Instagram Users’ Information Acceptance Process for Food-Content

: Taking pictures of food and sharing them on social networking services has now become a general consumer trend. In particular, many companies are interested in Instagram marketing due to the increase in users and word-of-mouth effect generated by using hashtags. Therefore, this paper aims to investigate Instagram users’ food-content acceptance processes by applying the information adoption model. Sample data for this study was collected by an online survey company and a total of 333 valid responses were analyzed. The study found signiﬁcant relationships between food-content quality—accuracy, relevance, and conciseness—and information usefulness. Among the factors constituting source credibility, source trustworthiness and hashtag scalability were found to have a signiﬁcant relationship with information usefulness. Information adoption mediated the relationship between information usefulness and continued use of food-content Instagram/information sharing intentions. The theoretical and managerial implications based on these ﬁndings can contribute to designing marketing strategies.


Introduction
Food is one of the most popular themes for users on social networking services (SNSs). Several food-related hashtags such as #food, #instafood, #foodporn, and #yummy ranked among the world's top 100 hashtags in 2020 [1]. According to a report, respondents between the ages of 18 and 35 browse food content approximately five days a week on Instagram, and 30% of respondents are reluctant to visit a restaurant if its Instagram presence is weak [2]. Sharing food-content is a way for individuals to express their identity, feelings, cultures, and memories [3]. In this way, posting food photos on SNSs and choosing restaurants to visit through SNSs has become not only a SNS trend but also a dining trend. Moreover, understanding users' information adoption behavior on SNSs would be helpful to the food industry management.
SNSs allow users to freely express their feelings about products, brands, and experiences [4]. They are a communication platform where users exchange opinions, argue, and co-create with each other [5] The information generated by interactions between SNS users helps customers reduce their uncertainty regarding intangible hospitality products before purchasing [6]. In addition, SNSs enable continuous communication between customers and companies and facilitate building sustainable relationships [7][8][9]. Therefore, it is necessary to understand what information-related variables drive consumers' information adoption and sharing behaviors on SNSs.
SNS users now appear to prefer viewing images rather than reading posts with text [10]. Accordingly, Instagram, an image-based SNS, has been gaining in popularity among teens and young adults in comparison with Facebook [11]. As Pittman and Reich 2. Literature Review 2.1. Information Adoption Model (IAM) IAM, proposed by Sussman and Siegal (2003) [21], has been widely adopted to explain how people are affected by computer-mediated communication. Petty and Cacioppo (1986) [22] postulated that individuals process persuasive arguments and transform their attitudes through two distinct routes-central and peripheral and introduced the elaboration likelihood model (ELM). Hence, Sussman and Siegal (2003) [21] suggested a new theoretical model incorporating ELM and technology acceptance theory. They asserted that information usefulness is critical to explain theories of informational influence since whether or not information is useful could be an important factor determining information acceptance processes. Further, both central and peripheral cues in persuasive messages affect an individual's information adoption but are mediated by information usefulness. Several studies have built a theoretical model based on IAM and verified the path among central cues, peripheral cues, information usefulness, and information adoption in the context of SNSs [23], online communities [24], and websites [25].

Food-Content Quality
Due to the dramatic growth of content (i.e., information) on SNSs, including Instagram, users potentially have access to vast amounts of information. Although the available amount of content is now enormous, the quality of that information is also important because users only adopt information that is helpful to them [26]. According to IAM, as well as ELM, information quality is a central cue in information processing [21,22].
Several researchers have identified various attributes of information quality. Accuracy, relevance, conciseness [27,28], timeliness, and accessibility [28] were identified as the attributes of information quality. Huang et al. (1999) [29] categorized elements of information quality discussed in the previous literature into four dimensions: intrinsic, contextual, representational, and accessibility information quality. In this study, variables representing three of the dimensions mentioned above (accuracy, relevance, and conciseness) were selected. However, accessibility was excluded because it had a relatively low correlation with information quality in Kang and Namkung's (2019) [30] study.
The most basic function of an information source, such as a web site or SNS, is to offer clear and correct information about a particular topic [31]. The accuracy of information is a major factor among the dimensions of information quality and affects the usefulness of information [17] and satisfaction [32]. Further, inaccurate information not only lowers trust but also raises concerns about the source [33] and eventually causes users to deviate [34].
Today's users often face the stress of information overload, so it is important to obtain information that meets user needs [35]. SNSs provide users with numerous opportunities to actively share content, but users also face the stress of information overload [35]. Many previous studies have suggested that relevance is an element of representational information quality that must be discussed when assessing information quality [27,28].
SNS users increasingly prefer viewing visual information over reading the text of their posts [13]. This is that users are likely to invest little efforts in reading, and the conciseness of information is becoming more important. Image information could deliver messages briefly [36] and make it easier to recognize and remember over times [37]. Conciseness would act as a critical factor in explaining the quality of image-centric information.

