝛽 1 = 0

H₁: Shape of fashion apparel has a SE on the PI of 20 to 24-year-old university women in Lima.

𝛽 1 ≠ 0

Specific Hypothesis 2

H₀: Fabric of fashion apparel does not have a SE on the PI of 20 to 24-year-old university women in Lima.

𝛽 2 = 0

H₁: Fabric of fashion apparel has a SE on the PI of 20 to 24-year-old university women in Lima.

𝛽 2 = 0

Specific Hypothesis 3

H₀: Color of fashion apparel does not have a SE on the PI of 20 to 24-year-old university women in Lima.

𝛽 3 = 0

H₁: Color of fashion apparel has a SE on the PI of 20 to 24-year-old university women in Lima.

𝛽 3 = 0

There were 196 valid surveys, that is, the number statistically determined for the generation of the model with the information obtained from the surveys. Data were organized according to five categories. The SPSS software was used to run the normality test (Table 5), and a summary of the model was obtained (Table 6), showing that 52.4% of the variable is explained by this model, followed by the ANOVA test (Table 7) and finally the multiple regression test (Table 8). The results were as follows:

Normality Test

H₀: The variable follows a normal distribution.

H₁: he variable does not follow a normal distribution.

Due to the volume of data, the Kolmogorov-Smirnov test is used.

The p-value obtained for each variable (> 0.05) means that H₀ cannot be rejected; therefore, the data normality assumption is valid. Verification of data normality is necessary to select the statistic to be used.

Hypothesis Testing

General Hypothesis

p-value = 0%, the null hypothesis (H₀) is rejected (see Table 7).

If p-value < 5%, then H₀ is rejected.

As the p-value < 5%, the H₀ is rejected; therefore, physical attributes of fashion apparel have a SE on the PI of 20 to 24-year-old university women in Lima.

Specific Hypothesis 1

p-value = 60.1% (see Table 8)

If p-value < 5%, then H₀ is rejected.

In this case, H₀ is accepted (p-value > 5%); therefore, the shape of fashion apparel does not have a SE on the PI of 20 to 24-year-old university women in Lima.

Specific Hypothesis 2

p-value = 0.1 % (see Table 8)

If p-value < 5%, then H₀ is rejected.

In this case, H₀ is rejected (p-value < 5%); therefore, the mechanical properties of fabric of fashion apparel have a SE on the PI of 20 to 24-year-old university women in Lima.

Specific Hypothesis 3

p-value = 0% (see Table 8)

If p-value < 5%, then H₀ is rejected.

In this case, H₀ is rejected (p-value < 5%); therefore, the color of fashion apparel has a SE on the PI of 20 to 24-year-old university women in Lima.

Regression Analysis No. 2

In the second analysis, the variable shape is removed to develop a model that can predict the purchase intention based on the two remaining variables: color and physical characteristics of the fabric (see Table 9).

All coefficients are statistically significant in this case. Color remains the most influential. This equation can explain the consumer’s purchase intention. Other extrinsic factors are known to influence this dependent variable; however, this study intends to focus only on the importance of the physical variables of which the fashion apparel is composed. A p-value of over 5% means there is no relationship of dependence between the variables.

DISCUSSION

Analysis of variance and multiple regression were used as analytical tools to test the hypotheses. Upon performing the ANOVA test for the general hypothesis, the results yielded a p-value = 0% (see Table 7); therefore, the null hypothesis was rejected and the alternative was accepted. Overall, physical attributes influence purchase intention.

Regarding the first specific hypothesis, the result of the multiple regression (Table 8) validates the null hypothesis stating that shape does not influence purchase intention with a p-value = 60.1%. Therefore, the alternative hypothesis stating that shape influences purchase intention is rejected. We can reflect on the circumstances in which young women of these ages make their purchases. They are very young women who are learning to shop, are entering the labor market, and have a different appreciation of what they buy, as opposed to more experienced shoppers. These are their first purchases, so they value new features in the products. This appreciation may change and consolidate for another age range, placing a different value on the shape of the product. This leads to a field of new research that can be proposed for other age groups. However, this result is at odds with results obtained in other studies, such as those of ^{De Klerk and Lubbe (2004}, ²⁰⁰⁸), who concluded that aesthetics was significant in determining product quality. Furthermore, ^{Davis (1985}) found a high correlation between physical attributes and the perception of product quality.

The multiple regression analysis (Table 8) validates the second specific hypothesis stating that fabric has a significant effect on purchase intention; the null hypothesis is rejected because the p-value = 0.1%, which is less than 5%. The results also show a coefficient of 0.302, which is the slope of the straight line of this variable in the model to be plotted; therefore, the specific hypothesis stating that the physical characteristics of the fabric influence the purchase intention of the product is accepted. The third specific hypothesis is also accepted because the null hypothesis is rejected as the p-value = 0%. This variable has a steeper slope of 0.529; therefore, the hypothesis stating that color influences purchase intention is accepted. The relationship between slopes leads to the conclusion that color has the greatest influence on this type of customer.

According to the results, garment shape does not have a decisive influence on the customer’s purchase intention; for this reason, a new model in which this variable is removed was proposed. This new model uses multiple regression (Table 9), but only with the variables mechanical properties of the fabric and color. The two specific hypotheses are validated in this analysis, where it is observed that color has a greater influence on purchase intention. This new model is not a general equation, but it provides some insight into how these variables work with each other. Many other variables could influence the dependent variable, such as extrinsic variables: price, packaging, image, etc. However, this paper has intended to address the analysis of the physical (intrinsic) variables of fashion apparel.

CONCLUSIONS

The following conclusions can be drawn from the analysis: