Ergo, researchers should also seriously consider giving research participants the option to answer, “Does not apply” in surveys and questionnaires.

Researchers make a big mistake if they do not give respondents the option to answer “Don’t know.”

‘Nuff said.

For example, consider the following sample survey question:

How do you feel about President Barack Obama winning the Nobel Peace Prize?

• Extremely unexcited
• Somewhat unexcited
• Somewhat excited
• Extremely excited

Unfortunately, writing a question in this manner does not give respondents the opportunity for a neutral response, such as “Neither unexcited or excited.”

This same rule also applies when using numbers as answers to a question. For example, the question above could be modified as follows:

On a scale of 1 to 10 (with 10 being high), how excited are you about President Barack Obama winning the Nobel Peace Prize?

1     2     3     4     5     6     7     8     9     10

The problem with such a numerical scale, or course, is that the “middle ground” is actually 5.5, which is not an option.

A better numerical scale would ask participants to rank their answer on a scale of 0 to 10 or 1-9, where (in both instances) the neutral answer is 5.

Simply put, when it comes to conducting research, the number of individuals contacted in any form of research study truly does matter.

When considering any target market it is extremely rare that you’ll be able to gather data from every single member of the targeted population. Hence, what you’ll do instead is gather info from a representative sample group of the entire population as a whole.

When defining a sample group, typically you can include people based upon a variety of factors, with demographic information (such as age, sex, education, etc.) and psychographic information (personality types, attitudes, lifestyles, etc.) the most common defining factors for inclusion.

However, even if you successfully select a representative sampling of your entire population, if you do NOT collect data from enough individuals, you run the risk that your research results will be statistically less and less valid (or applicable) across the population as a whole.

Ergo, the smaller the population, the greater your sample size needs to be (percentage-wise) to generate statistically valid results. Conversely, if you have a very large population to draw from (say the population of the United States), then the sample size can be much, much smaller on a percentage basis. That’s why you’ll often read about national U.S. surveys that have a confidence level of +/- 3% or 5% yet involved less than 5,000 participants.

So . . . at the end of the day, when it comes to research, remember that sample size does matter.

= = = = = = = = = =

Since many public relations campaigns start first with data gathering and analysis (and if you don’t, you should), I thought it might be appropriate to share the next several thoughts/ideas on ways to improve your primary research.

Primary Research Tip No. 1: Remove Emotion from your Surveys

Although you may begin a research undertaking with a preconceived notion of what your research results will show or prove, you must do all you can do to avoid inserting your own bias into your surveys and questionnaires.

For example, consider this question:

Given that recessions are a terrible time to look for new employment, what is the likelihood that you will begin graduate studies immediately after receiving your bachelor’s degree?

The bias inserted into the previous question is the opening phrase of the question, specifically: “Given that recessions are a terrible time to look for new employment . . .”

In truth, recessions may or may not be a terrible time to look for new employment. An unbiased way of rewriting this question is to simply remove the opening portion of the question so it might now read like this:

What is the likelihood that you will begin graduate studies immediately after receiving your bachelor’s degree?

Remove as much bias as possible from research studies; this is critical to improving the validity of your research results.

You’d think that crafting a Yes-No question in a questionnaire would be easy, right? Well, not really.

Think about it: Have you ever been asked a Yes/No question and realized that your true answer should be

• I Don’t Know / Undecided, or
• Maybe, or
• Does Not Apply?

When writing simple Yes or No questions in a research study, we recommend giving people the option of responding with at least one other answer besides Yes-No.

Why? Because very few population samples can be divided into two categories, such as Yes or No.

P.S.  We recommend a similar approach when crafting True-False questions as well.