{"id":"3c3b969c-8e11-4987-8e3c-7c4f422c65c9","arxiv_id":"2608.02952","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A census of all 39 Australian universities offering IT degrees shows that specialized and major-based degrees are now widespread, with data science and cybersecurity dominating.","lead":"This study counted and categorized every IT degree offered by Australian universities, splitting them into general, specialized, and major-based awards. It finds that technology-focused universities offer more specialized degrees, and that data science and cybersecurity are the most common specializations.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Census counts are supply-side data; the claimed 'insights into whether degrees address IT industry needs' lack any direct measure of employer demand.","rationale":"The authors have produced a genuinely useful descriptive census: no existing comprehensive inventory of Australian IT degree types was identified, and the tables are detailed. The reader's conditional verdict is reasonable. My stress-test focuses on a different clause of the stated contribution: the paper says it provides insights into whether degrees address industry needs. That requires a demand-side yardstick, and none is present. The observed counts can only show what universities supply; popularity among universities is not evidence of employer demand. The FinTech discussion is illustrative but not empirical. A job-ad comparison would settle it. If it showed alignment, the claim could stand; if not, the contribution should be re-scoped to 'status quo of offerings' only. Since the census itself remains informative, the current conditional verdict is appropriate, with the condition extended to cover the needs-claim as well as the data-sharing and classification concerns already noted by the reader.","tokens_in":16483,"tokens_out":13461,"duration_ms":125293,"concrete_test":"Build an external demand benchmark for the same period: extract Australian IT job advertisements (e.g., Seek, LinkedIn, ABS Labour Market Insights) and code required skill and title clusters (data science, cybersecurity, AI, software engineering, cloud, DevOps, FinTech, etc.). Rank these by vacancy volume and compare with the supply rankings in Tables 10 and 12, for instance with a Spearman rank correlation or an overlap analysis of the top-10 demanded versus top-10 offered areas. If the rank correlation is weak, or if high-demand areas such as cloud, DevOps, or full-stack are absent from the top offerings, the inference that the top four offerings reflect industry needs fails, and the abstract's needs-claim should be retracted or substantially weakened.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central contribution as stated in the abstract includes 'insights into whether these IT degrees properly address the IT industry's needs.' That claim is not supported by the data collected. RQ1-RQ9 are supply-side inventories: counts of GL/MA/SP degrees, research components, internships, professors, and research institutes. RQ10 counts industrial involvement mechanisms (advisory boards, co-design, co-teaching, internships), but these are not measures of industry need. The inference that the most popular specializations (data science/analytics, cybersecurity, AI, software engineering) 'properly address' industry needs relies on (a) the observation that these are the most frequently offered, and (b) a few generic citations rather than any systematic demand-side data. For example, the paper notes that no FinTech SP-degree or major exists, yet never measures whether FinTech employers actually demand such a degree from IT schools; the argument is purely illustrative. The descriptive census can stand alone, but the needs-assessment conclusion must either be removed or re-scoped, or supported with external labor-market and employer-demand evidence.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper reports a census study of all 39 Australian universities that offer IT degrees, categorizing their undergraduate and postgraduate coursework IT programs into general degrees (GL), specialized degrees (SP), and degrees with majors (MA). The authors pose ten research questions covering the prevalence of each degree type, differences by university category (universities of technology, Group of Eight, regional vs. metropolitan), the most popular specializations and majors, the presence of research components and internships, supporting professorial infrastructure and research institutes, and forms of industry involvement. They use counts, percentages, and population effect sizes (Cohen's d) to describe the status quo, and derive recommendations about career-centered curricula, cross-university co-teaching, and student co-creation of syllabi.","tokens_in":16689,"tokens_out":3475,"duration_ms":34220,"significance":"The study's main contribution is a systematic, population-level inventory of IT degree offerings in Australia, a useful reference for curriculum planners and policy discussions. The population approach is a strength: because all offering universities are included, the authors correctly avoid null-hypothesis significance testing and report effect sizes as descriptive magnitudes. The classification scheme (GL/MA/SP) is clearly defined and the counting conventions, though complex, are disclosed. However, the paper's stated conclusion that the data provide 'insights into whether these IT degrees properly address the IT industry's needs' is not supported by the supply-side counts alone; no demand-side or labor-market outcome data are collected. The post hoc exclusion of the University of Melbourne in RQ3 also changes the qualitative conclusion and needs stronger justification. These issues are fixable by re-scoping the claims and adding sensitivity analysis, but they affect the central interpretation.","major_comments":[{"comment":"The claim that the census provides 'insights into whether these IT degrees properly address the IT industry's needs' is