Source Credibility
Source credibility has widely been addressed as a peripheral cue that can change an individual's attitude [30,38]. Source credibility needs to be considered in virtual environments in particular because it plays an important role in decreasing uncertainty and establishing trust in the information, which contributes to decision making [39,40].
A variety of variables, such as trustworthiness [30,38,41], attractiveness [38], popularity [42], expertness, reputation [30], social tie strength, and informational social influence [38], have been suggested as sub-dimensions of source credibility. One factor that has been mentioned several times in previous studies as a sub-dimension of source credibility is trustworthiness [30,38,41,43]. Source trustworthiness refers to an individual's perception of an informant's honesty, reliability, and integrity [44]. Many studies have found that positive attitudes and behavioral changes are induced when the trustworthiness of an information source is higher rather than lower [45][46][47]. In addition, the Korea Fair Trade Commission's amendment to the Fair Labeling and Advertising Guidelines requires that celebrities must include the word/hashtag "sponsor" or "advertising" when they have been paid to post about certain products or brands on their SNSs. Under this revised guideline, which has been in effect since September 2020, users are able to access more reliable information [48].
This study attempted to reflect the function of hashtags as a component of source credibility. Hashtags, which often contain subjects or words relevant to the topic [49], hashtags have the advantage of making searching easier because they often consist of simple words related to the subject [50]. Users can acquire needed information more quickly, find people with common interests, and generate a positive attitude toward and trust in information sources that use hashtags with attractive topics or keywords [38]. Hashtags also make contents on SNSs more informative [38]. It could be assumed that the ability of information sources to expand and link to specific messages allows users to have confidence in the source. Although hashtags are an important factor influencing usage behaviors on SNSs, previous studies have not examined hashtags as antecedents affecting information processes. Given that hashtags have the power to connect people with common interests, disseminate information, and make information more easily searchable, they could be considered as a characteristic of information sources. Accordingly, this research attempted to determine the influence of 'hashtag scalability', which is defined as the function of hashtags that allows users to access a wide variety of information.
In online environments, popularity generally includes the number of followers and "Likes", which means that an information source has links with many other community members and receives wide support from them. Therefore, source popularity, the extent to which an information source is widely liked, accepted, or preferred, are dominant social influences on SNSs [51,52]. The number of "Likes" and comments [53], the number of followers [51], and the number of mentions by others [54] are all parameters used to evaluate popularity on SNSs. People are usually more likely to be persuaded by a message if other individuals agree with the message than if no one supports the message [55]. How popular and attractive an information source is thought to be is an important factor to consider for better understanding persuasive message processing on SNSs.
In sum, source trustworthiness, which a number of previous studies have suggested is a component of source credibility, is presented as a peripheral cue affecting information processing in the current study. Furthermore, to investigate the role of hashtags and social power, which are characteristics of SNSs, hashtag scalability and source popularity were included in the peripheral routes.

