How to beat sneaky survey bias!

Bias is the sly villain to all surveys. It has the power to send decision makers on a wild goose chase, trying to solve non-existent problems. All the while exhausting resources in the process.  However, sometimes we’re an unknowing accomplice, unaware that we invite bias in to wreak havoc with survey results.

Don’t worry, once you know what to look out for, it’s easy to avoid and below are the 4 main culprits.

1. Undercoverage bias

This happens when some members of the population are inadequately represented in the survey. It’s generally due to the method of collection not being accessible to the whole population. Limited survey channels mean responses are mainly received from patients with a preference or access to that mode. This is not inclusive to every demographic and integrity is compromised.

Solution: Use a mixture of methods to gather your data, e.g. Text, IVM, online, paper, Agent calls, for a true reflection of care experience. Remember, what’s important to one patient group, may be irrelevant to another, it’s impossible to pick up on these trends from just one survey channel.

2. Response bias 

This is when people respond differently to how they truly feel. This is common in a patient experience setting when the healthcare worker directly asks a patient how they felt about their care. It can also occur when a patient completes the survey in the provider’s environment, feeling their future care may be compromised if they suggest improvements or rank service poorly.

Solution: To minimise this bias, where possible, questions should be asked in a neutral environment and the identity of the questioner or questioning method should also be neutral.

For the most truthful responses, reinforce anonymity of feedback and create a space between the health service provider and the patient when surveying. (We recommend 48 hours after discharge)

3. Volunteer biasSurvey bias

 This occurs when we ask people if they would like to be included in the survey using an Opt-in approach. What may happen is that only people who are outgoing or have strong opinions are included in the sample and results are then skewed. Opt in methods result in lower response rates too!

Solution: Use an opt out process of sampling across all patient experience surveys.

4. Bias lurks in findings too!

There is one last hurdle to overcome before achieving the survey gold star of high quality data.

Confirmation bias is the inclination to selectively interpret data so the results align with current culture or ideology.  This is not on purpose, rather a sub-conscious mental error that can place more weight on data that supports one view whilst dangerously overlooking contradicting evidence.

It does happen, a survey creator can focus on findings that support their own opinion and discount the fact that it is only based on few respondents.

Solution: Take the human element out of interpretation and make sure you have the right software analytical and reporting tools to make sense of data without any emotional influence or predetermined mind set.

By avoiding these 4 common mistakes, you create a better survey experience for respondents, encourage participation and end up with honest and accurate feedback to confidently make service development decisions.

If you’d like to explore how to achieve the best quality survey data, we’re here to help! Phone 0845 9000 890 or email for a hassle free chat about different feedback options.