RESEARCH METHODOLOGY FOR UNDERGRADUATES.pptx

Embarking on the journey of research as undergraduates is both exciting and challenging. This presentation introduces the fundamental principles of research methodology, offering a roadmap for undergraduates to navigate the complexities of academic inquiry.

1. Research methodology for Undergraduates Dr. ROBIN T. VAVACHAN. MD. COMMUNITY AND FAMILY MEDICINE

2. What’s Research?

3. Objectives of research

4. Characteristics of a research Systemic and structured Objective and unbiased Empirical Logical and rational Replicable and verifiable Transparent and ethical Generalizable Incremental and cumulative Problem solving orientation Subject to peer review

5. Types of research

6. explanatory exploratory purpose aims to explain the relationships between variables, identify causal connections, and provide explanations for observed phenomena. Exploratory research aims to explore a topic, gain initial insights, generate hypotheses, and identify potential relationships or patterns. Nature builds on existing knowledge and theories to test specific hypotheses and establish cause-and-effect relationships typically conducted in the early stages of research when little is known about the subject of study focus Explanatory research seeks to provide answers to research questions and verify or support existing theories Exploratory research focuses on discovering new ideas, concepts, or areas of interest. methodology Quantitative methods such as experiments, surveys, or statistical analyses are commonly employed to collect data and test hypotheses Qualitative methods such as interviews, focus groups, observations, or case studies are often used to gather information and gain in-depth understanding. Sample size Explanatory research often requires larger sample sizes to ensure statistical power and generalizability usually small and not necessarily representative of the target population. findings The findings of explanatory research are analytical, statistical, and provide evidence for or against specific hypotheses descriptive and provide a foundation for further investigation or hypothesis testing.

7. A research proposal blueprint or roadmap for the research project, explaining the research objectives, methodology, and expected outcomes Writing a research proposal is both science and art, implies that it should be based on scientific facts and on the art of good communication

8. Every research study has two aspects: 1. Study population - • People: individuals, organizations, groups, communities ( they provide you with the information or you collect information about them) 2. Subject area- • Problems: issues, situations, associations, needs, profiles • Program : content, structure, outcomes , attributes , satisfactions, consumer, Service providers, etc. • Phenomenon: cause-and-effect relationships, the study of a phenomenon itself

9. Steps in research process

10. I. Formulating a research question

11. Steps of formulating a research Q

12. The FINER characteristics of a research question

13. F easible subjects; technical expertise; time; money; scope I nteresting to the investigator N ovel Confirms/refutes previous findings; Extends previous findings; Provides new findings E thical R elevant To scientific knowledge To clinical use, To public health or health policy To future research directions

14. PICOT Criteria in the development of a Good Research Question

15. Research Question Example

16. Few Realities About Research Research Question is always for the Target Population, True result in the target population is there but unknown to the Researcher Always, decision is taken based on only one sample/study i.e. results based on one sample findings are extrapolated to the defined target population Theoretically, all study results come with an “Open Ended Expiry Date” For all research findings, answer for two questions about the study result to be provided by the researcher Validity? (Internal and External); Reliability ?

17. Research Hypothesis Question – Is computer-assisted acetabular component better than freehand acetabular component placement in patients for total Hip arthroplasty?? Hypothesis – Statement – Single outcome - computer-assisted acetabular component placement leads to improved functional outcome or Composite outcome – computer-assisted acetabular component placement leads to both improved radiographic cup placement and improved functional outcome Based on a good research question at the start of study Drives data collection for the study Developed from the research question and then main elements of the study – sampling strategy, intervention, comparison and outcome variables

18. Whether three days workshop on Research Methods increases participants knowledge on research methods Whether such workshops Increase immediate understanding of research methods . The knowledge of research methods increases after such workshop when evaluated through scores on a pretest and posttest 19