Model Development and Hypotheses
In this research, a proposed model was developed to examine how food-content quality and source credibility impact food-content usefulness, information adoption, continued use of food-content Instagram, and information sharing intention based on the IAM. hashtags have the advantage of making searching easier because they often consist of simple words related to the subject [50]. Users can acquire needed information more quickly, find people with common interests, and generate a positive attitude toward and trust in information sources that use hashtags with attractive topics or keywords [38]. Hashtags also make contents on SNSs more informative [38]. It could be assumed that the ability of information sources to expand and link to specific messages allows users to have confidence in the source. Although hashtags are an important factor influencing usage behaviors on SNSs, previous studies have not examined hashtags as antecedents affecting information processes. Given that hashtags have the power to connect people with common interests, disseminate information, and make information more easily searchable, they could be considered as a characteristic of information sources. Accordingly, this research attempted to determine the influence of 'hashtag scalability', which is defined as the function of hashtags that allows users to access a wide variety of information.
In online environments, popularity generally includes the number of followers and "Likes", which means that an information source has links with many other community members and receives wide support from them. Therefore, source popularity, the extent to which an information source is widely liked, accepted, or preferred, are dominant social influences on SNSs [51,52]. The number of "Likes" and comments [53], the number of followers [51], and the number of mentions by others [54] are all parameters used to evaluate popularity on SNSs. People are usually more likely to be persuaded by a message if other individuals agree with the message than if no one supports the message [55]. How popular and attractive an information source is thought to be is an important factor to consider for better understanding persuasive message processing on SNSs.
In sum, source trustworthiness, which a number of previous studies have suggested is a component of source credibility, is presented as a peripheral cue affecting information processing in the current study. Furthermore, to investigate the role of hashtags and social power, which are characteristics of SNSs, hashtag scalability and source popularity were included in the peripheral routes.

Model Development and Hypotheses
In this research, a proposed model was developed to examine how food-content quality and source credibility impact food-content usefulness, information adoption, continued use of food-content Instagram, and information sharing intention based on the IAM. Figure

Food-Content Quality, Source Credibility, and Perceived Usefulness
Sussman and Siegal (2003) [31] developed the IAM, which posits that information usefulness mediates information quality/source credibility and information adoption, in

Food-Content Quality, Source Credibility, and Perceived Usefulness
Sussman and Siegal (2003) [31] developed the IAM, which posits that information usefulness mediates information quality/source credibility and information adoption, in order to better understand the complex knowledge adoption process. When users perceive that information quality is good, they consider the information to be useful for their task [17]. Cheung et al. (2008) [24] found a significant relationship between information quality (relevance and comprehensiveness) and information usefulness within online communities. Jin et al. (2009) [23] identified that information quality is an important determinant of information usefulness and influences satisfaction indirectly via information usefulness. Tseng and Wang (2016) [25] found the effect of information quality on information usefulness. Aghakhani et al. (2017) [17] presented empirical evidence that supports the effect of review accuracy on information usefulness. Fang et al. (2016) [56] mentioned that how easily a review can be understood is important for creating information value. Thus, the following hypotheses were proposed: Hypotheses 1a (H1a). The accuracy of food-content on Instagram positively affects information usefulness.
Hypotheses 1b (H1b). The relevance of food-content on Instagram positively affects information usefulness.
Information is usually generated by unknown individuals [57]. Accordingly, the information source would be key factor that helps users evaluate information [58]. Jin et al. (2009) [23] and Tseng and Wang (2016) [25] showed that source credibility had a positive effect on information usefulness. Ismagilova et al. (2020) [59] found that source trustworthiness and expertise have a positive effect on eWOM usefulness. Ayeh (2015) [60] identified the effects of source trustworthiness on usefulness when travelers are tasked with information seeking. Chang et al. (2015) [61] supported a positive relationship between popularity and usefulness. If users receive information that is a good fit with their given task, then they perceive the information as useful [62]. Since users can receive useful information for decision making via hashtags, the function of hashtags will have a significant influence on their perception of information usefulness [63]. Kim et al. (2019) [63] found that hashtags are associated with perceived information usefulness, although the degree of information usefulness varies according to the type of hashtag (informative/storytelling). Therefore, the following hypotheses were proposed: Hypotheses 2a (H2a). Source trustworthiness positively affects information usefulness.

Information Usefulness and Information Adoption
Usefulness is a user's perception that adopting new technology will contribute to improved performance [64]. When people perceive that information within an online environment is useful, they strongly tend to adopt the information to enhance their performance [24]. The relationship between information usefulness and information adoption is part of the IAM [21]. Numerous studies have confirmed that information usefulness is an important variable to determine whether an individual will adopt information [57,[64][65][66]. Therefore, the following hypothesis was formulated: Hypotheses 3 (H3). Food-content usefulness positively affects information adoption.

