5 Bias and confounding

Size: px
Start display at page:

Download "5 Bias and confounding"

Transcription

1 Bias and confounding 37 5 Bias and confounding Learning objectives In this chapter students learn about Most important factors that can systematically impair the results of a population based study, like selection and information bias Definitions and common examples of biases and confounders and Methods to avoid bias and confounding in the study design and analysis Study results drawn from samples can be influenced by random errors. This effect is estimated via statistical test procedures. In addition, there are a number of systematic biases which might lead to a misleading study result. Errors are created by chance whereas biases are systematic falsifications of data. Several factors can lower the quality of study results and some reduce the possibility to find a real risk factor - which lowers the power of a study. Other factors can skew the results. Sometimes the direction in which the data are skewed can be estimated - but this is not always the case. The design of the study or behavior of participants can impair and skew the results. Most interesting are systematic biases, which are usually divided into three broad categories: Selection bias occurs during the choice and selection of participants Information bias occurs when information, e.g. from a questionnaire or interview, is wrongly collected Confounders are additional factors which can distort the relationship between exposure and outcome 5.1 Selection bias As indicated above, this type of bias might occur during the sampling and selection of participants for the study. In the following example the researcher is interested in knowing how many women attended a mammography screening for breast cancer prevention during the last years. He is planning a cross-sectional study and sending invitation letters to a representative group of 1000 women between 40 and 60 years. From these 1000 women who were invited, only 300 participate in the study. This corresponds to a 30% response rate and it also means that 700 women did not answer. What happens if only those 300 questionnaires are analyzed? To answer the question it is important to find out how many women participate normally in a mammography screening. According to a large-scale Danish study about 70% of women participated in a mammography screening during the years in Copenhagen (Euler-Chelpin et al. 2006). What, in comparison, is the result of the new study? The percentage of participants actually looks quite dif-

2 38 Introduction in Epidemiology ferent. 95 to 97% of the 300 women report that they regularly attend a mammography screening. Due to the low response rate we can assume that mostly women who are familiar with and interested in the topic will answer the questionnaire. Or on the other hand, women that are not familiar with the screening might not be interested in answering a questionnaire about the topic. We expect an invalid result due to this selfselection process among the participants. This is a typical and mostly inevitable problem in epidemiological studies. Another problem might be the choice of a wrong participant group, which is mostly recognized just after finishing a study. Take for example a case control study on the relationship of alcohol consumption and lung cancer. A recruitment of lung cancer cases is done from a surgical department of a clinic. An easy identifiable control group would be non-lung cancer patients from an emergency department of the same hospital. There are enough patients and recruiting is easy because information can be gathered directly in the hospital. After analyzing the descriptive data we might discover that our control group is not suitable for comparison. We might have selected a group which consumes more alcohol than the population average since accidents regularly occur in connection with alcohol consumption. In this case the control group is not suitable to answer our study question because this group does not represent the alcohol consumption in the underlying population. Further examples for selection bias: In cross-sectional studies we want to show the whole profile of the population. Usually this is not feasible since not all parts of a population can be easily reached. Severely ill people for example are hard to reach when they are in hospital. In a cohort study, selection problems occur during followup - the longitudinal observation of a population. Some might leave the region or country where the study is conducted and cannot be traced and others refuse further participation. Another well-known type of selection bias is the "healthy-worker bias". In an occupational cohort there are usually more healthy people than in the general population, because ill, disabled and older persons mostly do not work anymore. 5.2 Information bias This kind of bias occurs when information is assessed wrongly by e.g. questionnaires or interviewers. What sounds like faulty planning (i.e. badly formulated or incomprehensible questions or questions that do not measure what you want to know) might rather be due to the fact that participants do not answer correctly. How can you use a questionnaire to find out if someone has diabetes? Are you injecting insulin? Are you taking pills? Did the doctor tell you not to eat sugar anymore? Which question is the right one? Even if the diagnosis is based on data from a general practitioner, this data must be available in the first place. In the beginning the researcher believes it is easy to collect relevant data, but during the planning process of a study he or she realizes that problems occur even when collecting "simple" information. Assessing the disease status usually is a simple question. But exposure status such as the dietary patterns like in the EPIC study or mobile phone use like in the INTER- PHONE study are quite difficult to objectively assess Therefore complex catalogues with questions are used but mistakes can be made here as well. It is important to think about that in advance. An important measure to assess the quality of a measurement instrument such as a questionnaire is its validity. It indicates whether the instrument really measures what it is supposed to measure or what the researcher is interested in.

3 Bias and confounding 39 This is a difficult task and one starts with assessing the instrument's repeatability - if it leads to the same results when used in the same persons repeatedly. A common example of information bias usually occurs in case-control studies. It is called recall bias. Cases and controls might not remember the exact information about the exposure if asked retrospectively. Another aspect of this bias is when cases and controls are concerned about the exposure and remember information about it to a different extent. 5.3 Confounding Study results can also be distorted by a third factor that occurs in a population and is connected to the disease under study. A good example: You might know the story about storks bringing babies. This relationship can even be shown statistically. Storks appear more frequently in regions with many children (Mathews 2000, Höfer et al. 2004). Should this statistical relationship be accepted and the biological knowledge be reconsidered? Instead we might consider a possible third factor which creates this association and allows consistency with current knowledge. We know that storks build their nests often in rural areas. Families who like to have more children more often live in rural areas, as well. There is a third factor (confounder) which is associated with the exposure (number of storks in the area) and the outcome (number of children born in the area). This factor can be accounted for in the analysis and we do not need to think about new biological theories to explain the increase in number of babies. In the following figure 5.1 this relationship is shown graphically: Exposure Outcome Confounder Figure 5.1: Relationship between exposure, outcome and confounder Requirements to identify a third factor as a confounder: 1. A confounder is associated with the exposure. 2. A confounder is associated with the outcome. 3. The real relationship between exposure and outcome will be skewed if the confounder is not considered. 4. The confounder is not an intermediate step or necessary requirement for the relationship between exposure and outcome. The risk for most diseases increases with older age. If you analyze the effect of an exposure which occurs more frequent in certain age groups then age can be a confounder. Another very common confounder is sex because some exposures as well as outcomes are differently distributed among men and women. Other common confounders are smoking, socio-economic status and nationality. Depending on the study question there might be other confounders as well.

4 40 Introduction in Epidemiology If you want to analyze the relationship between physical activity and high blood pressure you could divide the participants into two groups according to the amount of physical activity: one group with a lot and the other one with little activity. You will probably find higher blood pressure among the participants in the group with little activity. Does this prove a relationship? Furthermore it is well known that the blood pressure increases with higher age and that older people practice less physical activity than younger ones. Our division of the participants into groups leads also to a division of older and younger participants. Among the older people there will be higher blood pressure independent of their amount of physical activity. Age is a confounder here for the relationship between physical activity and blood pressure. In a study on the relationship between alcohol and lung cancer, smoking would have to be considered as a confounder since its effect on the outcome is well known. Among persons who drink alcohol, the proportion of smokers is higher than in the general population. Smoking and lung cancer are associated as well. 5.4 Avoiding bias and confounding To avoid all possible errors, bias and confounding is the primary goal of every epidemiological study. It should be considered as much as possible during planning of the study. Selection and information bias can only be considered in the study design and just to a very limited extent in the analysis. In contrast to selection and information bias, confounding can be considered in the study design but also in statistical analyses. Confounders must always be considered in population based studies to get a result which shows the real relationship between exposure and outcome. Information about a confounder must be known beforehand and sampled before. Even though you can consider confounders in statistical analyses, you must also consider them in the study design. If for example the smoking status and age of the participants were not assessed they cannot be considered as confounders in the analysis Selection and information bias in planning a study To avoid selection bias it is important to attain a response rate as high as possible. In the follow-up of a cohort study all information should be collected in detail. Data base management is a very important task in the study plan. The more accurately the data is documented, the easier it is to follow-up participants, evaluate study processes and detect errors. It is important to have comprehensive knowledge on how to develop a questionnaire to avoid information bias. If possible, it is recommended to use validated and widely used instruments. The process of data collection should be standardized and interviewers should be trained to conduct interviews in a standardized way. A comprehensive data base management is also important to avoid information bias. Data collection procedure must be assessed to allow finding and solving problems; when collecting data, the time of data collection and even identification numbers of used devices needs to be documented. If a device is not functioning properly the relevant data can be retrieved easily.

5 Bias and confounding Considering confounding in the study design In a study where it is known that smoking can be a confounder, it might be useful to just recruit non-smokers. By doing this, the confounding effect is eliminated because none of the participants smoke. This is also done when sex is a possible confounder and only male or only female participants are recruited. Another method to avoid confounding that can be applied is using a matched design. This method is frequently used in case-control studies and describes the process of aligning cases and controls according to certain factors. An example is individual matching according to age and sex of the participants. If a case is female and 30 years of age, then the control should be also a woman with a similar age. This way the confounders age and sex are equally distributed in the case and control groups and cannot impair the association between exposure and outcome in each of the groups Considering confounders when analyzing the data A variable should only be considered as confounder if it is differently distributed among the groups that are compared. An unequal distribution of these factors might impair the real association between the exposure and outcome of interest. An analysis with confounders always lowers the power of a study and therefore its significance and should be avoided if possible. However, if typical confounders like smoking status, age and others are differently distributed among cases and controls they must be included in the analysis. Otherwise the presented results are not valid. There are three methods to consider confounding: Sub-group analysis / stratification The study population is divided into sub-groups in which the analysis is done separately. This method is often called stratification and is a basic idea for simple cases of confounding. Standardization Mostly used as age-standardization in health reports. The two relevant methods of standardization are described in chapter 5. Multiple regression Regression analyses are an extension of sub-group analyses that can be used for more complex data and several confounders. To conduct a sub-group or stratified analysis, the study population has to be divided into groups according to categories in the confounding variable (e.g. smokers and non-smokers, older and younger participants; also more than two groups are possible like 10-year age groups). The analysis - for example of an OR - is done for the whole population and for each of the sub-groups (also called strata). By comparing those results it can be seen whether or not the third variable is a confounder and how strong the actual relationship between exposure and outcome is. The following tables show a stratified analysis of a fictive case-control study on malaria infection. In total, 150 cases (persons with malaria infection) and 150 controls (without malaria infection) were analyzed. The first step was to investigate the influence of the participants' sex on the risk of infection (see table 5.1).

6 42 Introduction in Epidemiology Table 5.1: Crude results of a case-control study on the effect of sex on malaria infection Cases Controls Total Exposure Men Women Total The OR is the measure of association in a case-control study. Using the numbers from table 5.1 the following OR can be calculated: OR= (88*82)/(62*68)=1.71. It seems that men have a higher chance of malaria infection than women. But might there be a confounder? A confounder would have to be associated with the exposure (sex) and the outcome (malaria infection). The work place of the participants might be such a confounding factor, since malaria infections occur more often among persons who work outside and in areas where the anopheles mosquito can be found. It is also known that in some regions men are usually working on the fields. A stratified analysis is shown in table 5.2. Table 5.2: Stratified results of a case-control study on the effect of sex on malaria infection Persons, who did not work on the field Persons, who worked on the field Cases Controls Total Cases Controls Total Men Men Women Women Total Total OR=1.00 OR=1.06 If the whole study sample (n=300) is divided into persons who did not work on the field (n=219) and persons who did (n=81), we have a different result. In the stratified analysis the association between sex and malaria infection disappears. How can we interpret these results? The crude OR for the whole study sample (n=300) shows the combined effect of sex and field work on malaria infection. When stratified according to field work there is no effect of sex on malaria infection. In this case stratified ORs show the isolated effect of sex on malaria infection. We can conclude that there is no difference in malaria infections between men and women but rather between persons who work on the field and persons who do not work on the field. The crude (or not stratified) analysis is wrong, because it is impaired by the variable working on the field. A confounding effect can be seen if the real effect is increased, decreased or hidden. One characteristic of a confounder is that the effects have basically the same size in the different strata. There are exceptions or specific situational circumstances, such as when you look at the case of asbestos exposure and risk of lung cancer. In this case smoking is a confounder but in the stratified analysis it can be seen that smokers have a much higher risk of developing lung cancer from asbestos than non-smokers. The OR in both strata are very different. This phenomenon is called interaction or effect modification.

