The Different Types on Bias in Research Studies with Specific Examples for Each of the Types - Assignment Example

Research Methods Task Introduction: As Joppe (2000) defined validity, the determination of whether true measures have been obtained by a research, as the study objectives were intended to, and whether the results obtained are true, is referred as validity. In other words it means determining whether the exact target of the research has been obtained. The test of validity is done by several questions being asked by the researcher to which he tries to find answers to determine whether the study is valid. Joppe defined reliability as the consistency of the outcomes of a research and whether the sample population of the research has been correctly presented, and the consistency of the results being reflected from other similar research following similar methods (Golafshani 2003). One of the major sources that can lead to invalidity and unreliability in a research is the bias of the researcher (Baumgarten 2012). Biases in research are mainly categorized into selection bias, measurement bias, and analysis bias (CEM - Research - Bias & Confounding n.d.). The present study focuses on learning about the different types on bias in research studies with specific examples for each of the types. Bias and Types of Bias in Research Studies: Bias is said to have occurred in a research when the outcomes of the research are different from the actual results. For example, if a research is focused on determining the risk of a disease and the results of the research reflect consistent overestimation of the risks, then the study can be said to have been based on bias. Design of a research is largely responsible for a bias and bias in general presents error in the same direction for a repeated number of the experiments for a particular research sample and work (World Health Organization 2001). Thus biases need to be reduced in order to produce a reliable and valid research. Types of Bias with Examples: Selection Bias: Selection bias is defined as the bias when some participants are favored and selected over others while sampling is conducted and the research outcomes are based on experiments conducted on such samples, thereby affecting the external validity of the research (Gravetter and Forzano 2010). It is also referred to as sampling bias. As a result of such bias, the outcome of the research reflects a systematic difference between those subjects that are included and those that are not. Thus either the results are underestimated or overestimated in terms of their statistics, rates, and ratios (Peat, Mellis and Williams 2002). This means that the study cannot be considered as reliable for use. Moreover, with different outcomes than the actual results, the validity of the research also gets reduced. An example of selection bias in research is a comparison of patients in one hospital where a new wound dressing method is applied on them, with patients of another hospital who are offered the standard method of treatment. In such a case the possibility of selection bias is significantly high since different patients in different hospitals differ in their social and biological characteristics. Thus such a research is exposed to selection bias (Stommel and Wills 2004). Allocation bias is a type of selection bias that occurs when there is a difference in the characteristics of the different samples put in different trial groups of a particular research experiment, and has significant influence on the outcomes of the research (Peat, Mellis and Williams 2002). Measurement Bias: Measurement bias is defined as the bias that occurs when biases or errors occur in the process of measurement and calculations of the data based on which the results are obtained. Such biases can result from erroneous tools of measurement, errors in calculations, errors while measurements are recorded, participant bias, recall bias, and so on (Howlett, Rogo and Shelton 2013). For example, if dietary recalls are considered for a 24 hours time period, then measurement biases can occur with the methods of the interview, the assumptions of coding, or through errors in the consumption of databases based on selection of nutrition, and the number of days for which the interviews should be conducted. Biases in such cases can affect the national diet intake process of several individuals (Monsen and Horn 2008). Thus when a research involves measurement biases, the study cannot be considered as reliable and valid as the true results would not be obtained with wrong data collections or measurements. Another example is in the calculations of BMI (Body Mass Index) of individuals. The calculation requires the measurement of exact height and weight of the body. In case of slight biases or errors in any of the two measurements, the overall calculation of the index would be affected. It is possible in many cases that patients are unable to stand, thus likely to give a wrong measurement of their heights, affecting the calculation of BMI. For such patients, only estimates of their heights are used that can lead to errors in the outcomes of the results (Howlett, Rogo and Shelton 2013). Analysis Bias: Analysis bias is defined as the bias that occurs at the stage of analysis of collected data when the researcher presents positive outcomes to avoid negative reflections (Pannucci and Wilkins 2011). For example, a research involves performing surveys and interviews and hence collecting data and all necessary information for the study. Risks are still involved in the study since all