DOCTORAL DISSERTATION
In the context of advertising design, an investigation of the stopper effect of social media applications: Instagram example.
Serdar Şenel, PhD · Bahçeşehir University · July 2023
Supervisor: Prof. Dr. Haluk Gürgen · PhD Program in Advertising and Public Relations
Design at the moment of attention
Before an advertising message can be considered, it must first be noticed. Serdar Şenel’s doctoral dissertation examines this initial encounter in a social media feed: the moment a visual interrupts scrolling and prompts a person to stop or click. The research approaches stopping power as a question of visual design and observable behavior.
Conducted in an Instagram simulation, the study investigates how color, design density and background darkness relate to participants’ selections. It connects practical decisions about advertising images with an empirical method for examining attention under brief viewing conditions.
The research questions
The dissertation asks how different color treatments and compositions affect the stopping power of social media advertising. It also examines whether preferences vary with demographic characteristics, including gender, age, education and employment status. These questions place visual form and audience characteristics within the same research setting.
A further objective is methodological: to evaluate Stopper Analyze as a tool for investigating visual advertising. Alongside the design findings, the thesis considers how a simulated feed can support research into rapid choices and what improvements are needed for subsequent studies.
A simulated feed and 460 participants
The study collected data from 460 participants through an online application accessible from phones and computers. Recruitment used convenience sampling. Participants were not randomly selected from the wider population, an important consideration when interpreting how broadly the results may apply.
Stopper Analyze combined a vertical scrolling environment with a behavioral test of visual selection. Participants encountered advertising images with different colors and design densities. The system recorded responses associated with viewing, stopping and clicking, allowing the researcher to compare the visual alternatives within a shared test format.
The automatic scrolling condition allocated approximately 1.10 seconds to each image to encourage rapid responses to visual characteristics. This was an experimental setting, not a universal measure of human attention span.
Color, contrast and visual density
The color test compared responses to alternative color treatments. A separate design test examined variations in density and background darkness through eight advertising images. The dissertation uses NPS terminology to discuss the relationship between negative and positive space, alongside the organization of the background and focal object.
A white milk bottle served as the central product image in the design comparisons. Changing the surrounding composition and background made it possible to investigate how presentation related to selection. The design therefore focused the inquiry on identifiable visual differences rather than treating each advertisement as an entirely unrelated creative execution.
What the study found
Blue, purple and red emerged as prominent choices in the color test, together accounting for roughly half of the selections. These results describe preferences among the alternatives presented in the experiment. They provide a basis for further testing rather than a prescription that the same colors will outperform others in every campaign or context.
In the eight-image design test, the low-density composition on a dark background was the most frequently selected individual option, accounting for approximately 34.8% of participants’ choices. Its clear central object and contrasting background made it a useful example for discussing visual hierarchy and figure-ground separation.
The result was not a simple victory for minimalism. Two highly dense compositions together attracted approximately 28.7% of selections. The thesis therefore draws attention to both clear, contrasting arrangements and complex visual treatments. Their performance invites closer examination of how different forms of visual distinction operate in a scrolling environment.
Reading the evidence
The analysis used frequency tables, cross-tabulations and chi-square tests to examine distributions and relationships. Several demographic comparisons of color and design choices did not identify statistically significant differences. This finding is specific to the sample and measures; it does not establish that demographic characteristics never matter in visual communication.
Stopping and clicking are also distinct from understanding, remembering or purchasing. The experiment investigates an early stage of engagement. Assessing the meaning of the message, its longer-term effects or commercial outcomes would require additional measures and research designs.
Implications for design and communication
For designers, advertisers and communication researchers, the study offers a way to investigate assumptions before turning them into creative rules. Color, contrast and density can be treated as variables to compare, with the surrounding context and the intended audience considered alongside the visual itself.
Limits and future research
The dissertation identifies several limitations: convenience sampling, the demographic composition of the participant group, the selection and ordering of images, scrolling speed and the difference between participating in a test and using social media naturally. These factors shape the interpretation of the findings.
Future directions include larger and more balanced samples, comparisons across cultural settings and further development of the testing environment. The thesis also discusses potential integration with eye tracking and AI-supported analysis. These are directions for extending the research, separate from the behavioral evidence collected in the doctoral study.
Read the doctoral dissertation
Read the complete thesis in the PDF viewer below. The Zenodo record provides a separate academic source for the dissertation.
Related research