Information Adoption, Intentions to Continue Using Instagram, and Information Sharing Intentions
SNSs, including Instagram, are emerging as innovative sources for people seeking information. In fact, it is considerably important as an information search tool [67]. If a technology satisfies users as an information search tool, then it will induce positive attitudes and users will continue to use the technology [68][69][70]. Hence, it could be assumed that the more satisfied users are with information adopted from Instagram, the greater the likelihood is that they will continue to use Instagram. Imlawi (2017) [71] confirmed the significant relationship between information adoption and sharing intention, and mentioned that when users accept useful information, they want to actively participate in SNS activities by sharing the information. Thus, the following hypotheses were proposed: Hypotheses 4 (H4). Information adoption positively affects continued use of food-content Instagram.

Data Collection
This study employed quota sampling method to examine Instagram users' information acceptance processes. Prior to conducting the main study, a pilot-test was performed on 40 SNS users to ensure reliability and validity of the items. To collect the data for the main study, an online survey was conducted via the online research company, from 5 February to 12 February 2020. Online questionnaires were sent out to users who had experiences with food-content on Instagram in the last three months. The age ratio of the respondents was set by taking into consideration the actual age ratio of Instagram users in 2019 as presented in Statista's report [72]. Based on the report, data were collected by assigning 30% to respondents 18-23 years old, 30% to respondents 24-29 years old, 25% to respondents 30-39 years old, 10% to respondents 40-49 years old, and 5% to those over 50 years old. The questionnaires were e-mailed to respondents and 1059 respondents participated in the survey. 370 responses were screened using specific items (e.g., What SNS do you usually use? and what information did you get through SNSs in the last 3 months?). Lastly, 37 answers were removed due to incompleteness, and a total of 333 valid answers were used in the analysis. A detailed demographic profile of the respondents is presented in Table 1.

Measurement Development
The constructs for the proposed model were measured with multiple questionnaire items using a 7-point Likert ("1 = strongly disagree" and "7 = strongly agree") The information quality of food-content, which consists of accuracy, relevance, and conciseness, was measured with four items for each factor. The measurement items are presented in the Table 2. Accuracy (e.g., "The food-content on Instagram is accurate." and "The food-content on Instagram is correct.") and relevance (e.g., "The food content on Instagram is relevant." and "The food content on Instagram is suitable.") of food-content were measured using  [24]. The items for assessing conciseness of food-content (e.g., "The food-content on Instagram is readable." and "The food-content on Instagram is formatted compactly.) were adopted from Kim and Kyung (2018) [73] and Lee et al. (2002) [74]. Source trustworthiness was assessed with four items (e.g., "The source of food-content is reliable." and "The source of food-content is accurate.") adopted from Cheung et al. (2008) [24]. Four items (e.g., "Hashtags help me get additional information." and "Hashtags help me get updated food-content.) from Chang (2010) [77]. Food-content usefulness was assessed with four items (e.g., "The food-content on Instagram is informative." and "The food-content on Instagram is valuable.") adapted from Bailey and Pearson (1983) [78] and Chu and Kim (2011) [79]. Information adoption was measured with four items (e.g., "The food-content made it easier for me to make decision." and "The food-content helped me effectively make a purchase decision.") adopted from Cheung et al. (2009) [22], Erkan and Evans (2018) [66], and Wu and Shaffer (1987) [80]. Continued use of Instagram for food-content was measured with four items (e.g., "I intend to continue using Instagram to search for food-content." and "I will frequently use Instagram.") adopted from Bhattacherjee (2001) [81] and Hasim and Tan (2015) [82]. Information sharing intentions were measured with three items (e.g., "I would like to continue sharing food-content with others on Instagram in the future." and "I expect to continue sharing food-content on Instagram in the future.") adopted from Fang and Chiu (2010) [83] and Lee and Ma (2012) [84].

Data Analysis
To examine the theoretical model, this study conducted descriptive analysis for profiling the respondents and used the two-step method suggested by Anderson and Gerbing (1988) [85] to test the hypotheses. Further, this study evaluated the validity and reliability of the measurement models [86] and then analyzed the structural model by using AMOS 18.0 software.