7 Bias and confounding 43 Remember In every population based study there is a potential for systematic bias due to characteristics of the study population or methods of data collection. The three forms are selection bias, information bias and confounding. Careful planning and conducting of the study help to avoid or reduce selection bias, information bias and confounding. Confounding factors can be accounted for by sub-group or stratified analyses, regression analyses or standardization. For further reading: Aschengrau A, Seage GR. Essentials of Epidemiology in Public Health. 2 nd edition. Jones and Bartlett Publishers 2008 relevant chapters: 10, 11 Bhopal R: Concepts of Epidemiology. 2 nd edition, Oxford University Press 2009 relevant chapter: 4 Gordis L: Epidemiology. 4 th edition. Saunders 2009 relevant chapter: 15 Scientific publication: Euler-Chelpin M, Olsen AH, Njor S, Vejborg I, Schwartz W, Lynge E. Women s patterns of participation in mammography screening in Denmark. Eur J Epidemiol 21(203-9, Hoefer T, Przyrembel H, Verleger S: For the Classroom: New evidence for the Theory of the Stork. Pediatr Perinat Epidemiol. 18:88-92, Further reading and online resources

8 44 Introduction in Epidemiology

Higher Psychology RESEARCH REVISION

Higher Psychology RESEARCH REVISION Higher Psychology RESEARCH REVISION 1 The biggest change from the old Higher course (up to 2014) is the possibility of an analysis and evaluation question (8-10) marks asking you to comment on aspects

More information

To evaluate a single epidemiological article we need to know and discuss the methods used in the underlying study.

To evaluate a single epidemiological article we need to know and discuss the methods used in the underlying study. Critical reading 45 6 Critical reading As already mentioned in previous chapters, there are always effects that occur by chance, as well as systematic biases that can falsify the results in population

More information

Confounding, Effect modification, and Stratification

Confounding, Effect modification, and Stratification Confounding, Effect modification, and Stratification Tunisia, 30th October 2014 Acknowledgment: Kostas Danis Takis Panagiotopoulos National Schoool of Public Health, Athens, Greece takis.panagiotopoulos@gmail.com

More information

ISC- GRADE XI HUMANITIES ( ) PSYCHOLOGY. Chapter 2- Methods of Psychology

ISC- GRADE XI HUMANITIES ( ) PSYCHOLOGY. Chapter 2- Methods of Psychology ISC- GRADE XI HUMANITIES (2018-19) PSYCHOLOGY Chapter 2- Methods of Psychology OUTLINE OF THE CHAPTER (i) Scientific Methods in Psychology -observation, case study, surveys, psychological tests, experimentation

More information

Trial Designs. Professor Peter Cameron

Trial Designs. Professor Peter Cameron Trial Designs Professor Peter Cameron OVERVIEW Review of Observational methods Principles of experimental design applied to observational studies Population Selection Looking for bias Inference Analysis

More information

Confounding Bias: Stratification

Confounding Bias: Stratification OUTLINE: Confounding- cont. Generalizability Reproducibility Effect modification Confounding Bias: Stratification Example 1: Association between place of residence & Chronic bronchitis Residence Chronic

More information

Introduction. Step 2: Student Role - Your Plan of Action. Confounding. Good luck and have fun!

Introduction. Step 2: Student Role - Your Plan of Action. Confounding. Good luck and have fun! Confounding Introduction You have learned that random error and bias must be considered as possible explanations for an observed association between an exposure and disease. This week we will examine the

More information

11 questions to help you make sense of a case control study

11 questions to help you make sense of a case control study Critical Appraisal Skills Programme (CASP) making sense of evidence 11 questions to help you make sense of a case control study How to use this appraisal tool Three broad issues need to be considered when

More information

INTERNAL VALIDITY, BIAS AND CONFOUNDING

INTERNAL VALIDITY, BIAS AND CONFOUNDING OCW Epidemiology and Biostatistics, 2010 J. Forrester, PhD Tufts University School of Medicine October 6, 2010 INTERNAL VALIDITY, BIAS AND CONFOUNDING Learning objectives for this session: 1) Understand

More information

Main objective of Epidemiology. Statistical Inference. Statistical Inference: Example. Statistical Inference: Example

Main objective of Epidemiology. Statistical Inference. Statistical Inference: Example. Statistical Inference: Example Main objective of Epidemiology Inference to a population Example: Treatment of hypertension: Research question (hypothesis): Is treatment A better than treatment B for patients with hypertension? Study

More information

Stratified Tables. Example: Effect of seat belt use on accident fatality

Stratified Tables. Example: Effect of seat belt use on accident fatality Stratified Tables Often, a third measure influences the relationship between the two primary measures (i.e. disease and exposure). How do we remove or control for the effect of the third measure? Issues

More information

Observational Studies and Experiments. Observational Studies

Observational Studies and Experiments. Observational Studies Section 1 3: Observational Studies and Experiments Data is the basis for everything we do in statistics. Every method we use in this course starts with the collection of data. Observational Studies and

More information

Handout 11: Understanding Probabilities Associated with Medical Screening Tests STAT 100 Spring 2016

Handout 11: Understanding Probabilities Associated with Medical Screening Tests STAT 100 Spring 2016 Example: Using Mammograms to Screen for Breast Cancer Gerd Gigerenzer, a German psychologist, has conducted several studies to investigate physicians understanding of health statistics (Gigerenzer 2010).

More information

Define the population Determine appropriate sample size Choose a sampling design Choose an appropriate research design

Define the population Determine appropriate sample size Choose a sampling design Choose an appropriate research design Numbers! Observation Study: observing individuals and measuring variables of interest without attempting to influence the responses Correlational Research: examining the relationship between two variables

More information

Bias. Sam Bracebridge

Bias. Sam Bracebridge Bias Sam Bracebridge Bias Errors in epidemiological measurements Based on the slides of Sam Bracebridge By the end of the lecture fellows will be able to Define bias Identify different types of bias Explain

More information

Psychology 205, Revelle, Fall 2014 Research Methods in Psychology Mid-Term. Name:

Psychology 205, Revelle, Fall 2014 Research Methods in Psychology Mid-Term. Name: Name: 1. (2 points) What is the primary advantage of using the median instead of the mean as a measure of central tendency? It is less affected by outliers. 2. (2 points) Why is counterbalancing important

More information

Bias. A systematic error (caused by the investigator or the subjects) that causes an incorrect (overor under-) estimate of an association.

