The primary purpose of this study was to develop a technological affordance scale based on the constructs identified by the MAIN Model. The manuscript presents a 12-item scale for the individual technological affordances and a 2-item scale for perceived credibility of social media platforms. Three social media platforms, Facebook, Twitter and Snapchat were evaluated for perceived credibility and the technological affordances; Modality, Agency, Interactivity and Navigability. A composite scale was developed for platform credibility and platform technological affordance across the three social media platforms. The manuscript concludes that Agency played a moderate role in predicting credibility across all three social media platforms, where the other constructs of the MAIN model were not significant factors in predicting credibility.
Keywords: main model, technological affordances, digital media, credibility, perceptions
Digital advertising is a billion-dollar industry that continues to grow due to mobile device and Internet adoption. Social Media is also becoming an increasingly contributing factor to the growth advertising expenditures. A prominent issue facing advertising scholars, marketers and publishers is source credibility in online environments. The overall ambiguous nature of Internet-based communication has left researchers, marketers and consumers with little to measure in terms of source, platform or feature credibility. Specifically, as social media modalities gain strength in strategic communication efforts, researchers will seek to uncover the affordances, features and credibility within an online environment while uncovering a theory that helps explain the complexities of source credibility in online environments. Source credibility is an important factor as it can lead consumers toward product, service or message consumption and economic implications for organizations, businesses and enterprises.
With the growing number of sources, platforms and advertisers growing each day, credibility remains a complex set of variables to measure in an online environment. This research explores the utility of the MAIN Model (Sundar, 2008) for measuring technological affordances and heuristic cues available in online social media platforms and how it coalesces with a new breed of credibility measures. The MAIN Model is a descriptive term for evaluating the modality, agency, interactivity and navigability of a message or platform. There is no current research or scholarly work that has attempted to evaluate the technological affordances and the accuracy of the MAIN Model for explaining credibility of messaging in an online environment. This research is attempting to evaluate the MAIN model beyond the conceptual proposition and identify empirical evidence to its claim that increases in technological affordances help predict credibility in online environments.
The study explores the MAIN Model for assessing the perceived technological affordances available on the Facebook, Twitter and Snapchat. Additionally, participants have provided their perceptions of the credibility associated with each social media platform. This assessment will be used to determine and qualify acts of credibility based on two dimension; expertise and trustworthiness (McCroskey et. al, 1974). This research proposes psychometric scale for measuring the MAIN model across social media platforms and other web-based properties. Additionally, once this scale is proposed and tested, analysis can be produced to determine the predictive power of the MAIN model, or its constructs. The implications for identifying features that predict credibility for social media platforms have profound implications for future research, media-buyers, marketers, advertisers and scholars.
In a review of the literature, the MAIN model was used as a supportive instrument across the majority of the studies. In-depth analysis has been conducted on the elements contained within the MAIN model, such as heuristic cues, individual affordances (Lee & Sundar, 2013; Kim & Sundar, 2015; Kim & Sundar, 2011) but an evaluation of the MAIN model as an aggregate or composite construct for predicting credibility has not been found in a review of the literature.
The research questions, hypotheses and arguments relied on the MAIN model as a surface-level argument for heuristic cues identified in digital media platforms and technologies, such as websites and social media platforms. The MAIN model as a construct for identifying heuristics, cues and affordances were not evaluated, measured or tested. Overwhelmingly, the authors relied on the heuristics identified by Sundar (2008) to formulate a research question or hypothesis and in some cases, tie it back to the results. For example, Lee and Sundar (2015) utilized the heuristic properties associated with the MAIN model taxonomy for evaluating the bandwagon cue and the authority cue. What isn’t evident in the research is a consistent representation of heuristics and taxonomy for evaluating digital media and technological affordances nor the use of the MAIN model in its entirety to evaluate or predict credibility in digital media properties. The taxonomy seems to be subject to the interpretation of the researcher for identifying the most convenient heuristic that best meets the needs of their research question or digital media platform.
