Study Designs

This short article gives a brief guide to the different study types and a comparison of the advantages and disadvantages. See also Levels of Evidence

These study designs all have similar components (as we’d expect from the PICO):

  • A defined population (P) from which groups of subjects are studied
  • Outcomes (O) that are measured

And for experimental and analytic observational studies:

  • Interventions (I) or exposures (E) that are applied to different groups of subjects

Overview of the design tree

Figure 1 shows the tree of possible designs, branching into subgroups of study designs by whether the studies are descriptive or analytic and by whether the analytic studies are experimental or observational. The list is not completely exhaustive but covers most basics designs.

Study designs

Figure: Tree of different types of studies (Q1, 2, and 3 refer to the three questions below)

Download a PDF by Jeremy Howick about Study Designs

Our first distinction is whether the study is analytic or non-analytic. A non-analytic or descriptive study does not try to quantify the relationship but tries to give us a picture of what is happening in a population, e.g., the prevalence, incidence, or experience of a group. Descriptive studies include case reports, case-series, qualitative studies and surveys (cross-sectional) studies, which measure the frequency of several factors, and hence the size of the problem. They may sometimes also include analytic work (comparing factors “” see below).

An analytic study attempts to quantify the relationship between two factors, that is, the effect of an intervention (I) or exposure (E) on an outcome (O). To quantify the effect we will need to know the rate of outcomes in a comparison (C) group as well as the intervention or exposed group. Whether the researcher actively changes a factor or imposes uses an intervention determines whether the study is considered to be observational (passive involvement of researcher), or experimental (active involvement of researcher).

In experimental studies, the researcher manipulates the exposure, that is he or she allocates subjects to the intervention or exposure group. Experimental studies, or randomised controlled trials (RCTs), are similar to experiments in other areas of science. That is, subjects are allocated to two or more groups to receive an intervention or exposure and then followed up under carefully controlled conditions. Such studies controlled trials, particularly if randomised and blinded, have the potential to control for most of the biases that can occur in scientific studies but whether this actually occurs depends on the quality of the study design and implementation.

In analytic observational studies, the researcher simply measures the exposure or treatments of the groups. Analytical observational studies include case””control studies, cohort studies and some population (cross-sectional) studies. These studies all include matched groups of subjects and assess of associations between exposures and outcomes.

Observational studies investigate and record exposures (such as interventions or risk factors) and observe outcomes (such as disease) as they occur. Such studies may be purely descriptive or more analytical.

We should finally note that studies can incorporate several design elements. For example, a the control arm of a randomised trial may also be used as a cohort study; and the baseline measures of a cohort study may be used as a cross-sectional study.

Spotting the Study Design

The type of study can generally be worked at by looking at three issues (as per the Tree of design in Figure 1):

Q1. What was the aim of the study?

  1. To simply describe a population (PO questions) implies descriptive
  2. To quantify the relationship between factors (PICO questions) implies analytic.

Q2. If analytic, was the intervention randomly allocated?

  1. Yes? impliesRCT
  2. No? impliesObservational study

For observational study the main types will then depend on the timing of the measurement of outcome, so our third question is:

Q3. When were the outcomes determined?

  1. Some time after the exposure or intervention? impliescohort study (‘prospective study’)
  2. At the same time as the exposure or intervention? impliescross sectional study or survey
  3. Before the exposure was determined? impliescase-control study (‘retrospective study’ based on recall of the exposure)


Advantages and Disadvantages of the Designs

Randomised Controlled Trial

An experimental comparison study in which participants are allocated to treatment/intervention or control/placebo groups using a random mechanism (see randomisation). Best for study the effect of an intervention.


  • unbiased distribution of confounders;
  • blinding more likely;
  • randomisation facilitates statistical analysis.


  • expensive: time and money;
  • volunteer bias;
  • ethically problematic at times.

Crossover Design

A controlled trial where each study participant has both therapies, e.g, is randomised to treatment A first, at the crossover point they then start treatment B. Only relevant if the outcome is reversible with time, e.g, symptoms.


  • all subjects serve as own controls and error variance is reduced thus reducing sample size needed;
  • all subjects receive treatment (at least some of the time);
  • statistical tests assuming randomisation can be used;
  • blinding can be maintained.


  • all subjects receive placebo or alternative treatment at some point;
  • washout period lengthy or unknown;
  • cannot be used for treatments with permanent effects

Cohort Study

Data are obtained from groups who have been exposed, or not exposed, to the new technology or factor of interest (eg from databases). No allocation of exposure is made by the researcher. Best for study the effect of predictive risk factors on an outcome.