Bias. A systematic error (caused by the investigator or the subjects) that causes an incorrect (overor under-) estimate of an association. Bias A systematic error (caused by the investigator or the subjects) that causes an incorrect (overor under-) estimate of an association. Here, random error is small, but systematic errors have led to

More information

Objectives. Quantifying the quality of hypothesis tests. Type I and II errors. Power of a test. Cautions about significance tests

Objectives. Quantifying the quality of hypothesis tests. Type I and II errors. Power of a test. Cautions about significance tests Objectives Quantifying the quality of hypothesis tests Type I and II errors Power of a test Cautions about significance tests Designing Experiments based on power Evaluating a testing procedure The testing

More information

Villarreal Rm. 170 Handout (4.3)/(4.4) - 1 Designing Experiments I

Villarreal Rm. 170 Handout (4.3)/(4.4) - 1 Designing Experiments I Statistics and Probability B Ch. 4 Sample Surveys and Experiments Villarreal Rm. 170 Handout (4.3)/(4.4) - 1 Designing Experiments I Suppose we wanted to investigate if caffeine truly affects ones pulse

More information

Chapter Three Research Methodology

Chapter Three Research Methodology Chapter Three Research Methodology Research Methods is a systematic and principled way of obtaining evidence (data, information) for solving health care problems. 1 Dr. Mohammed ALnaif METHODS AND KNOWLEDGE

More information

How to use this appraisal tool: Three broad issues need to be considered when appraising a case control study:

How to use this appraisal tool: Three broad issues need to be considered when appraising a case control study: CASP Checklist: 11 questions to help you make sense of a Case Control Study How to use this appraisal tool: Three broad issues need to be considered when appraising a case control study: Are the results

More information

Handout 1: Introduction to the Research Process and Study Design STAT 335 Fall 2016

Handout 1: Introduction to the Research Process and Study Design STAT 335 Fall 2016 DESIGNING OBSERVATIONAL STUDIES As we have discussed, for the purpose of establishing cause-and-effect relationships, observational studies have a distinct disadvantage in comparison to randomized comparative

More information

You can t fix by analysis what you bungled by design. Fancy analysis can t fix a poorly designed study.

You can t fix by analysis what you bungled by design. Fancy analysis can t fix a poorly designed study. You can t fix by analysis what you bungled by design. Light, Singer and Willett Or, not as catchy but perhaps more accurate: Fancy analysis can t fix a poorly designed study. Producing Data The Role of

More information

Psychology: The Science

Psychology: The Science Psychology: The Science How Psychologists Do Research Ex: While biking, it seems to me that drivers of pick up trucks aren t as nice as car drivers. I make a hypothesis or even develop a theory that p/u

More information

Two-sample Categorical data: Measuring association

Two-sample Categorical data: Measuring association Two-sample Categorical data: Measuring association Patrick Breheny October 27 Patrick Breheny University of Iowa Biostatistical Methods I (BIOS 5710) 1 / 40 Introduction Study designs leading to contingency

More information

STATISTICS 8 CHAPTERS 1 TO 6, SAMPLE MULTIPLE CHOICE QUESTIONS

STATISTICS 8 CHAPTERS 1 TO 6, SAMPLE MULTIPLE CHOICE QUESTIONS STATISTICS 8 CHAPTERS 1 TO 6, SAMPLE MULTIPLE CHOICE QUESTIONS Circle the best answer. This scenario applies to Questions 1 and 2: A study was done to compare the lung capacity of coal miners to the lung

More information

Welcome to this series focused on sources of bias in epidemiologic studies. In this first module, I will provide a general overview of bias.

Welcome to this series focused on sources of bias in epidemiologic studies. In this first module, I will provide a general overview of bias. Welcome to this series focused on sources of bias in epidemiologic studies. In this first module, I will provide a general overview of bias. In the second module, we will focus on selection bias and in

More information

8/10/2012. Education level and diabetes risk: The EPIC-InterAct study AIM. Background. Case-cohort design. Int J Epidemiol 2012 (in press)

8/10/2012. Education level and diabetes risk: The EPIC-InterAct study AIM. Background. Case-cohort design. Int J Epidemiol 2012 (in press) Education level and diabetes risk: The EPIC-InterAct study 50 authors from European countries Int J Epidemiol 2012 (in press) Background Type 2 diabetes mellitus (T2DM) is one of the most common chronic

More information

Bias and confounding. Mads Kamper-Jørgensen, associate professor, Section of Social Medicine

Bias and confounding. Mads Kamper-Jørgensen, associate professor, Section of Social Medicine Bias and confounding Mads Kamper-Jørgensen, associate professor, maka@sund.ku.dk PhD-course in Epidemiology l 7 February 2017 l Slide number 1 The world according to an epidemiologist Exposure Outcome

More information

observational studies Descriptive studies

observational studies Descriptive studies form one stage within this broader sequence, which begins with laboratory studies using animal models, thence to human testing: Phase I: The new drug or treatment is tested in a small group of people for

More information

Challenges of Observational and Retrospective Studies

Challenges of Observational and Retrospective Studies Challenges of Observational and Retrospective Studies Kyoungmi Kim, Ph.D. March 8, 2017 This seminar is jointly supported by the following NIH-funded centers: Background There are several methods in which

More information

UNIT 5 - Association Causation, Effect Modification and Validity

UNIT 5 - Association Causation, Effect Modification and Validity 5 UNIT 5 - Association Causation, Effect Modification and Validity Introduction In Unit 1 we introduced the concept of causality in epidemiology and presented different ways in which causes can be understood