the collected information will have to be analyzed by the researcher. Analysis is the sole responsibility of the researcher and his bias can actually affect the outcomes of the research (Market Research Challenge: Analysis Bias 2013). Another example of analysis bias is participants not remaining in the trial groups that they are allocated in the beginning of the research. If participants do not remain in their groups and complete the study, biases would follow. Researchers need to follow up continuously, which when they ignore, lead to analysis bias in research (CEM - Research - Bias & Confounding n.d.). Thus if the analyses are not true, the study becomes unreliable and invalid. Conclusion: From the above study it can be concluded that bias is a major concern in relation to the validity and reliability of a research that has significant effects on the actual results of the research. As discussed above, selection bias, measurement bias, and analysis bias are the three major types of bias that need to be reduced in order to achieve effective and accurate research outcomes. With such biases, the actual results of a research cannot be obtained, hence making the study unreliable and invalid in nature. Task 2 Introduction: There is generally a certain level of deception in researches as researchers are constantly focused on reducing their biases that they otherwise reflect through providing excess data before completion of the study (Nestor and Schutt 2011). The present study focuses on learning about the concepts of cross-sectional studies, cohort study, and case-control study with examples. Cross-sectional Study: A cross-sectional study is defined as a study which is descriptive in nature that involves measurement of the status of disease and exposure in a chosen sample population simultaneously. It allows the researcher to have an image of the frequencies of the exposure and the traits of the diseases collected from data at a single point of time. Acute or chronic conditions of diseases in individuals can be detected with such studies. However since data is collected at a single point of time, it is difficult to determine whether the exposure has been proceeded by or followed from the disease (What is a cross sectional study? 2009). An example of such studies could be to determine the opinions of people in regard to the performance of the Secret Service for protection of the president. Instead of participating in a survey, all people could be questioned at a single point of time (Bailey 2008). Another particular example of cross sectional study is the census of the United States. This research studied the population of the country at a single point of time. Studies that are explanatory or descriptive are often cross sectional in nature. If a researcher performs survey on a national scale to determine the foundations leading to discriminations of races and religions, then he would have to perform the survey considering a single point of time, which would represent a cross-sectional study. Another example is of a survey conducted in China, by a researcher named Yanjie Bian, who wanted to study stratification in the modern society prevailing in the country. While conducting the survey he had to keep in mind the effects of all the national campaigns that could influence the survey and hence focus on cross-sectional study (Babbie 2012). Cohort Study: Cohort study is defined as a study based on analysis where individuals are identified having a suspected factor with varying levels of exposure who are observed and followed over a certain period of time, that can vary between weeks to months, to determine the effects of the disease on the health. Relation between the estimated exposures and the rates of occurrence of the disease under study, in the cohorts, is measured and analyzed. Historical studies can be used potentially and retrospectively for cohort studies (What is a cohort study? 2001). An example of cohort study is a national survey that takes place in every twenty years, intended to determine the nature and behavior of cohorts born during the Second World War based on the involvement of the United States in international affairs. One survey could involve individuals aged between 15 and 20 years who were interviewed in 1960. A second sample could include individuals of ages 35 to 40 years and this survey was conducted in the year 1980. A third sample survey was conducted in 2000 that included individuals of 55 to 60 years old (Babbie 2010). The study thus involved comparisons of the cohorts born during different times, although each of them differing from one another. This enabled the researcher to study the political orientation that occurred in the country over the years. As a result, a trend reflecting the changes in the issues of races, religions and discriminations could be studied from such a cohort study, explaining the status of the cohorts during different times over a certain period of time that reflected the time of the research (Babbie 2010). The significance of a cohort study is that while the researcher decides on the sampling and control groups, and process of data collection, the outcomes of the surveys are not known to him. Thus outcomes are obtained through observations and survey on individuals over time (Riegelman 2005). Case-control Study: A case-control study is defined as a study where individuals having a disease are considered as ‘cases’ and those without the disease are considered as ‘controls’, and a comparison of the two groups is conducted determine the effects of hypothesized exposures on them. Retrospective data is generally needed for this kind of study, involving the chances of recall bias in the study. With a recall bias, the tendency of the subjects is to report occurrences in a way that proves to be different between the two groups. The problem of recall bias may be reduced with the use of biological markers that are being developed, and enabling newer ways to determine the traits of exposures (What is case-control study? 