not supported by the data actually collected. RQ1–RQ9 are supply-side inventories (counts of degree types, research components, internships, professors, and research institutes), and RQ10 counts only the presence of industry advisory boards, co-design, co-teaching, and internships. None of these measures captures employer demand, graduate employment outcomes, skill shortages, or the fit between curricula and job requirements. For example, the observation that no university offers a FinTech specialization is used to argue a gap, but no evidence shows that employers demand such a degree from IT schools. I recommend removing or explicitly re-scoping the needs-assessment claim to 'supply-side landscape,' or adding external labor-market data (e.g., job advertisement analyses, employer surveys) if the claim is retained.","section":"Abstract and Introduction; RQ5/RQ7 discussions; Conclusion"},{"comment":"The exclusion of the University of Melbourne from the RQ3 analysis is post hoc and changes the qualitative conclusion. Initially, the Go8 comparison for bachelor's SP-degrees yields Cohen's d = -0.14 (marginally fewer than other universities); after removing UoM, the text reports d = -0.09 and concludes that 'Go7 offers at least the same mean number of (and often more) MA-degrees and SP-degrees as other universities at both degree levels.' This is a data-dependent decision: the exclusion is motivated by the Melbourne Model, which is a legitimate structural reason, but the manuscript does not present a pre-specified outlier criterion or a sensitivity analysis showing both Go8 and Go7 results side by side with interpretation. I recommend reporting both sets of effect sizes in full and clearly labeling the Go8 result as driven by UoM, rather than switching to Go7 without transparent comparison.","section":"RQ3, Table 7"},{"comment":"The dataset for this census is described as 'available upon reasonable request,' but no raw data, degree lists, classification codebook, or per-university worksheet is provided. Because every result in the paper derives from manual coding of university websites, an independent reader cannot verify the central counts in Tables 2, 4, 6, 8, 10, 12, 14, 15, or the classification of degrees into GL, MA, and SP. I strongly encourage making the full dataset and coding protocol publicly available as supplementary material (e.g., a spreadsheet listing each university, each program title, the assigned category, and the webpage source and access date). Without this, the census is not auditable, and the manuscript's reproducibility claim is weakened.","section":"Data availability statement and Study settings"}],"minor_comments":[{"comment":"The notes contain typos: 'Gohen's d' should be 'Cohen's d' in all three table notes.","section":"Tables 5, 7, 9"},{"comment":"In the RQ3 discussion, 'G07' should be 'Go7', and 'formers' should be 'former's' in the sentence about 'the formers' scale of operations.'","section":"RQ3 text"},{"comment":"The header rows of these tables are visually ambiguous: the repeated 'No. of MA-degrees' columns are not clearly separated into UoTs vs CUs (or Go8 vs others, metropolitan vs regional) and bachelor's vs master's. Adding subheaders or blank columns would improve readability.","section":"Tables 4, 6, 8"},{"comment":"The standard deviations reported for SP-degrees (bachelor's mean 1.2, SD 4.3; master's mean 1.8, SD 8.7) are implausibly large for count-like data and suggest a highly skewed distribution. Reporting medians and interquartile ranges, or a direct frequency comparison, would be more informative than mean and Cohen's d for such data.","section":"RQ6"},{"comment":"The definition of 'percentage of coverage' says the authors checked whether a university has professors whose expertise covers a major or specialization, but the method for determining 'expertise' from websites (e.g., research profile keywords, faculty directory categories) is not described. Please add a sentence specifying the coding rule.","section":"RQ9"},{"comment":"The reference 'Kumari, A. (2026)' appears to be a blog post with a publication date after the data collection period (October–November 2025). Verify the date and source; if it is a blog, consider replacing it with a peer-reviewed curriculum theory reference.","section":"References"},{"comment":"The footnote for Table 13 correctly notes that bachelor's-level counts exclude UoM because it offers no bachelor's IT degree, but the table row label 'Bachelor's or master's' is ambiguous. Please rephrase to make clear that the 84.6% refers to either level among all 39 universities.","section":"Table 13 footnote"}],"recommendation":"major_revision","confidential_remarks":"The paper is a descriptive census that could be a useful reference for the journal's readership, but the authors overreach when they claim to assess whether IT degrees meet industry needs. The RQ3 outlier handling is the clearest methodological weakness; the authors should be asked to present both Go8 and Go7 analyses and to temper the needs-assessment language. I would not reject the paper, but the revision requires substantive changes to the framing and, ideally, public release of the dataset."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The new thing here is the dataset and the categorization. Nobody else has systematically classified every Australian university's IT bachelor's and coursework master's offerings into general degrees, specialized degrees, and degrees with majors, then compared counts across university types. That is a real gap, and the authors filled it. The counting conventions are disclosed, and using Cohen's d on a population is defensible because they are measuring the magnitude of group differences, not inferring from a sample. The RQ-by-RQ structure makes the paper easy to check. I believe the descriptive findings: UoTs offer more specialized degrees than comprehensive universities; regional universities are closer to metropolitan than intuition suggests; the top specializations are data science/analytics and cybersecurity; master's degrees have fewer internships.