When companies fund research they obviously hope that the research will demonstrate the superior effectiveness of their products. Consequently, publication of negative results might make companies less enthusiastic about supporting such research in the future. Since researchers who use industrial support are likely to be eager to continue with this funding, some believe that this might increase the risk biased reporting. This is the premise debated in this monthˈs Point/Counterpoint. As humans, we are hard-wired toward implicit bias.1 The Washington Post published on biased reporting in pharmaceutical research,2 which triggers some self-examination about potential publication bias in medical physics. Let us assume that vendors and researchers have been able to avoid conscious bias exerted by pressures of market shares and up-or-out research faculty appointments which depend on securing ever scarcer grants. The issue of implicit bias still remains. To conduct good science, we must address all factors affecting the quality of science; publication bias is a major known factor to affect research quality, and with shrinking NIH budgets will only gain in influence. A large body of research is available on publication bias. To summarize: (1) there is publication bias in medical journals toward positive outcomes,3 (2) the incidence of editors or reviewers rejecting negative studies is small for JAMA,4 but unknown for most other journals, and (3) published reports from industry-funded studies show a larger bias toward positive results.5 Unless medical physics journals (e.g., Medical Physics, PMB, JACMP) publish data to the contrary, my working hypothesis is for an existing editorial/reviewer bias of unknown size toward rejecting papers which report negative study outcomes. No reviewer guideline, journal review submission websites, or our Code of Ethics6 address implicit bias toward positive study outcomes. One proposed way to remove positive publication bias is to require all federally funded-research to be published independent of the outcome, provided the scientific method and statistical analysis meet quality standards. It remains to be seen if industry would commit to this solution as well because, for such a commitment to be meaningful, vendors would need to provide means of independent verification such as a publicly accessible listing of all outside research contracts. Increased reliance on industry funding also increases the risk for comparative effectiveness research on technology to remain unfunded. Few institutions can afford the cost and redundancy of operating two or more technologies designed to perform the same function. For vendor-funded research, there is no incentive to compare clinical effectiveness of competing technologies. Instead, the implicit bias is toward research on the new technology vs older technology. One clinical example is respiratory motion compensation. Several vendors provide solutions for each of the four techniques: compression, breath-hold, gating, and real-time tracking. Despite the widespread acceptance of these technologies and a large variation of cost to implement and use them, there is not a single prospective clinical trial which would provide data on patient outcomes based solely on the technology used for treatment. As scientists we have a mandate to generate new scientific knowledge. As medical physicists we perform and publish work that can improve how we detect, image, diagnose, and treat disease. However, our academic integrity struggles against biased reporting for any publication independent of its funding source: we have an inherent self-interest in having articles published that help us get and keep grants, help with promotion and career prospects, help with invitations to give talks at interesting places, and many other benefits. As a result, for many reasons, a few of us “behave badly.”7 Industry-funded research plays an important role in improving health outcomes, typically supporting medium-to-late stage research aligned with product roadmaps. Often late-stage research requires engineering support to allow clinical testing that otherwise would not be possible. The potential for conflicts-of-interest exists. Fortunately, to avoid any undue influence of industry pressure on the outcomes of research, there are a number of mechanisms to protect and isolate researchers from external influences and, therefore, reduce the risk of biased reporting. Three protection mechanisms reducing the risk of biased reporting are: (1) increased accountability from medical journals regarding ethics and conflicts of interest, (2) greater academic freedom in industrial-university agreements, and (3) stronger governmental regulation of commercially sponsored research. All reputable medical journals now have conflict of interest policies. Many journals follow the ICMJE conflict-of-interest policy in which each author must submit a written signed disclosure. Medical Physics requires that “Each author of a paper is required to disclose any and all potential conflicts of interest that could be perceived to bias the results reported in the paper.” This policy raises awareness for authors submitting the work, reviewers, and readers, and increases authors’ accountability. Over time, universities have taken a much stronger stance with respect to research conduct and publication control. An example from a University of Sydney agreement states: “As a matter of basic academic policy, the University retains the right to publish in it discretion material relating to the conduct and conclusions