More information

Vocabulary. Bias. Blinding. Block. Cluster sample

Vocabulary. Bias. Blinding. Block. Cluster sample Bias Blinding Block Census Cluster sample Confounding Control group Convenience sample Designs Experiment Experimental units Factor Level Any systematic failure of a sampling method to represent its population

More information

Purpose. Study Designs. Objectives. Observational Studies. Analytic Studies

Purpose. Study Designs. Objectives. Observational Studies. Analytic Studies Purpose Study Designs H.S. Teitelbaum, DO, PhD, MPH, FAOCOPM AOCOPM Annual Meeting Introduce notions of study design Clarify common terminology used with description and interpretation of information collected

More information

CRITICAL APPRAISAL SKILLS PROGRAMME Making sense of evidence about clinical effectiveness. 11 questions to help you make sense of case control study

CRITICAL APPRAISAL SKILLS PROGRAMME Making sense of evidence about clinical effectiveness. 11 questions to help you make sense of case control study CRITICAL APPRAISAL SKILLS PROGRAMME Making sense of evidence about clinical effectiveness 11 questions to help you make sense of case control study General comments Three broad issues need to be considered

More information

University of Wollongong. Research Online. Australian Health Services Research Institute

University of Wollongong. Research Online. Australian Health Services Research Institute University of Wollongong Research Online Australian Health Services Research Institute Faculty of Business 2011 Measurement of error Janet E. Sansoni University of Wollongong, jans@uow.edu.au Publication

More information

Controlling Bias & Confounding

Controlling Bias & Confounding Controlling Bias & Confounding Chihaya Koriyama August 5 th, 2015 QUESTIONS FOR BIAS Key concepts Bias Should be minimized at the designing stage. Random errors We can do nothing at Is the nature the of

More information

Clinical problems and choice of study designs

Clinical problems and choice of study designs Evidence Based Dentistry Clinical problems and choice of study designs Asbjørn Jokstad University of Oslo, Norway Nov 21 2001 1 Manipulation with intervention Yes Experimental study No Non-experimental

More information

3. Factors such as race, age, sex, and a person s physiological state are all considered determinants of disease. a. True

3. Factors such as race, age, sex, and a person s physiological state are all considered determinants of disease. a. True / False 1. Epidemiology is the basic science of public health. LEARNING OBJECTIVES: CNIA.BOYL.17.2.1 - Define epidemiology. 2. Within the field of epidemiology, the term distribution refers to the relationship

More information

Module 2: Fundamentals of Epidemiology Issues of Interpretation. Slide 1: Introduction. Slide 2: Acknowledgements. Slide 3: Presentation Objectives

Module 2: Fundamentals of Epidemiology Issues of Interpretation. Slide 1: Introduction. Slide 2: Acknowledgements. Slide 3: Presentation Objectives Slide 1: Introduction Issues of of Interpretation in in Epidemiologic Studies Developed through APTR Initiative to Enhance Prevention and Population Health Developed Education through in collaboration

More information

Sampling. (James Madison University) January 9, / 13

Sampling. (James Madison University) January 9, / 13 Sampling The population is the entire group of individuals about which we want information. A sample is a part of the population from which we actually collect information. A sampling design describes

More information

Chapter 1 - Sampling and Experimental Design

Chapter 1 - Sampling and Experimental Design Chapter 1 - Sampling and Experimental Design Read sections 1.3-1.5 Sampling (1.3.3 and 1.4.2) Sampling Plans: methods of selecting individuals from a population. We are interested in sampling plans such

More information

Science is a way of learning about the natural world by observing things, asking questions, proposing answers, and testing those answers.

Science is a way of learning about the natural world by observing things, asking questions, proposing answers, and testing those answers. Science 9 Unit 1 Worksheet Chapter 1 The Nature of Science and Scientific Inquiry Online resources: www.science.nelson.com/bcscienceprobe9/centre.html Remember to ask your teacher whether your classroom

More information

Confounding and Effect Modification. John McGready Johns Hopkins University

Confounding and Effect Modification. John McGready Johns Hopkins University This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike License. Your use of this material constitutes acceptance of that license and the conditions of use of materials on this

More information

Effect measure modification. Outline. Definition. Gustaf Edgren, PhD Karolinska Institutet

Effect measure modification. Outline. Definition. Gustaf Edgren, PhD Karolinska Institutet Effect measure modification Gustaf Edgren, PhD Karolinska Institutet Outline Definition and terminology Effect modification vs. confounding Scale dependence Assessment of effect measure modification Examples

More information

In the broadest sense of the word, the definition of research includes any gathering of data, information, and facts for the advancement of knowledge.

In the broadest sense of the word, the definition of research includes any gathering of data, information, and facts for the advancement of knowledge. What is research? "In the broadest sense of the word, the definition of research includes any gathering of data, information, and facts for the advancement of knowledge." - Martyn Shuttleworth "Research

More information

Case-Control Studies

Case-Control Studies Case-Control Studies Marc Schenker M.D., M.P.H Dept. of Public Health Sciences UC Davis Marc Schenker M.D., M.P.H, UC Davis 1 Case-Control Studies OBJECTIVES After this session, you will be familiar with:

More information

In the 1700s patients in charity hospitals sometimes slept two or more to a bed, regardless of diagnosis.

In the 1700s patients in charity hospitals sometimes slept two or more to a bed, regardless of diagnosis. Control Case In the 1700s patients in charity hospitals sometimes slept two or more to a bed, regardless of diagnosis. This depicts a patient who finds himself lying with a corpse (definitely a case ).

More information

HIGH LEVELS OF PREVENTABLE CHRONIC DIEASE, INJURY AND MENTAL HEALTH PROBLEMS

HIGH LEVELS OF PREVENTABLE CHRONIC DIEASE, INJURY AND MENTAL HEALTH PROBLEMS HIGH LEVELS OF PREVENTABLE CHRONIC DIEASE, INJURY AND MENTAL HEALTH PROBLEMS Let s look at CANCER AS A WHOLE. What is the nature of the problem? Well, cancer is the growth of cells within the body. We

More information

An Introduction to Epidemiology

An Introduction to Epidemiology An Introduction to Epidemiology Wei Liu, MPH Biostatistics Core Pennington Biomedical Research Center Baton Rouge, LA Last edited: January, 14 th, 2014 TABLE OF CONTENTS Introduction.................................................................

