GenAI-Enabled accounting information systems competence: Antecedents, perceived usefulness, AI ethics awareness, and career readiness outcomes
Data files
Sep 18, 2026 version files 74.43 KB
-
Analysis_Data_471.xlsx
65.15 KB
-
README.md
9.29 KB
Abstract
This study examines GenAI-enabled accounting information systems competence (GAISC) among accounting and auditing students and its relationships with GenAI readiness (GAR), perceived ease of use (PEOU), digital learning self-efficacy (DLSE), facilitating conditions (FC), perceived usefulness (PU), AI ethics awareness (AIEA), and career readiness (CR). The model is anchored in Socio-Technical Systems Theory (STS) and Social Cognitive Career Theory (SCCT), with technology readiness, TAM, self-efficacy, and facilitating-condition logic used as construct-level underpinning mechanisms. Survey data from 471 students at five Vietnamese universities were analyzed using PLS-SEM, complemented by a common latent factor diagnostic and Gaussian copula endogeneity assessment. The baseline PLS-SEM model showed statistically significant positive associations for all nine hypothesized relationships and explained 53.1% of the variance in GAISC, 45.6% in AIEA, and 52.9% in CR. Substantively, PU was the strongest baseline antecedent of GAISC (β = .427, f2 = .244), while GAISC showed a particularly strong association with AIEA (β = .675, f2 = .838). Endogeneity adjustment indicated that seven relationships remained robust; however, the relationships between FC and GAISC and between AIEA and CR became non-significant. Accordingly, these two paths are interpreted as baseline-supported but not robust after adjustment. The findings move beyond generic GenAI acceptance by positioning GAISC as an accounting-specific competence construct linked to ethical awareness and career readiness. Because GAISC and CR were measured through student self-reports and the evidence is cross-sectional, the results are interpreted as association-based rather than as demonstrations of objectively verified competence, employability, or causal effects.
Associated journal: PLOS ONE
Associated manuscript ID: PONE-D-26-30921
Data description
This dataset contains the final analytical survey data used in the associated study of GenAI-enabled accounting information systems (AIS) competence among undergraduate accounting and auditing students in Vietnam.
The public dataset contains 471 valid analytical cases. Respondents were third- and fourth-year undergraduate accounting and auditing students from five participating universities in Vietnam. Eligibility required completion of foundational accounting courses and prior exposure to AIS, digital accounting tools, or GenAI-supported learning activities.
The survey was administered using a paper-based questionnaire. The public data contain only the final analytical sample. Pilot-test responses and questionnaires excluded during data screening are not included in this public file.
The dataset is anonymized. It contains no names, student identification numbers, email addresses, telephone numbers, university identifiers, survey-start dates, questionnaire collection dates, or other direct personal identifiers. The Case_ID variable is a newly assigned sequential anonymous identifier used only to distinguish the 471 public analytical cases and has no substantive meaning.
Ethical approval for the study was granted by University of Finance – Marketing on 20 October 2025. The approved protocol permitted verbal informed consent for the anonymous, minimal-risk student survey. Only students who verbally agreed to participate received the questionnaire.
All 32 substantive survey variables are measured on a five-point Likert scale:
- 1 = Strongly disagree
- 2 = Disagree
- 3 = Neither agree nor disagree
- 4 = Agree
- 5 = Strongly agree
The public analytical dataset contains no missing values for the 32 survey items.
Files and variables
Exact data filename
The data file included in this Dryad submission is:
Analysis_Data_471.xlsx
Important: The exact filename is Analysis_Data_471.xlsx, including the underscores.
The workbook contains two worksheets:
Analysis_Data_471Codebook
Worksheet: Analysis_Data_471
This worksheet contains:
- 471 data rows, one row per analytical case
- 33 columns
- 1 anonymous case identifier (
Case_ID) - 32 survey-item variables
- no pilot cases
- no excluded questionnaires
- no respondent-level date variables
- no personally identifiable information
Worksheet: Codebook
This worksheet provides variable-level documentation for all 33 columns in Analysis_Data_471, including:
- variable name
- construct
- full item wording
- data type
- allowed values
- coding information
Variable and construct dictionary
| Variables | Construct | Description |
|---|---|---|
Case_ID |
Anonymous case identifier | Sequential anonymous identifier from 1 to 471; no substantive meaning and no link to respondent identity |
GAR1–GAR4 |
GenAI Readiness | Readiness and adaptability to use GenAI in accounting and AIS-related learning |
PEOU1–PEOU4 |
Perceived Ease of Use of GenAI in AIS Learning | Perceived ease, clarity, prompt interaction, and effort associated with GenAI use in AIS learning |
DLSE1–DLSE4 |
Digital Learning Self-Efficacy for AIS Learning | Confidence in learning and solving AIS-related problems in digital environments |
FC1–FC4 |
Facilitating Conditions for GenAI-Enabled AIS Learning | Perceived infrastructure, instructor guidance, practice opportunities, and responsible-use guidance |
GAISC1–GAISC4 |
GenAI-Enabled AIS Competence | Self-assessed ability to process accounting information, verify AI outputs, identify errors, and apply professional judgment |
PU1–PU4 |
Perceived Usefulness of GenAI in AIS Learning | Perceived usefulness of GenAI for AIS learning, accounting information evaluation, and preparation for professional tasks |
AIEA1–AIEA4 |
AI Ethics Awareness in GenAI-Enabled AIS Use | Awareness of inaccuracy, bias, privacy/confidentiality, transparency, explainability, accountability, and human oversight |
CR1–CR4 |
Career Readiness for GenAI-Enabled Accounting Work | Self-reported readiness to work in GenAI-enabled accounting environments and evaluate AI-supported accounting outputs |
The full wording of every individual survey item is provided in the Codebook worksheet of Analysis_Data_471.xlsx.
