Survey Design Best Practices

The most crucial part of survey research, aiming for a clear, concise instrument that accurately measures intended aspects.

Include measures that evaluate sentiment regarding the subject being studied, assisting with tracking over time and correlation analysis.

Essential for most surveys, though unrelated to the primary research objective. They aid in describing sample representation and result segmentation.

A survey structure that begins with general questions and moves to more specific inquiries, useful for getting initial opinions before delving deep.

A survey structure by experience, organizing the questions in the order of experiences (e.g., a customer journey).

Questions placed before core questions to filter out respondents who don't qualify for the survey.

Questions that limit responses to specific options, making surveys quantitative.

Tools used commonly for conducting surveys, facilitating their accessibility.

Questions asked conditionally based on previous answers, only applied to a subset of respondents.

Ideally placed at the end of a survey, unless used as screeners, to avoid initial respondent fatigue.

The duration of a survey should be short (5-7 minutes max) for better response completion rates and reduced drop-off.

A long survey broken into smaller parts, dispersed over time.

A survey distributed through multiple platforms for wider reach or enhanced engagement.

Values representing distinct groups based on a characteristic or label.

Categorical data with no inherent order.

Categorical data with a clear order or ranking.

A scale from 1 to 5 used to measure how familiar customers are with a company or product. 1 represents not at all familiar, and 5 represents very familiar.

Open-ended questions used to understand customer perceptions of a company (e.g., "What words would you use to describe [company]?" ).

Questions measuring the degree to which customers believe a brand possesses specific attributes (e.g., "How much do you agree that [company] is focused on quality?" ).

Questions that suggest a particular answer or contain biases.

Vague or unclear questions that can be interpreted in different ways leading to inaccurate results.

Survey components include overall measures and demographic questions. Survey approaches include funnel (broad to specific) and pathway (experience-ordered).

Scales of measurement (categorical and continuous) apply to quantitative data. Categorical scales include nominal and ordinal. Continuous scales include interval and ratio.

Considerations for survey responses include: scale type (categorical or continuous), balancing scales, consistent comparisons, visual weight, response order, avoiding overlap, and including "not applicable/don't know" options.

Avoid complex, leading, loaded, ambiguous, double-barreled survey questions.

Surveys can gather feedback on new product directions (development), brand perceptions (unbiased), and customer/employee satisfaction (engagement).

A measurement scale with clear order and exact differences between values.

A measurement scale with equal intervals and an absolute zero point.

A survey response scale with options ranging from one extreme to another, sometimes with a neutral midpoint.

A Likert scale ensuring equal representation of positive and negative options.

The tendency for respondents to choose options at the top of a list more often.

Variables that can be ordered, but intervals between values are not equal.

Representations of formal education attainment, like High School, Bachelor's, Master's or Doctoral.

The perceived strength or importance a response option has based on presentation.

Scales that measure intensity (e.g., Likert scale). They can measure from positive to negative or negative to positive.

Include options like 'not applicable', 'don't know', or 'unsure' when appropriate.

Avoid using options that overlap. Each option should be exclusive; so respondents can pick only one valid option.

Questions designed to gather detailed answers, not just a simple choice from options.

Questions with multiple parts or stipulations; don't confuse respondents.

Questions that subtly suggest a particular answer; they bias the results.

Questions that might encourage a socially desirable answer (often influencing or pressuring the respondant).