\n\nCredit where due: the authors notice when their own expectations fail (e.g., regional vs metropolitan) and offer concrete mechanisms. The Go8 analysis is handled honestly: UoM's Melbourne Model is a real outlier, and they rerun as Go7 and report both. The limitation statement about website currency is candid.\n\nNow the soft spots. The biggest one is the gap between the supply-side census and the industry-needs conclusion. The abstract, RQ10, and the recommendations repeatedly infer that popular specializations mean the sector \"properly addresses\" industry needs, but none of the data measure employer demand, hiring difficulty, or graduate outcomes. The FinTech example cuts the other way: the absence of FinTech degrees is taken as a curriculum gap, but the paper does not show that employers demand FinTech-specific IT degrees; arguably a data science or cybersecurity graduate is what is actually wanted. That claim should be re-scoped to \"these are the most common offerings, which suggests what universities believe the market wants,\" or backed with external labor-market data. The post-hoc UoM exclusion is a soundness wrinkle: the Go8-vs-others bachelor's comparison flips from mixed to positive after exclusion. The authors are transparent about it, which I respect, but the conclusion in the abstract and conclusion (\"Go8 universities (except UoM) offer ... more\") is fragile with n=7/8. Better to present both and emphasize the sensitivity. The data availability statement (\"upon reasonable request\") is weak for a census whose whole value is the dataset; they should deposit the degree-level counts in a repository. I would also like to see inter-rater reliability on the GL/MA/SP classification if more than one person coded, or at least a sample audit. Minor: some discussion paragraphs lean on speculative explanations, but they are clearly labeled as speculation.\n\nWho is this for? Computing education researchers, Australian universities planning programs, and policy makers. It is not causal, not theoretical, but it is the missing descriptive base layer. A serious referee should engage with it. My recommendation: send to peer review, with the note that the industry-needs language needs to be pulled back and the data published.","headline":"A genuinely useful descriptive census of Australian IT degrees that deserves peer review, but the industry-needs claim outruns the supply-side data and the dataset should be shared openly.","tokens_in":17151,"tokens_out":1927,"would_cite":true,"duration_ms":17945,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"A complete census of Australian universities shows that the general IT degree is no longer the sole norm: most institutions now offer specialized or major-based degrees, with data science, cybersecurity, AI, and software engineering…","keywords":["information technology curriculum","general degree","specialized degree","degree with majors","census study","effect size analysis","graduate employability","Australian universities"],"falsifier":"Contact the universities and compare the online course data against official handbooks and program regulations; if the counts of specialized degrees, majors, internships, research components, or research institutes change materially, the census picture is not stable. A repeat census at a different date (for example, six months later) that shifts the top-four specializations or reverses the effect-size directions would also contradict the reported status quo.","tokens_in":16322,"feed_emoji":"🎓","tokens_out":9528,"duration_ms":70594,"temperature":0.7,"pith_summary":"This paper reports a complete census of information-technology degree offerings at every Australian university that teaches IT, classifying each degree as general, specialized, or a degree with majors. The central finding is that offering only an unspecialised general IT degree has become unusual: over 92% of bachelor's-level and over 94% of master's-level institutions offer a specialized or major-based alternative, and the most common specializations and majors are data science and analytics, cybersecurity, artificial intelligence, and software engineering. Using effect sizes rather than significance tests, the authors show that technology-focused universities offer more specialized degrees than comprehensive universities, that the prestigious Group of Eight (Go8) research-intensive universities offer at least as many once one outlier is set aside, and that master's degrees embed internships less often than bachelor's degrees. The census also exposes gaps: formal industry co-design and co-teaching are nearly absent, and career-oriented specializations such as FinTech do not appear in IT schools. The paper thus provides the first systematic baseline for asking whether Australian IT curricula are producing graduates who match industry needs.","feed_headline":"Australian IT degrees shift decisively toward specialized offerings","feed_subtitle":"Nearly every university offers a specialized or major-based IT degree; data science, cybersecurity, and AI lead.","key_machinery":"The central object is the three-way classification of IT degrees into general (GL), specialized (SP), and major-based (MA) degrees, where an MA-degree contains a named major inside a broader degree and therefore sits on a spectrum between the general and the specialized. The argument is carried by a hand-collected census dataset of all 39 Australian universities offering IT study, built from university websites over October–November 2025. The analytic machinery is Cohen's d effect size applied to the entire population, used to compare mean numbers of offerings between groups such as technology-focused versus comprehensive universities; because the dataset is a census, the paper deliberately avoids hypothesis testing and instead reports the magnitude of differences.","core_discovery":"The paper's central discovery is the status quo