Throughout my career, the survey research landscape has continually evolved in response to a changing world, and the pace of that change has accelerated dramatically over the past couple of decades. Response rates have been declining for decades, budgets and timelines continue to tighten, and the way people communicate has fundamentally changed: landlines have nearly disappeared, mail is often discarded or goes unopened, unknown numbers go to voicemail, and a survey request is just one more email or text competing for attention.

Accordingly, no single survey mode is adequate for reaching most target populations anymore. That’s why mixed-mode design, which strategically combines multiple ways of reaching sample members and collecting their responses, has become a cornerstone of modern survey research.

The logic behind mixed-mode survey design is straightforward, even if the execution isn’t. Mixed-mode approaches are important because:

• Web-only designs exclude people with limited internet access or lower digital literacy.
• Phone-only designs miss the growing share of households without landlines and people who do not answer calls from unknown numbers.
• Mail-only designs lose anyone who doesn’t open an unfamiliar envelope.

By offering multiple ways for respondents to be contacted and to respond, mixed-mode surveying improves representation. It also reduces nonresponse bias by giving people accessible and convenient ways to participate. Meeting people where they are and making it easier for them to respond is often the difference between a respondent sample that resembles the population and one that doesn’t.

When designing studies, it helps to treat contact and data collection as two separate levers that work together rather than a single design choice. Contact is about how we reach out to and motivate sample members to respond, whether by mail, email, text, phone, or in-person outreach. Data collection is about how people complete the survey, whether through the web, paper questionnaire, phone interview, or in-person interview.

There are three commonly used design strategies for combining contact modes, data collection modes, or both.

1. Sequential designs introduce one mode at a time, typically starting with a lower-cost option like email or mail before escalating to phone or in-person modes for nonrespondents.
2. Concurrent designs offer multiple modes at once, which makes sense when a field period is too short to sequence modes or simply to accommodate respondent preference.
3. Within-mode strategies allow respondents to move between modes mid-interview, such as switching to a self-administered module for sensitive questions during an otherwise interviewer-led interview, to encourage more honest reporting by allowing respondents to answer sensitive questions privately.

Mixed-mode designs come with tradeoffs. While they can increase response rates, improve representation, and reduce nonresponse bias, they can also introduce measurement differences across modes. The same person may answer questions differently depending on whether an interviewer is present or whether a question is read aloud versus presented on a screen.

Research has shown, for instance, that survey participants tend to give more socially desirable answers in the presence of an interviewer. This is a design consideration that deserves careful attention. That said, the accumulated evidence is consistent: concerns about mode effects are typically outweighed by the gains in representation from reaching more of the target population.

Rather than avoiding mixed-mode designs, researchers should deliberately plan for these differences through careful instrument development and testing across modes. In practice, many mode-related differences can be mitigated through deliberate questionnaire design choices, such as standardizing question wording and presentation across modes and using self-administered formats for sensitive questions. Rigorous testing before fielding helps ensure respondents interpret and answer questions consistently across modes.

These examples show that there is no universally optimal mixed-mode design. Effective designs are tailored to the population, the study goals, and the stage of participation. Success comes not from offering more modes, but from using the right combination of modes at the right time. Mixing modes requires matching a deliberate combination of contact and data collection strategies to a study’s specific target population, topic, timeline, and budget.

As communication habits keep shifting and pressure on budgets keeps rising, that kind of tailored, evidence-based design, grounded in real experiments rather than a one-size-fits-all playbook, is what will keep surveys representative, credible, and sustainable for years to come.