The current research was guide by these questions:
RQ1a: Can reliable measures of each of the four technological affordances associated with the MAIN model be developed?
RQ1b: Do the psychometric properties of the technological affordance scales vary as a function of social media platform?
RQ2a: How does the MAIN model and related technological affordances predict perceived credibility of Twitter?
RQ2b: How does the MAIN model and related technological affordances predict perceived credibility of Facebook?
RQ2c: How does the MAIN model and related technological affordances predict perceived credibility of Snapchat?
Respondents completed an instrument that included demographic questions. Additionally, an instrument evaluated the participant’s social media usage and familiarity with social media platforms. From the social media utilization instrument, social media channels were selected based on the usage and popularity scores across the population as reported by the participants. Facebook, Twitter, Instagram and Snapchat, respectively were the highest utilized social media platforms across the sample population. Instagram was not included in the study due to measurement and validity issues. Utilizing the MAIN Model for technological affordances, a repeated measures instrument evaluated perceptions based on each unit of the MAIN Model, as they exist across four social media platforms; Twitter, Facebook and Snapchat. In addition to measuring the platform specific technological features or affordances, perceived platform credibility was also evaluated for the same social media platforms. The credibility instrument included a modified, two-pronged version of McCroskey’s (1974) credibility measures based on expertise and trustworthiness. Using this methodology, participants completed the instrument in reference to their perceptions of technological affordance and credibility for each social media platform.
The Technological Affordance Scale
The technological affordance scale is a 12-item instrument that asks respondents to report their perceptions on the individual technological affordances available in a specific social media platform. This instrument is multidimensional and shared across three sections that measured perceived technological affordance and features in Twitter, Facebook and Snapchat. The result of each section created a summary of the perceived technological affordances reported for each platform with respect to the feature affordances as described by the MAIN model (Modality, Agency, Interactivity and Navigability) (Sundar, (2008). This MAIN model composite calculation will help the researcher evaluate the perceived technological affordance levels for each social media platform. All responses were solicited using a 7-point Likert Scale ranging from strongly disagree (1) to strongly agree (7).
The modality unit is the most structural of the four affordances and the most apparent on an interface. Computer based media has complicated traditional modalities and now use the term multimedia (Sundar, 2008). Modality was measured within the 12-item technological affordance scale as a 5-item sub-scale measuring modality perceptions on social media channels. This asked respondents to report their perceptions on the different modality affordances available in social media platforms such as video, animation, text, images or multi-modal (containing all modalities).
The agency affordance of digital media helps make possible the assignment of sourcing the particle entities in the chain of communication from the computer to an online source location or multiple locations (Sundar, 2008). Agency was measured within the 12-item technological affordance instrument scale as a 2-item sub-scale measuring agency perception on social media channels. Each statement prompted the respondents to report their perception on the different agency, source, or authoritative-like qualities in each social media platform.
Interactivity affordances in digital media are capable of cueing a wide variety of cognitive heuristics and are the most distinctive affordance of digital media (Sundar, 2008). The interactivity unit of the MAIN Model was measured within the 12-item technological affordance instrument as a 2-item sub-scale instrument measuring interactivity perception on social media channels. Each statement prompted the respondent to report their perception on the interactive features such as user-control and user-participation in each social media platform.
The navigability affordance has the dual ability to directly trigger heuristics with different navigational aids on the user-interface, as well as to transmit cues through the content that it generates (Sundar, 2008). The navigability unit of the MAIN Model was measured within the 12-item technological affordance instrument as a 3-item sub-scale instrument measuring navigability perception on social media channels. Each statement prompted respondents to report their perception on the navigability features such as menu items, scaffolding, buttons, and links in each social media platform.
