CoChat
Sign Up
Cited by 88Digital Marketing and Social MediaMarketing and Advertising StrategiesSMEs Development and Digital Marketing

Influencer Marketing on Instagram: Empirical Research on Social Media Engagement with Sponsored Posts

Open access
Gross, Jana, von Wangenheim, Florian·2022·Journal of Interactive Advertising
DOI 10.1080/15252019.2022.2123724
Open this paper in CoChat — free accountOpen PDF

Abstract

While sponsored posts by social media influencers (SMIs) on Instagram have dramatically increased as an advertising strategy, empirical results on their efficiency have yet to emerge. These sponsored posts are placed next to nonsponsored posts (NSPs) on SMIs’ Instagram profiles. In the current study, we investigate how consumers engage with sponsored posts compared to NSPs and how engagement varies when such are posted by microinfluencers compared to macroinfluencers. Moreover, we explore engagement with sponsored posts containing different advertising appeals based on the advertising strategy, which determines whether sponsored posts contain appeals based on factual and rational information (i.e., informational appeals) or emotional and personal information (i.e., transformational appeals). Using a binomial regression on a unique data set of 64,438 sponsored Instagram posts by SMIs, we show that (1) users engage more with sponsored compared to NSPs, (2) particularly by microinfluencers, and (3) users engage more with sponsored posts with informational appeals by macroinfluencers than by microinfluencers. These findings provide meaningful insights for both advertisers interested in employing influencer marketing and SMIs interested in cooperating with advertisers for sponsored posts; the findings also advance the literature on influencer advertising and marketing.

AI snapshot

Written from the paper’s full text.

TL;DR

Using a field dataset of 64,438 Instagram posts, this study finds that users engage more with sponsored posts than nonsponsored posts, especially for microinfluencers, and that informational appeals by macroinfluencers attract more engagement than those by microinfluencers. The paper reports no support for the hypothesis that informational appeals generally outperform transformational appeals.

Key findings

  • Sponsored posts received more engagement than nonsponsored posts: the SPONSOR coefficient was b = 0.167, p < 0.001 for total engagement; likes b = 0.156, p < 0.001 and comments b = 0.556, p < 0.001. Descriptive means were 1,559.21 engagements for sponsored posts versus 1,157.38 for nonsponsored posts.
  • Higher follower counts weakened engagement with sponsored posts: the FOL × SPONSOR interaction was negative and significant for total engagement (b = −0.329, p < 0.001), likes (b = −0.327, p < 0.001), and comments (b = −0.316, p < 0.001).
  • Informational appeals did not significantly increase engagement overall: INFO was positive for total engagement (b = 0.061, p < 0.05) and likes (b = 0.061, p < 0.05) but not comments (b = 0.028, p > 0.1); the paper states Hypothesis 3 was not supported. Descriptively, informational sponsored posts averaged 1,481.38 engagements versus 1,609.18 for transformational sponsored posts.
  • Higher follower counts strengthened engagement with informational appeals: FOL × INFO was positive and significant for total engagement (b = 0.187, p < 0.001), likes (b = 0.195, p < 0.001), and comments (b = 0.053, p < 0.05).

Methods

  • Field data from an influencer marketing agency focused on sports, lifestyle, and health: 64,438 Instagram posts from 2015–2019, of which 8.71% were sponsored; 39% of sponsored posts had informational appeals. The advertising-appeal analysis used a subsample of 5,617 sponsored posts.
  • Sponsorship was identified via a sponsorship-label list built from three prestudies and word detection; advertising appeal was operationalized as a binary variable (INFO = 1 for informational, 0 for transformational) using keyword detection. A qualitative prestudy interviewed six professional Instagram SMIs.
  • Main estimation used negative binomial regression because engagement was overdispersed (M = 1,192.41, VAR = 9,542,609; overdispersion test α = 1,074.95, p < 0.001), with heteroskedasticity-consistent standard errors. Robustness checks included Poisson regression, linear regression, an engagement-quantile subsample, and an alternative dependent variable weighting comments as five likes.

Limitations

  • The paper states that it did not include other relevant KPIs such as sentiment, conversion, or click rates; excluded unpaid sponsored posts; used a dichotomous informational/transformational appeal classification even though posts may contain both; limited source characteristics to follower count and did not examine SMI gender; and could not make causal conclusions about hedonic value because the prestudy was qualitative.
  • The paper also states that some posts had incomplete information due to data extraction, privacy, and security, and that a small risk of unobserved heterogeneity persists.
  • Evident from the text: the data come from one influencer marketing agency in sports, lifestyle, and health, and sponsorship labels included German and English terms, so generalizability to other agencies, content categories, or languages is not established. The design is observational field data, so the reported relationships are not causal.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1080/15252019.2022.2123724

Cites

73 of 73 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

19 of 75 cited sources checked · 8 supported · 1 flagged · accuracy not yet measured

  • Premisecitation

    Likes and comments are KPIs in influencer marketing.

