It is best to keep your
survey as short as possible. The closer the relationship you have with the
group you wish to survey, the more tolerant they will be of longer surveys.
If their interactions with you are relatively limited and impersonal, such as in
surveys of satisfaction with a customer or a technical support interaction,
then you should consider a very brief survey of perhaps only dozen questions or
so. The closer your relationship is, as with large business-to-business
customers, your employees, or your dealers, the more likely they will be to
complete surveys and the less put-off they may be by surveys that are longer.
Our longest surveys range up to 100 items (including demographics), but we
generally do not recommend surveys that long. Fifty items tends to be
comfortable for an important survey, 75 is pushing the envelope, and more than
that is really asking a lot of your respondents.
Having said this, we
should also note that most people who open a survey will go on to complete it
and most of those who do not intend to complete it will never open it and see it
at all. So, length is only one factor in determining your response rate.
More important is the relationship you have with the intended survey recipients.
Consider that and stay within the boundaries of what you can reasonably expect
of someone you interact with at the level of the people you intend to survey.
There are techniques for
programming surveys that can allow you to present a common set of items to all
respondents and specific subsets of items to sub-samples of respondents.
Skips or branching can be programmed into Web surveys so that each person gets a
survey that, although it shares items with everyone else, also has items that are
only relevant to people in his or her subgroup or with his or her experience with your
company. Skips are when people who say "Yes" or "No" to a question skip
over items that follow but are not relevant. Branching is when they are
directed to separate pages that are specific to their subgroup (e.g., decision
makers) or experience (e.g., use of a particular class of products). Skips
are usually used on single page surveys and branching is used on multi-page
surveys. These solutions can make longer surveys appear shorter to an