IMHO Radio buttons should retire. The ComboBox (Drop-Down list mode) should always be preferred.
Drop-Down list takes minimal screen space, and you can add/remove items programmatically.
No need to resize anything (hard), or disable irrelevant options (ugly).
Can you think of a situation when a Radio button is still useful?
Interfaces where there's not much on the screen, such as Wizard pages or 10-foot UIs (such as Media Center).
Also, for touch interfaces, it's easier to tap a radio button than to navigate a drop-down list.
But I agree that radio buttons take up potentially valuable screen real estate - so I'd only use them if there are very limited number of options.
When there is a limited number of options, radio buttons are usually preferred. Things like choosing your sex where there'll only be two options in the foreseeable future, I much prefer radio buttons since it's quick for you to see your options.
In most other cases, I agree, dropdowns are to be preferred.
Radio buttons are a great way to get one option from a user.
So you make a report manager for your users that lets them see sales this year, last year, this month or last month.
Put these at the top of your manager so you user chooses only one. Then fills in drop downs etc for extra filtering
Radio buttons should be used:
Generally, according Jakob Nielsen's guidelines, you should prefer usage of radio buttons to comboboxes, which are much less accessible to distinctive clear.
From rant:
Dropdowns should be used when the options in the dropdown follows a sequential pattern of some kind - ie dates, months, placenames, counters, booleans, phone number type (mobile/home/work) etc.
Where they shouldn't be used is when each option doesn't nicely fit a sequence, or the user can't 'guess' straight away what option to select - i.e. when selecting different 'states' which will usually require the user having to select the dropdown and reading through each option before they can determine what to put in.
(Meaning: radio-button are more appropriate in this case)
If the dropdown contains options that aren't obvious what they will be, every time someone hits that page, they may have to select the dropdown anyway to be able to scan the available options, and make sure they HAVE got the right one.
Radio buttons are quicker and easier to use because you can immediately see all options without interaction, and you can select you choice using a single click. If the number of options are limited (say 5 or less), radios are the better choice.
If finding screen space for five radio buttons are a problem, the UI probably needs a redesign.
In brief the radio buttons are very helpful when there are few choices available. What immediately comes to my mind are Yes/No questions or questions in surveys, for example ranges from Stongly Agree to Strongly Disagree. Since all the options are laid out on the screen (most likely horizontally) it is sort of a scale on which the user can provide the answer. I don't think the same will look as nice with a combo box
You should still use them when:
*) From a programmers point of view, I very much prefer the combo box, though ;-)
* Yes, I do mean that we are stuck with it, whether other solutions are better or not...
Mark "gender" example is just proving my point.
After the product is shipped, customer / users demand allowing them a third selection: “Leave Unspecified”.
And while radio button allow you to easly see aviable options, Drop-Down list allow you to easly see what is selected. Most of the time users are just looking at settings, changes are less frequent.
Radio buttons have two chief advantages over combo boxes:
They’re faster for the user, being selected with one click rather than two. A typical click takes 1.2 to 2.4 seconds. Translate that into equivalent CPU cycles and you see that a little work optimizing the UI is worth a lot of work optimizing the code.
They provide better self-documentation of the control without having to click anything. Seeing all the options rather than just the default tells the users more about whatever they’re seeing or setting. For example, seeing “Answer Type: Right” is ambiguous while seeing “Answer Type: “(o) Right – ( ) Left” is better. Seeing “Priority: High” has different implications than seeing “Priority: ( ) Critical – (o) High – ( ) Medium – ( ) Low”
Both radio buttons and combo boxes show the user the current selection.
From the user’s standpoint, the only disadvantage of radio buttons are the screen real estate they consume. Thus, radio buttons are preferred over combo boxes unless the real estate is better used for something else. This holds irrespective of the number of options, although, obviously, the more options you have, the greater the chance there’s something better you can use the real estate for.
The only other consideration is that if you have a large number of options (say, 10 or more), combo boxes make it easier to read the current value, since the user doesn’t have to visually search a good portion of the page/window to find it.
As for the work necessary for programmatically adding options, my philosophy is that programmers should work hard so users don’t have to.
Disabling provides vital information to the user and should never be avoided because it’s “ugly.” Disabling is not equivalent to removing. Disabling suggests something unavailable can be made available, while removing suggests it’s never available. That it’s not possible to disable an item in a combo box list is another disadvantage of combo boxes over radio buttons.
Use radio buttons when
You have a small list of choices
The choices are mutual exclusive
You need to let the user see all of the choices all the time.