Observational (or non-experimental) studies have no intervention and purely observe changes in one characteristic (e.g. smoking) compared to another (e.g. lung cancer) in the natural world. These studies can be classed into analytical studies, which aim to determine correlation and causality, or descriptive which merely reports a certain situation.

Analytical studies

Analytical (analytic) studies examine populations and analyze if there is a causal correlation between an exposure (i.e. risk factor, e.g. smoking) and an outcome (e.g. disease). Analytical studies can be prospective (following subjects from a start point forward in time), retrospective (investigating the past of subjects) or cross-sectional (at a certain point or period in time). There are several types:

Case-control study

Case-control studies look for differences in previous risk factors (e.g. level of exercise) between “case” subjects (e.g. those with a condition of interest) versus a suitable “control” group (e.g. those who resemble the cases in aspects like age and sex but do not have the condition of interest). They observe if an attribute changes in frequency or level when comparing, for example, the diseased and non-diseased. Such studies must be retrospective because it starts after the onset of the disease or condition of interest and looks back to the postulated causal factors.

Advantages	Disadvantages
Can investigate multiple exposures	Unclear timing between exposure & outcome
Good for rare outcomes	Bad for rare exposures
Rapid (no-follow up)	Can’t estimate incidence
Efficient/less costly

Cohort study

Cohort studies (sometimes follow-up or longitudinal prospective studies) classify a defined population into those exposed or not exposed (or to different degrees) to a potential risk factor (which can be beneficial or harmful). They follow these two groups over time to see if different outcomes develop (e.g. disease occurrence). Unlike case-control studies which compare different outcomes groups to determine risk factors, cohort studies compare populations with different exposure to risk factors to look for differences in outcome. Cohort studies generally observe large groups over a long period (commonly years) with a comparison of incidence rates in groups that differ in exposure levels. Such studies can be performed either prospectively or retrospectively.

Advantages	Disadvantages
Can estimate incidence	Large sample size
Good for rare exposures	Bad for rare outcomes
Can investigate multiple outcomes & exposures	Can take a long time (for prospective studies)
	Loss to follow-up

Cross-sectional study

Cross-sectional studies look at the occurrence of outcome and exposure at at one point in time (e.g. over a week). It’s a snapshot that will likely not be generalizable to other time periods or populations. It is likely unable to ascertain causality between a factor and an outcome as temporality is unknown. Some cross-sectional studies (known as incidence studies) are descriptive instead of analytical.

Advantages	Disadvantages
Quick & cheap	Cannot confirm causality
Good for hypothesising risk factors	Unclear timing between exposure & outcome

Ecological study

Ecological studies look at the outcome and exposures at the population level instead of an individual level unlike the other studies (e.g. averacy pollution level vs prevalence of lung cancer in different cities). It is unable to ascertain causality between a factor and an outcome as confounding is a serious problem (i.e. pollution level and lung cancer prevalence may be known, but not smoking uptake). Furthermore ecological fallacy needs to be taken into account, where trends that occur in populations (i.e. at an aggregate level) may not represent what happens in individuals (e.g. countries with increased breast cancer rates have a fattier average diet, but it is unclear if the individuals with breast cancer themselves ate a fattier diet)

Advantages	Disadvantages
Quick & cheap	Cannot confirm causality
Good for hypothesising risk factors	Confounding is a serious problem
	Ecological fallacy

Descriptive studies

Descriptive studies report the existing distribution of variables without regard to causal or other hypotheses, generally using published data (secondary sources). They include studies like case-reports and case series which are in-depth examinations of patients (or a group of patients) and their clinical conditions. There are also incidence studies which describe the number of new cases of a condition in a population.