Lakhasly

Material and Methods: Exploring Chinese Consumer Acceptance of Smart Food Packaging

This study investigated Chinese consumer acceptance of active (AP) and intelligent (IP) food packaging through quantitative surveys conducted in Beijing. To ensure a well-founded basis for the surveys, the research team conducted a comprehensive literature review and a qualitative pre-study. This pre-study comprised 16 individual interviews with packaging experts and five focus groups (n=32) in China, exploring consumer perception and acceptance of smart food packaging.

Two paper-and-pencil surveys, focusing separately on AP and IP, were administered to a convenience sample of participants aged 18 and older, living in China, and primarily responsible for household food management decisions. Participants were recruited using an intercept method in shopping malls, convenience stores, supermarkets, and tourist sites.

The surveys included three sections: demographics, packaging, and trust. The packaging section, specific to each survey type (AP or IP), examined consumer perceptions, behaviors, and attitudes towards smart food packaging. It included questions on existing food packaging practices, satisfaction levels, and willingness to consume products utilizing AP or IP. Participants also indicated their acceptance of specific product types packaged with smart technology.

Trust questions addressed participants' trust in various institutions regarding food safety responsibility, including agricultural industry, food industry, science/research field, pharmaceutical industry, government agencies, and consumer organizations.

Data analysis utilized SPSS software (version 25) with a chi-square (χ2) test for independence to analyze the relationship between novel packaging acceptance and consumer characteristics. Post hoc testing was employed using contingency tables to identify the variable contributing to statistical significance.

The results of this study shed light on the factors influencing Chinese consumer acceptance of AP and IP technologies and contribute to the understanding of consumer behavior within this emerging market.

2 | MATERIAL AND METHODS
2.1 | Literature research and prestudy
Due to the limited research on consumer acceptance, behavior, and knowledge regarding smart food packaging technologies, the quantitative research was preceded by intensive literature research and a qualitative prestudy to attain a well-founded basis for the

