Complaints about the poor quality of science education are a familiar refrain in many countries, as are national anxieties about falling behind the rest of the world. What's not so familiar is that pretty much everyone feels this way. Nature Publishing Group's educational division, Nature Education, last year conducted a survey of 450 university-level science faculty members from more than 30 countries. The first report from that survey, freely available at http://go.nature.com/5wEKij, focuses on 'postsecondary' university- and college-level education. It finds that more than half of the respondents in Europe, Asia and North America feel that the quality of undergraduate science education in their country is mediocre, poor or very poor.

Despite agreeing that inadequate secondary-school science education is the major problem, respondents concurred on how they could help contribute to a solution: by having professors provide better college-level teaching. Moreover, 77% of respondents indicated that they considered their teaching responsibilities to be just as important as their research — and 16% said teaching was more important.

Yet although there was general agreement about the low quality of undergraduate education, a substantial majority of the respondents felt that their own teaching was highly effective. This suggests that at least some of the respondents are fooling themselves.

Certainly, that apparent complacency would help to account for the notably slow uptake of pedagogical innovations in the teaching of science to undergraduates. But there is strong evidence that talking at students isn't nearly as effective as engaging them with cooperative, hands-on learning activities. A prime example of the latter approach is Process Oriented Guided Inquiry Learning (http://pogil.org), which originated in US college chemistry departments in 1994, and which is now used in many other subject areas.

But the biggest barrier to improvement is the pervasive perception that academic institutions — and the prevailing rewards structure of science — value research far more than teaching. That perception was apparent in one of the survey's most striking contradictions: despite their beliefs that teaching was at least as important as research, many respondents said that they would choose to appoint a researcher rather than a teacher to an open tenured position.

To correct this misalignment of values, two things are required. The first is to establish a standardized system of teaching evaluation. This would give universities and professors alike the feedback they need to improve. Undergraduate student outcomes can already be measured in a variety of innovative ways, such as the 'concept inventory' system developed in physics. But more research is needed in this area.

The second requirement is to improve the support and rewards for university-level teaching. For example, universities and professional societies could offer staff systematic training in how to teach well — something less than a two-year degree, but more than a two-hour workshop. Universities could encourage donors to endow professorships based on teaching excellence. And funding agencies could make teaching more of a financial priority, as does the private Howard Hughes Medical Institute in Chevy Chase, Maryland, which offers scientists up to US$1 million over four years to innovate in science education.

Correcting the misalignment will be neither quick nor easy. But by showering so many rewards on research instead of on teaching, universities and funding agencies risk undermining the educational quality that is required for research to flourish in the long term. They need to find a more balanced way to allocate their resources — and in the process allow the majority of academic scientists to act on their conviction that teaching and research are equally important.