Primacy and recency effects describe how the position of an answer option in a survey response list can influence the likelihood that it will be selected. The primacy effect is more common when respondents read options, while the recency effect often occurs when options are heard, making response-order design a material source of survey bias.
What are primacy and recency effects?
Primacy and recency effects are response-order effects observed when survey participants choose from a list of answer categories, brands, statements, product attributes or other alternatives. They occur because respondents do not always process every option with equal attention, memory and cognitive effort. Instead, the position of an option in the list can affect its probability of selection independently of its substantive relevance.
The primacy effect occurs when options presented at the beginning of a list are selected more frequently than options placed later. This pattern is particularly relevant in self-completion surveys, including online questionnaires, mobile surveys and paper forms, where respondents can see several answer options at once. Early options may receive greater attention, be processed more carefully or be chosen before the respondent fully evaluates the remainder of the list.
The recency effect occurs when options presented at the end of a list are selected more often. It is especially associated with interviewer-administered telephone surveys, voice interfaces and other auditory formats. When respondents hear a sequence of options, the most recently presented alternatives may be easier to retain in short-term memory at the moment of answering.
In market research, primacy and recency effects are not interpreted as evidence that respondents have changed their actual preferences. They are measurement effects. A respondent may express a different preference solely because an option appeared first, last or in a more cognitively accessible position. This is why how response order affects survey results is an essential methodological question in questionnaire design, fieldwork planning and data interpretation.
These effects are most likely to matter when response lists are long, options are similar, respondents are under time pressure, the subject is of limited personal relevance or the question requires comparison across many alternatives. Examples include lists of competing brands, reasons for purchase, perceived product benefits, communication messages, service improvements and barriers to adoption.
Application of primacy and recency effects in practice
Understanding primacy and recency effects is necessary whenever a survey uses a closed-ended question with multiple non-ordered response categories. The objective is usually not to exploit the effects, but to reduce their influence so that the measured distribution reflects respondents’ views rather than the construction of the questionnaire.
In quantitative studies, the most common preventive procedure is answer-option rotation. The order of eligible response options is changed between respondents, either randomly or according to a controlled rotation scheme. As a result, no single answer consistently benefits from appearing first or suffers from appearing last. The original position of an option can also be retained in the data file, allowing analysts to check whether position affected selection rates.
Primacy and recency effects should be considered in particular in the following research situations:
- Brand awareness and consideration studies – lists of brands should usually be rotated, provided that the category does not have a meaningful natural order.
- Concept and message testing – alternative product claims, advertising messages or package descriptions should be presented in balanced sequences to avoid systematic advantages for the first or last concept.
- Customer experience surveys – lists of service attributes, improvement areas or reasons for dissatisfaction require careful ordering when respondents can select several items.
- B2B decision-maker research – lists of supplier-selection criteria, technology barriers or expected business benefits may be long and cognitively demanding, increasing the risk of position bias.
- Telephone and interviewer-administered surveys – recency effect risk is high when answer lists are read aloud, especially if respondents cannot see a showcard or screen.
Not every list should be randomized. Some answer categories have a logical, chronological or legally defined order. Examples include frequency scales from “never” to “very often”, age ranges, stages of a purchase process or financial ranges. Randomizing such lists may reduce comprehension and create a different measurement problem. In these cases, the order should remain stable, while researchers should ensure that labels are clear, mutually exclusive and easy to process.
Hume’s Institute may account for primacy and recency effects during questionnaire programming by rotating nominal answer lists, fixing logically ordered scales and checking response distributions by displayed position. Such controls are especially relevant in studies where small differences between brands, propositions or segments may influence business decisions.
Primacy and recency effects and related research methods
Primacy and recency effects belong to the broader category of survey response effects. They are related to questionnaire design, mode effects and cognitive processes in answering, but they should not be treated as the same phenomenon as other forms of order bias.
Most importantly, primacy and recency effects differ from the question order effect. A question order effect concerns the influence of earlier questions on answers to subsequent questions. For example, asking about recent service problems before measuring overall satisfaction may lower satisfaction ratings because negative experiences become more salient. Primacy and recency effects concern the order of answer options within one question, not the sequence of questions across the questionnaire.
They are also distinct from acquiescence bias, satisficing and social desirability bias:
- Acquiescence bias is the tendency to agree with statements regardless of their content.
- Satisficing occurs when respondents select an acceptable answer rather than making a fully considered choice, often because the task is demanding.
- Social desirability bias occurs when respondents give answers perceived as more socially acceptable or favourable.
- Mode effects refer to differences caused by the survey channel, such as web, telephone, face-to-face or mobile research. Primacy and recency effects can be one mechanism behind mode effects.
In mixed-methods research, quantitative evidence of a possible response-order effect can be examined through qualitative follow-up. Cognitive interviews, for example, help identify whether respondents understood all options, remembered the full list or selected early answers before hearing or reading the complete set. This approach does not eliminate primacy and recency effects by itself, but it can improve the design of subsequent questionnaire waves.
How to reduce response-order bias
Managing primacy and recency effects requires deliberate questionnaire design, appropriate programming and quality control during analysis. The best solution depends on the survey mode, question type and whether the response categories have a meaningful fixed order.
The following practices reduce the risk that response position will distort findings:
- Rotate or randomize nominal response options, such as brands, purchase drivers or campaign messages.
- Keep logically ordered scales in a stable sequence and use clearly differentiated labels.
- Use showcards, visual lists or screen displays in interviewer-led research where feasible, reducing reliance on auditory memory alone.
- Limit the length of response lists or group long lists into understandable thematic sections.
- Place “Other”, “None of these” and “Don’t know” options separately when appropriate, rather than rotating them with substantive categories.
- Test questionnaires across the intended devices, particularly mobile screens where respondents may not view all answer options simultaneously.
- Analyse selection rates by displayed position when response order has been randomized.
Primacy and recency effects cannot always be removed completely, because survey participation involves attention, memory and practical decision-making shortcuts. They can, however, be anticipated and controlled. For market researchers, the central principle is that answer-list order should reflect the logic of the measurement task, not become an unintended factor shaping the reported result.