Media Effects Research Lab - Research Archive
Capitalizing on Clickbait: Identifying the Effects of Listicle Type, Length, and Sponsorship
Student Researcher(s)
Jason Freeman (Ph.D Candidate);
Yiting Chai (Masters Candidate);
Christen Buckley (Ph.D Candidate);
Faculty Supervisor
INTRODUCTION
Entire journalistic empires have been built upon sensational headlines designed to elicit consumer attention and provide advertising partners with what they want most, eyeballs on content. The listicle, or article as numbered list, is a popular format for delivering news and information in online spaces and is considered by some to be a form of clickbait. Little empirical research has sought to test the efficacy of various listicle formats on subsequent user evaluations. Thus, to fill this gap, this study proposes the following questions: what are the psychological mechanisms behind a user’s experience with various listicle forms? Are clickable or shorter listicle better received? Does sponsorship of such content matter?
RESEARCH QUESTION / HYPOTHESES
RQ1: Controlling for age, gender, and ethnicity, what are the effects of scrollable and clickable listicles on communicative outcomes?
H1: The longer listicle will lead to greater frustration.
*H2: The clickable listicle (vs. scrollable) will lead to greater perceived ease of use which will in turn positively influence communicative outcomes.
*H3: The clickable listicle (vs. scrollable listicle) will lead to greater frustration which will in turn negatively influence communicative outcomes.
RQ2: Controlling for age, gender, and ethnicity, how do type and length together influence the communicative outcomes?
RQ3: Controlling for age, gender, and ethnicity, when the listicle is sponsored, does the perception of the sponsor moderate the influence of listicle type and length?
*Competing Hypotheses
METHOD
This study employed a 2 (listicle type) x 2 (listicle length) x 2 (presence of sponsorship) between-subject factorial experiment using the Qualtrics online survey platform. Participants were randomly assigned to browse one of eight websites representing each of the possible listicle conditions. Once participants had finished reading the assigned listicle, they answered several questions related to their reading experience. Participants were asked to list the thoughts generated while they were reading the listicle, recall the details of the listicle they read with some prompts and answer some recognition multiple-choice questions. Additionally, participants were asked to report their feelings, including their degree of frustration, interest, and how much they enjoyed the reading experience. Participants in the conditions where a sponsor was present were also asked to rate their attitudes toward the sponsor, and rate how much they perceived the sponsor as credible. As a manipulation check, participants were asked to indicate if they were aware of the sponsor, Delta Airlines, prior to reading the listicle. Finally, all the participants rated their attitudes toward the reading experience, the extent to which they thought the listicle was easy to use, and how likely they would be to share this listicle with other people.
RESULTS
A t-test was performed to answer H1, which predicted that the longer listicle would elicit greater frustration. The t-test revealed a significant relationship between listicle length and frustration such that the longer listicles (M = 3.18, SD = .15) led to more frustration than did the shorter length listicles (M = 2.62, SD = 1.37), t(380) = -3.85, p < .001.To test the hypotheses 2 and 3 and research question 1, we built a structural equation model (SEM) in AMOS
To test the indirect effects of listicle type and length on communicative outcomes, bootstrapping procedures using 2000 bootstrap samples and bias-corrected confidence intervals were employed. Results revealed several significant indirect effects of listicle type on communicative outcomes through perceived control. Specifically, the clickable listicle led to more use of pictures in information processing (ß = 01, p = .03), an increase in the number of correct recall (ß = .01, p = .05), a higher score of recognition questions (ß = .02, p = .02), higher level of enjoyment (ß = .02, p = .02), more interest (ß = .02, p = .02), more positive attitude towards the listicle (ß = .03, p = .02), and higher sharing intention (ß = .04, p = .01). However, the indirect effect of listicle type on number of thoughts generated through perceived control was not significant (ß = .00, p = .43). Therefore, hypothesis 2 was partially supported. The indirect effects of listicle type on communicative outcomes, including number of thoughts (ß = .00, p = .97), use of pictures in information processing (ß = .00, p = .95), number of correct recall (ß = .00, p = .96), score of recognition (ß = .00, p = .94), enjoyment (ß = .00, p = .97), interest (ß = .00, p = .97), attitude towards the listicle (ß = .00, p = .97), and sharing intention (ß = .00, p = .97), through frustration were all not significant, therefore, hypothesis 3 was not supported.
Research question 2 asked the combined effects of listicle type and length. Based on the analyses, it can be concluded that type and length both influence one’s cognition, feelings, and behavioral intentions. Specifically, type of listicle has significant impacts on communicative outcomes through perceived control, but not frustration. To analyze whether the pre-existing evaluations of sponsorship moderates the effects of type and length of listicles on communicative outcomes (RQ3), a MANCOVA test and several regression tests were performed with only the participants in the conditions where the sponsorship was present. No main effects of listicle type or length were found. However, results revealed that the perceived credibility of Delta Airlines significantly predicted how much participants enjoyed reading the listicle (ß = .37, t = 5.43, p <.001), how much they were interested in the listicle (ß = .35, t = 5.01, p < .001), their attitude towards the listicle (ß = .47, t = 7.19, p < .001), and how likely they would be to share this listicle with other people (ß = .32, t = 4.63, p < .001). However, the perceived sponsor credibility did not predict how many thoughts were generated while reading the listicle (ß = .14, t = 1.87, p = .06), how many recognition questions participants correctly answered (ß = .04, t = .60, p = .55), or how many details in the listicle they could recall (ß = .08, t = 1.14, p = .26). To analyze the effects of listicle type and length on whether participants used pictures to assist information processing when sponsorship was present, two chi-square tests were performed, showing no significant influence of listicle type on visual processing.
CONCLUSIONS/DISCUSSION
The scale and depth of this study yielded compelling findings that will contribute to future scholarly and industry considerations of listicles. The findings from our study establish that listicle type and length do matter. Perceived control and frustration were also established as important variables through which certain communicative outcomes can be affected. From our findings, we can also infer that clickability is more preferred for listicles and highlight the critical role of modality in online interface design. Perceived credibility of the sponsor overrode the effects of listicle type and length, indicating the importance of considering presence of sponsorship when making design decisions.
For more details regarding the study contact
Dr. S. Shyam Sundar by e-mail at sss12@psu.edu or by telephone at (814) 865-2173