Model
CFA was conducted to assess the model fit. Table 2 presents the fit indices of the measurement model, which indicates a good fit between the theoretical model and the data (χ 2 = 1295.192; df = 620; χ 2 /df = 2.089; GFI = 0.826; IFI = 0.931; TLI = 0.921; CFI = 0.931; RMSEA = 0.057). By examining the Cronbach's alpha coefficients, which ranged from 0.868 to 0.909, the reliabilities of each construct were verified. The CR for all constructs ranged from 0.864 to 0.911, which was above the threshold of 0.700 [87]. AVE values for all variables ranged between 0.615 and 0.906, which were greater than the recommended threshold of 0.500. Therefore, the convergent validity of the measurement scale was identified.
To assess discriminant validity this study employed the constrained phi approach [88] and the overlapping confidence intervals method [85]. First, a chi-square difference test was conducted between the unconstrained model and the constrained model [88]. When the chi-square statistic is more than 3.84 based on a freedom difference of 1 between the unconstrained model and the constrained model, then the two models are regarded as distinguishable and discriminant validity is confirmed. In this research model, the chi-square difference existed at a significance level of 0.01 (∆χ 2 = 7.896, ∆df = 1), which presented evidence of discriminant validity. Discriminant validity was also observed by developing a confidence interval for each pair of constructs [88]. The high end of the confidence interval between Instagram information usefulness and intentions to continue using Instagram, ranged from 0.973 to 0.669 and did not include the value of 1.0, which supported the existence of discriminant validity.

Discussions
The empirical findings that can be interpreted based on the results of statistical analysis are discussed below. First, the results provide evidence that the quality of food-content on Instagram, consisted of accuracy, relevance, and conciseness, was a determinant of information usefulness. The relevance of food-content had a relatively stronger effect than other factors, which is consistent with the findings of Cheung et al. (2008) [26]. This result means that whether information meets an individual's needs and situation is a key factor to assess usability. The significant influence of the conciseness of food-content was also revealed and is in line with previous research by Fang et al. (2016) [56]. This result and the high mean value of the conciseness items in the descriptive analysis imply that users perceived information that was easy and simple-to-understand as more useful than information requiring literacy and deep comprehension.
Second, source trustworthiness and hashtag scalability were found to be related to information usefulness. These findings are in line with the results of Ismagilova et al. (2020) [59] and Kim et al.'s (2019) [63] studies. Considering that hashtag scalability had a relatively greater influence than other factors, it could be assumed that hashtags on SNSs is a vital component in determining usability. These results show that the characteristics of Instagram are well reflected in the research. However, the relationship between source popularity-how many people follow the source-and information usefulness was insignificant, which is not consistent with the findings of Chang et al. (2015) [61]. These results imply that the popularity of an information source does not necessarily satisfy utilitarian motives such as useful information. This means that sources with a large number of followers do not always provide information that meets the needs of users.
Finally, the more useful the food-content is, the more likely it is to be adopted. Furthermore, as the adoption of information through Instagram increases, the continued use of food-content Instagram and the intention to share information also increases. The results of this study supports findings form Sussman and Siegal (2003) [21], Ifinedo (2018) [68], Sun et al. (2020) [69], and Imlawi (2017) [71]. In the future, the amount of information on SNSs will grow even more massive and how useful and helpful that information is will be crucial to users. If users obtain benefits and value from a specific SNS by receiving useful information, then they will continue using that SNS as a channel to search for information and actively participate in SNS activities.

Theoretical Implications
The findings of this research provide several theoretical implications. First, this study targeted an image-based SNS that has increased its number of users in recent years. In addition, it is also very meaningful that the hashtag, one of the main functions of Instagram, is reflected in the model. Despite how much has been written about information on SNSs, there was a paucity of research focusing on image information or image-intensive SNSs in the field of the food service industry. Therefore, this study contributes to the literature by offering a theoretical basis for future research related to food-content.
Second, this study recognized the importance of hashtags, one of the main functions of Instagram, and included them as a factor, referred to as scalability, that constitutes source credibility. Hashtag scalability had the second greatest influence on information usefulness after the relevance of food-content. The findings provide useful additional evidence that hashtags act as a peripheral cue in message processing. This implies that hashtag scalability needs to be treated as a critical construct in user's information processing on SNSs.
Third, this study examined post-information (e.g., continuous intentions to use Instagram and information sharing intentions) behavior that occurs after the information adoption process explained by the IAM. The existing literature has so far set 'information adoption' as a dependent variable and verified causal relationships with other constructs. In contrast, by extending the dependent variable to include intentions to continue using Instagram and information sharing intentions, the current study was able to verify post-information adoption behaviors and obtain broader implications.
Finally, previous research typically examined information quality and source credibility as a single dimension [89,90]. In contrast, this study was able to acquire specific implications by separately verifying the influences of sub-factors composing information quality and source credibility on information usefulness.