itself: as of late 2025, offering a plain general IT degree as the only option is no longer the norm in Australia. Across the 39 universities that offer IT study, 92.1% of bachelor's-level institutions and 94.7% of master's-level institutions offer either a specialized degree or an IT degree with one or more majors. The four most popular specializations and majors are the same list – data science and analytics, cybersecurity, artificial intelligence, and software engineering – and together they account for 78.3% of bachelor's and 85.9% of master's specialized offerings. The authors interpret this as universities responding to the employment market, but they also find that curricula are largely technology-centered rather than career-centered, that master's degrees embed internships less often than bachelor's degrees, and that formal industry involvement beyond internships is minimal.","pith_inferences":["The general/specialized/majors trichotomy could serve as a comparative lens for IT education in other countries, enabling cross-national comparisons of how curricula specialize.","If the population-level effect sizes hold, they give any university a direct baseline: its own counts of specialized and major-based degrees can be compared with the national means reported here.","The finding that internships are more common in bachelor's than in master's degrees is likely to change as professional master's programs grow, and could be tracked by repeating the census over time.","The paper's explanation for regional universities narrowing the gap – city campuses that behave like metropolitan universities – suggests that physical campus location, rather than institution type, may be the better predictor of curriculum breadth."],"forward_implications":["A prospective IT student in Australia can expect nearly every university to offer a specialized or major-based degree, with data science/analytics, cybersecurity, AI, and software engineering the most common focal areas.","Universities of technology offering more specialized degrees than comprehensive universities implies that institutional mission and STEM focus shape curriculum breadth, not just the level of resources.","The near-absence of FinTech as an IT specialization and only one quantum-computing degree suggest these curricula are aligned with today's job market more than with emerging technology careers.","Low rates of industry co-design and co-teaching imply that internships are currently the dominant, and often the only, formal industry link in most IT degrees."],"supporting_citations":[{"why":"Supplies the 2021 census data showing IT qualifications in Australia grew 36% since 2016, establishing the population trend the study is set against.","marker":"Australian Bureau of Statistics, 2022"},{"why":"Provides the thresholds for interpreting Cohen's d effect sizes (0.2 small, 0.5 medium, 0.8 large) used to compare university groups.","marker":"Magnusson, 2025"},{"why":"Source for classifying universities as regional or metropolitan based on head-campus location.","marker":"Good Universities Guide, 2025"},{"why":"Source for the list of regional universities used in the RQ4 comparison.","marker":"Regional Universities Network Australia, 2025"},{"why":"Supports the claim that IT curricula are technology-centered rather than career-centered, grounding the paper's main recommendation.","marker":"Herbert et al., 2013"},{"why":"Explains why specialized master's programs are more likely than general ones to include internships.","marker":"IMFS, 2025"},{"why":"Links internships to employability in the ICT sector, underpinning the internship findings.","marker":"Chillas et al., 2015"}],"fun_headline_variants":["92% of Australian IT degrees now specialized or major-based","Australia's IT degrees shift to specialized tracks, census shows","Data science, cybersecurity, AI lead new wave of Australian IT degrees","Census finds Australian IT curricula tech-centric, industry-light","General IT degrees nearly gone in Australia; specializations rule"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The study assumes that the information posted on each university's official website is accurate, current, and complete enough to classify every IT degree and to count the presence of majors, internships, research components, professors, and research institutes correctly.","fun_headline_variants_meta":{"raw":{"variants":["92% of Australian IT degrees now specialized or major-based","Australia's IT degrees shift to specialized tracks, census shows","Data science, cybersecurity, AI lead new wave of Australian IT degrees","Census finds Australian IT curricula tech-centric, industry-light","General IT degrees nearly gone in Australia; specializations rule"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000414,"raw_usage":{"total_tokens":2102,"prompt_tokens":867,"completion_tokens":1235,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":483,"completion_tokens_details":{"reasoning_tokens":1152}},"tokens_in":483,"tokens_out":1235,"duration_ms":10962,"temperature":1.0,"reasoning_tokens":1152,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T14:53:09.379964+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Contact the universities and compare the online course data against official handbooks and program regulations; if the counts of specialized degrees, majors, internships, research components, or research institutes change materially, the census picture is not stable. A repeat census at a different date (for example, six months later) that shifts the top-four specializations or reverses the effect-size directions would also contradict the reported status quo.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the thresholds for interpreting Cohen's d effect sizes (0.2 small, 0.5 medium, 0.8 large) used to compare university groups."},{"cited_title":"obviously less","cited_arxiv_id":null,"evidence_quote":"Source for the list of regional universities used in the RQ4 comparison."}],"review_version":2}