In part one of Josh Glendinning's research tips he discussed how to be creative with research - in part he begins by explaining the importance of testing, testing, testing.

5. Test, don’t ask

People lie, both in real life and in surveys. However, by constructing your survey in the right way, you can start to see through some of these deceptions, providing you with unparalleled insights.

There’s two ways to go about using research to test ideas. The simpler version is to use a quiz format. You can look at the results work out the biggest gaps in public knowledge – whether it’s about British history or the prices of supermarket products.

The second, more sophisticated approach usually involves splitting your sample in two or more ways and comparing the differences. For example, rather than asking people whether they think chocolate influences their decisions, why not ask them to think about chocolate and then measure the difference it makes to their decisions compared to those who haven’t thought about it?

If you can show a real difference between your groups, you’re onto a really powerful story that not only describes, but explains.

Key tips

• This kind of approach is a bit fiddly and requires a careful construction, but if you have an idea, get expert help to find the right methodology
• Think about words – naturally, we can’t ask a survey respondent to do anything in the real world, so this approach works best when you’re thinking about the impact of words and communication on beliefs and attitudes

Some examples

People change their looks depending on their name, study finds (Hebrew University in Jerusalem)

Public more likely to ignore experts if science is too easy: study (University of Münster)

6. One survey, many stories

Every piece of research is a huge opportunity to produce content. While you may cherry-pick the juiciest headlines for your main launch release, there will often be a wealth of data available that never sees the light of day.

For most surveys, people’s views are unlikely to change significantly over time, so the results are still perfectly valid even a year after the poll is conducted, barring any drastic events in the meantime. By dividing your survey into distinct sections, you can ensure you have multiple pieces of content for months or even years to come.

This works especially well if you’re reaching a niche or powerful audience – most of the cost comes from contacting them in the first place, so why not invest a little more up front to significantly increase the ROI over time?

Key tips

• Be distinctive – try to make sure different sections are sufficiently distinct: if journalists perceive you to be issuing the same story, they might switch off

• Snackable data – a press release can only hold so much information, but if you package each interesting data point into a tweet, it can sustain a social media account for a week

7. Think visually

Surveys have come a long way since the days they could only be conducted by phone. With online questionnaires, you can display multimedia prompts, such as videos or images, adding an extra dimension to your research and improving the quality of your data.

Images can be used as a stimulus, which is especially useful if you’re testing things that might not be familiar to respondents – such as a new product or piece of technology. If you’re asking about the attractiveness of celebrities, why not include a picture of them to help jog people’s memories?

Other possibilities include asking how visual stimuli make people feel or whether they recognise an image, such as a face or a brand logo.

Key tips

• A picture is worth a thousand words – if you’re describing something in a questionnaire, have a think about how it could be expressed as an image
• Source material – we can suggest appropriate images to test but it’s always best if you can supply them yourself or have a clear idea of what they should be

Some examples

How Audrey Hepburn is the ultimate beauty icon, according to new survey (Superdrug)

Four in ten Brits can't find London on a map, and 85% struggle to find Birmingham (Ordnance Survey)

8. How big is that?

One of the biggest difficulties in communicating research is explaining what a number means: is 26% big or small? What about 54%? Percentages on their own are pretty difficult to comprehend. One of the best ways around this is to try to find a point of comparison – either within the survey or afterwards.

If your survey finds that people spend 1.5 hours a day on their computer, how does this compare to what they said about talking with friends or family? Or looking after their children? Alternatively, what else could you compare 1.5 hours to? A football match? The time it takes to make a cake?

Be creative. Sometimes the best comparisons are the most unexpected.

 Key tips

• Leave yourself plenty of options for points of comparison in your survey rather than just hoping that one or two turn out like you hoped
• Make sure you’re comparing apples with apples: questions that are phrased differently won’t make for valid comparisons
• The more unexpected the better:

Some examples

Three-quarters of UK children spend less time outdoors than prison inmates – survey (Persil)

REVEALED: Dog owners spend more time photographing their dogs than walking them (Kennel Club)

Average Brit will spend one year and eight months of their life in a bad mood, study finds (A. Vogel Herbal) Remedies)

So there you go, eight simple steps for creating research that will get you, and your brand, featured in all the right places. So be original, be thorough and do your research!

Josh Glendinning is research manager at Opinium Research