Size of Survey (sample size)
Make sure the sample is robust enough to support the claims you are making. Please note this is different from making sure that the sample is large enough to impress the journalist! Statistical reliability (OK I know it’s duller than watching paint dry) is generally determined by the size of the sample and not the size of the universe (population) you are seeking to represent. So, rather than automatically assuming you need 1,000 interviews for any study, start by thinking about the way you might want to use the data. If, for example, you want to be able to compare the attitudes of IT Managers in France, Germany and the UK you need a minimum number of completed interviews from each country. Technically you could do this with as few as 50 completed interviews in each country but in reality 100 or 200 would make the results significantly more reliable. In short start from the bottom up i.e. decide how you want to use the results and structure the sample accordingly.
Be cautious of using research to arrive at “absolute” numbers rather than percentages
Although specific numbers e.g. UK industry wastes £200 million a year on energy, can make great copy it’s really hard to substantiate claims of this sort (i.e. to say how confident we are that the figure is accurate) through market research unless you use very large samples (literally 1,000’s). Again this is a function of statistical reliability but in simple terms it’s much harder to be confident about a result when respondents have an almost limitless range of options. e.g. How much do you spend on energy per annum (Spend on energy for example could range from £100 to millions in the case of a steel plant) rather than a question with a limited number of options e.g. Is Chlamydia
a. A flower
b. A Latin verb
c. A sexually transmitted disease
d. A member of the Spice Girls
In the latter case there are only 4 options and consequently we can be far more confident about the accuracy of the answer we get and justify it with statistics. In fact if you feel you need statistics on specific numbers it’s probably better to use information that’s already available as you can reference them, quote the source but ultimately you’re not responsible for the accuracy of the figure you quote.
Use Research to uncover original information
To be honest a survey isn’t “bogus” just because it fails to come up with anything that’s genuinely new or exciting but it will have more value, particularly from a PR perspective, if it is original and journalists can’t simply dismiss it with a shrug of the shoulders and a “tell me something I didn’t know”! So it’s worth spending some time before a study is commissioned looking to see what’s already been done and trying to come up with an original angle.
Don’t conduct research that directly promotes the company
Another “Golden rule” is to avoid conducting research that’s obviously been designed to put the sponsor of the research on a pedestal. Most journalists see through this pretty quickly and a lot more value can be derived from a survey by tackling an issue, raising its profile and then associating the company sponsoring the study with that issue e.g. The Terrence Higgins Trust and the education that’s need on sexually transmitted diseases . It works just as well, has more credibility and can help to develop a debate around the topic which the company can benefit from in the medium to longer term.
Make sure the research focuses on specific issues rather than the opinions of a few- Avoid “Thought Leaders”.
All too frequently polls are conducted that simply solicit the opinions of a few so called “thought leaders” with their comments then being used as definitive findings. No doubt they have a view, their view might be representative and indeed their views on issues might be an accurate reflection of the way in which it or a market might develop. However, their opinions are not statistically reliable and they are not necessarily an accurate reflection of the situation or issue. Rather than conduct opinion polls amongst supposed “thought leaders” and then represent these as a definitive truths it’s quite often better to find a way of using market research to look at an issue in an original fashion, come up with new findings and then use these to stimulate debate. Of course there’s nothing to stop you asking thought leaders for their opinions on the findings of a study but fundamentally you’ve used the research to come up with something new and that, after all, is the purpose of research.
Don’t use the research to replace serious thought and analysis.
There is a danger that research can become an end in its own right with results being published simply to support a point e.g. 9 out of 10 cats prefer Whiskas. As an advertising slogan it works but in truth it doesn’t promote much thought and analysis (in our experience cats are generally pretty poor conversationalists) and it’s better to use the research as a focal point for genuine thought and analysis. Of course we’re not saying that statistics and sound bites don’t work – they do – as the Chlamydia article example proves but this is a classic example of how coverage can be maximised with its sponsors (The Terrence Higgins Trust) using the research to engender debate and highlight what needs to be done.