of the Research”, meaning that the academic staff have the right to publish and interpret their own results without industry direction or oversight. The US FDA, NIH, and other government bodies have a vested interest in ensuring that publications of studies represent the actual results. There is new and proposed regulation for conflicts of interest, assessing conflicts prior to the start of a study to potentially recuse investigators, avoid data falsification, and provide data storage and data access. An as example, NIH policy states: “This complexity, as well as a need to strengthen accountability, led to changes that expand and add transparency to Investigators’ disclosure of Significant Financial Interests (SFIs), enhance regulatory compliance, and effective institutional oversight and management of Investigators’ financial conflicts of interests.” In summary, medical journals, universities, and governments are actively working to protect investigators from external influences and, therefore, decrease the risk of biased reporting in journals such as Medical Physics. Dr. Keall highlights the increased efforts to address biased reporting in science. The questions we need to evaluate are: How effective are we in enforcing these rules, and have we done enough to cause a change in our culture? To use an analogy, the posted speed limit is the rule, but the unwritten culture (on most highways in the USA, Germany, and Australia at least) is for traffic to proceed at 10 miles/h above the posted speed limit without fear of repercussion by the highway patrol. Dr. Keall cited a very good paper published in 2005,7 which I have examined for data pertaining to funding source influence on outcome reporting, i.e., violation of scientific integrity standards. Table I, entry 10 in this paper7 lists the incidence of “Changing the design, methodology, or results of a study in response to pressure from a funding source” as 15.5% on average (9.5% for early career and 20.6% for mid-career scientists). Given that the study plausibly argues the under-reporting of results, these percentages hardly constitute just a few of us behaving badly, as my opponent states. I was unable to find data suggesting this percentage had decreased with the implementation of the three protection mechanisms Dr. Keall listed. To be effective in making science less prone to bias, the interventions must (1) remove the motivation for biased behavior, (2) implement a means of identifying researchers who do not follow scientific standards, and (3) increase the stakes for being found in violation of scientific standards. None of the three interventions listed by Dr. Keall remove the motivation for biased behavior. Indeed, decreasing support through outcome-neutral funding sources has increased the pressure. The second intervention, greater freedom to publish, looks good on paper but in reality will not protect the researcher from losing continued industry sponsorship should negative outcome reporting be undesired by industry. In summary, while gross scientific misconduct through fabrication, falsification, and plagiarism is indeed committed by very few of us, I (pessimistically) maintain that our scientific culture has not yet changed sufficiently to remove the significant pressure toward biased reporting. We neither have data to allow us to make a conclusive statement that biased reporting is not an issue, nor do we have any auditing procedures in place to raise the stakes for breaking the rules. Dr. Dieterich has some very good points and suggestions to improve the amount of negative-result research being published and the impact of declining Federal funding on clinically important research, such as comparative effectiveness studies. She asserts that (1) humans are subject to bias, (2) there is evidence supporting publication bias, (3) federally funded (and ideally industry-funded) research be published regardless of positive/negative results, and (4) the impact of increased reliance on industry funding means that some important research areas are unfunded. These observations are all valid. However, they do not lead to the conclusion that increasing dependence on industry-funded research creates higher risk of biased reporting in medical physics. Moreover, several of the references used to support her statement pertain to the pharmaceutical world; none specifically address biased reporting in medical physics research. To further the debate, the pathway to impacting patients on a large scale is necessarily through industry. Having ideas proceed from bench to bedside is one of the most rewarding professional accomplishments in our field. Academic/industrial interactions are essential for this translational research. AAPM TG109 (Ref. 6) states “There is nothing inherently wrong with a conflict of interest, but it should be acknowledged to eliminate the perception of possible impropriety. The best protection against conflict of interest accusations is full disclosure and the acquisition, interpretation, and publication of research findings in a manner that is transparent and above suspicion.” To navigate these interactions, in addition to the three protections reducing the risk of biased reporting detailed in my Opening Statement, researchers receive ethics education throughout their lives, along with a growing ethics component of graduate8 and residency9 medical physics programs and in our profession.6 Prof. Keall wishes to acknowledge his lab group for lively discussion on this debate and particularly Brendan Whelan and Julie Baz for comments and suggestions.