What are you hoping to learn from running this survey and who are you hoping to learn this from?
- Defining goals is an often overlooked initial step in planning your survey. To get valuable data with insights that you can take action on, you first need to define exactly what you’re looking to get out of running this survey, why, and from whom. Ask yourself these three questions to get started:
- What data am I hoping to collect from this survey and why?
- Who is my target audience?
- What do I want to be able to accomplish when I have my final results?
- Defining your goals may also involve seeing what other information is out there. Doing preliminary research into the question you’re trying to answer could save you time and money. It can also give you ideas for what’s missing from the current data and what specifically you’d like to answer.
Now that you’ve defined your goals and target audience, take careful steps to set up your survey.
Who are you trying to reach?
- Targeting options allow you to reach a variety of different audiences. Whether you’re looking to get feedback about your website, poll the general population, or target small business owners and managers, selecting the right audience during the survey set up process will determine who sees and responds to your survey. You should know, based on your defined goals, what’s important for you in an audience:
- Do I want a representative sample of a country’s population for a policy poll? If so, use general population targeting.
- Do I want to reach users of Android phones about a new product I’m developing? If so, target to Android smartphone users on the Google Opinion Rewards app.
How many responses do you need and how much are you willing to pay?
- Pricing varies based on the number of questions you ask. The number of responses needed is up to you.
- Carefully select the right amount of responses and screening variables to match your budget and data needs.
- It’s important to ask questions that are concise and specific (within the recommended character limit of 125 characters). To get clean data, make sure to use simple, clear phrasing that avoids jargon. This includes using familiar terms instead of technical ones, such as water instead of H2O.
- Example: Specify a time range
- Try asking How many times have you bought orange juice within the last 12 months? rather than Did you buy orange juice? as a way of finding specifics about consumer purchasing behavior.
- Avoid providing examples to the respondent. For example, consider the following two questions: How often do you eat vegetables? and How often do you eat vegetables (e.g. spinach, Brussels sprouts)? In the second question, the respondent is far more likely to only consider how often she eats those two types of vegetables, instead of how often she eats vegetables.
- It’s important to define the audience you’re interested in. You’ve already taken steps to do this while defining your goals, but now is the time where you can use questions to hone in even further.
- Ascribe Surveys offers several custom targeting options to help you ask questions to the perfect audience. To find out more about Ascribe Surveys, contact us.
- Example: If you only want answers from pet owners, start with a screening question such as Do you own a dog or cat?Then ask How many times did you take your pet to the groomer last year?
With web surveys, it’s a balance between providing enough answer options to get a valid sample and keeping the list small enough to avoid a lot of user scrolling. It’s important that you get to know all of our question types in order to find the best type for what you’re looking to ask and make sure that you’re providing enough answer options to include all respondents.
Avoid binary questions
- Try to turn questions with two opposite answers, like “yes” and “no”, into a scale. For example, instead of asking “Do you agree with the following statement? I like apples, ask, How much do you like apples? and give respondents options from “Not at all” to “Very much.” Validation tests have shown that incidence accuracy is better when binary answers are not used. Having three or more answers helps eliminate confusion for the respondents, provides a better user experience, and improves survey quality.
Have an opt out
- Some questions need an opt out for respondents whom they may not apply to. For example, you want to ask respondents if they have had their phone stolen. If you provide only two answer choices – “Yes” and “No” – you assume that all respondents have a phone. To make this question apply to everyone, you need to add an opt out, such as “I do not have a phone” or “I prefer not to answer”.
Selecting the right settings for each question is important to receive accurate data.
Find the right question type
- Understand the different question types offered and have a plan for which ones you’ll use. You’ll want to keep in mind that certain types of questions convey different meanings.
Understand partial ordering
- Survey questions have a maximum number of answer choices that can be displayed at one time to a respondent; seven answers for multiple choice questions and two answers for image choice questions. When the number of answer choices for a question exceeds the number that can be displayed at one time, a subset of the answer choices will be shown. If you want all respondents to always see the same number of answer choices, consider narrowing down your answer options or choosing an open end question that presents the respondent with a text box. If you’re considering having more answer choices than can be shown at once, first think about what is need to know versus nice to know. Ask yourself these two questions: Do I really need these extra response options? and What extra information will they give me?
- In a vertical list of multiple choice answers, people tend to click more on the top responses. By default, multiple choice answers will be randomized to reduce this effect. You can also randomly reverse the answer order if you want to maintain some order (e.g., ascending or descending numeric answer options). To avoid bias in your results, we recommend you keep your answers randomized in some way unless these randomization options don’t make sense for your question.
Pin specific answers
- There are times when you want to randomize answers but ensure that one option is always shown in a specific location. For example you might want to “pin” the answer option “I don’t know” so that it always shows at the bottom of a list.
- Make sure all questions comply with the Google Surveys program policies, which prohibit questions about certain topics and require adherence to editorial standards.