What are the best practices when defining components that, when later loaded into a sandbox, end up in unique folders inside the sandbox root folder?

FYI I'm a relative newcomer to RTC, coming from extensive experience with CM Synergy.

I created a project area with two components. In one scenario, I accomplished the results I expected: when I look at the components' files in the repository, I see them grouped inside a containing folder that has the same name as the component, and naturally they load into a sandbox with the same organization. In another scenario, the components' files land in the repository without that containing folder, and when I load both components into a sandbox, the top-level files of each component are intermingled in the sandbox root folder.

The trouble is, I don't know exactly what I did differently in the two scenarios.

I understand how to load components such that they end up in the component-specific folders that I want, even though they are not organized that way in the repository. But I would really like to know how to ensure I get the repository organization the way I want.

Can anyone point me to documentation on the relevant concepts? For example of what doesn't help is the information on loading/unloading at http://publib.boulder.ibm.com/infocenter/clmhelp/v3r0/topic/com.ibm.team.scm.doc/topics/t_scm_eclipse_load.html, which tells me "if you want to do this, select the option to do this." and if you want to do that, select the option to do that". I am looking for something that will explain the difference, behaviorally and conceptually, between this and that.

Thanks in advance,
Scott