  • ethically safe;
  • subjects can be matched;
  • can establish timing and directionality of events;
  • eligibility criteria and outcome assessments can be standardised;
  • administratively easier and cheaper than RCT.


  • controls may be difficult to identify;
  • exposure may be linked to a hidden confounder;
  • blinding is difficult;
  • randomisation not present;
  • for rare disease, large sample sizes or long follow-up necessary.

Case-Control Studies

Patients with a certain outcome or disease and an appropriate group of controls without the outcome or disease are selected (usually with careful consideration of appropriate choice of controls, matching, etc) and then information is obtained on whether the subjects have been exposed to the factor under investigation.


  • quick and cheap;
  • only feasible method for very rare disorders or those with long lag between exposure and outcome;
  • fewer subjects needed than cross-sectional studies.


  • reliance on recall or records to determine exposure status;
  • confounders;
  • selection of control groups is difficult;
  • potential bias: recall, selection.

Cross-Sectional Survey

A study that examines the relationship between diseases (or other health-related characteristics) and other variables of interest as they exist in a defined population at one particular time (ie exposure and outcomes are both measured at the same time). Best for quantifying the prevalence of a disease or risk factor, and for quantifying the accuracy of a diagnostic test.


  • cheap and simple;
  • ethically safe.


  • establishes association at most, not causality;
  • recall bias susceptibility;
  • confounders may be unequally distributed;
  • Neyman bias;
  • group sizes may be unequal.

Sign up for the EBHC Bulletin. Course news and blogs straight to your inbox.

8 comments on “Study Designs

  1. Pingback: The Ideal Diet

  2. Pingback: MS, Diet And Research; One Piece Of The Puzzle At A Time | Food Matters 365

  3. AvatarNagendra Dudi-Venkata

    Thanks for explaining this so succinctly. Can I ask one quick question – where does “retrospective cohort study” in – would they be considered as Case-control study?

  4. In response to the question, “Where does [the] retrospective cohort study come in?”- I’ll (rather lengthily, and hopefully, usefully) explain it like this. In epidemiology, a “cohort” is used to describe/define a group or set of people (subjects, cases, patients etc.), who are investigatively followed for a certain period to observe what changes have occurred to the members of that cohort over that particular time period. John Last speaks about a ‘cohort’ in the following way: “The term “cohort” has broadened to describe any designated group of persons who are followed or traced over a period of time” (the “or traced” is relevant here).
    Temporally, a study can be either prospective or retrospective. A prospective study takes a group of subjects, and monitors the group for outcomes (e.g. disease occurrence) during the study period, and relates this (mathematically) to factors (exposures) that influence the likelihood of an outcome (appearing, existing or changing). In a prospective study, the distribution of the disease outcomes (i.e. how many are ill, who is ill, how ill they are etc.) is not known at the beginning of the study period – this becomes apparent with time.
    A retrospective study takes a group of subjects, and looks backwards to examine exposures thought likely to influence an outcome, and to determine in what way these exposures relate (mathematically) to the outcome, which is known – and has been verified – at the beginning of the study period. In a retrospective study, the distribution of the disease outcomes (i.e. how many are ill, who is ill, how ill they are etc.) is known at the beginning of the study period.
    By comparison, case-control studies take a representative group of subjects, and consider known outcomes in that group, and will, by dint of study design, and careful choice of controls, attempt to refer backwards to determine what bearing any number of possible exposures (and the degree to which subjects were exposed to them) had on these outcomes. In a case-control study, the distribution of the disease outcomes (i.e. how many are ill, who is ill, how ill they are etc.) is also known at the beginning of the study period. The difference with retrospective cohort studies, is that, because we cannot include everyone who has this disease (outcome) we have to choose a representative sample – a small subset of the entire population of people with the disease – and compare that will controls
    So, as we said before, cohort studies can be either prospective or retrospective. However, because the great majority of cohort studies follow a cohort (frequently longitudinally) to its outcome development, and are therefore prospective, it has (rather sloppily) become almost shorthand to say ALL cohort studies look forward in time from exposure to outcome, and are therefore prospective. Retrospective cohort studies are not uncommon, and are particularly (but not exclusively) encountered in one specific area of epidemiology; the investigation of outbreaks (especially foodborne) of infectious disease.
    Cohort studies are classified as being prospective or retrospective, depending on the point at which, the outcomes were verifiably identified. So, taking as an example, a concocted foodborne outbreak as illustrative; let’s say 50 young adults attended a party on a Saturday evening at 6:00PM. During the party, tuna, salad, chicken and beef sandwiches were served. And, after a period of about 12-18 hours, a number of people fall ill with signs of gastroenteritis (i.e. cases of illness start to appear on the Sunday).
    By Wednesday (for the sake of argument), we local epidemiologists have cottoned on to the fact that there was ‘something up’ at that party (we were getting reports from party goers, from General Practitioners/Family Physicians etc. that there is a cluster – a defined collection – of linked cases of gastrointestinal illness, all of whom had attended the same function on Saturday evening). Since the possibility exists, that a single source might be responsible for this outbreak, it follows that – if the source of the infection is still in existence – it may be still possible, to prevent further cases of illness. And even if it is not possible to prevent further cases of illness as a result of THIS outbreak, we may, by undertaking a study, learn something that might help us in prevent ANOTHER, similar outbreak, in the future.
    A hypothetical Outbreak Control Team is convened and meets on the Wednesday afternoon. Our investigation tells us that, in total, 22 people became ill with a clinical picture that strongly suggests salmonellosis (giving an attack rate of 44% – 22/50) and from a couple of the attendees, Salmonella Enteritidis had been detected in their stools. We track down all 50 party-goers, we apply an outbreak-specific questionnaire to all of them (100% response rate – cases AND non-cases – and we now have a cohort), we undertake a wee bit of analysis and we construct some 2×2 contingency tables, which we analyse. Following our analysis, we find that, among all the food consumed, it is the chicken sandwiches that appear noteworthy. We find that those party-goers who consumed the chicken sandwiches were (again for the sake of argument) 3.3 times more likely to develop salmonellosis, if they had consumed the chicken sandwiches, than if they had NOT consumed the chicken sandwiches. This measure of association is the relative risk. The relative risk of illness among chicken eaters, as compared with non-chicken eaters is 3.3. And, since we have a cohort, we can measure the actual risk; we do not need to derive a proxy measure such as the odds ratio (the derived measure of association in case-control studies).
    The study we have undertaken is a cohort study; we had a cohort of partygoers (comprising all cases and non-cases), this entire group (all of which were identified and questioned systematically) attended the party. Some ate chicken sandwiches, some got ill, but the entire cohort went through a process of potentially being exposed (becoming infected), and some of these then went on to develop an outcome (salmonellosis).
    This is a retrospective cohort study in that, by the time the epidemiologist has learned of the food poisoning incident, the outcome was already known (gastroenteritis, which turned out to be salmonellosis). So, the outcome of the outbreak was actually the piece of information that investigators were beginning with. The subsequent design of the study then sought to gather historical information on all possible relevant exposures, in order to look back in time, and to retrospectively determine if there was a statistical association between these food exposures and the outcome.

