Bias and Confounding

Differentiate between bias and confounding. Discuss the criteria necessary to establish a
factor as a confounder and provide an example applying these criteria. What is one way to
adjust for a confounding relationship in the study design or the analysis

Bias and Confounding

Mainly, when examining the existing relationship between an explanatory factor and an
outcome, there is always a great need to identify the main effect modifiers on the expected
outcome from the study findings. This is carried out in the aim of knowing the potential bias and
confounding factor which may lead to a misleading effect of the outcome. Notably, bias and
confounding have a distinctive difference although they usually have a related effect in
measurement and study design. Glasser (2008), argues that bias usually consists of systematic
errors in study design, the target population recruitment, and data collection which usually have a
significant effect on the true effect of the results. On the other hand confounding effect refers to a
situation whereby the outcome and exposure of the expected resulted are usually distorted due to
the existence of positive and negative confounding variables. The paper aims to differentiate
between bias, confounding by discussing the criteria necessary to establish a factor as a
confounder, and providing one way for a confounding relationship in study analysis.
Bias always limits the validity of information whereby the truth of information tends to
be unmeasurable as well as limiting the obtained results from generalization. Glasser (2008),
argues that there exist two types of bias which usually consist of selection error that is mainly

identified during the selection of the targeted population. For instance, during the process of
identifying cases injured shoulders, the old people may have a higher probability to have such
cases while a few may have been treated. However, people who may have been treated such
cases may be not be easily identified when compared to people who are not treated. Therefore, if
the study is interested in selecting only the individuals who are treated, this may lead to selection
bias. Besides, information bias may also be established due to the use of instrumentation, which
may offer inaccurate measurement of different diseases to the patients.
Confounding and confounders
A confounding situation refers to different cases whereby the relationship between the
exposures of different results is distorted by the existence of different variables. There exist two
types of confounding results which are mainly represented by positive and negative variables.
Glasser, (2008), argues that a positive confounding usually consists of an observed association
that is usually biased away from the null results while the negative confounding is usually
observed when a certain association is biased toward the expected null results of the study.
The confounders usually have extraneous variables, which mainly accounts on different
types of risks to a suggested patient. However, Glasser, (2008), argues that the presence of a
confounder always have a negative effect on the expected result as it leads to inaccurate findings.
Besides, a confounder usually meets several conditions, which consist of putative risk factor,
independent existence of putative factor and the main causative agent of the risk factor. For
instance, if a hypothesis states that Diabetes acts as a positive risk factor, which causesa coronary
attack, then the study will require having an efficient cross-sectional study aimed at investigating
the suggested hypothesis results.


In conclusion, there exist two types of bias, which consist of selection bias and
information bias. Besides, confounding error is also attributed by the distortion of the expected
exposure and the outcome of results based on the positive and the negative confounding. Lastly,
bias and confounding usually have a negative effect on decision-making policies that may have a
negative effect on the patients.


Glasser, S. P. (2008). Bias, Confounding, and Effect Modification. In Essentials of Clinical
Research (pp. 295-302). Springer,