Spotting INcredible Research (forum)

Hi Josh

You actually highlight a few important points, let me try and get to them here. Note, I did not read the study in detail, and have focused on the issues you have raised here. I did see quite a few more reading through (note, it took less than two months to go from submission to publication). - Correct. In an experimental study you would keep as many components between two training groups as you can to try and control for 'third party confounders'. These confounders are other things that could explain the results observed in the study. If these are not controlled for (sometimes this can be done statistically but often this is done through design) you really cannot say whether the findings are the result of the 'targeted intervention' or any other differences between the conditions. Given this is a two arm between subjects trial (two groups engage in a diferrent intervention rather that all participants engaging in both interventions, ie a within-subjects trial) this is even more important. In this study, this does not seem to be accounted for.

• Also correct. The more systematic bias we introduce, the more likely it is that the results can be explained by another variable than the one being studied. In the context of blinding (how much people are aware of in terms of the intervention) there are three types of blinding. Studies with a low risk of bias have addressed all three of these sources of bias (bias as a result of insufficient blinding of those delivering the intervention, bias as a result of insufficient blinding of participants, and bias as a result of insufficient blinding of those analysing the results). In this study it does not seem that any of these forms of blinding were done (or at least not reported).

Job

Hi Josh. It is not you. You are right. Many people, especially in sport, do not declare their conflict of interest. Here is what you need to know. Journals generally require authors to declare if they have conflicts. When they do, an author can do one of two things: declare a conflict or declare no conflict. The latter is by far the most popular option, and unless the editor of the journal has a keen understanding of the private and professional life of the authors, they can almost never police whether this declaration was accurate. So, in short, not declaring a conflict of interest when there is one (or at least a perceived conflict of interest which should also be declared) is one of the most pertinent issues related to credibility in our area. For example, when you read through papers, have a look at how often people have multiple affiliations next to their name, including sometimes their own company or external funders, but declare that no conflict exists. Just FYI, this is what a declaration of conflict of interest looks like when I have written work that can potentially be affected by owning SkillACQ: "Declaration of conflict of interest: The author declares that he works as a skill acquisition specialist in professional sports practice. There may be a (perceived) conflict of interest as a result."