  5. In your explanation of retrospective cohort study above, I am curious whether you are aware of the article Dekkers, Olaf M., et al. “Distinguishing case series from cohort studies.” Annals of internal medicine 156.1_Part_1 (2012): 37-40.

    In Dekkers’ article, the case is made to define cohort studies as studies that satisfy the following requirements: 1. whose starting point is the exposure, 2. there is follow-up, and 3. risk calculation is possible.

    I agree with your statement that the “foodborne outbreak” study is a cohort study but I believe that Dekkers would take issue with the explanation why. You write that the starting point of the study is the outcome (which Dekkers would define as a case-series study). However, the study design does not call for a questionnaire of only those with the outcome. It sends a questionnaire to ALL people exposed (people who ate at the party) and the outcome of each exposure risk was unknown until the “follow up” of receiving the questionnaire answers. All outcomes are actually considered – both those with salmanella and those without – and because of this a risk calculation is possible.

  6. The issue here is the confusion about the “outcome”. In a retrospective cohort study, the outcome includes, BOTH, the occurrence of disease and the non-occurrence of disease. The diseased patients are just the “trigger” (and not the outcome) to look back at ALL the exposed patients, which actually form the “cohort” – this is why it is retrospectively looking at and starting from exposure, rather than from outcome. Further, the outcome is known only AFTER we follow ALL the exposed patients, from the past to the present, to know which patients developed the outcome of interest (disease) and which did not – to calculate the risk ratio.
    In a case-control study, the group of subjects with outcome and the group of subjects without outcome is already well-defined and separately known a priori. Here, the outcome of ALL subjects is known before the start of the study. Hence, we try to move from outcome to exposure, backwards in history, to elucidate any and all possible exposures. The starting point here is outcome.
    I hope this helps!

  7. Hi, I’ve usually thought of an outbreak investigation as a typical a case control study? Normally I would assume that we round up all the people who are sick, give them a survey questionnaire and then find approximately the same number of people who are not sick and do a food questionaire and do a simple t-test to see if there is a significant difference in their food consumption history? Its more for efficiency more than anything else, especially if you have a large numbers of people eating at a canteen for example.

Leave a Reply

Your email address will not be published. Required fields are marked *