Research uses control groups and strategically chosen situations, people and events. The theoretical and practical aspects of the research may provide great results and the outcome may prove the experiments you have been carrying out. External validity answers the query: are the experiments applicable in the real world? Your outcome may have been successful with the setting you had but will it be viable in real-life situations?
External Validity - FAQ
External validity involves propagation of results and it is the degree to how the results of a study can be applied to other contexts and categories of people.
Tip: If you are done with your academic work, we can help you with dissertation printing!
External validity has two classifications: population validity (Applying your results to a different category of the population) and ecological validity (Propagating your outcome to different contexts).
Internal validity is how far an experiment supports a reliable cause and effect relationship between research and results.
External validity is the degree to how your outcomes can be applied to other situations and contexts.
An experimental design should have external and internal validity. The most important requisite that must be present in an experiment is interal validity. Only then can conclusions be drawn about treatment effects. If your goal is to establish internal validity, external validity should be monitored.
External validity: Definition
External validity is the degree to how your outcomes can be generalized to other situations, groups and contexts. External validity helps you understand whether the results of your research can be applied to a wider context. Scientific studies aim to create inferable facts about the real world. A high external validity is needed to apply your theoretical and experimental outcomes to other groups of people and in real-life circumstances.
Classifications of external validity
Population validity
Population validity is concerned with whether your results can be spread from your sample to different categories of people and the population at large. Population validity is dependent on the selection of population and the degree to which the research sample reflects the population.
Example: low population validity
You want to test the theory that students are academically motivated by different things and how that determines their performance.
You enlist about 150 participants, most of them female, between 18-21 years and residing in big cities. For a laboratory experiment, you give out a questionnaire that lets them describe their backgrounds, lifestyle and goals in life. You discover that most participants are motivated by desiring a fancy life rather than a simple, minimalist life. Can you presume that most people want an expensive and fancy lifestyle
Explanation
Your sample does not reflect the whole student population in the institution. The results can only be applied to the groups of people with the same attributes as the participants. That population caters to people from a high socioeconomic class of the general public.
For a higher population validity, your experiment would have to include other students with different attributes. Other than people from urban areas and densely populated areas, the study will need to encompass students from rural areas, male students, those from different socioeconomic backgrounds and students from different countries.
Ecological validity
The concern of ecological validity is whether the research outcome can be propagated to other contexts and situations in reality. It is used to determine whether you will get the same outcome if you take your research to a different location outside the experiment setting. Laboratory experiments are usually controlled and the environment is set without distractions so that the participants may focus on their role.
Example: low ecological validity
You want to assess the hypothesis that when others are speaking, the driving response time becomes slower.
For a laboratory experiment, you configure an easy computer-based exercise to analyse the response times. The participants have to assume that they are driving through a racecourse and whenever they see a red icon on the screen, they have to double-tap the mouse. In one round, the participants listened to a show. For the second period, the participants did not listen to anything. After analysing the outcome, you discover that their response time is slower when they are listening to the show. Can you presume that when someone is speaking, the driving response times become slower?
Explanation
In consideration of the example above, generalizing the results to emulate the driving conditions in the real world is hard. A computer-based exercise that requires the participants to utilize a mouse during the simulation is not a depiction of how you drive in reality. Also, the static red icon representing obstacles when driving does not mimic the common conditions of actual driving.
To achieve a prominent ecological validity, try utilizing a real car or an advanced simulation of a car with gears, a steering wheel and foot pedals. That will increase the resemblance of driving in real-life and improve the outcome of your research.
Trade-off between internal and external validity
An improved external validity is usually achieved by sacrificing internal validity and improving internal validity comes with the sacrifice of external validity. There is an implicit trade-off between external validity and internal validity.
Trade-off example
In the driving response times experiment, you can control the research setting by ensuring no external elements could affect the findings. The experiment has a higher internal validity, making you conclude that listening to a show causes slower response times.
When you move the research to real-life driving situations, the external validity greatly increases compared to internal validity. That is because numerous external elements like weather, which will influence the results, will be considered.
Threats of external validity
During a research experiment, some elements may affect its external validity.
Example
An analyst wants to examine the hypothesis that people diagnosed with mental disorders may gain from being mindful daily for two months. They enlist people identified with depression. Those participating were trained on mindfulness and requested to practice being mindful daily. A pre-test and a post-test were administered to evaluate the number of times they got anxious.
The anxiety levels were reduced between the tests, making the analyst infer that being mindful could be useful to the medical population.
Threats facing external validity
| Type | Explanation | Example |
| Testing effect | Issuing a test before or after a research influences the results. | The questionnaire about academic motivation used in the tests prompts the participants to engage themselves on what motivates them. |
| Sampling bias | The sample does not represent the general population. | The students participating in the research are mostly female and come from densely populated areas. The male students and those from rural areas are not significant, meaning the outcome cannot be applied to the public. |
| Hawthorne effect | Participants change their behaviour when they are aware that there is research being done on them. | The students may try to be more involved with their studies and be unusually invested in their academic life. That may be because they are aware they are part of an experiment. |
In a Nutshell
- A research with a better external validity involves being aware of the elements that will influence the results of your research.
- When you make your research more applicable in a wider setting, you will be less able to control the external elements of your research.
- Threats have to be identified and countered for a powerful study.
- Develop a robust external validity experiment to achieve a high external validity for better conclusions.