Managerial Implications
This research offers useful practical implications. First, many hospitality companies, including food service companies, have been interested in SNS marketing to communicate with both existing customers and potential customers. This study found that hashtag scalability is one of the major variables influencing perceptions of the usefulness of foodcontent. Therefore, companies should establish Instagram marketing tactics to facilitate consumer searches for the brand or service and concisely identify key information through hashtags. For example, organically exposing customers to a company's brand, services, and products through an event hashtag could help raise brand awareness and the company's image. In addition, because the latest consumer trend has shifted from enjoying textoriented content that needs to be read to preferring content that can be viewed [10] brief and short hashtags that make a strong impression can provide useful information for users.
Second, the relevance of food-content was founded to be considerably important and could be related to hashtags because they are used as an inducement to attract customers. However, attracting customers through keywords that are irrelevant to the service or product may cause customers to be dissatisfied with a firm. Thus, it is critical to use product/service-related hashtags and provide accurate and relevant content for customers.
Third, in order to increase the conciseness of food-content, it is recommended that firms provide information using images or short videos. When firms provide images and short messages (i.e., captioned photos) on SNSs, users are able to quickly and easily understand the service or brand and perceived content usefulness can be enhanced.
Fourth, in terms of the accuracy of food-content, the visual information presented on SNSs should be similar to the actual image of products and excessive modification of these images should be avoided. Further, presenting information in the form of stars, scores, or ratings can compensate for the limitations of subjective evaluations of service products compared to tangible products and allow users to have a more accurate grasp of the service.
Fifth, in order to increase source trustworthiness, companies need to present accurate product information on SNSs. Additionally, the previous literature [91] shows that users find reviews generated by individuals' who actually used the products/services more trustworthy than marketer-generated content. Accordingly, if companies facilitate customers actively sharing their experiences on SNSs, then higher marketing performance will be achieved. Besides, sincerely addressing both a service's strengths and weaknesses in a review can enhance source trustworthiness.
Finally, encouraging users to adopt and share food-content is important to increase the effects of marketing. Therefore, this study proposes that firms should offer various benefits (e.g., free beverage, coupon, discount, etc.) to users who share their restaurant experiences through Instagram posts. Further, the eWOM effect can be reinforced if information is further disseminated through Instagram events so that accepted content can be shared with third parties.

Limitations and Future Research
There are several limitations requiring attention in future research. First, this study did not cover all possible factors that make up information quality and source credibility. Further research could explore additional sub-factors that sufficiently reflect the characteristics of Instagram, an image-intensive SNS, to better understand users' behaviors.
Second, although this study focused on food-related content and users' information acceptance processes, the food-related content mentioned by respondents might not only deal with the topic of food, but could also include a variety of topics, such as travel, places, interiors, animals, etc. Hence, future research needs to develop a method for selecting and quantifying the subject of content and apply it to grasp the structural relationship of consumer's information adoption.
Third, the current study did not confirm a significant relationship between source popularity and food-content. However, considering that previous literature has shown that how popular an information source influences a user's beliefs and attitude formation [92], the importance of source popularity cannot be overlooked. Thus, it remains important to revalidate the effect of source popularity on usefulness by targeting other service areas or exploring other dependent variables influenced by source popularity.
Fourth, although information processing via the central and peripheral routes might vary depending on the user's knowledge, ability, and motivation, the moderating effects of individual variables were not examined in the study. In future research, further impli-cations could be obtained if the moderating effects of factors, such as frequency of SNS usage, duration of using SNSs, and degree of SNS participation, are verified across these dual paths.
Fifth, this study's results lack generalizability due to the sample. Although Instagram is a SNS used by people all over the world, research was conducted only in Korea. And the sample was small compared to the number of Instagram users. Therefore, future research needs to be carried out by extracting a larger number of samples from various countries.
Lastly, the research subjects in this study were limited to Instagram users. Future research should comprehensively examine other image-oriented SNSs such as YouTube and Pinterest as well so that broader implications can be obtained.