    “Likes and comments are KPIs in influencer marketing (Hughes, Swaminathan, and Brooks 2019).”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites Hughes, Swaminathan, and Brooks 2019 source
  • Premisecitation

    In low-involvement environments, users put a greater emphasis on peripheral cues, such as the number of followers.

    “In low-involvement environments, users put a greater emphasis on peripheral cues, such as the number of followers (Hughes, Swaminathan, and Brooks 2019).”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites Hughes, Swaminathan, and Brooks 2019 source
  • Premisecitation

    SMIs have control of the ultimate advertising message to be communicated in sponsored posts.

    “SMIs have control of the ultimate advertising message to be communicated in sponsored posts (Hughes, Swaminathan, and Brooks 2019).”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Hughes, Swaminathan, and Brooks 2019 source
  • Premisecitation

    Post characteristics have been found to affect engagement.

    “In line with this, post characteristics have been found to affect engagement (De Vries, Gensler, and Leeflang 2017; Sabate et al. 2014).”
    Quote in the paperSubstantial claim1 of 2 sources checked
    • cites De Vries, Gensler, and Leeflang 2017 source

      The available passages from our copy are bibliographic material and keywords (labeled Abstract, no page number), not findings about post characteristics or engagement. Searches for those terms and for likes and comments found no passage addressing the claim, so this copy cannot verify it. (LLM, conf=0.70)

      Experimental — accuracy not yet measured
    • cites Sabate et al. 2014 source
  • Findingown data

    The positive relationship between sponsored posts (vs. nonsponsored posts) and social media engagement is weaker for SMIs with more followers (macroinfluencers) than for those with fewer followers (microinfluencers).

    “(2) particularly by microinfluencers”
    Quote in the paperSubstantial claim
  • Classified as probably not a substantial claim. 54
  • May combine two claims. 45
  • The quote only partly states this claim. 1

  • Classified as probably not a substantial claim.

    Premisecitation

    Facebook influencer campaigns that included incentives significantly increased the number of post comments.

    “Hughes, Swaminathan, and Brooks (2019) found that Facebook influencer campaigns that included incentives significantly increased the number of post comments.”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites Hughes, Swaminathan, and Brooks (2019) source

      The Dependent Variables section (p. 17) measures Facebook engagement by post likes, not comments. The Model Results (p. 23) report that campaigns with giveaways significantly decreased Facebook post likes; the paper instead associates incentives with increased engagement on blogs (Model Results, p. 22). (LLM, conf=0.89)

      Checked by a reasoning model — may be wrong.

      Experimental — accuracy not yet measured

  • Classified as probably not a substantial claim.

    Premisecitation

    Previous research suggests that hedonic value can have an impact on post interaction and popularity.

    “Previous research suggest that hedonic value can have an impact on post interaction and popularity (De Vries, Gensler, and Leeflang 2012;Hughes, Swaminathan, and Brooks 2019;Sabate et al. 2014).”
    Quote in the paperSubstantial claim2 of 3 sources checked
    • cites De Vries, Gensler, and Leeflang 2012 source

      The available passages labeled Abstract contain bibliographic metadata and the paper’s title, but no finding about hedonic value or its effect on post interaction or popularity. Searches for “hedonic,” “entertainment,” and likes and comments found no passage addressing that relationship, so the claim cannot be verified from our copy. (LLM, conf=0.70)

      Experimental — accuracy not yet measured
    • cites Hughes, Swaminathan, and Brooks 2019 source
    • cites Sabate et al. 2014 source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Brand posts with higher level of vividness are more popular.

    “De Vries, Gensler, and Leeflang (2012) found that brand posts with higher level of vividness are more popular.”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites De Vries, Gensler, and Leeflang (2012) source

      The supplied passages, labeled Abstract, contain citation metadata but no finding about vividness. Searches of our copy for “vividness,” “higher level of vividness,” and related terms returned no passage addressing whether more vivid brand posts are more popular. (LLM, conf=0.75)

      Experimental — accuracy not yet measured

  • Probably not a substantial claim: may not pass S2 (checkable).

    Premisecitation

    Social media engagement is an important key performance indicator (KPI) to measure the success of sponsored posts on social media.

    “social media engagement is an important key performance indicator (KPI) to measure the success of sponsored posts on social media (Hughes, Swaminathan, and Brooks 2019;Lammenett 2019).”
    Quote in the paperNot a substantial claim1 of 2 sources checked
    • cites Hughes, Swaminathan, and Brooks 2019 source
    • cites Lammenett 2019 source

  • Classified as probably not a substantial claim. May combine two claims.

    Premisecitation

    Transformational appeals positively affect brand liking, brand knowledge, brand attitude, and word of mouth.