19. Research Problems/Questions Helps in Framing Title of the Research Proposal and Hypothesis

20. II. Extensive Literature Revi ew

21. Essential preliminary task in order to acquaint yourself with the available body of knowledge in your area of interest. Literature review is integral part of entire research process and makes valuable contribution to every operational step. Reviewing literature can be time-consuming, daunting and frustrating, but is also rewarding. Its functions are: a. Bring clarity and focus to your research problem b. Improve your methodology c. Broaden your knowledge d. Contextualise your findings

22. Procedure for reviewing the literature

23. Literature review (Books, Journals, Scientific Reports, Newspaper etc.) Helps in Writing Introduction and rationale, and references of the research proposal

24. MeSH Search MeSH (Medical Subject Headings) is the NLM controlled vocabulary thesaurus used for indexing articles for PubMed. BOOLEAN SEARCH It involves the use of logical operators (AND, OR, NOT) to create more specific and targeted search queries

25. AND : The operator "AND" narrows down the search results by retrieving only the docu.ments or webpages that contain both search terms. For example, "diabetes AND exercise" will retrieve results that include both terms OR : The operator "OR" broadens the search results by retrieving documents or webpages that contain either of the search terms. For example, "diabetes OR hypertension" will retrieve results that include either term. NOT : The operator "NOT" excludes specific terms from the search results. It helps to refine the search by eliminating irrelevant or unwanted results. For example, "diabetes NOT type 2" will retrieve results related to diabetes but exclude any results related to type 2 diabetes. BOOLEAN SEARCH

26. II. The formulation of objectives

27. Objectives are the goals you set out to attain in your study. They inform a reader what you want to attain through the study. It is extremely important to word them clearly and specifically. Objectives should be listed under two headings: a) main objectives ( aims); b) sub-objectives. • The main objective is an overall statement of the thrust of your study. It is also a statement of the main associations and relationships that you seek to discover or establish. The sub-objectives are the specific aspects of the topic that you want to investigate within the main framework of your study. -They should be numerically listed. -Wording should clearly, completely and specifically

28. Characteristics of a good objective

29. Introduction to study designs

30. A specific plan or protocol for conducting the study, which allows the investigator to translate the conceptual hypothesis into operational hypothesis The procedures and methods, predetermined by an investigator, to be adhered to in conducting a research project Methods used to obtain valid data to answer a research question (or prove/refute a hypothesis)

31. association Association refers to a statistical relationship or connection between two or more variables t indicates that there is a consistent pattern or correlation between variables, meaning that changes in one variable are related to changes in another variable. can be + (both variables increase or decrease together), - (one variable increases while the other decreases), or null (no relationship observed), do not imply a cause-and-effect relationship cross-sectional, case-control, or cohort studies. causality refers to a cause-and-effect relationship between variables, where changes in one variable directly cause changes in another variable. Causal relationships are considered stronger evidence for understanding the underlying mechanisms and making predictions about the effects of interventions or treatments Experimental studies, such as randomized controlled trials

32. Case report

33. A case report described and discusses an instance of disease in a patient – Rare Indications The essential characteristic of a publishable case report is educational value Writing case report is one of the best ways to get started in medical writing. They are little mysteries that hold readers’ interest and take less time to prepare Ex: Case report of Kaposi’s sarcoma in a young homosexual man -> development of AIDS Drawbacks False alarms can be raised   Less citable (Max- meta-analysis and min-case reports) Reduce Impact Factor, hence editors do not like case reports   Publication bias (90%) reporting successes versus 10% reporting failure Methodology is not robust   Most of the once-popular discarded therapies are based on case reports

34. Case series

35. Case Series (also known as a clinical series ) Type of medical research study that tracks subjects with a known exposure, such as patients who have received a similar treatment, or examines their medical records for exposure and outcome May be consecutive or non-consecutive , depending on whether all cases presenting to the reporting authors over a period were included, or only a selection. drawbacks Lack of comparison group Selection bias Limited generalizability Lack statistical power Unable to establish temporal relationships

