The research assessment exercise that consumed enormous institutional energy throughout 2024 and early 2025 finally concluded with results released in mid-2025. Universities spent thousands of staff hours preparing submissions, selecting their best research outputs, and crafting impact case studies. Whether this massive effort produced value proportionate to cost is deeply questionable.

The Excellence in Research for Australia framework assessed research quality across disciplines and institutions. Universities submitted four research outputs per eligible researcher, plus impact case studies demonstrating how research influenced policy, practice, or culture beyond academia. Panels of senior researchers evaluated everything, rating research quality from one (well below world standard) to five (well above world standard).

The results were predictably hierarchical. Go8 universities dominated the highest ratings, regional universities performed modestly, and everything in between filled the middle. Anyone familiar with Australian research could have predicted the rank ordering without conducting the expensive assessment. Whether the precision of quantification justifies the cost is the relevant question.

Universities immediately issued press releases celebrating their ratings, selectively highlighting areas where they performed well while downplaying weaker results. Marketing departments translated complex ERA ratings into simplistic “world-class research” messaging that obscures more than it illuminates. The public relations exercise probably matters more for universities than the actual substance.

Research quality assessment did reveal some interesting patterns. Several regional universities achieved high ratings in specific niches where they’ve concentrated expertise. A few disciplines at non-elite universities outperformed prestigious institutions, demonstrating that research quality doesn’t perfectly correlate with overall institutional reputation. These nuances mostly disappeared in public communication.

The impact case studies proved more interesting than quality ratings. Researchers documented how their work influenced government policy, changed industry practices, shaped legal frameworks, or affected cultural understanding. Some case studies demonstrated impressive real-world impact. Others strained credibility, claiming influence that seemed overstated relative to evidence provided.

Measuring research impact is inherently difficult. Causation is hard to establish—policy changes often result from multiple factors beyond single research projects. Time lags between research and impact can span decades. Attribution is complex when research builds cumulatively on many contributors’ work. The ERA framework acknowledges these challenges without solving them.

Gaming opportunities existed and were likely exploited. Universities could be strategic about which outputs to submit, selecting only their strongest work. Researchers with several high-quality publications could spread them across assessment cycles, while those with weaker records might be excluded from submissions entirely. Whether this represents smart strategy or gaming depends on your perspective.

The workload was extraordinary. Research administrators spent months managing submission processes. Senior researchers evaluated submissions internally before external assessment. Academic staff provided supporting documentation and verified claims. For large universities, the total effort probably exceeded 10,000 staff hours. At staff cost of $50-80 per hour, that’s $500,000-800,000+ in direct labor costs per institution.

The external assessment process involved hundreds of senior researchers serving on panels. That’s weeks away from their own research, teaching, and supervision. The opportunity cost is substantial but difficult to quantify. Some panelists reported the experience as professionally valuable, learning about research across their fields. Others viewed it as tedious obligation.

International comparison shows mixed approaches. The UK’s Research Excellence Framework is similar to ERA and faces similar criticisms. The US has no equivalent national assessment, relying instead on decentralized peer review through grant funding and institutional reputation. European countries vary widely in assessment approaches. No consensus exists on best practice.

The utility of ERA results is debated. Government doesn’t use ERA ratings directly for funding allocation, unlike some assessment exercises elsewhere. Universities use results for internal performance management and external marketing. Researchers might reference ratings when applying for jobs or grants, though whether ERA ratings actually influence these decisions is unclear.

Critics argue ERA entrenches conventional research while potentially disadvantaging innovative work that doesn’t fit established frameworks. Interdisciplinary research, community-engaged scholarship, and non-traditional outputs may be undervalued by assessment processes designed around traditional academic publications. The framework acknowledges these issues without fully addressing them.

The five-year cycle means universities immediately began preparing for the next ERA shortly after this one concluded. Research management becomes partially oriented around assessment cycles rather than purely driven by research questions. That’s arguably distortionary—research should happen for its own sake or practical application, not to perform well in assessment exercises.

Discipline-specific differences matter substantially. Some fields have clear quality indicators through journal rankings and citation metrics. Others rely more on book publications or practice-based research where quality is harder to quantify. The ERA framework attempts to accommodate disciplinary differences but necessarily applies somewhat standardized approaches across diverse fields.

The transparency of ERA is both strength and limitation. Results are publicly available, enabling scrutiny and comparison. But public availability also creates pressure for universities to perform well regardless of whether that serves their actual mission. A regional university serving its community well might rationally prioritize teaching and engagement over research ratings, but ERA scores create competitive pressure that makes that difficult.

Several universities reported morale problems during ERA preparation. Researchers whose work wasn’t selected for submission felt devalued. Those included faced pressure to ensure their outputs rated highly. The competitive internal processes created tension in research communities that ideally would be collaborative. Whether these problems were worth the assessment benefits is questionable.

Indigenous research faced particular challenges in ERA framework. Much Indigenous research operates outside conventional academic publication channels and prioritizes community outcomes over academic metrics. The assessment framework attempted accommodation but struggled to value Indigenous research appropriately within structures designed for Western academic conventions.

For early career researchers, ERA had ambiguous effects. High-performing early career researchers benefited from visibility when their work was included in institutional submissions. But ERA also emphasized established researchers with proven track records, potentially disadvantaging innovative early career work that hasn’t yet accumulated conventional metrics.

The counterfactual is unknowable—what would Australian research look like without ERA assessment exercises? Possibly very similar, with funding allocated through existing mechanisms and quality evaluated through peer review. Or possibly worse, without systematic assessment identifying strengths and weaknesses. The answer depends partially on whether you trust decentralized evaluation or prefer centralized frameworks.

Looking ahead, the next ERA cycle looms in 2028-2029. Universities will prepare submissions again, panels will assess again, results will be released again, and institutions will market their ratings again. Whether the cycle produces sufficient value to justify its costs remains contested. The assessment happens because policy requires it, not because there’s consensus it’s worth doing.

For researchers, ERA is mostly something to tolerate rather than engage with enthusiastically. Submit your strongest work when selected, participate in assessment if asked, then return to actual research. The exercise serves administrative and political functions more than research functions. That’s neither good nor bad, just reality about how research systems operate.