    “transformational appeals have been found to positively affect brand liking, brand knowledge, brand attitude, and word of mouth (WOM) (Naylor et al. 2008;Teichert, Hardeck, Liu, and Trivedi 2018).”
    Quote in the paperSubstantial claim1 of 2 sources checked
    • cites Naylor et al. 2008 source
    • cites Teichert, Hardeck, Liu, and Trivedi 2018 source

      The Hypothesis section reports that emotional appeals increase brand liking, and the Overall framework section reports stronger effects on positive attitude change and information-search intention. It treats information search as an indicator of knowledge accumulation, not a measured increase in brand knowledge. The paper does not report an effect on word of mouth. (LLM, conf=0.87)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Premisecitation

    The number of followers is an indicator of popularity and reflects an SMI's network and group size.

    “The number of followers is an indicator of popularity and reflects an SMI's network and group size (De Veirman, Cauberghe, and Hudders 2017).”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites De Veirman, Cauberghe, and Hudders 2017 source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Internet search and social media activities are higher during the week than on weekends.

    “This is in line with previous research that showed Internet search and social media activities are higher during the week than on weekends (De Vries, Gensler, and Leeflang 2012; Rutz and Bucklin 2011).”
    Quote in the paperNot a substantial claim2 of 2 sources checked
    • cites De Vries, Gensler, and Leeflang 2012 source

      The available passages labeled Abstract contain citation metadata and social-media keywords, but no finding about weekday versus weekend activity. Searches for weekends, weekdays, day of the week, and the citing sentence found no passage addressing that comparison. (LLM, conf=0.72)

      Experimental — accuracy not yet measured
    • cites Rutz and Bucklin 2011 source

      The “Indicator variables for day of week and month” section reports lower weekend search activity, with a negative Saturday effect and a positive Monday effect. The “Test for Reverse Spillover” section likewise describes a steep drop in search activity beginning Friday. These passages support the Internet-search portion of the claim, but not the social-media portion. (LLM, conf=0.88)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Premisecitation

    Social media platforms represent low-involvement, high-distraction, and hedonic environments.

    “Social media platforms represent low-involvement, high-distraction, and hedonic environments (Hughes, Swaminathan, and Brooks 2019; Schulze, Sch€ oler, and Skiera 2014).”
    Quote in the paperSubstantial claim1 of 2 sources checked
    • cites Hughes, Swaminathan, and Brooks 2019 source

      The paper describes Facebook as a low-involvement, high-distraction platform (Influencer Marketing on the Facebook Platform, pp. 14–15) and discusses the value of hedonic content there. It does not characterize social media platforms generally as low-involvement, high-distraction, and hedonic: its comparison finds blogs more involving and less distracting than Facebook (Influencer Marketing on the Blogging Platform, pp. 8–9). (LLM, conf=0.88)

      Experimental — accuracy not yet measured
    • cites Schulze, Sch€ oler, and Skiera 2014

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Classified as probably not a substantial claim. May combine two claims.

    Premisecitation

    Advertising appeals are not mutually exclusive categories of advertising, but the weight of informational and transformational appeals in advertising messages varies.

    “While advertising appeals are not mutually exclusive categories of advertising, the weight of informational and transformational appeals in advertising messages varies (Teichert et al. 2018).”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites Teichert et al. 2018 source

  • Classified as probably not a substantial claim.

    Premisecitation

    Social media activities and engagement might vary between weekend and weekdays.

    “First, social media activities and engagement might vary between weekend and weekdays.”
    Quote in the paperSubstantial claim2 of 2 sources checked
    • cites De Vries, Gensler, and Leeflang 2012 source

      The available passages labeled Abstract contain bibliographic information and keywords, not findings about differences between weekends and weekdays. Searches for weekend, weekday, and posting-time effects found no passage in our copy addressing the claim. (LLM, conf=0.74)

      Experimental — accuracy not yet measured
    • cites Rutz and Bucklin 2011 source

      The paper’s “Indicator variables for day of week and month” section reports lower weekend search activity, with negative Saturday and positive Monday effects. This supports a weekday–weekend difference in paid-search activity, but the paper does not establish that social media activity or engagement varies in the same way. (LLM, conf=0.84)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Premisecitation

    People in larger groups feel less responsible and addressed and then stand by rather than act.

    “Despite macroinfluencers' sponsored posts being exposed to a larger number of users (De Veirman, Cauberghe, and Hudders 2017), people in larger groups feel less responsible and addressed and then stand by rather than act (Darley and Latan e 1968).”
    Quote in the paperSubstantial claim1 of 2 sources checked
    • cites Darley and Latan e 1968 source
    • cites De Veirman, Cauberghe, and Hudders 2017 source

  • Classified as probably not a substantial claim.

    Premisecitation

    Vivid content characteristics receive more likes and comments.

    “Indeed, research has found that vivid content characteristics receive more likes and comments (De Vries, Gensler, and Leeflang 2012).”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites De Vries, Gensler, and Leeflang 2012 source

      The indexed passage (Abstract; page not given) says vivid brand-post characteristics enhance the number of likes. It attributes an increase in comments to an interactive characteristic—a question—not to vividness, so the comments part of the claim is unsupported. (LLM, conf=0.86)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Findingown data

    Users engage more with sponsored posts by microinfluencers, particularly when the sponsored posts use transformational appeals.

