Part II Collecting and analysing data 7.5.1. Step one: Defining the research objectives The first step in conducting a survey is to define what the objectives of the research are; in other words, what does the researcher want to learn? Researchers should be able to state, clearly and concisely, their overall research goals, as well as the specific research questions to be answered. When designing the survey, you should refer back to these objectives on a regular basis to ensure that the questions to be asked are the right ones. Finally, in defining the research objectives, information that supports the NHRI’s advocacy objectives should be prioritized to ensure the efficient use of resources. 7.5.2. Step two: Defining the population samples In the next step, the population sample must be defined. In other words, the researcher must determine who they will survey to answer their research questions. This will involve choosing the target population, or what kind of people the researcher wants to survey, as well as the sample size, or how many people the researcher will survey. Determining the target population will depend on the parameters of the study; what sector of society or geographic region, for example, that the researcher is interested in. To help ensure the accurate reflection of relevant subgroups in the target population, quotas of people may be used. Determining the sample size will depend on factors such as the time available, the research budget and the necessary degree of precision. Generally, the larger the sample, the more precisely the sample reflects the target population. This can be calculated with the help of online tools, however, it is important to be familiar with two key concepts: confidence intervals and confidence levels. Confidence intervals – also known as the “margin of error” – is the plus-or-minus figure usually reported following results. For example, say you have a confidence interval of 5%. If 38% of your sample answers “no” on a survey then you can be “certain” that between 33% (38-5) and 43% (38+5) of the entire population would also answer “no”. How “certain” you are of this result depends on your confidence level. This is a percentage that tells you how often the true percentage of the population would pick an answer that lies within the confidence interval. For example, a 95% confidence level means you can be 95% certain that the true population would choose an answer within your confidence interval. Most researchers in the social sciences use a 95% confidence level. 7.5.3. Step three: Designing survey questions Developing well-designed survey questions is critical. Researchers generally ask three basic types of questions. The most common type of question is multiple choice. These are easy to tabulate and compare. Multiple-choice questions can be basic, factual questions (e.g. “Where do you live?”), with a choice of answers listing different locations. Multiple choice questions can also work as rating scales and agreement scales; for example, respondents choose on a scale from “strongly disagree” to “strongly agree”, to rate how much they agree with a statement. Survey questions can also be numeric openended (e.g. “How old are you?”). Finally, survey questions can be text open ended or “verbatim” (e.g. “How can the company improve its working conditions?”). Chapter 7: Collecting primary data | 77