What the study analysed
The study, published in the IEEE ACDSA 2026 proceedings, examined AI literacy, attitude toward AI, and perception of employment prospects among final-year undergraduates. A sequential mixed-methods approach ran qualitative interviews first, then a structured questionnaire measured the same three constructs on a six-point Likert scale.
Interviews opened with demographics and AI coursework, then moved to open-ended questions on experience with AI tools, perceived competence of those tools, applicability of AI knowledge in the job market, and optimism about employment prospects. All were audio recorded and transcribed verbatim.
Two methods, one model
The interview data went into Evidano for thematic interpretation. The survey data went into SmartPLS 4 for variance-based structural equation modelling, run as a pilot test of the conceptual model before the study expands to a larger sample.
The thematic analysis surfaced a pattern the quantitative model could then be read against: AI literacy shapes attitude toward AI, which in turn shapes how employment prospects are perceived. Respondents felt positive about AI enhancing their prospects while voicing uncertainty about job security.
How Evidano fit the workflow
A qualitative strand used to add richness to the numbers.
- 1Interviews recorded and transcribed verbatim
Open-ended questions covering AI experience, competence, applicability, and job-market outlook.
- 2Transcripts entered into Evidano
Thematic interpretation of the interview data.
- 3Survey modelled in SmartPLS 4
Measurement model checked for reliability and validity, with structural paths explored as a pilot.
- 4Strands triangulated
The qualitative themes were used as triangulation to add richness to the quantitative findings.
“The qualitative interview data were entered into AILYZE and were interpreted thematically, which was used as triangulation to add richness to the quantitative data.”
“The pilot test results show that the identified constructs exhibit reliability and validity, showing potential relationships on the path proposed in the conceptual framework, and the respondents felt positive about AI in enhancing employment prospects, yet expressed simultaneous uncertainty about job security.”
Mixed methods without the bottleneck
The qualitative strand is usually the one that slips, so mixed-methods designs often lean on the survey and treat interviews as colour.
Themes come back fast enough that the qualitative strand can actually triangulate the model rather than trail behind it.
Made for transcripts, focus groups, and open-ended responses.
Used alongside SmartPLS 4 in a sequential design, not in place of it.
The paper names the tool and the role it played.
Published by IEEE in the ACDSA 2026 proceedings, available on IEEE Xplore.