Participants were 420 undergraduate students (292 female) enrolled in communication courses at a south-central state university. The respondents’ ages ranged from 18 to 43 years (M = 20.22, SD = 2.59). 113 of the respondents were freshmen, 77 were sophomores, 121 were juniors, 105 were seniors and four specified as either non-degree seeking visiting/transient students. The demographic composition was similar to that of the university student population, with 344 (81.9% White, 44 (10.5%) African American, 14 (3.3%) Asian and the remaining 18 (4.3%) participants reported another ethnicity. Participants completed the survey for course credit in the College of Communication and Information using the SONA system.
The dependent variable in the study is credibility and the independent variable is technological affordance. Technological affordance was measured as a part of the 12-item scale and credibility was measured as a part of the 2-item scale. Sampling procedures were conducted based on the completion of the survey, qualified respondents that were at-least 18 years old. The measures were individual affordances related to the MAIN model and the technological affordance scale aggregated across each social media channel. Additionally, a credibility scale was aggregated for each platform.
The instrument was repeated measures and was designed to ask the same items across each social media platform. The 12-item technological affordance scale was repeated each time across Facebook, Twitter and Snapchat. The two item credibility scale was also presented across each social media platform as well. This allowed for an aggregate measure for technological affordance per social media channel and an aggregate credibility scale across each social media channel.
After calculating an aggregate score for technological affordance for each social media platform and social media credibility Pearson-Product moment correlation was conducted along with means, standard deviations and reliability. The participants indicated their agreement with these items using a 7-point Likert scale ranging from 1 (strongly disagree) to 7 (strongly agree). All items were positively correlated. Table 1 shows the correlation between the two scales as measured by composite for perceived social media credibility and the composite for perceived technological affordances for each social media platform.
Table 1. Means, standard deviations, internal consistency reliability and coefficients between social media credibility and technological affordances for social media platforms
A standard multiple regression analysis was performed between the dependent variable (credibility) and the dependent variable (technological affordance per platform). Analysis was performed using SPSS REGRESSION. The following are results for social media platform affordances and social media platform perceptions of credibility.
Regression analysis revealed that the model significantly predicted credibility as a function of agency and navigability for the Facebook platform, F(2, 410) = 33.98 p <.001. R2 for the model was .33 and adjusted R2 was .14. Table 1 displays the unstandardized coefficient (ß), intercept and regression coefficient (b) for each variable.
Regression analysis revealed that the model significantly predicted credibility as a function of agency and navigability for the Twitter platform, F(2, 410) = 25.1, p <.001. R2 for the model was .39 and adjusted R2 was .151. Table 1 displays the unstandardized coefficient (ß), intercept and regression coefficient (b) for each variable.
Regression analysis revealed that the model significantly predicted credibility as a function of modality, agency and interactivity for the Snapchat platform, F(2, 410) = 41.48, p <.001. R2 for the model was .238 and adjusted R2 was .232. Table 1 displays the unstandardized coefficient (ß), intercept and regression coefficient (b) for each variable.
Using the 12-item scale used for evaluating the technological affordances identified by the MAIN model were factor analyzed using principal component analysis with varimax (orthogonal) rotation. This analysis addresses RQ1a and RQ1b to confirm the measurement of the factors associated with the MAIN model. Table 2 displays a factor analysis for the Twitter platform. Similar loadings were found for both Facebook and Snapchat. Two question relating to the credibility associated with the social media platform were factor analyzed using principal component analysis with varimax (orthogonal) rotation on an individual unidimensional basis.
Table 2. Factor analysis for MAIN model technological affordances for Twitter **
A reliability analysis was conducted to determine based on RQ1 and RQ1b if a reliable measure of each of the four technological affordances (modality, agency, interactivity and navigability) and aggregate scores for credibility in each of the social media channels were present in the scale. The results in Table () are the descriptive statistics for the mean, standard deviation and Cronbach's alpha for the four affordances and credibility across twitter, facebook and snapchat.