More information

The following are questions that students had difficulty with on the first three exams.

The following are questions that students had difficulty with on the first three exams. The following are questions that students had difficulty with on the first three exams. Exam 1 1. A measure has construct validity if it: a) really measures what it is supposed to measure b) appears, on

More information

Psychology - MR. CALLAWAY Mundy s Mill High School Unit RESEARCH METHODS

Psychology - MR. CALLAWAY Mundy s Mill High School Unit RESEARCH METHODS Psychology - MR. CALLAWAY Mundy s Mill High School Unit 2.1 - RESEARCH METHODS Intro to Research How do psychologists ask & answer questions? Differentiate types of research with regard to purpose, strengths,

More information

GCSE PSYCHOLOGY UNIT 2 FURTHER RESEARCH METHODS

GCSE PSYCHOLOGY UNIT 2 FURTHER RESEARCH METHODS GCSE PSYCHOLOGY UNIT 2 FURTHER RESEARCH METHODS GCSE PSYCHOLOGY UNIT 2 SURVEYS SURVEYS SURVEY = is a method used for collecting information from a large number of people by asking them questions, either

More information

TRACER STUDIES ASSESSMENTS AND EVALUATIONS

TRACER STUDIES ASSESSMENTS AND EVALUATIONS TRACER STUDIES ASSESSMENTS AND EVALUATIONS 1 INTRODUCTION This note introduces the reader to tracer studies. For the Let s Work initiative, tracer studies are proposed to track and record or evaluate the

More information

Study design. Chapter 64. Research Methodology S.B. MARTINS, A. ZIN, W. ZIN

Study design. Chapter 64. Research Methodology S.B. MARTINS, A. ZIN, W. ZIN Chapter 64 Study design S.B. MARTINS, A. ZIN, W. ZIN Research Methodology Scientific methodology comprises a set of rules and procedures to investigate the phenomena of interest. These rules and procedures

More information

Designing Studies of Diagnostic Imaging

Designing Studies of Diagnostic Imaging Designing Studies of Diagnostic Imaging Chaya S. Moskowitz, PhD With thanks to Nancy Obuchowski Outline What is study design? Building blocks of imaging studies Strategies to improve study efficiency What

More information

Running head: FALSE MEMORY AND EYEWITNESS TESTIMONIAL Gomez 1

Running head: FALSE MEMORY AND EYEWITNESS TESTIMONIAL Gomez 1 Running head: FALSE MEMORY AND EYEWITNESS TESTIMONIAL Gomez 1 The Link Between False Memory and Eyewitness Testimonial Marianna L. Gomez El Paso Community College Carrie A. Van Houdt FALSE MEMORY AND EYEWITNESS

More information

Handout 16: Opinion Polls, Sampling, and Margin of Error

Handout 16: Opinion Polls, Sampling, and Margin of Error Opinion polls involve conducting a survey to gauge public opinion on a particular issue (or issues). In this handout, we will discuss some ideas that should be considered both when conducting a poll and

More information

Confounding and Interaction

Confounding and Interaction Confounding and Interaction Why did you do clinical research? To find a better diagnosis tool To determine risk factor of disease To identify prognosis factor To evaluate effectiveness of therapy To decide

More information

AP Statistics Exam Review: Strand 2: Sampling and Experimentation Date:

AP Statistics Exam Review: Strand 2: Sampling and Experimentation Date: AP Statistics NAME: Exam Review: Strand 2: Sampling and Experimentation Date: Block: II. Sampling and Experimentation: Planning and conducting a study (10%-15%) Data must be collected according to a well-developed

More information

The Research Enterprise in Psychology Chapter 2

The Research Enterprise in Psychology Chapter 2 The Research Enterprise in Psychology Chapter 2 This multimedia product and its contents are protected under copyright law. The following are prohibited by law: any public performance or display, including

More information

Quitting. Study Guide. Information for teachers. The accompanying factsheets: The main resource:

Quitting. Study Guide. Information for teachers.   The accompanying factsheets: The main resource: www.nosmokes.com.au Quitting Study Guide Information for teachers This section looks at quitting. It explains the process of addiction and looks at changing your thinking about smoking. It explores ways

More information

Design and Analysis of Observational Studies

Design and Analysis of Observational Studies review article Design 10.5005/jp-journals-10016-1079 and Analysis of Observational Studies 1 KP Suresh, 2 MR Gajendragad, 3 H Rahman ABSTRACT Appropriate study design forms the basis of any successful

More information

Tobacco, Alcohol and Drug Use

Tobacco, Alcohol and Drug Use Tobacco, Alcohol and Drug Use Healthier Community Assessment 2005 101 Tobacco Use Why It Is Important Cigarette smoking is the most preventable cause of disease and death in the United States. 37 Lung

More information

Patrick Breheny. January 28

Patrick Breheny. January 28 Confidence intervals Patrick Breheny January 28 Patrick Breheny Introduction to Biostatistics (171:161) 1/19 Recap Introduction In our last lecture, we discussed at some length the Public Health Service

More information

MODULE 1: BASIC DRUG KNOWLEDGE

MODULE 1: BASIC DRUG KNOWLEDGE MODULE 1: BASIC DRUG KNOWLEDGE 1.1 How Drugs Work A drug is a medicine or substance (but not food or water) which when ingested (or otherwise introduced into the body) alters the way the body works physically

More information

Epidemiological study design. Paul Pharoah Department of Public Health and Primary Care

Epidemiological study design. Paul Pharoah Department of Public Health and Primary Care Epidemiological study design Paul Pharoah Department of Public Health and Primary Care Molecules What/why? Organelles Cells Tissues Organs Clinical medicine Individuals Public health medicine Populations

More information

UNIT I SAMPLING AND EXPERIMENTATION: PLANNING AND CONDUCTING A STUDY (Chapter 4)

UNIT I SAMPLING AND EXPERIMENTATION: PLANNING AND CONDUCTING A STUDY (Chapter 4) UNIT I SAMPLING AND EXPERIMENTATION: PLANNING AND CONDUCTING A STUDY (Chapter 4) A DATA COLLECTION (Overview) When researchers want to make conclusions/inferences about an entire population, they often

More information

Experimental Research in HCI. Alma Leora Culén University of Oslo, Department of Informatics, Design

Experimental Research in HCI. Alma Leora Culén University of Oslo, Department of Informatics, Design Experimental Research in HCI Alma Leora Culén University of Oslo, Department of Informatics, Design almira@ifi.uio.no INF2260/4060 1 Oslo, 15/09/16 Review Method Methodology Research methods are simply

More information

Clever Hans the horse could do simple math and spell out the answers to simple questions. He wasn t always correct, but he was most of the time.