    “We provide first findings that users engage more with sponsored posts by microinfluencers, particularly when they use transformational appeals.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingreasoning

    SMIs feel creatively inspired and challenged to embed the advertising appeal in a credible and compelling manner into their posts.

    “SMIs feel creatively inspired and challenged to embed the advertising appeal in a credible and compelling manner into their posts.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Findingown data

    Agencies provide several labels to SMIs, show examples of how to label posts, and recommend doing so at the beginning of the caption.

    “Agencies provide several labels to SMIs, show examples of how to label posts, and recommend doing so at the beginning of the caption.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: Sponsorships are labeled in the captions of Instagram posts.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Findingown data

    All posts had at least one like or comment, which led to excluding the zero-inflated regression model.

    “All posts had at least one like or comment, which led us to exclude this regression estimation.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    In a linear regression with square-root-transformed dependent variable, results remained the same in direction and significance, except for the coefficient of INFO, which was negative and insignificant.

    “The results remained the same in direction and significance, except for the coefficient of INFO, which was negative and insignificant.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Significance and direction of coefficients remained unchanged using a Poisson regression.

    “Nonetheless, significance and direction of coefficients remained unchanged using a Poisson regression.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Sponsored posts receive more likes, comments, and engagement than nonsponsored posts.

    “SPONSOR | 0.102 *** | 0.521 *** | 0.112 ***”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    The square root transformation reduced the dependent variable's skewness from 12.05 to 3.06.

    “We used a square root transformation to reduce the dependent variable's skewness from 12.05 to 3.06.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    A Breusch-Pagan test showed heteroskedasticity is present in the main sample of 64,438 posts (BP = 4,767.4, DF = 13, p < 0.001) and in the subsample of 5,617 sponsored posts (BP = 255.52, DF = 13, p < 0.001).

    “To test heteroscedasticity, we conducted a Breusch-Pagan test (Breusch and Pagan 1979) which showed that heteroskedasticity is present in the main sample of 64,438 posts (BP ¼ 4,767.4,DF ¼ 13, p < 0.001) and in the subsample of 5,617 sponsored posts (BP ¼ 255.52,DF ¼ 13, p < 0.001).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Breusch and Pagan 1979 source

  • May combine two claims.

    Findingown data

    There was no evidence that users significantly engaged more with sponsored posts with informational appeals; hypothesis 3 was not supported.

    “There was no evidence that users significantly engaged more with sponsored posts with informational appeals; hypothesis 3 was not supported.”
    • Also stated in: § Discussion · plain
      “We do not find any empirical evidence that the type of advertising appeal in sponsored posts affects social media engagement.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Sponsored posts received more likes (b = 0.156, p < 0.001) and more comments (b = 0.556, p < 0.001) than nonsponsored ones.

    “Sponsored posts received more likes (b ¼ 0.156, p < 0.001) and more comments (b ¼ 0.556, p < 0.001) than nonsponsored ones.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The interaction coefficient FOL × INFO was positive and highly significant for number of likes (b = 0.195, p < 0.001) and positive for number of comments, with a significance level of 5% (b = 0.053).

    “The interaction coefficient (FOL Ã INFO) was positive and highly significant for number of likes (b ¼ 0.195, p < 0.001) and positive for number of comments, with a significance level of 5% (b ¼ 0.053).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Users engage more with sponsored posts than NSPs, particularly for those posts by microinfluencers.

    “Our findings show that users engage more with sponsored posts than NSPs, particularly for those posts by microinfluencers.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    A higher number of followers negatively affected the number of likes (b = -0.327, p < 0.001) and comments (b = -0.316, p < 0.001).

    “The results also showed that a higher number of followers negatively affected the number of likes (b ¼ À0.327, p < 0.001) and comments (b ¼ À0.316, p < 0.001).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The link between informational appeals and number of likes was positive and significant (b = 0.061, p < 0.05), and the link between informational appeal and number of comments was positive and insignificant (b = 0.028, p > 0.1).

    “The link between informational appeals and number of likes was positive and significant (b ¼ 0.061, p < 0.05), and the link between informational appeal and number of comments was positive and insignificant (b ¼ 0.028, p > 0.1).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The data contained 8.71% sponsored posts; of these, 39% had informational appeals.

    “The data contained 8.71% sponsored posts; of these, 39% had informational appeals.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: Sponsorships are labeled in the captions of Instagram posts.

  • The quote only partly states this claim.

    Findingown data

    In a negative binomial regression on a subsample excluding posts in the upper and lower 20% engagement quantiles, estimation results supported the hypotheses.

    “Again, the estimation results supported our hypotheses.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Premisecitation

    Entertaining brand posts positively affect likes and comments of a post more than nonentertaining brand posts do.

    “Luarn, Lin, and Chiu (2015) showed that entertaining brand posts positively affect likes and comments of a post more than nonentertaining brand posts do.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Luarn, Lin, and Chiu (2015) source

  • Classified as probably not a substantial claim.