Cronbach’s alpha’s had acceptable internal consistency for the psychometric properties in determining Modality for Twitter (α = .87, M = 6.11, SD = .96). Facebook Modality, (α = .91, M = 6.38, SD = .76) and Snapchat Modality (α = .83, M = 6.04, SD = .94).
Cronbach’s alpha was lower in internal consistency for the psychometric properties in determining Agency for Twitter (α = .43, M = 3.56, SD = 1.34). Facebook Agency (α = .36, M = 3.79, SD = 1.30) and Snapchat Agency (α = .46, M = 3.28, SD = 1.45).
Chronbach’s alpha was again lower in internal consistency for the psychometric prosperities associated with determining Interactivity for Twitter (α = .51, M = 5.76, SD = 1.06). Facebook Interactivity (α = .59, M = 6.14, SD = .94) and Snapchat Interactivity (α = .64, M = 5.40, SD = 1.41).
Chronbach’s alpha had acceptable internal consistency for the psychometric prosperities associated with determining Navigability for Twitter (α = .86, M = 5.74, SD = 1.15). Facebook Navigability (α = .88, M = 6.02, SD = .92) and Snapchat Navigability (α = .87, M = 3.62, SD = 1.87).
Chronbach’s alpha had acceptable internal consistency for the psychometric prosperities associated with determining credibility for Twitter (α = .77, M = 4.01, SD = 1.19).
Chronbach’s alpha had acceptable internal consistency for the psychometric prosperities associated with determining credibility for Facebook (α = .81, M = 3.99, SD = 1.27).
Chronbach’s alpha had acceptable internal consistency for the psychometric prosperities associated with determining credibility for Snapchat (α = .86, M = 4.07, SD = 1.52).
Table 3. Descriptive statistics (mean, SD) and internal consistency (Cronbach's alpha) for MAIN and credibility per platform
Development of a Technological Affordance Scale
RQ1a was supported in an attempt at creating a psychometric scale based on the feature components of the MAIN model at a unidimensional level. When evaluated across dimensions, Agency and Interactivity as a construct were less reliable and have concerns for internal consistency when analyzed across the MAIN model in its entirety. Factor loadings were internally consistent when analyzed unidimensionally.
RQ1b was supported as the psychometric properties of the technological affordances varied across social media platforms. For example, Modality as a variable was internally consistent across Facebook, twitter and Snapchat, whereas Agency held less internal consistency across all three platforms. This might be due to the relative psychometric properties for how individuals assign perceptions based on the concepts of Agency and Interactivity as formulated in the scale. For example, some participants might not understand the concept of agency in terms of source credibility, as it might be a factor considered when clicking, viewing or engaging with a post. Furthermore, interactive features again might not be readily perceived by the user across any social media platforms. These features might be available, but participants might not be active agents in evaluating them as a part of the psychometric properties of the proposed scale. Additionally, the factor analysis states that the scale was uniformly distributed based on the constructs or factors for the MAIN model. This provides data that might suggest that the components of the MAIN, primarily Agency and Interactivity might be problematic to measure or record on behalf of a novice or standard user of social media. The model, assumes that the features are used, recognized and perceived, in order to assign credibility measures to the digital media property. This psychometric test might indicate that not all features of MAIN are in an individuals’ consideration set when experiencing a digital media property, in this case a social media platform. Future research should be conducted to evaluate the psychometric value and properties of both Agency and Interactivity and how it appropriately fits into a technological affordance theory. Are these two constructs perceived too similarly, or are there other feature affordances that could be integrated that explain how credibility is assigned across a platform?
MAIN Model for Predicting Credibility
In the regression analysis it was found that agency was an overwhelming predictor of credibility across Facebook, Twitter and Snapchat. The aggregate technological affordance measure helped explain how much of the MAIN model is represented in each platform. This aggregate score was then regressed on the credibility measure for each social media platform. RQ2a, RQ2b and RQ3c all were concerned with how the MAIN model predicted credibility across each social media platform. Agency performed highest in terms of predicting credibility the three social media platforms. Agency was an important variable for every platform as the loads are higher for predicting the variance in credibility in social media platforms.
