What is the criteria used in review?
The IRB considers many complex issues, but the majority of these issues fit the following categories. In planning research, it helps to keep in mind the golden rule: treat your subjects as you would wish to be treated!
A primary issue of any IRB is the ratio of risk to benefit. A study may be very simple and have minimal risks, but if it is poorly designed and the knowledge gained is of minimal use, IRB approval may be withheld. Conversely, a study may have great risks to subjects (e.g., a study of a very toxic new cancer drug brought before a hospital IRB) but if the potential benefits are high, IRB approval may be given.
A major issue of any IRB is consent: whether or not the subjects have the knowledge to fully agree to participate in the study. There is a standard consent form template on this web-page that you can download. Using this form will allow you to cover the major issues of consent. The subjects need to know: how they were selected for participation, what will happen to them as participants, what are the risks and benefits of this research, where to get more information, how their privacy will be protected, and that their participation is entirely voluntary. This form needs to be written in grammatically correct style, in simple, jargon-free language, and in an easily readable font (especially important for elderly subjects). When putting this document together, consider how you would react upon first seeing this document if you were a research subject. Before you put it in your IRB application, have someone who knows nothing about your research read it for clarity. Note that IRB committee members spend much of their time making sure this form is done well. You should also.
A third major issue of any IRB is how subjects are recruited for a research study. Some common problems with recruitment include compromising subjects' privacy and recruiting subjects who cannot easily decline. An example of compromising subjects' privacy is getting a list of subjects from a medical clinic and then calling to recruit them (the clinic should call them first to see if they agree to be contacted). An example of recruiting subjects who cannot easily decline is having nursing home staff recruit their own nursing home clients (better to request co-workers to ask their clients if they would be willing to be contacted about the research). Think of how you would want any of your vulnerabilities protected.
A fourth major IRB issue is privacy. Subject privacy impacts on recruiting subjects, keeping individual data private when doing the research, and destroying data that is tied to individual subjects when the study is over. Think how you would like your privacy protected if you were a research subject.