36. Clinical based Clinical case-series - usually a coherent and consecutive set of cases of a disease (or similar problem) which derive from either the practice of one or more health care professionals or a defined health care setting e.g., a hospital or family practice. A case-series is, effectively, a register of cases. Analyse cases together to learn about the disease. Clinical case-series are of value in epidemiology. Studying symptoms and signs. Creating case definitions. Clinical education, audit and research. Population based clinical case-series is, effectively, a population-based case-series consisting of a population register of cases. Epidemiologically the most important case-series are registers of serious diseases or deaths, and of health service utilisation, e.g., hospital admissions. Usually compiled for administrative and legal reasons

37. Cross-sectional study

38. Cross-sectional studies are observational studies that aim to measure the prevalence of an outcome or condition within a population at a specific point in time Measure prevalence Identify associations Generate hypotheses a representative sample from the target population selected to ensure the findings can be generalized to the broader population Questionnaire/surveys unable to Establish Causality Recall bias Studies only “survivors” and “stayers” –Survivor Bias Lack of temporal information

39. AETIOLOGICAL RESEARCH PREVALENCE ODDS RATIO PREVALENCE RATIO POR VS PR NEYMAN BIAS REVERSE CAUSATION MISS-CLASSIFICATION BIAS SOCIAL-DESIRABILITY BIAS SELECTION BIAS

40. OUTCOME + OUTCOME - EXPOSURE + a b EXPOSURE - c d Odds Ratio = (a/b) / (c/d) = ad / bc

41. Interpreting the odds ratio >1. positive association - individuals with the exposure have higher odds of experiencing the outcome compared to those without the exposure. =1 . no association - The odds of having the outcome are equal for both exposed and unexposed groups. <1 . negative association - individuals with the exposure have lower odds of experiencing the outcome compared to those without the exposure.

42. Matching Controls Confounding bias. Enhances Efficiency: efficient use of the sample size by reducing the variability within each group making it easier to detect a significant effect or association with a smaller sample size. Pair matching Propensity score matching Stratified matching Time period matching Exact matching

43. Overmatching on too many variables can reduce the generalizability Incomplete matching can introduce bias if the unmatched characteristics are related to the outcome variable It aims to control for observed confounders but does not address unmeasured confounders. M atched sample may be more homogeneous, limiting the generalizability of the findings to broader populations it can be challenging to find suitable matches for all participants in the study It can be an inefficient process, particularly when the number of potential matches is limited.

44. Propensity score matching

45. Case control study

46. Case-control studies are observational studies designed to compare individuals with a specific outcome (cases) to those without the outcome (controls) and identify potential risk factors. Identify Risk Factors : The primary purpose Comparing Cases and Controls: To determine whether certain exposures are more prevalent in cases, indicating a potential association Matching to minimize confounding and enhance comparability. Ideal set of cases would be all the new (incident) ones in the population under study (but very difficult??)

47. Sampling Methods : Cases are typically identified from existing records or disease registries, while controls are selected from the same population from which the cases arose. Controls should be representative of the population from which the cases were derived Studying Rare Diseases ( Statistically efficient, Logistical & financially practical A llow researchers to investigate multiple potential risk factors or exposures simultaneously Assessing Strength of Associations Selection bias : researchers should ensure that controls are selected from the same population and use appropriate sampling techniques Information Bias : Standardized data collection methods, blinding of investigators, and quality control measures Recall bias: use standardized and structured questionnaires to minimize recall bias. Odds ratio biased when the disease is rare

48. Cohort studies

49. Cohort studies are observational studies that aim to investigate the relationship between exposure to a risk factor and the subsequent development of a disease. Investigate Relationship : The primary purpose of cohort studies is to examine the association between exposure to a specific risk factor or intervention and the subsequent occurrence of a disease or outcome. Identify Risk Factors: Cohort studies help identify potential risk factors that contribute to the development of diseases. Study Disease Incidence : Cohort studies allow for the assessment of disease incidence and the calculation of measures such as relative risk and incidence rates.