It was found that for RQ2a, which was concerned with the MAIN model predicting perceived credibility in Facebook that Agency (r =.317) and Navigability (r = .178) accounted by 14% of the variance. The remaining technological affordances were not significant or were weak for predicting the variance in credibility for Facebook.
RQ2b was concerned with predicting the perceived credibility for the MAIN model in Twitter. The data suggested that Agency (r = .237), Interactivity (r =.121) and Navigability (r = .202) accounted for 15% of the variance for perceived credibility. As with Facebook, Modality was not a factor for predicting the variance on the Twitter platform.
RQ2a was concerned with predicting the perceived credibility for the MAIN model found on Snapchat. The results of the data suggested that, alongside Twitter three factors predicted the variance in perceived credibility for Snapchat. Modality (r = .165) was the only platform that Modality was a factor, Agency (r = .385) was the highest of the beta weights for technological affordances within the regression analysis and finally, Navigability (r = .141) accounted for 23% of the variance for perceived credibility.
Social media is made up of individual users, who rely on other users to share, post and produce content for consumption. Agency as a construct is an important component of the social construction of media found on socially driven platforms. This and the data suggest that, when considering a social media campaign, the features associated with Agency should be considered when deploying a message. Who, what and the verifiability of the information or sender as it relates to the audience all play a critical factor in how a message will be received as credible or not. Additional research should be conducted to evaluate how individually agency predicts credibility on social media channels and if there are any additional components that contribute towards perceived credibility for consuming content on a socially driven network.
The data suggested by Snapchat, the newer of the social media platforms, revealed that modality, agency and interactivity were a function of credibility. This could be based on the simple feature sets presented on Snapchat. Alternatively, it could be based on the new features that Snapchat had made available at the time of this study. As a result, the users of the platform maintained a high level of awareness of those new features. This leads to the question, do all users of a social media platform use the affordances offered? It might be the case that some are unaware of the affordances presented on Facebook, but might be very well-versed in the affordances on Snapchat because of a recent introduction to new features in the product.
Limitations and Future Directions
This current study has several limitations. We only focused on social media platforms as a means to evaluate credibility of the constructs of the MAIN model. Future research should evaluate web experiences, and other digital media platforms beyond social media platforms. Additionally, the participants were asked to report their perceptions of the platform, rather than the content, source or message found on the platform. This presents difficulties in making assumptions about individual content derived from each channel. Users could have pre-conceived understandings, opinions and values associated with one social media platform over the other.
Additionally, the low internal validity and reliability for the multidimensional constructs for Interactivity and Navigability presents opportunities for furthering the creation of the psychometric scale. The reliability measures were evaluated unidimensionally across each social media channel is it related to Modality, Agency, Interactivity and Navigability separately. However, when ran as a composite, reliability issues were present, this can be seen in Table 3. This will help to determine if the scale is the source of the low levels of internal consistency or if the MAIN model constructs for Interactivity and Navigability are problematic for scale development or psychometric testing. If the latter, additional research should be conducted to determine alternative psychometric constructs that can be more accurately articulated or identified as feature affordances in a web or digital media environment.
In order to understand how individual features, promote, guide and predict credibility on digital media properties, more empirical research should be conducted on evaluating the propositions made by the MAIN model for technological affordances. Agency has been shown to be moderately critical in predicting credibility in three social media platforms, but little research has been presented on how the remaining constructs within the MAIN model account for credibility. Additionally, the MAIN model presents researchers with is the best conceptually designed construct for identifying all of the proper feature affordances for a user experience in a web environment. Can a better conceptual model be developed?
