3. Methods
3.1. Research Design
A mixed-methods design combined statistical data and interviews to compare responsive and fixed designs under the same conditions. This approach follows the standards of usability research [
1,
2]. To improve the study’s reliability, a triangulation strategy was used, where quantitative and qualitative data complement each other. The experimental design controlled for variables such as lighting, browser version, and internet speed to ensure that differences were due to the design type itself, rather than external factors. Furthermore, a pilot test was conducted with six participants to verify the clarity of the instructions and eliminate potential errors. The results of the pilot phase allowed us to refine the wording and adapt the tasks to the participants’ skill level.
When selecting the sample, factors such as device type, website frequency, and preferred website categories were also taken into account. This helped to obtain a more complete picture of how responsive design is perceived by different user groups. This approach enhances the external validity of the study and allows for the application of its findings to real-world settings.
The study also employed a standardization strategy for the environment. All participants used identical browser versions, and screen brightness was adjusted before testing. This reduced the influence of external factors on the results. Furthermore, participants were prohibited from using additional tools, such as voice search or built-in magnification features, to avoid skewed task completion time measurements.
3.2. Participants
Sixty university students (ages 18-30 years, average = 22.4) participated—half used laptops, and the other half used smartphones. All were frequent internet users. Participation was voluntary and unpaid.
3.3. Sampling Method
The sampling procedure had two phases. First, we identified frequent web users from the population. Second, we randomly selected participants from this group to reduce selection bias and obtain a sample that reflects typical internet users. In order to use the sampling method, which showed the balanced amount of users of laptops and smartphones in experimental groups. The first selection was with surveys with users who spend 10+ hours on the internet. This kind of selection minimizes the selection mistake and increases the validity of the investigation.
In addition to the main steps described above, the sampling procedure included additional checks to improve the representativeness of the sample. Following the initial screening, respondents were asked to complete a short demographic questionnaire, including questions about age, internet experience, device type, and frequency of visiting various website categories. This ensured that the sample truly reflects the diversity of modern users’ habits and skills.
Device diversity was also considered: participants were required to actively use at least two types of digital devices (e.g., a smartphone and a laptop), allowing for the inclusion of more flexible users. The next step involved stratified random sampling: participants were divided into categories based on their frequency of smartphone and laptop use, after which respondents were randomly selected from each category. This approach ensured an even distribution across groups and avoided any bias toward users of one particular device type.
To ensure sample quality, an additional criterion was introduced: participants had to have minimal experience with online surveys. This avoided problems related to a lack of understanding of the study structure or difficulties completing the tasks. Participants who failed the pre-screening were excluded, as their inclusion could bias the data.
Furthermore, a motivation check was used: participants were asked control questions to identify disinterested respondents or those who responded randomly. Only those participants who completed all verification stages were included in the final sample.
The use of a combined sample, consisting of random, stratified, and purposive selection elements, significantly increased the accuracy and reliability of the data obtained. This approach also ensured a balanced comparison of the groups participating in the adaptive and fixed-design comparisons, making the study results more objective and valid.
3.4. Materials and Devices
Two websites were used: one with a responsive design and the other with fixed design. Devices included a 14-inch Windows laptop (1920×1080) and a 6.5-inch Android phone (1080×2400). Data were collected using Google Chrome, OBS Studio (for screen recording), and surveys with 5-point scales.
To more accurately record interactions, an additional software module was used to record touch and click coordinates. This helped identify areas of the interface where users most frequently made errors. In addition, the Lighthouse tool was used to evaluate the technical optimization of websites, including speed, accessibility, and interface stability when resizing screens.
3.5. Procedure
Each participant:
Signed the consent form and received instructions.
Was randomly assigned a device and a version of the website.
Completed three tasks: learn about the main page, find a particular paragraph, and read it.
Filled out the survey and answered short interview questions.
Participants were also asked to verbally comment on their actions while completing the tasks. This method allowed us to capture emotional reactions, moments of confusion, and indicators of satisfaction. If participants encountered difficulties, screen recordings were analyzed to identify usability issues, such as incorrect button placement, unclear hierarchy, or excessively dense arrangement of elements.
After completing all tasks, a semi-structured interview lasting 5-7 minutes was conducted. Questions focused on usability, emotional comfort, and suggestions for improvement. All responses were later grouped into thematic categories for further analysis.
After performing the main series of tasks, participants were asked to complete an additional attention test. This included searching for hidden interface elements and completing timed tasks. This allowed us to determine how adaptive design impacts users’ ability to quickly perceive elements, even those not centered on the screen.
An important stage of the study was assessing psychoemotional responses. Participants rated their emotional state on a 7-point scale, assessing parameters such as stress level, fatigue, engagement, and comfort. Comparison of these data between groups showed that the adaptive website significantly reduced stress during task completion.
Finally, participants completed a memory task: they had to reproduce the website’s structure. Participants using the adaptive version demonstrated more accurate recall of the structure, indicating a better cognitive organization of the interface.
3.6. Data Collection and Analysis
Quantitative data (task completion time, ratings, and clicks) were analyzed with SPSS using ANOVA testing. The interviews were transcribed and coded in NVivo to identify recurring themes.