Clever Hans the horse could do simple math and spell out the answers to simple questions. He wasn t always correct, but he was most of the time. Clever Hans the horse could do simple math and spell out the answers to simple questions. He wasn t always correct, but he was most of the time. While a team of scientists, veterinarians, zoologists and

More information

Chapter 13. Experiments and Observational Studies

Chapter 13. Experiments and Observational Studies Chapter 13 Experiments and Observational Studies 1 /36 Homework Read Chpt 13 Do p312 1, 7, 9, 11, 17, 20, 25, 27, 29, 33, 40, 41 2 /36 Observational Studies In an observational study, researchers do not

More information

How do we know that smoking causes lung cancer?

How do we know that smoking causes lung cancer? How do we know that smoking causes lung cancer? Seif Shaheen Professor of Respiratory Epidemiology Centre for Primary Care and Public Health Blizard Institute Barts and The London School of Medicine and

More information

Lecture 4. Experiments and Observational Studies

Lecture 4. Experiments and Observational Studies Lecture 4 Experiments and Observational Studies Thought Question 1: In a study to relate two conditions, researchers often define one as the explanatory variable and other as the outcome or response variable.

More information

RESEARCH. Breast cancer mortality in organised mammography screening in Denmark: comparative study

RESEARCH. Breast cancer mortality in organised mammography screening in Denmark: comparative study 1 The Nordic Cochrane Centre, Rigshospitalet, University of Copenhagen, Denmark 2 Norwegian Institute of Public Health, Oslo, Norway Correspondence to: K J Jørgensen kj@cochrane.dk Cite this as: BMJ 2010;340:c1241

More information

INTRODUCTION TO EPIDEMIOLOGICAL STUDY DESIGNS PHUNLERD PIYARAJ, MD., MHS., PHD.

INTRODUCTION TO EPIDEMIOLOGICAL STUDY DESIGNS PHUNLERD PIYARAJ, MD., MHS., PHD. INTRODUCTION TO EPIDEMIOLOGICAL STUDY DESIGNS PHUNLERD PIYARAJ, MD., MHS., PHD. 1 OBJECTIVES By the end of this section, you will be able to: Provide a definition of epidemiology Describe the major types

More information

A Guide to Understanding Self-Injury

A Guide to Understanding Self-Injury A Guide to Understanding Self-Injury for Those Who Self-Injure What is Non-Suicidal Self-Injury? Non-Suicidal Self-Injury (NSSI), also referred to as self-injury or self-harm, is the deliberate and direct

More information

Measuring impact. William Parienté UC Louvain J PAL Europe. povertyactionlab.org

Measuring impact. William Parienté UC Louvain J PAL Europe. povertyactionlab.org Measuring impact William Parienté UC Louvain J PAL Europe povertyactionlab.org Course overview 1. What is evaluation? 2. Measuring impact 3. Why randomize? 4. How to randomize 5. Sampling and Sample Size

More information

CRITICAL APPRAISAL AP DR JEMAIMA CHE HAMZAH MD (UKM) MS (OFTAL) UKM PHD (UK) DEPARTMENT OF OPHTHALMOLOGY UKM MEDICAL CENTRE

CRITICAL APPRAISAL AP DR JEMAIMA CHE HAMZAH MD (UKM) MS (OFTAL) UKM PHD (UK) DEPARTMENT OF OPHTHALMOLOGY UKM MEDICAL CENTRE CRITICAL APPRAISAL AP DR JEMAIMA CHE HAMZAH MD (UKM) MS (OFTAL) UKM PHD (UK) DEPARTMENT OF OPHTHALMOLOGY UKM MEDICAL CENTRE MINGGU PENYELIDIKAN PERUBATAN & KESIHATAN PPUKM Lecture content Introduction

More information

TEACHERS TOPICS. The Role of Matching in Epidemiologic Studies. American Journal of Pharmaceutical Education 2004; 68 (3) Article 83.

TEACHERS TOPICS. The Role of Matching in Epidemiologic Studies. American Journal of Pharmaceutical Education 2004; 68 (3) Article 83. TEACHERS TOPICS American Journal of Pharmaceutical Education 2004; 68 (3) Article 83. The Role of Matching in Epidemiologic Studies Kevin W. Garey, PharmD College of Pharmacy, University of Houston Submitted

More information

Small Cell Lung Cancer Causes, Risk Factors, and Prevention

Small Cell Lung Cancer Causes, Risk Factors, and Prevention Small Cell Lung Cancer Causes, Risk Factors, and Prevention Risk Factors A risk factor is anything that affects your chance of getting a disease such as cancer. Learn more about the risk factors for small

More information

Section 6.1 Sampling. Population each element (or person) from the set of observations that can be made (entire group)

Section 6.1 Sampling. Population each element (or person) from the set of observations that can be made (entire group) Section 6.1 Sampling Population each element (or person) from the set of observations that can be made (entire group) Sample a subset of the population Census systematically getting information about an

More information

AP Psychology -- Chapter 02 Review Research Methods in Psychology

AP Psychology -- Chapter 02 Review Research Methods in Psychology AP Psychology -- Chapter 02 Review Research Methods in Psychology 1. In the opening vignette, to what was Alicia's condition linked? The death of her parents and only brother 2. What did Pennebaker s study

More information

Non-Small Cell Lung Cancer Causes, Risk Factors, and Prevention

Non-Small Cell Lung Cancer Causes, Risk Factors, and Prevention Non-Small Cell Lung Cancer Causes, Risk Factors, and Prevention Risk Factors A risk factor is anything that affects your chance of getting a disease such as cancer. Learn more about the risk factors for

More information

STUDY DESIGNS WHICH ONE IS BEST?