50. Studying Rare Exposures: Cohort studies are effective for studying rare exposures or risk factors that are uncommon in the general population. Since exposure status is determined before the development of the outcome, researchers can evaluate rare exposures over a longer period. Assessing Disease Incidence: Cohort studies allow for the direct calculation of disease incidence rates, providing valuable information on disease occurrence within a defined population. Identifying Risk Factors: Cohort studies help identify potential risk factors by examining the relationship between exposure and disease outcomes over time.

51. Potential Biases and Strategies to Minimize Them: Loss to Follow-up: Participants may drop out of the study or become lost to follow-up, leading to biased results. Strategies to minimize loss to follow-up include maintaining good participant engagement, providing incentives, and using multiple contact methods. Confounding: Confounding occurs when the association between exposure and outcome is influenced by a third variable. Researchers can minimize confounding by design (randomization, matching), statistical adjustment, or stratification during data analysis. Information Bias: Information bias can occur if there are errors or inaccuracies in exposure or outcome assessment. Standardized protocols, training of data collectors, and rigorous data quality control measures can minimize information bias.

52. Prospective Cohort Study Retrospective Cohort Study Timing of Data Collection Participants are identified and classified into exposed and unexposed groups based on their exposure status at the beginning of the study. Participants are identified based on their exposure status in the past, and researchers assess their subsequent outcomes retrospectively using historical records or existing databases Timing of Exposure Assessment Exposure information is typically collected at the beginning, allowing for real-time or near-real-time assessment of exposures. This helps to minimize recall bias may introduce recall bias as participants may have difficulty accurately recalling past exposures Temporality long-term follow-up rely on existing records for outcome assessment. Follow-Up establish a clear temporal relationship between exposure and outcome. rely on existing records, which may not always provide a clear temporal sequence of exposure and outcome

53. Disease Developed Disease Not Developed Total Incidence Rate Exposed a b a+b a/ a+b ( Ie ) Not exposed c d c+d c/ c+d ( Iue )

54. M easures of association in prospective cohort studies Relative Risk (RR): The relative risk is the ratio of the risk of developing the outcome in the exposed group compared to the unexposed group. It provides an estimate of the strength of the association between exposure and outcome. The formula for calculating relative risk is: RR = (Risk of outcome in exposed group) / (Risk of outcome in unexposed group) Ie / Iue Risk Difference (RD): The risk difference represents the absolute difference in the risk of developing the outcome between the exposed and unexposed groups. It provides an estimate of the excess risk associated with the exposure. The formula for calculating risk difference is: RD = (Risk of outcome in exposed group) - (Risk of outcome in unexposed group Ie - Iue

55. A ttributable risk (AR) is a measure that quantifies the proportion of disease incidence in the exposed group that can be attributed to the exposure. It provides an estimate of the excess risk associated with the exposure in the study population AFE = (R1 – R0) / R1 = (RR -1)/RR if RR>1 If the attributable risk is zero, it suggests that there is no excess risk associated with the exposure. The incidence rate of the outcome in the exposed group is the same as that in the unexposed group. If the attributable risk is positive, it indicates that there is an excess risk of the outcome in the exposed group compared to the unexposed group. The exposure is associated with an increased risk of the outcome. If the attributable risk is negative, it suggests a protective effect of the exposure. The incidence rate of the outcome is lower in the exposed group compared to the unexposed group

56. Population Attributable Risk (PAR): The population attributable risk is the proportion of the risk of developing the outcome in the entire study population that can be attributed to the exposure. It represents the potential impact of the exposure on the occurrence of the outcome at a population level. Hazard Ratio (HR): The hazard ratio is a measure of the instantaneous risk of developing the outcome over time in the exposed group compared to the unexposed group. It is commonly used in survival or time-to-event analysis in prospective cohort studies. The hazard ratio can be estimated using Cox regression analysis.

