After identifying your research problem and formulating your research question and objectives, you need to decide how you will answer your question. In other words, what data will you be collecting and how will you collect and measure it (what is your study design)? Will you use a survey? Make clinical observations? Use an intervention? If so, which intervention is this? Will you implement new training or new use of an existing medicine? A diagnostic tool? Or are you going to extract your answer from a data set you already have, such as your clinical records? ​ 

If your study includes human participants you need to also decide who your study participants will be (who your study population is, whom you will include and exclude), and where your study will take place (e.g. hospital ward, outpatient clinic, participants households, etc.).  

What you’re measuring in your study is usually the outcome; you’d like to see how your outcome differs based on what intervention your population received in comparison to some existing standard. Refer to the “Research Question” module for more details. Your population may help you define your study setting. For example, if your population consists of patients being admitted to your hospital, then that hospital would be your study setting. 

But how do you measure your outcome? One of the first decisions you should make is what sort of study design you would like to undertake. Different study designs are better for different types of research questions.  

  • Descriptive studies are better if you want to understand the current state of things. For example, if I wish to understand the current rates of handwashing in my clinic, I will do a descriptive study.  

  • Analytic studies are better if you want to understand how things can change, given some intervention. For example, if I wish to understand how rates of handwashing in my clinic change after I put up signs to encourage handwashing, I will do an analytic study.  

You can get more specific too. Within descriptive studies, there are different study types, and within analytic studies, there are different study types. The image below describes the different study types: 

Figure 1: Types of study design 

Now, you should decide what it means to “measure” your outcomes. Do you want to measure your outcomes in terms of numbers? Or is it more meaningful to collect descriptive information (i.e. words)? Numbers and descriptions have different advantages. For example, collecting numbers makes sense if your outcome is weight, blood pressure, or the presence/absence of some disease. These outcomes are absolute and have tools to help measure them. On the other hand, your outcome might be knowledge of appropriate bandaging practices for wounds. This isn’t so easy to measure and doesn’t have a readily available tool. Instead, it would probably be more helpful to collect a descriptive outcome by interviewing nurses and assessing whether they have adequate knowledge of bandaging practices through those interviews.  

The chart below can help you decide what it means to measure your outcomes:  

https://www.simplypsychology.org/qualitative-quantitative.html 

Now that you have defined what you want to study, your target population, and the most appropriate study design, the next question is how you will study your target population. For example, if your study objective is to determine the effect of regular exercise on type 2 diabetics, then your target population is all patients diagnosed with type 2 diabetes, but since it is impractical to include your entire target population of type 2 diabetics in your study, the best thing to do is to conduct your study on a smaller group that is representative of your target population. This process is known as sampling.  

There are two types of sampling methods: Probability sampling and non-probability sampling. 

  1. Probability sampling is when all the individuals in your target population have an equal chance of being selected as participants in your study. This method ensures that your sample is representative of the target population, and the results of your study would be generalisable to it.  

  1. Non-probability sampling is when the selection of your study participants is done intentionally, i.e. you select your participants from the target population according to specific criteria, which means that not everyone in your target population has the same chance of being selected to participate.  This is often a more convenient method, but its limitations should be kept in mind. 

Do you need more help with this? Please find links to additional resources below.   

  • Study designs: resources for different types of studies on TGHN  

  • Who will you include in your study (your population and sample)? 

  • Where will your study take place (your study settings)? 

Need more support?