LIet aL.
| 3979
quantitative survey. Previous research has illustrated the commonal- ity of supplementing focus groups with quantitative research, which is generally required to obtain the information needed to make sub- stantive conclusions (Manstan & McSweeney, 2020). This manner of approach has been employed in previous research exploring packag- ing elements and further studies on consumer perceptions toward novel technology (Greenbaum, 1998; Huang, Qiu, Bai, & Pray, 2006; Lindh, Olsson, & Williams, 2016).
First, 16 packaging experts were interviewed individually on the topic of smart food packaging, consumer demand, and consumer acceptance of smart food packaging. The experts comprised of academics, scientists, specialist advisors, private researchers, and industry stakeholders. The goal was to obtain an understanding of the key developments in food packaging and expert opinion on con- sumer acceptance of these novel technologies. Secondly, five focus groups (n = 32) were conducted in China to determine consumer perception and acceptance of smart food packaging technologies. Data collected from focus groups are particularly sensitive to cul- tural variables and have accordingly been used in multiple cross-cul- tural studies (Dolgopolova, Teuber, & Bruschi, 2015; Ger & Belk, 1996; Lazear, Pires, Isaacs, Chaulk, & Huang, 2008; Perrea, Grunert, & Krystallis, 2015). The semi-structured focus groups provided an exploration into current behaviors related to food packaging and purchasing, and exploration into levels of acceptance of smart food packaging. Thus, expert opinions and consumer perception could be compared and investigated in terms of differences, knowledge gaps, and misconception. The quantitative survey reported in this manu- script was developed based on the findings of these prestudies.
2.2 | Design and sample
This study utilized two paper-and-pencil surveys carried out in Beijing, China, to obtain quantitative data on consumer accept- ance of the two forms of smart packaging: AP and IP. Respondents were selected based on convenience sampling, utilizing an intercept method for recruitment, as outlined by Lavrakas (2008). This non- probability sampling method was employed as a wide range of par- ticipants were sought on a limited budget. The intercept method was conducted in the following four types of areas in Beijing: shopping malls, convenience stores (200 m2), and tourist sites (parks, walking areas). Participants were required to be 18 years or older, currently living in China, and be primarily or jointly responsible for food management decisions (e.g., food shopping, storage decisions, food preparation) within the household. Only sur- veys that met the inclusion criteria were retained for analysis.
2.3 | Materials
The survey instrument was composed in stages, using previous prestudy findings as a foundation. Three distinct sections were de- veloped: demographics, packaging, and trust. AP and IP were split
into two surveys, as initial tests indicated that inclusion of questions relating to both technologies resulted in an extended length of time (averaging 35 min) and risked participant fatigue. The two versions of the survey contained identical sections (part 1, part 2, and part 4). A differing section (part 3) of the survey explored the acceptance of either AP or IP. Both surveys were prepared in Chinese.
2.4 | Demographic variables
Demographic questions were developed with the aid of Hughes, Camden, and Yangchen (2016) and Chan (1999). Demographic infor- mation included the following: gender, age, marital status, province, income, education, and employment. In addition to standardized demographic information, participants were asked if they had a “qualification in the field of science and technology” and “dietary re- quirements” in line with previous studies on novel food technologies (Ceccoli & Hixon, 2012; Grimsrud et al., 2004; Huang et al., 2006; Hudson, Caplanova, & Novak, 2015).
2.5 | Food packaging questions
This survey sought to find relationships between existing behaviors and acceptance/rejection of smart packaging. To explore an individu- al's perception, opinion statements were formulated in the first per- son and language such as “I like” and “it is important to me” was used to encourage answers without overthinking. Prior to the introduction of smart packaging, questions regarding existing behaviors, knowledge, and current satisfaction with food packaging were presented. This order of introduction reduced the potential for response bias.
An established scale that determined health consciousness and measured consumer sensitivity to health issues was adapted for this survey. The original scale questions were developed by Kraft and Goodell (1993) and later adapted by Jayanti and Burns (1998). This study employed two questions verbatim and updated the remaining, result- ing in four questions concerning consumer behavior. The questions are presented in the first column of Table 2. Responses were measured on a five-point Likert scale ranging from “never” to “almost always.”
Depending on the survey, an introduction to either AP or IP was presented. This included a definition of the technology and two di- agrams, demonstrating the packaging's primary method of action. Lay terminology was used to define technologies. Following these definitions, participants were required to respond to the following: “I am willing to consume products that use active/intelligent packag- ing.” Responses were measured on a five-point Likert scale rang- ing from “strongly disagree” to “strongly agree.” The inclusion of this question enabled later segmentation and analysis of participants.
The participants were further asked to indicate product-specific acceptance. The statement “I would accept active/intelligent packaging for...” was presented. This question allowed participants to tick any relevant answer. The options included the following: dairy; fruit and vegetables; meat; and drink products.

3980 |

2.6 | Trust in institutions
Questions regarding trust were adapted from Pliner and Hobden (1992), Siegrist (2000) and Roosen et al. (2015). The scale prompted participants to indicate trust in respective institutions, regarding food safety responsibility. Six institutions were presented for evaluation: the agricultural industry, the food industry, the science/research field, the pharmaceutical industry, government agencies, and consumer organizations. The five-point Likert scale used to measure responses ranged from “extremely suspicious” to “extremely trustworthy”.
2.7 | Data analysis
Data analysis was performed using SPSS software (version 25). In order to examine the association between novel packaging acceptance and Chinese consumer characteristics, a single item question “I am willing to consume products that use active/intelligent packaging” was included in the packaging section of the survey. Respondents that “agreed” or “strongly agreed” were grouped, and others that “disagreed” or “strongly disagreed” were grouped. All participants that responded “neither agree nor disagree” were omitted in further analysis. A chi-square (χ2) test for independence was utilized. This method of analysis was chosen as other methods, such as comparisons of means and t tests required normally distributed results (Kinnear, 2004). When reviewing the results from this study, it was de- termined that the categorical outcome assumption of t tests did not hold. SPSS Statistics' Exact Module was used in result examination as not all results had an expected count greater than or equal to five.
Further post hoc testing utilized contingency tables. Proof and calculations of this method of post hoc testing are outlined by Beasley and Schumacker (1995) and García-pérez and Núñez-antón (2003). This required adjusted residuals from previous analysis to be transformed into chi-square values, and further calculations were then utilized to determine p-values. In doing so, the variable causing the statistical significance could be identified.