Canmark will assist you in developing appropriate survey content based on your specific needs and will provide continuous feedback and support in taking your survey concept to the field. Our expertise, developed through hundreds of survey design and implementation projects, will result in contributions to your survey project that cannot be quantified or invoiced, rather, they are simply a benefit of partnering with Canmark.
Sample Survey Questions
Survey questions can be of any format conducive to obtaining the data needed, however, the question should be carefully constructed so as to not impact the integrity of the data acquisition process.
While there is much to be learned from
open-ended response type questions, there is often a need for numeric scores or for groupings based on survey responses. Questions with a fixed number of response options are therefore necessary. Survey questions should be considered for the current need, as well as for likely tracking and comparison needs in the future.
For ease of use and to support your analytical needs, the majority of survey questions will likely be limited to response ranges of positive-neutral-negative. Respondents select their answers by clicking an item, checking a box next to the item, circling the item, etc.
Unless there is a compelling need otherwise, questions requiring a rating response should use a balanced scale. Response options should follow the basic model of:
| Very Good |
Good |
Neutral |
Bad |
Very Bad |
Alternately, the commonly used positively biased scale may be used if it is thought that respondents are negative-averse, meaning that respondents would tend to force answers to the positive side of the scale, distorting more meaningful gradations of score.
| Excellent |
Very Good |
Good |
Fair |
Poor |
From left to right, the numeric values typically associated with the rating terms are 5, 4, 3, 2 and 1. If the mean (average) score is 3.0, this is "Neutral" (neither good nor bad) on the balanced scale, while on the positively biased scale that same 3.0 is "Good".
Note that "Not Applicable" or "Not familiar to me" response options are included as appropriate, but do not constitute a rating.
Sliding scale survey questions can also support your data gathering needs. These usually use a horizontal line with Minimum, Center/Neutral and Maximum points indicated. The survey respondent simply indicates where their response falls with a mark on the line. This is typically an "X" in traditional paper surveys, and a mouse-click for online versions.
The Canmark data capture processes then simply identifies where the mark falls on the line, and since the line is of a known length, a numeric score is generated. This example would yield a score of 7.6 out of 10. Furthermore, with the use of this methodology, the score can be easily converted to any desired rating scale (five-point, seven-point, etc.). For example, this response is 6.8 on a nine-point scale, and 3.0 on a four-point scale.
Ranking questions are used extensively in some surveys. The respondent is asked to rank order a group of items. This is usually done by listing the items followed by a blank line in which to indicate rank position (1st, 2nd, 3rd, etc.), or by providing a randomized list of items and asking the respondent to write them out in rank order.
Another type of ranking question is a summing-type, relative rating, question. Here the respondent is asked to allocate some number of total points among a number of attributes or options based on a characteristic, such as importance.
The resultant data gives a clear indication of perceived or stated customer needs, and not only addresses attribute ranking, but also indicates the relative significance of each item with respect to the others. A tremendous amount of data packaged in one simple question!
Other question types commonly seen on surveys include basic Yes/No questions, as well as standard multiple-choice and pick-from list questions. Canmark online methodologies allow us to provide a constantly changing randomized list to remove any response order bias, as well as provide a more exhaustive list of options when this is in the best interest of the project.
