In my previous post, we talked about the “Demographic” or “Breakout” questions you could ask in your employee survey, and now I’d like to talk about the “how”; how to ask, and where to put them.
(To refresh, “demographic” or “breakout” or “segmentation” questions are used to segment groups within the organization for reporting; such as “Department”, “Location”, “Role”, “Gender”, etc.)
Should I make some or all of my demographic questions “required”?
By “required” here I mean that the person cannot submit the survey without choosing an answer to the question.
But here’s why you should be cautious in pulling the “required” trigger; if you make a question “required”, and they really don’t want to answer, (usually because they are concerned about anonymity), they might either:
- not submit the survey at all – which means you’ve lost ALL their data
- obfuscate the response by purposely selecting the wrong answer – which is of course far worse
Because of this, I would suggest that you make a demographic question required only if the results are useless without this information. For example a value for “Location” is critical, otherwise, the individual’s response will not appear in the report at all.
How about I make it required, but add “I would prefer not to answer” to the options?
We generally try to stay away from adding this or “n/a” or anything like it as a response option to any question, as it’s just WAY to easy to select as a choice. I mean, why not?
If you want to explicitly indicate that you are ok with them not answering a given question though, then by all means go ahead. Just be aware that this means you might not get as much detail from the question as you might have had without this option.
Where should I put my demographic questions in the survey?
First off, put them all in one place – don’t scatter them through the survey. Why? You’ll want your respondent to stay focused within the mindset they will be in when they are thinking about their work environment, and jumping back and forth between job-related questions and demographics will just disconcert and confuse them.
Given that, do you put them at the beginning or the end? (Putting them all midway would just confuse them again.)
We tend to lean slightly toward putting them at the end as the pros often (but not always) outweigh the cons, particularly in the case of employee surveys.
We’ve done both given the circumstances, and it will depend on how important or “obtrusive” your demographics are seen to be, but here’s what to consider:
Pros to putting them at the end:
- They’ve already got their “head in the game” after reading your intro, so it’s best to keep them in that headspace (re thinking about their work environment) as they start the questions.
- This is the most common layout for employee surveys, so it may be more familiar to the respondent.
- By the time they get to the end, they’re feeling more comfortable about the survey and what it’s about; so where they might have hesitated on seeing the demographics right up front, they might be more inclined to fill them out now.
Pros to putting them at the beginning:
- It can be a gentle lead-in to the survey (an ice-breaker so to speak) as the questions are easy to answer – but as mentioned above for this type of survey it might be more disconcerting than helpful.
- If they stop halfway through at least you have the demographics for what they HAVE done.
- Putting the demographics at the end “might” result in them going through the entire survey and then realizing you are asking questions that they are not willing to answer. They might then feel frustrated or “tricked” and just abandon the survey altogether.
It can be a judgment call either way, but if you’re not sure which is better, I’d suggest you default to putting them at the end.
How should I ask them? What options should I use?
There are two types of demographic questions – categorical (ex. “Location”, “Department”), and scalar (ex. “Years with the organization”, “Age”).
Categorical: In the case of categorical questions, the only consideration really is whether the resulting groups will be big enough to report out for anonymity’s sake. (We usually draw that line at seven responses – below that we don’t report out at all.) If you think a group is too small, you might need to merge the group with another when that makes sense.
Scalar: If the scales are small enough, say 7 or less, just use the scale – otherwise you are going to want to set up ranges that make sense in regard to whether they are actionable.
This is particularly true when addressing “Age”. Beyond that fact that it’s unlikely that you are going to develop one program for 35-year-olds and another for 36-year-olds, the more granular you get, the more nervous the respondent is going to get – and then they will either not answer or lie to you.
So in this case, “Generation” works well, or use multiple ranges for “date of birth”. (Yes, in my experience people are more comfortable giving a DOB than age.)
A bonus suggestion on scroll boxes: You know those things you use on a bank or other website to show your year, month, and day of birth? Don’t use them. In the days of wheel mice (mouses), it is WAY too easy to think you are scrolling to the next page, when in fact you’re changing the answer in the scroll box that is now off the top of the screen.
So the short version?:
- Don’t make demographic questions required unless you absolutely have to (there is a cost)
- Don’t use “I would prefer not to answer” unless you HAVE made the question(s) required.
- When in doubt put these questions at the end of the survey.
- Don’t get too granular in your categorical or scalar response options. Ask yourself:
- are the groups are too small to be actionable?
- and will the respondent be freaked out by the detail and so not answer? (I mean, how many 36-year-olds ARE there in Accounting?)
Hopefully, that answers your questions on setting up your demographic questions, but as always give me a shout if you would like to discuss.
… and remember as mentioned in the last post to communicate to them that you are not going to cross-reference the demographics.
Next up: Pro Tip Series 20 – What are “Double-Barrel” questions, and how do I avoid them?