Agency, as an overwhelming predictor of credibility has positive implications for how advertisers, media-buyers and digital media strategist position and select content to display. But how can the affordances predict credibility in social media? Furthermore, noting the regression analysis, different affordances are predicting credibility at different strengths as the social media platform changes. For example, generating content that elicits cues of agency on the Facebook platform might be more positively perceived, whereas eliciting cues of modality or interactivity might not yield different results. Likewise, for Twitter, which Agency is an overwhelming predictor for credibility, but navigability is not as important. According to these results and the predictive power of the MAIN model, the results are inconclusive for positing the predictive power for the MAIN model as an aggregate construct for understanding more about how credibility is a factor in a digital media environment. It still can be said that media planners can plan to deploy different content with separate or distinct cues across various social media platforms, however, to claim that the MAIN model is a distinct predictor for credibility in specific social media channels cannot be supported based on the data collected in this study. As such, additional research needs to be conducted to determine how the MAIN model as an aggregate construct can provide evidence for predicting credibility.
The MAIN model as it is currently positioned has made propositions that it cannot fully support. Meaning, no empirical research or evidence has been developed that explains the feature affordances identified by Sundar (2008). Furthermore, the heuristic taxonomy used across research that utilizes the MAIN model is troublesome. Meaning, the convenient approaches to cues, heuristics and affordances are too precarious and extensible across studies to be useful towards and actionable body of knowledge. This study attempts to qualify the propositions presented by the MAIN model for predicting credibility in the presence of technological affordances in web-based environments and in this case, social media platforms.
The MAIN model is a useful mechanism for evaluating the technological affordances in a digital media environment. What does this say about technological affordances? How do the technological affordances in a digital media environment differ, or relate to the affordances presented to humans in a physical or ecological environment? The theory of affordances, proposed by Gibson (1977) states,” The affordances of the environment are what it offers to the animal, what it provides or furnishes either for good or ill.” (Gibson, 1977). Gibson lists four properties for affordances; horizontal, flat, extended and rigid as physical properties of a surface. Gibson states that these affordance properties are relative and must be measured by the animal. As such, an affordance cannot be measured as it is in physics. Gibson provides this as an illustration that affordances are subjective and objective to the individual. That is, to perceive an affordance is not to classify an object. This does not discount the taxonomy given by the MAIN model for classification of features of a digital media, but rather provides a lens for further evaluation and research how affordances are perceived and assigned. This could lead towards a theory of technological affordance, rather than a predefined set of constructs that are problematic to evaluate, measure and have weak predictive power.
Chemero, A. (2003). An outline of a theory of affordances. Ecological psychology, 15(2), 181-195.
Fogg, B. J. (2003). Prominence-interpretation theory: Explaining how people assess credibility online. Paper presented at the CHI'03 extended abstracts on human factors in computing systems.
Gibson, J. J. (1977). The theory of affordances. Hilldale, USA.
Kim, H.-S., & Sundar, S. S. (2011). Using interface cues in online health community boards to change impressions and encourage user contribution. Paper presented at the Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, Vancouver, BC, Canada.
Kim, K. J., & Sundar, S. S. (2015). Mobile Persuasion: Can Screen Size and Presentation Mode Make a Difference to Trust? Human Communication Research.
Lee, J. Y., & Sundar, S. S. (2013). To Tweet or to Retweet? That Is the Question for Health Professionals on Twitter. Health Communication, 28(5), 509-524. doi: 10.1080/10410236.2012.700391
McCroskey, J., Holdridge, W., & Toomb, K. (1974). An instrument for measuring the source credibility of basic speech communication instructors. The Speech Teacher, 23(1). Retrieved from http://www.tandfonline.com/doi/citedby/10.1080/03634527409378053
Sundar, S. S. (2008). The MAIN model: A heuristic approach to understanding technology effects on credibility. In M. J. Metzger & A. J. Flanagin (Eds.), Digital media, youth, and credibility (pp. 72-100). Cambridge, MA: The MIT Press