STUDY DESIGNS WHICH ONE IS BEST? STUDY DESIGNS WHICH ONE IS BEST? David Nunan, PhD Research Fellow and Tutor Nuffield Department of Primary Care Health Sciences and Oxford Centre for Evidence Based Medicine University of Oxford Exercise

More information

DEPARTMENT OF EPIDEMIOLOGY 2. BASIC CONCEPTS

DEPARTMENT OF EPIDEMIOLOGY 2. BASIC CONCEPTS DEPARTMENT OF EPIDEMIOLOGY EXIT COMPETENCIES FOR MPH GRADUATES IN GENERAL EPIDEMIOLOGY 1. DEFINITION AND HISTORICAL PERSPECTIVES 1. Historical Trends in the most common causes of death in the United States.

More information

Choosing the right study design

Choosing the right study design Choosing the right study design Main types of study design Randomised controlled trial (RCT) Cohort study Case-control study Cross-sectional study Case series/case note review Expert opinion BEST QUALITY

More information

Objectives. Data Collection 8/25/2017. Section 1-3. Identify the five basic sample techniques

Objectives. Data Collection 8/25/2017. Section 1-3. Identify the five basic sample techniques Section 1-3 Objectives Identify the five basic sample techniques Data Collection In research, statisticians use data in many different ways. Data can be used to describe situations. Data can be collected

More information

Examining Relationships Least-squares regression. Sections 2.3

Examining Relationships Least-squares regression. Sections 2.3 Examining Relationships Least-squares regression Sections 2.3 The regression line A regression line describes a one-way linear relationship between variables. An explanatory variable, x, explains variability

More information

Lecture 4. Confounding

Lecture 4. Confounding Lecture 4 Confounding Learning Objectives In this set of lectures we will: - Formally define confounding and give explicit examples of it s impact - Define adjustment and adjusted estimates conceptually

More information

Social Research (Complete) Agha Zohaib Khan

Social Research (Complete) Agha Zohaib Khan Social Research (Complete) Agha Zohaib Khan What is Research? Research is the systematic process of collecting and analysing information (data) in order to increase our understanding of the phenomenon

More information

Table of Contents. Introduction. 1. Diverse Weighing scale models. 2. What to look for while buying a weighing scale. 3. Digital scale buying tips

Table of Contents. Introduction. 1. Diverse Weighing scale models. 2. What to look for while buying a weighing scale. 3. Digital scale buying tips Table of Contents Introduction 1. Diverse Weighing scale models 2. What to look for while buying a weighing scale 3. Digital scale buying tips 4. Body fat scales 5. Is BMI the right way to monitor your

More information

Australian Longitudinal Study on Women's Health TRENDS IN WOMEN S HEALTH 2006 FOREWORD

Australian Longitudinal Study on Women's Health TRENDS IN WOMEN S HEALTH 2006 FOREWORD Australian Longitudinal Study on Women's Health TRENDS IN WOMEN S HEALTH 2006 FOREWORD The Longitudinal Study on Women's Health, funded by the Commonwealth Government, is the most comprehensive study ever

More information

CHAPTER 15: DATA PRESENTATION

CHAPTER 15: DATA PRESENTATION CHAPTER 15: DATA PRESENTATION EVIDENCE The way data are presented can have a big influence on your interpretation. SECTION 1 Lots of Ways to Show Something There are usually countless ways of presenting

More information

Strategies for Data Analysis: Cohort and Case-control Studies

Strategies for Data Analysis: Cohort and Case-control Studies Strategies for Data Analysis: Cohort and Case-control Studies Post-Graduate Course, Training in Research in Sexual Health, 24 Feb 05 Isaac M. Malonza, MD, MPH Department of Reproductive Health and Research

More information

Evaluating Social Programs Course: Evaluation Glossary (Sources: 3ie and The World Bank)

Evaluating Social Programs Course: Evaluation Glossary (Sources: 3ie and The World Bank) Evaluating Social Programs Course: Evaluation Glossary (Sources: 3ie and The World Bank) Attribution The extent to which the observed change in outcome is the result of the intervention, having allowed

More information

Observational Study Designs. Review. Today. Measures of disease occurrence. Cohort Studies

Observational Study Designs. Review. Today. Measures of disease occurrence. Cohort Studies Observational Study Designs Denise Boudreau, PhD Center for Health Studies Group Health Cooperative Today Review cohort studies Case-control studies Design Identifying cases and controls Measuring exposure

More information

Mobile Phones & Cancer How does epidemiology investigate this?

Mobile Phones & Cancer How does epidemiology investigate this? Mobile Phones & Cancer How does epidemiology investigate this? Anthony Swerdlow ICNIRP MEMBER/ CHAIR OF THE ICNIRP STANDING COMMITTEE ON EPIDEMIOLOGY INSTITUTE OF CANCER RESEARCH There are now >5 billion

More information

UN Handbook Ch. 7 'Managing sources of non-sampling error': recommendations on response rates

UN Handbook Ch. 7 'Managing sources of non-sampling error': recommendations on response rates JOINT EU/OECD WORKSHOP ON RECENT DEVELOPMENTS IN BUSINESS AND CONSUMER SURVEYS Methodological session II: Task Force & UN Handbook on conduct of surveys response rates, weighting and accuracy UN Handbook

More information

Hal's decision making model. Hal's decision making model. Educating people about risks & benefits of drugs

Hal's decision making model. Hal's decision making model. Educating people about risks & benefits of drugs Hal's decision making model Educating people about risks & benefits of drugs Facts Values Steven Woloshin, MD, MS & Lisa M. Schwartz, MD, MS and H. Gilbert Welch, MD, MPH VA Outcomes Group, White River

More information

Learning to use a sign language

Learning to use a sign language 85 Chapter 8 Learning to use a sign language It is easy for a young child to learn a complete sign language. A child will first begin to understand the signs that others use, especially for people and

More information