Research models are used across many fields, including medicine, psychology, social sciences, education, business and pharmaceutical research. The model may be presented as a diagram, a set of relationships, or a combination of concepts and assumptions. It does not have to be complicated. In many cases, a simple model can make a research question much easier to understand.
What is the model of research?
At its simplest, a research model depicts the relationship between different elements of a study. These elements can be independent variables, dependent variables, mediating factors, outcomes or other conditions that can influence the results.
For example, a researcher may want to consider whether sleep has an impact on the academic performance of university students. A simple model of research could show:
Sleep quality -> concentration -> school performance
In this case, one of the things that is studied is sleep quality. Concentration can be the intermediate factor, and academic performance is the result.
The model enables the researcher to go beyond the question of “Does sleep affect academic performance?” It makes them think about how the relationship might happen and what else could affect it.
Why is a research model important?
A research model can be useful before the study starts as it helps to organise the researcher’s thinking. It may highlight missing elements of the research question or indicate that there are too many variables.
It can also help when deciding what information needs to be collected. If a model suggests that age, treatment exposure and baseline health may influence an outcome, the researcher can consider these factors when designing the study.
In published research, a model can also make a study easier for readers to follow. Someone reading the paper can see the proposed relationships rather than having to work them out from several pages of text.
That said, a research model is not proof that a relationship exists. It is a representation of the relationships the researcher proposes or expects. The findings might in fact support the model, might partially support the model or might reveal that the relationships are different from what was predicted.
Read the full explanation
Research model and conceptual framework
The terms ‘research model’ and ‘conceptual framework’ are sometimes used interchangeably but are not always exactly the same.
The conceptual framework describes the main ideas in a study and how they are believed to be related. A research model can be more specific and can focus on relationships that can be tested with data that has been collected.
For example, a conceptual framework can describe the impact of stress, workload, social support, and sleep on employee well-being. So a research model could take those ideas and turn them into specific relationships that could be studied statistically.
This distinction is not universal across academic disciplines, so researchers should therefore use the terminology used in their field or by the journal.
What are the key components of a research model?
What exactly goes into it varies by study, but there are a few common components.
Independent variables are the factors thought to influence another variable. This could be a treatment or a dose in a clinical study. Or in an education study, it could be a teaching method.
The outcomes that are measured are dependent variables. Examples are blood pressure, test scores, severity of symptoms, and customer satisfaction.
Mediating variables explain how one factor may influence another. Exercise may affect wellbeing partly through improvement in sleep. Sleep would be a potential mediator in that model.
Moderating variables can change the strength or direction of a relationship. Age, for example, might affect the relationship between physical activity and a particular health outcome.
Researchers may also include confounding variables. These are factors associated with both the exposure and outcome that can distort the apparent relationship if they are not properly considered.
Not every research model needs all of these components. Adding variables simply to make a model look detailed can make the research harder to conduct and interpret.
How is a research model developed?
Developing a useful model usually starts with the research question. Then the researcher looks at the literature to see what is already known and where there is uncertainty.
For example , the research question is : Does a new intervention improve recovery post-surgery? The researcher might view the intervention as the primary independent variable and recovery time as the outcome. Existing research may also suggest a role for age, baseline health and complications to influence recovery. These factors may therefore need to be considered in the model.
The researcher can then map out the proposed relationships and decide which ones can actually be tested.
A good model should be connected to the research question. It should also be realistic. If the available data cannot measure a particular variable reliably, including it may not be useful.
How do you test research models?
The testing of a research model depends on the type of study.
Quantitative research can use statistical techniques such as regression analysis, correlation, structural equation modelling or other methods suitable for the data. These techniques may help researchers to investigate the strength and direction of relationships between variables.
Models can be used in qualitative research in various ways. A researcher studying patients’ experiences of long-term treatment might develop a model from themes identified during interviews. In this situation, the model may evolve as new information is analysed rather than being fully defined before data collection.
In both cases, the model should remain linked to the evidence. A model should not be treated as established simply because it looks logical on paper.
What are the limitations of a research model?
Research models simplify reality. Many factors often influence human behaviour, biological processes, and social systems at the same time. A model cannot always capture every relevant influence.
There is also a risk of building a model around assumptions that have not been properly tested. A relationship between two variables may appear plausible but may disappear when other factors are taken into account.
For this reason, researchers should be clear about what their model proposes and what the evidence actually demonstrates.
In simple terms
A research model is a map of a research study. It brings together the key variables or concepts and illustrates how the researcher expects them to relate.
It can help to inform the research question, guide data collection and make findings easier to interpret. But that’s still a representation of a more complex reality. Its value is dependent on the strength of the support of existing evidence, the clarity of the definition of the variables and the appropriateness of the testing of the proposed relationships.
Research Use Only. This glossary explains terminology used in published research. It is not medical advice and does not describe any product for human or animal use.