Survey guide

Setting up the survey...

• Have you included all of the profile points needed to define your target audience?
• Will this population yield significant results (e.g., is it the correct target. is it the appropriate N)?

A good survey yields good results and one that...

• Reads well and is easy to understand
• Has no spelling or grammar errors
• Has the right amount of info (not too much and not too little)
• Does not include industry jargon that might not be readily understandable by a typical consumer, or where necessary terms are explained
• Has the appropriate verb tenses
• Each question is straightforward for a typical person to answer (i.e. put yourself in the shoes of different people and try to answer)

 

Regarding questions and question types...

• If the question has images / videos, do they match the description and answer options
• There is appropriate routing both on a question and answer option level
• The type of question is correct (e.g. 'choose many' OR 'choose one')
• Is matched with appropriate instructions in the question (e.g. 'Select all that apply' OR 'select one')
• Answer options are randomized if appropriate (note that the 'None of the above', 'Not sure / Other' options are always automatically fixed at the bottom)
• Appropriate scales are used (e.g., very positive, somewhat positive, neutral, somewhat negative, very negative) and appear in logical order (i.e. not randomized)
• The answer options cover all possibilities or they need a 'None of the above' or 'Not sure / other' option
• If the question will not apply to everyone, then a 'Not applicable' option is needed
• When including 'None of the above' and/or 'Not sure / Other' options, please use the buttons provided so that those options are fixed at the bottom, rather than adding those options manually in the response option list.
• If a respondent could answer about more than one experience or example, there is an instruction on which to think about when answering (e.g. think about your main bank account)
• If the question is personal/ sensitive, you should include a 'Prefer not to say' option
• Please don't make assumptions about people's situations (e.g., not everyone owns a smartphone, not everyone has a sibling)
• Avoid overlapping scales, especially for frequency questions (e.g., weekly, several times per month, once every other month)
• Remember that if YouGov Surveys already has a profiling attribute, you do not need to ask that question in the survey - it will be available as a filter in the results
• If the question is measuring sentiment, use an equal number of positive and negative statements
• Requirements when asking about Sensitive Personal Data: YouGov Surveys allows clients (you) to ask its members (people who respond to surveys) about Sensitive Personal Data, but a “prefer not to say” answer option must always be included. Sensitive Personal Data is defined as any question relating to the following topics: racial or ethnic origin, political opinions, religion or philosophical beliefs, trade union membership, health, a person’s sex life and sexual orientation.

 

 

Reviewing the survey as a whole...

• How is the flow? Does the ordering seem reasonable? There is no 'right' or 'perfect' survey, but make sure you are comfortable with it.
• Does the survey avoid any biased or leading questions?
• You may want to have a colleague review your draft survey - a new set of eyes can sometimes pick up on things that you might overlook after working on the survey for a while! 

 

Ending the survey...

• Did you check how it would look on the preview function?
• How many responses are you looking to receive based on the estimate at the top?
• Are you happy with the budget needed to fulfill this?
• Do you understand that your account will be charged the 'budget' amount shown?
• Do you want to save the survey as a draft for now to edit later or do you want to publish it immediately?