    Premisecitation

    The informational value of influencer-generated content positively affects trust in it.

    “Lou and Yuan (2019) found that the informational value of influencer-generated content positively affected trust in it.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Lou and Yuan (2019) source

  • Classified as probably not a substantial claim.

    Premisecitation

    Number of likes is significantly positively affected by calls to action in brand messages.

    “Schultz (2017) found that number of likes is significantly positively affected by calls to action in brand messages.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Schultz (2017) source

  • Classified as probably not a substantial claim.

    Premisecitation

    Entertainment is a main motivation for contributing to brand-related content.

    “Muntinga, Moorman, and Smit (2011) identified entertainment as a main motivation for contributing to brand-related content.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Muntinga, Moorman, and Smit (2011) source

  • May combine two claims.

    Premisecitation

    Both informational and transformational appeals positively affect advertising attitude, purchase intentions, and consumer experience.

    “Both appeals positively affect advertising attitude, purchase intentions, and consumer experience (Akbari 2015;Naylor et al. 2008)”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites Akbari 2015 source
    • cites Naylor et al. 2008 source

  • Classified as probably not a substantial claim.

    Premisecitation

    Vividness positively affects interaction with brand posts.

    “Sabate et al. (2014) found that vividness positively affects interaction with brand posts.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Sabate et al. (2014) source

  • May combine two claims.

    Premisecitation

    Likes and comments are significant predictors of companies' financial and social performance, and they positively influence sales.

    “likes and comments are significant predictors of companies' financial and social performance, and they positively influence sales (Lee, Lee, and Oh 2015;Yoon et al. 2018)”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites Lee, Lee, and Oh 2015 source
    • cites Yoon et al. 2018 source

  • May combine two claims.

    Premisecitation

    When users pay less attention to advertising, they put less mental effort into evaluating the message, and they rely more on heuristics, signals, and cues.

    “When users pay less attention to advertising, they put less mental effort into evaluating the message, and they rely more on heuristics, signals, and cues (Petty, Cacioppo, and Heesacker 1981;Tellis 2003).”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites Petty, Cacioppo, and Heesacker 1981 source
    • cites Tellis 2003

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • May combine two claims.

    Premisecitation

    Likes and comments positively affect brand trust, involvement, and overall company reputation.

    “beyond financial success, they positively affect brand trust, involvement, and overall company reputation (Ji et al. 2017;Phua and Ahn 2016).”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites Ji et al. 2017

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

    • cites Phua and Ahn 2016

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductreasoning

    The model incorporated covariates to rule out alternative explanations of the modeling results.

    “Besides the explanatory variables, we incorporated a few covariates into our model to rule out alternative explanations of the modeling results.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim. May combine two claims.

    Premisecitation

    People in larger groups feel less responsible and less personally addressed, which creates a bystander effect.

    “Darley and Latan e (1968) showed that people in larger groups feel less responsible and less personally addressed, which creates a bystander effect.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Darley and Latan e (1968) source

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

    Interpretationreasoning

    Companies can guide SMIs into using either informational or transformational appeals when advertising brands and products.

    “They can guide SMIs into using either informational or transformational appeals when advertising brands and products.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation

    Both types of appeals can occur in the same advertising message, but one of them will dominate.

    “Both types of appeals can occur in the same advertising message, but one of them will dominate (Cutler, Thomas, and Rao 2008).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Cutler, Thomas, and Rao 2008

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    While both types of messages can occur in the same post, one of the approaches will dominate.

    “While both types of messages can occur in the same post, one of the approaches will dominate (Cutler, Thomas, and Rao 2008) in practice.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Cutler, Thomas, and Rao 2008

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Classified as probably not a substantial claim. May combine two claims.

    Premisecitation

    According to Puto and Wells (1984), for an ad to be informational it must present factual, relevant information about the brand; present information immediately and obviously important to the potential consumer; and present data the consumer accepts as verifiable.

    “According to Puto and Wells (1984), for an ad to be informational it must contain the following characteristics: (1) It must “present factual, relevant information about the brand,” (2) “present information which is immediately and obviously important to the potential consumer,” and (3) “present data which the consumer accepts as being verifiable.””
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Puto and Wells (1984)

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Probably not a substantial claim: may not pass S4 (based).

    Interpretationreasoning

    A dichotomous classification of the advertising appeal is in line with practices in the influencer industry.

    “A dichotomous classification of the advertising appeal is in line with practices in the influencer industry.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationreasoning

    The discount code used in Figure 1 contains all three characteristics of informational advertising.

    “The discount code used in the caption presented on Figure 1 contains all three characteristics.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Message length may either positively or negatively affect outcome measures such as click-through rates or engagement.

