Sunteți pe pagina 1din 5

1

Experimental versus Observational studies


There are 2 major types of analytical studies. Each of these types has several sub-types
within itself as well.
The first major type of study is experimental. In experimental studies, an explanatory
variable is used to evaluate the effect on a specific outcome. The explanatory variable is varied
by the researcher in order to observe for any change in the response variable. Data is collected by
defining metrics that can accurately show changes in variables. This type of study often involves
control groups and placebos to account for bias due to the presence of researchers and/or
knowledge of treatment. In most cases, these studies need to be conducted in labs or in isolated
locations to properly account for controlled variables. Experimental studies are usually used to
evaluate a precise relation between 2 variables while all else is kept constant. These studies are
considered very useful in determining a cause and effect relationship between 2 variables which
have a known correlation.
The second major type of analytical study is an observational study. An observational
study does not make use of complex experimental designs as it is carried out by little or no
intervention. Data is collected by simply recording what is observed. Participants are often not
aware of being observed and no variable is deliberately manipulated by the researchers. There
are no control groups in this type of study and random sampling is also difficult to ensure. Since
participants are often not made aware of the presence of researchers, there is no psychological
effect to account for. These studies are usually conducted in open and public spaces as little to no
intervention is needed. The goal of an observational study is to verify the existence of a
relationship between 2 variables rather than defining the exact parameters and framework of the
relation. This makes observational studies a very good tool to make prediction models.

Advantages and Disadvantages of Observational studies


Advantages Disadvantages
1 Can be conducted in the natural setting of a Can not be randomized.
participant.
2

2 Can be conducted for a larger number of Tend to overrepresent one cluster of the
participants at one time than experimental population over another.
studies.
3 Can be cheaper in setting up. Data is of poorer quality as observations are
made from a distance.

4 They are more generalizable as they are Offer researchers little control over the
normally conducted within a subset of the setting of the study.
actual population.
5 They are very representative of the actual Make it difficult to eliminate the possibility
population. of an unaccounted variable affecting the
outcome.
6 They have greater external validity than They lack internal validity.
experimental studies.

7 Can evaluate multiple variables of interest Cannot be used to define new relationships
within the same study. between any variables.

8 Can be used to generalize known relations to Often mistake correlation for causation
new settings and locations. leading to weaker results.

9 The study design process is not very time- Often result in overfitting and confounding
consuming. of data.

1 There are fewer psychological biases that Have a greater chance of having observer
0 may influence the outcomes. bias than experimental studies.

1 Generally, do not have as many ethical Observational methods do not provide as


1 considerations as an experimental design. much detail as experiments.
3

1 Can be easily improvised and adjusted Often require aligning study design to match
2 during the study the nature of available data.

1 Pose less liability as there is no intervention


3 on part of the researcher

Circumstances which warrant use of observational studies


Observational studies are best used in circumstances when we are interested in gathering
evidence which verifies the existence of a relationship (Sells and Ellis 1951). We require a model
robust enough to handle uncertainty in our data and come up with a good estimation of our
desired outcome. It is also more suitable in situations when there are ethical concerns associated
with the intervention of a researcher. In analytical health studies, for example, we might need to
know whether the quantity of a specific protein A in the blood is greater for people who are
infected with disease B. In this case we are not interested in knowing what is the exact chain of
events which could lead to an increase in A. We are also not concerned with knowing if B causes
A or if there is a stimulant which causes them both. All we need to know is if there is a strong
correlation between A and B which would help us use A as a proxy for determining the existence
of B in the future. Moreover, it would be ethically wrong and legally dangerous to manipulate
variables in a way which could worsen the extent of the disease within our patient so we may not
want to intervene. In such a scenario, observational methods provide us with the best way of
tackling these problems.

Qualitative and Quantitative examples of observational studies


Qualitative observational studies make use of non-statistical techniques to develop a
richer perspective on a given topic. These studies evaluate variables that take on distinct values
and properties. Qualitative studies provide researchers with a way of explaining complex and
disproportionate relations. A famous example of such studies is Darwin’s study of Tortoises.
Darwin recorded characteristics of tortoise colonies on a different island in the region of
Galapagos (Sulloway 2009). By analyzing how similar and how different each colony was to
other colonies, Darwin realized the tortoises had at some point belonged to the same species
4

(Sulloway 2009). Once they were separated on different islands, they evolved into different
species which helped Darwin come up with the idea of evolution.
Quantitative observational studies use measurable processes to gather data, which is then
ordered for analysis. The goal of these studies is to provide statistical evidence to support or
reject a hypothesis. An example of this is a study that evaluates the chance of experiencing
cancer as time progresses for those subjects who have had contact with asbestos and are past
smokers. This study hypothesizes that people who have a history of smoking have a much
greater probability of developing cancer in their lungs than people who do not engaging in
smoking when the 2 groups have had the same level of contact with asbestos (Reid et al. 2006).
This hypothesis is supported by the result which shows people who smoke have a greater chance
of developing cancers in the lungs for 20 years after giving up smoking.

Example of naturalistic and participant observation in healthcare


Naturalistic observation involves observing participants in their natural surroundings
while participant observation involves the researchers stepping into the participant’s environment
to observe them. Both of these can be used in a healthcare study which determines if greater
patient-doctor interaction leads to extended hospital stays. A naturalistic setup would involve the
researchers simply observing how long and how frequently a doctor interacts with a patient while
a participant observational approach might involve researchers themselves being the doctors who
interact with the patients. A quantitative perspective would be to determine the difference in
length of stay of a patient with higher interaction as compared to one with a lower interaction
with the doctor. A qualitative perspective would be to determine if the measured difference in
length of stay is significantly different for senior patients than it is for younger ones.
5

Bibliography

Reid, A., De Klerk, N., Ambrosini, G., Berry, G., & Musk, A. (2006). The Risk of Lung Cancer
with Increasing Time since Ceasing Exposure to Asbestos and Quitting
Smoking. Occupational and Environmental Medicine, 63(8), 509-512. Retrieved August
2, 2020, from www.jstor.org/stable/27732778
Sells, S., & Ellis, R. (1951). Observational Procedures Used in Research. Review of Educational
Research, 21(5), 432-449. Retrieved August 2, 2020, from www.jstor.org/stable/1168307

Sulloway, F. (2009). Tantalizing Tortoises and the Darwin-Galápagos Legend. Journal of the


History of Biology, 42(1), 3-31. Retrieved August 2, 2020,
from www.jstor.org/stable/40271531

S-ar putea să vă placă și