2008). It has been observed that if the frequency of contact is more for the cases than for the controls, then an increased risk of the result of the research can be associated with the contact or exposure of the population. The study and analysis of this type of research involves the research to go backwards and compare the results with other similar experimental studies (Dumitrescu 2012). An example of a case-control study would be the investigation to determine whether zinc oxide proves to a better measure for prevention of skin cancer. For the study, two groups of former lifeguards were selected – one group including those who developed cancer on their noses and cheeks, and the other group involving those who did not suffer from the cancer. These groups were exposed to zinc oxide and absorbent sunscreen lotions to determine whether cancer on the control group could be achieved. The study could be either a matched or an unmatched one since the use of the sunscreen lotions was dependent on the memories of the former lifeguards as they used to use. Differences could be in their ages, and also in the manner in which they were exposed to factors causing skin cancer (Case Control Study 2011). For research outcomes that require longer periods of time and are rare in nature, case-control study proves to be the most effective method of research. The cost of such research methods is also less as follow up of the participants at particular time intervals is not needed in this type of study. The effects of the exposures are better studied with this method, although this might prove to be harmful in certain cases as well. For example, effect of exposure of groups of participants to smoke of tobacco might prove to be an unhealthy research approach. However, in order to study risks associated with the exposures and if there are multiple risk factors, the case-control study proves to be most effective for research (Dumitrescu 2012). Conclusion: The different types of observational studies and their concepts could be learnt that may be suitably applied by the researcher based on the need of the research, such as in the case of identification of facial expression during deception. Such concepts are also essential to determine research results that are valid and reliable. References Babbie, E.R. (2010) The Practice of Social Research. Connecticut: Cengage Learning. Babbie, E.R. (2012) The Basics of Social Research, 6th ed. Connecticut: Cengage Learning. Bailey, K. (2008) Methods of Social Research, 4th Edition. New York: Simon and Schuster. Baumgarten, M. (2012) Paradigm Wars - Validity and Reliability in Qualitative Research (Google eBook). Germany: GRIN Verlag. Case Control Study (2011) Gwumc. [Online]. Available at: http://www.gwumc.edu/library/tutorials/studydesign101/casecontrols.html [Accessed 12 August 2013]. CEM - Research - Bias & Confounding (n.d.) Collemergencymed. [Online]. Available at: http://www.collemergencymed.ac.uk/CEM/Research/technical_guide/biasconfound.htm [Accessed 9 August 2013]. Dumitrescu, A.L. (2012) Understanding Periodontal Research (Google eBook). New York: Springer. Golafshani, N. (2003) Understanding Reliability and Validity in Qualitative Research. The Qualitative Report, 8.4, pp.597-607. Gravetter, F.J. and L.B. Forzano (2010) Research Methods for the Behavioral Sciences (Gravetter). Connecticut: Cengage Learning. Howlett, B., Rogo, E.J. and T.G. Shelton (2013) Evidence-based Practice for Health Professionals: An Interprofessional Approach. Massachusetts: Jones & Bartlett Publishers. Market Research Challenge: Analysis Bias (2013) Researchrockstar. [Online]. Available at: http://www.researchrockstar.com/market-research-challenge-analysis-bias/ [Accessed 10 August 2013]. Monsen, E.R. and L.V. Horn (2008) Research: Successful Approaches. Chicago: American Dietetic Association. Nestor, P.G. and R.K. Schutt (2011) Research Methods in Psychology: Investigating Human Behavior. London: SAGE. Pannucci, C.J. and E.G. Wilkins (2011) Identifying and Avoiding Bias in Research. Plastic and Reconstructive Surgery. NCBI. [Online]. Available at: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2917255/ [Accessed 11 August 2013]. Peat, J., Mellis, C. and K. Williams (2002) Health Science Research: A Handbook of Quantitative Methods. London: SAGE. Riegelman, R.K. (2005) Studying a Study and Testing a Test: How to Read the Medical Evidence. Philadelphia: Lippincott Williams & Wilkins. Stommel, M. and C. Wills (2004) Clinical Research: Concepts and Principles for Advanced Practice Nurses. Philadelphia: Lippincott Williams & Wilkins. What is case-control study? (2008) EHIB. [Online]. Available at: http://www.ehib.org/faq.jsp?faq_key=34 [Accessed 15 August 2013]. What is a cohort study? (2001) EHIB. [Online]. Available at: http://www.ehib.org/faq.jsp?faq_key=37 [Accessed 15 August 2013]. What is a cross sectional study? (2009) EHIB. [Online]. Available at: http://www.ehib.org/faq.jsp?faq_key=41 [Accessed 15 August 2013]. World Health Organization (2001) Health Research Methodology:  A Guide for Training in Research Methods. Switzerland: World Health Organization. Read More