    “Moreover, previous research found that message length may either positively or negatively affect outcome measures such as click-through rates or engagement (Baltas 2003;Robinson et al. 2007).”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites Baltas 2003 source
    • cites Robinson et al. 2007 source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Companies differentiate two major types of sponsored posts: those aiming to increase long-term consumer actions and those aiming to increase short-term consumer actions.

    “Companies differentiate two major types of sponsored posts: Those that aim to increase long-term consumer actions (e.g., brand awareness, loyalty, and image) and those that aim to increase short-term consumer actions (e.g., sales, leads, and traffic).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Funke 2019

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    In practice, one advertising appeal is more dominant than the other.

    “In line with this, Cutler, Thomas, and Rao (2008) argued that, in practice, one appeal is more dominant than the other.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Cutler, Thomas, and Rao (2008)

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Classified as probably not a substantial claim.

    Premisecitation

    Transformational advertising is affect based rather than cognitive based.

    “Second, Puto and Wells (1984) state that transformational advertising is affect based rather than cognitive based.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Puto and Wells (1984)

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • May combine two claims.

    Premisecitation

    Users recognize SMIs' posts as sponsored and willingly engage with them.

    “users recognize SMIs' posts as sponsored and willingly engage with them (Evans et al. 2017).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Evans et al. 2017 source

  • May combine two claims.

    Premisecitation

    Sponsored posts by macroinfluencers are perceived as less authentic and more biased, while those by microinfluencers are perceived as more authentic and honest.

    “While sponsored posts by macroinfluencers are perceived as less authentic and more biased, those by microinfluencers are perceived as more authentic and honest (Gross 2020).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Gross 2020 source

  • May combine two claims.

    Premisecitation

    When users pay more attention to advertising, they put more mental effort into evaluating the message, and they do so more carefully.

    “When users pay more attention to advertising, they put more mental effort into evaluating the message, and they do so more carefully (Cacioppo and Petty 1979;Petty, Cacioppo, and Heesacker 1981;Petty, Cacioppo, and Schumann 1983).”
    Quote in the paperSubstantial claim0 of 3 sources checked
    • cites Cacioppo and Petty 1979 source
    • cites Petty, Cacioppo, and Heesacker 1981 source
    • cites Petty, Cacioppo, and Schumann 1983 source

  • Classified as probably not a substantial claim.

    Premisecitation

    Informational appeals dominate when the advertising is focusing on brands and products themselves.

    “Informational appeals dominate when the advertising is focusing on brands and products themselves (Puto and Wells 1984).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Puto and Wells 1984

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationreasoning

    Promotional incentives such as discount codes represent informational appeals in sponsored posts to consumers.

    “Hence, these promotional incentives represent informational appeals in sponsored posts to consumers.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim. May combine two claims.

    Interpretationreasoning

    Users may feel less responsible and addressed by macroinfluencers' sponsored posts compared to those by microinfluencers, resulting in a looking effect.

    “Users may feel less responsible and addressed by macroinfluencers' sponsored posts compared to those by microinfluencers, resulting in a looking effect.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Premisecitation

    Transformational appeals dominate when the advertising is focusing on the experience of brands and products.

    “In contrast, transformational appeals dominate when the advertising is focusing on the experience of brands and products (Puto and Wells 1984).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Puto and Wells 1984

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Probably not a substantial claim: may not pass S4 (based).

    Interpretationreasoning

    SMIs either use informational or transformational appeals in their advertising messages in sponsored posts.

    “Therefore, SMIs either use informational or transformational appeals in their advertising messages in sponsored posts.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable).

    Interpretationreasoning

    Informational appeals should play an important role in eliciting social media engagement between micro- and macroinfluencers.

    “this implies that informational appeals should play an important role in eliciting social media engagement between micro-and macroinfluencers.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Interpretationreasoning

    In the study context, informational and transformational appeals are used independently.

    “In our study context, informational and transformational appeals are used independently.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S4 (based). May combine two claims.

    Premisenone given

    Sponsored posts with informational appeals contain incentives that immediately benefit users, whereas sponsored posts with transformational appeals do not.

    “Sponsored posts with informational appeals contain incentives that immediately benefit users.This is not the case for sponsored posts with transformational appeals.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S4 (based).

    Interpretationnone given

    SMIs spend more time, effort, and thought to produce sponsored posts because they receive something in return from the company.

    “SMIs spend more time, effort, and thought to produce sponsored posts because they receive something in return from the company.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Study design and conductnone given

    The authors demonstrated the reliability of the measurements and robustness of the abovementioned effects.

    “We demonstrated the reliability of the measurements and robustness of the abovementioned effects.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationown data

    The estimates from a set of robustness checks offered further empirical evidence on the proposed econometric model.

    “In sum, the estimates from a set of robustness checks offered further empirical evidence on the proposed econometric model.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationreasoning

    SMIs spend more time, effort, and thought in sponsored posts, resulting in richer content with higher hedonic value for social media users.

    “They spend more time, effort, and thought in sponsored posts, resulting in richer content with higher hedonic value for social media users.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    Comments allow companies to evaluate whether and how users perceive the advertised brands and products in a post, while likes do not.