Variable order in Analysis_Data_471
The columns appear in the following exact order:
Case_ID, GAR1, GAR2, GAR3, GAR4, PEOU1, PEOU2, PEOU3, PEOU4, DLSE1, DLSE2, DLSE3, DLSE4, FC1, FC2, FC3, FC4, GAISC1, GAISC2, GAISC3, GAISC4, PU1, PU2, PU3, PU4, AIEA1, AIEA2, AIEA3, AIEA4, CR1, CR2, CR3, CR4.
Data quality and processing
The file Analysis_Data_471.xlsx contains the final dataset used for the study's statistical analyses.
Quality checks applied before the final analytical dataset was produced included screening for completeness, eligibility, duplicate participation, straight-lining, and consistency with the target population. Only the final 471 retained cases are included in the public file.
For the released file:
Case_IDvalues are unique integers from 1 to 471.- All 32 item variables contain integer values from 1 to 5.
- There are no missing item responses.
- No formulas are required to interpret the data.
- No respondent identifiers or respondent-level survey dates are included.
- Pilot-test data are not included.
The constructs are modeled as reflective constructs in the associated study. GAISC and career readiness are self-reported constructs and should not be interpreted as objectively verified task performance, employability, or workplace performance.
Code and software
No analysis code or executable scripts are included in this Dryad deposit.
The data are provided in Microsoft Excel Open XML format (.xlsx). The file can be opened with software capable of reading XLSX files, including Microsoft Excel and LibreOffice Calc, and can be imported into common statistical environments such as R or Python.
The analyses reported in the associated manuscript used PLS-SEM in SmartPLS, with a common latent factor diagnostic estimated in AMOS/CFA and Gaussian copula analysis used as an endogeneity robustness assessment. The public file contains the item-level data required for independent reanalysis using appropriate statistical software.
Access information
Dryad dataset DOI: https://doi.org/10.5061/dryad.1g1jwsvdw
The dataset is associated with the manuscript:
GenAI-Enabled Accounting Information Systems Competence: Antecedents, Perceived Usefulness, AI Ethics Awareness, and Career Readiness Outcomes
Journal: PLOS ONE
The data file in this submission is exactly:
Analysis_Data_471.xlsx
No external data sources are required to interpret the released survey data. The Codebook worksheet and this README provide the documentation needed to understand the variables and coding.
Human-subjects and privacy statement
The released dataset contains anonymized quantitative survey responses only. No direct personal identifiers are included. University identifiers and respondent-level survey dates are not included in the public dataset. The sequential Case_ID values were created for the public analytical file and cannot be used to identify participants.
The public file contains only the final analytical sample of 471 cases. Pilot responses and questionnaires excluded during screening are not part of Analysis_Data_471.xlsx.
Human subjects data
All participants provided informed consent prior to participating in the study, including consent for the anonymous research data to be used for research purposes and made publicly available. The survey was conducted anonymously from the outset and no personally identifiable information, such as names, email addresses, telephone numbers, identification numbers, or precise addresses, was collected. The archived dataset therefore contains no direct personal identifiers. Prior to deposit, the dataset was additionally reviewed to ensure that no information capable of identifying individual participants was included. The data made available through Dryad are fully anonymous and cannot reasonably be linked back to individual participants.
The main analysis is conducted using SmartPLS and follows a two-stage PLS-SEM logic. First, the reflective measurement model is assessed to ensure that each construct is measured reliably and distinctly. Indicator reliability is evaluated using outer loadings, with values of 0.70 or higher considered acceptable. Internal consistency reliability is assessed using Cronbach’s alpha, composite reliability, and rho_A. Convergent validity is evaluated using average variance extracted (AVE), and discriminant validity is assessed using the heterotrait-monotrait ratio (HTMT), with values below the conservative threshold of 0.85 interpreted as supporting discriminant validity (Hair et al., 2021; Henseler et al., 2015).
After the measurement model is established, the structural model is assessed. Collinearity among predictors is examined using inner VIF values. The statistical significance of direct relationships is assessed using bootstrapping with 5,000 resamples. Path coefficients, standard errors, t-values, and p-values, are used to evaluate the hypotheses. Explanatory power is assessed using R-square and adjusted R-square for endogenous constructs, while local effect sizes are evaluated using f-square.
The hypothesized structural model includes nine direct relationships: GAR -> GAISC, PEOU -> GAISC, DLSE -> GAISC, FC -> GAISC, PU -> GAISC, GAISC -> AIEA, GAISC -> CR, AIEA -> CR, and PU -> CR. This specification aligns the estimation strategy with the research model presented in Figure 1. Although indirect effects may be reported as supplementary mechanism checks, the main hypotheses focus on the direct paths specified in Figure 1.