I think cross-sectional surveys must be representative of the targeted population (often leading to important sample sizes). Is it the same for case-control study ?
Cross- sectional is a descriptive study where we just take a snap shot of the situation and describe it in term sof frequencies mostly the prevalance while case control is an obsevational analytical study and is retrospective in nature where by we compare two groups one already have the disease and other free of disease but both having the same attributes other then disease to assess the presence or absence of the exposure
Case contro study starts from cases and controls i,e. from diseased patients and absolutely disease free controls and we have to go back and collect the information about risk factors responsible for occurance of disease in patients and controls. while in cross sectional study we have to collect the information on certain study variables or disease under study from defined study population in defined geographic area and in defined time of period and information will be collected once or single shot from population.cross sectional study also known as prevalence study.
In my epidemiology class my memory of the simple definition is that case control is retrospective and as Mohite suggest you start with outcome and one looks back to determine the exposure. whereas in cross sectional you both exposure and outcome have taken place
Sarbjeet Khurana is spot-on. As you will see from her ppt presentation, a cross-sectional study is also an analytical study, it is not only a simple descriptive study as suggested by Umer. You will always end up with at least 4 "groups" to enter into your 2x2 table - (1) those with the outcome and exposed; (2) those with the outcome and unexposed; (3) those without the outcome and exposed; (4) those without the outcome and unexposed. As for representivity, in a case control study one often cannot take a representative sample of all cases, especially if the outcome of interest is uncommon or rare - one has to take every case that one can find. Then one must try to find a control group that resembles the case group in every way possible, except that the control group does not have the outcome of interest - this is not always easy, and can be a source of bias. To reduce random errors in case control studies of rare or uncommon outcomes where the number of cases is relatively small, one should increase the ratio of controls to cases.
A cross-sectional study measures both the exposure and the disease/outcome of interest at the same point in time. This is generally useful when little or no research has been reported on the question of interest. More sophisticated analyses, such as case-control studies, typically come along after the prevalence questions that a cross-sectional study can address are answered. I agree with Rose that a cross-sectional investigation is not simply a descriptive study.
Case-control: You choose (bias) applying randomization/sistematic/blind/or not randomization the cases and controls (can be matched or not matched with cases)...(not in a cohort study) and can be retrospective or prospective or mixed, BUT level of evidence is low (ej: IIIb,IIIc), but when you select CASES you know they´re CASES (the patients HAVE the disease).
Cohort study: You select a lot of people, BUT you DO NOT select cases, you have to WAIT for ocurrence of the disease in time, some individuals will not have the disease and others yes, then you make comparisons in your computer organizing CASES vrs NO CASES (DISEASE vrs NO DISEASE)
BOTH METHODS CAN BE APPLIED USING RETROSPECTIVE (for cohort methodology you have to randomize) or PROSPECTIVE (randomize or not) or MIXED (risk method calculating may be a challenge...!!!!)
Several of the discussions of control selection are problematic "Then one must try to find a control group that resembles the case group in every way possible, except that the control group does not have the outcome of interest - this is not always easy, and can be a source of bias." and "we compare two groups one already have the disease and other free of disease but both having the same attributes other then disease to assess the presence or absence of the exposure". Both comments gloss over the function of controls and form follows function . Plus both seem to suggest that controls should be similar to cases for correlates of exposure -which bakes a bias to the null into the study.
The purpose of controls is to measure the prevalence of exposure in the source population from which the cases arose. It can be tricky to conceptualize this sometimes, but in a simple example, cases of lung cancer among adult residents of New York City. You can conceptualize a valid control by asking the question - if this prospective control subject had of developed disease would they have been selected to be a case in my study. If yes - they are from the same source population that generated the cases and would be a valid control subject. Ideally you recruit controls randomly from the source population, and if employing matching, the selection process is a stratified random sample. With control recruitment stratified by the matching factor, lets say gender, so that the % of men in the final control sample matches the % of men in the case sample.
I am sending the chapter Overview of Epidemiologic Study Designs from the book Essentials of Epidemiology in Public Health, as an attached document, where you can find the explanation of all epidemiological studies
These are primarily used to determine prevalence. Prevalence equals the number of cases in a population at a given point in time. All the measurements on each person are made at one point in time.
CASE-CONTROL STUDIES
In contrast with cohort and cross sectional studies, case-control studies are usually retrospective. People with the outcome of interest are matched with a control group who do not. Retrospectively the researcher determines which individuals were exposed to the agent or treatment or the prevalence of a variable in each of the study groups