    “Comments allow companies to evaluate whether and how users perceive the advertised brands and products in a post, while likes do not (Lammenett 2019).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Lammenett 2019 source

  • May combine two claims.

    Interpretationreasoning

    Consumers receive higher hedonic value from consuming sponsored posts, which positively affects social media engagement.

    “Our results suggest that consumers receive higher hedonic value from consuming sponsored posts, which positively affects social media engagement.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    A comment has the same value as seven likes.

    “Moreover, Kim and Yang (2017) stated that a comment has the same value as seven likes.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Kim and Yang (2017) source

  • May combine two claims.

    Interpretationreasoning

    Users evaluate advertising messages in sponsored posts followed by more people as more interesting and relevant, and thus pay more attention to it compared to microinfluencers.

    “We argue that users evaluate advertising messages in sponsored posts followed by more people as more interesting and relevant, and thus pay more attention to it compared to microinfluencers.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Premisecitation

    Writing a comment requires more effort than clicking the Like button.

    “Yoon et al. (2018) argued that writing a comment requires more effort than clicking the Like button.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Yoon et al. (2018) source

  • May combine two claims.

    Interpretationown data

    Both content and SMI characteristics are indispensable for effective influencer marketing.

    “While previous research focuses on SMI characteristics influencing effectiveness (Erdogan 1999;Ohanian 1990), the current study suggests that both content and SMI characteristics are indispensable for effective influencer marketing.”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites Erdogan 1999 source
    • cites Ohanian 1990 source

  • Classified as probably not a substantial claim.

    Study design and conductnone given

    The authors carefully prepared, cleaned, and tested the data to provide a reliable data set of highest quality.

    “Nonetheless, we carefully prepared, cleaned, and tested the data to provide a reliable data set of highest quality.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim. May combine two claims.

    Study design and conductreasoning

    The prestudy findings are based on a qualitative study, which does not allow causal conclusions.

    “However, these findings are based on a qualitative study, which does not allow us to make causal conclusions.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S1 (asserted), S3 (load-bearing). May combine two claims.

    Interpretationcitation

    Users interact more with posts that contain dynamic content compared to static content because dynamic content triggers more emotional reactions.

    “It might be that users interact more with posts that contain dynamic content, such as videos, compared to static content, such as images, because they trigger more emotional reactions.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Premisecitation

    Users may find posts more attractive depending on type of content.

    “Third, users may find posts more attractive depending on type of content.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S1 (asserted), S3 (load-bearing). May combine two claims.

    Premisecitation

    Female users might be more likely to build parasocial interaction with female social media influencers, while male users might be more likely to build parasocial interaction with male social media influencers.

    “Female users might be more likely to build parasocial interaction with female SMIs, while male users might be more likely to build parasocial interaction with male SMIs (Hudders and Jans 2022).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Hudders and Jans 2022 source

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Interpretationreasoning

    Users give first or even more attention to the SMIs' follower counts rather than the type of advertising appeal when evaluating sponsored posts.

    “One potential explanation for this finding might be that users give first or even more attention to the SMIs' follower counts rather than the type of advertising appeal when evaluating sponsored posts.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationreasoning

    Users evaluate macroinfluencers' sponsored posts with informational appeals more carefully compared to those by microinfluencers, which results in higher engagement.

    “Hence, users evaluate macroinfluencers' sponsored posts with informational appeals more carefully compared to those by microinfluencers, which results in higher engagement.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationreasoning

    Engagement with sponsored posts decreases with a higher number of followers because users passively engage rather than actively engage with macroinfluencers' sponsored posts.

    “It is not surprising to observe that engagement with sponsored posts decreases with a higher number of followers; this may be because users passively engage (i.e., looking effect) rather than actively engage (i.e., liking and commenting) with macroinfluencers' sponsored posts.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationreasoning

    When exposed to sponsored posts, users might first evaluate the source, which then influences how they evaluate advertising messages in sponsored posts.

    “When being exposed to sponsored posts, users might first evaluate the source (i.e., macroinfluencers versus microinfluencers), which then influences how they evaluate advertising messages in sponsored posts.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

    Study design and conductnone given

    A small risk of unobserved heterogeneity persists.

    “While we are confident that the results remain unaffected, a small risk of unobserved heterogeneity persists.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductown data

    Some posts included incomplete information due to data extraction, data privacy, and security.

    “Some posts included incomplete information due to data extraction, data privacy, and security.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Sponsored posts might contain both informational and transformational appeals.

    “However, sponsored posts might contain both types of appeals (Puto and Wells 1984).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Puto and Wells 1984

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductcitation

    The gender of the social media influencer is a characteristic that might confound results.

    “However, gender of the SMIs is another characteristic that might confound results.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationreasoning

    Both content and SMIs' characteristics explain how social media users engage with sponsored posts.

    “From a theoretical perspective, our study implies that both content and SMIs' characteristics explain how social media users engage with sponsored posts.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationown data

    Findings from the prestudy suggest that heightened hedonic value in sponsored posts might be a potential explanation for higher social media engagement compared to nonsponsored posts.

    “Findings from our prestudy suggest that heightened hedonic value in sponsored posts might be a potential explanation for higher social media engagement compared to NSPs.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Study design and conductown data

    Heteroskedasticity-consistent standard errors were reported in all models.

    “Thus, we reported heteroskedasticity-consistent standard errors in all models.”
    Quote in the paperSubstantial claim
87 of 87

  • The authors state this as an assumption.

    Users positively evaluate informational appeals due to the informational value in sponsored posts, which increases engagement compared to transformational appeals.

    “It is reasonable to assume that users positively evaluate informational appeals due to the informational value in sponsored posts which, in turn, increases engagement compared to those with transformational appeals.”

  • The authors state this as an assumption.

    Users experience higher hedonic value from sponsored posts, which results in more social media engagement, compared to nonsponsored posts.

    “we assume that users experience higher hedonic value from sponsored posts, which results in more social media engagement, compared to NSPs.”

  • The paper relies on this without support.

    Sponsorships are labeled in the captions of Instagram posts.

    “Sponsorships are labeled in the captions of Instagram posts.”

  • The paper relies on this without support.

    There are no strict rules or fixed regulations regarding the labels that must be used to disclose sponsorship.

    “While labeling sponsored posts is mandatory, there are no strict rules or fixed regulations regarding the labels that must be used to disclose sponsorship.”

  • The paper relies on this without support.

    Total engagement (the sum of likes and comments) serves as a pricing scheme for companies and SMIs.

    “We use total engagement as DV because the total number of likes and comments serves as a pricing scheme for companies and SMIs.”

  • The paper relies on this without support.

    SMIs are paid up to five times as much for a comment than for a like.

    “Indeed, SMIs are paid up to five times as much for a comment than for a like.”

  • The paper relies on this without support.

    Instagram posts allow for captions of a maximum of 2,200 characteristics.

    “Second, Instagram posts allow for captions of a maximum of 2,200 characteristics.”

  • The paper relies on this without support.

    This study is among the first in understanding the determinants affecting social media engagement with sponsored posts on Instagram.

    “This study is among the first in understanding the determinants affecting social media engagement with sponsored posts on Instagram.”
8 of 8

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

  • Significant appeal effect is described as unsupported

    Major

    The results report a positive effect of informational appeals on engagement significant at the 5% level, then say there was no evidence of significantly greater engagement and reject the corresponding hypothesis.

  • Total sample is described as entirely sponsored

    Minor

    The abstract calls all 64,438 posts sponsored, but the methods say only 5,617 posts in that sample were sponsored.

2 of 2

Cites

Flagged first, then by citations

Research works printed but not linked: 6

  1. Brewster M., 2020, Pivoting for the Pandemic, 77
  2. Cacioppo J. T., 1984, ACR North American Advances, 673
  3. Childers C. C., 2018, Journal of Current Issues & Research in Advertising, 2018, 1
  4. Funke S.-O., 2019, Influencer-Marketing: Strategie, Briefing, Monitoring
  5. Rossiter J. R., 2018, Marketing Communication: Objectives, Strategy, Tactics
  6. Wojdynski B. W., 2016, In the New Advertising: Branding, Content and Consumer Relationships in a Data_Driven Social Media Era, 2016, 203

Cited by

Sorted by citations

Related

Authors
Gross, J., & von Wangenheim, F.
Affiliations
  • Kedge Business School (FR)
  • ETH Zurich, Zurich, Switzerland
Corresponding author
Gross, Jana
ORCID
Venue
Journal of Interactive Advertising
Publisher
Informa UK Limited
Published
2022-09-02
Volume
22
Issue
3
Pages
289-310
ISSN
1525-2019 (ISSN-L)
DOI
10.1080/15252019.2022.2123724
OpenAlex
W4306411632
Semantic Scholar
c0962905d9c89260702bc0c7f484c002c406144e
Cited by
88
References
70
Citations by year
2026: 18 · 2025: 31 · 2024: 27 · 2023: 10 · 2022: 2
FWCI
25.09
Type
journal article
Language
en
Open access
Licence
Locations
  • Journal of Interactive Advertising (publishedVersion, cc-by-nc-nd)
  • Repository for Publications and Research Data (ETH Zurich) (submittedVersion, cc-by-nc-nd)
  • ETH Zürich Research Collection
Topics
  • Social Sciences › Sociology and Political Science › Digital Marketing and Social Media
  • Social Sciences › Business, Management and Accounting › Marketing › Marketing and Advertising Strategies
  • Social Sciences › Demography › SMEs Development and Digital Marketing
Keywords
influencer marketing, Instagram, sponsored posts, social media engagement, micro-influencers, advertising appeals, informational appeals, influencer advertising, customer engagement

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-14.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-14.

Cited sources are checked by gpt-6-sol.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review