57. Allow complete information on the subject’s exposure, including quality control of data, and experience thereafter. Provide a clear temporal sequence of exposure and disease. Give an opportunity to study multiple outcomes related to a specific exposure. Permit calculation of incidence rates (absolute risk) as well as relative risk. Methodology and results are easily understood by non-epidemiologists. Enable the study of relatively rare exposures. Not suited for the study of rare diseases . Not suited when the time between exposure and disease manifestation is very long, although this can be overcome in historical cohort studies. Exposure patterns, for example the composition of oral contraceptives, may change during the course of the study and make the results irrelevant. Maintaining high rates of follow-up can be difficult . Expensive to carry out because a large number of subjects is usually required. Baseline data may be sparse because the large number of subjects does not allow for long interviews.

58. Experimental study designs

59. Non-Randomized • Pre-Post Study • Pragmatic Clinical Trial Randomised Randomized Controlled Trial Field Trial/Community Trial

60. What is a clinical trial? A clinical trial is defined as a prospective study comparing the effects and value of intervention (s) against a control in human beings A properly planned and executed clinical trial is a powerful experimental technique for assessing the effectiveness of an intervention Phase I studies Preclinical information – in-vitro studies or animal models, early data must be obtained in humans Assessment of Toxicity – Maximum Tolerated Dose Step-up approach Human volunteers who have already tried and failed to improve on the existing standard interventions

61. Phase II Studies Depends upon the quality and adequacy of Phase I study Major aim is to evaluate whether the drug has any biologic activity and to estimate the rate of adverse effect Dose-escalation study No standard control – non-comparative Phase III Clinical Trials /randomized controlled trials Define a study population suitable for answering the question. Divide the study population into two or more groups. The control group may be offered the best-known alternative. In the ideal trial, the study and control populations are similar in characteristics impacting on disease outcomes. To achieve this similarity individuals in the study are assigned randomly to the groups. This is a randomised, controlled trial. “Best known alternative" is sometimes an intervention which is “psychologically” of similar impact to the study intervention

62. Randomization Randomization is the process of assigning participants to different study groups (e.g., treatment group and control group) randomly. It involves using a randomization method, such as a computer-generated random number sequence or randomization tables, to ensure that each participant has an equal chance of being assigned to any group. Randomization helps minimize selection bias and ensures that the study groups are comparable in terms of both known and unknown confounding factors. Reduces Bias: reducing the risk of confounding. Enhances Generalizability

63. Types of Sampling Probability (Random) Sampling Simple random sampling (SRS) Stratified random sampling Systematic random sampling Cluster sampling Multistage sampling Non-Probability Sampling Convenience sample Purposive sample Snowball Sampling Respondent Driven Sampling Quota Sampling 69

64. Blinding , also known as masking, is a method used in clinical trials to minimize bias and ensure the integrity of the study results. It involves keeping certain individuals or groups unaware of the treatment assignment to prevent their knowledge from influencing the study outcomes. Single-Blind : In a single-blind trial, either the participants or the investigators are unaware of the treatment assignment.. Double-Blind : In a double-blind trial, both the participants and the investigators or healthcare providers involved in the study are unaware of the treatment assignment.

65. Biases in medical research Selection bias W hen there is a systematic difference in the selection of participants or controls that is related to the exposure or outcome being studied C oncern in case-control studies, cohort studies, and randomized controlled trials (RCTs) if there are differences in the characteristics or eligibility criteria of participants in different groups. RECALL BIAS Refers to a systematic error in participants' ability to accurately recall past exposures or events It can introduce bias in case-control studies where cases may recall past exposures differently compared to controls

66. INFORMATION BIAS E rrors or inaccuracies in the measurement or assessment of exposure or outcome variables It can arise in any study design. CONFOUNDING BIAS Confounding bias arises when an extraneous factor is associated with both the exposure and outcome, and it distorts the observed association between the two. Confounding variables can introduce bias in any study design, but they are particularly important in observational studies such as cohort studies and case-control studies Proper study design, randomization, matching, and statistical adjustment can help minimize confounding bias

67. PERFOMANCE BIAS Performance bias occurs when there are differences in the care or treatment provided to participants in different study groups, leading to biased results This bias can be a concern in RCTs if there are variations in the delivery or adherence to interventions among different groups. PUBLICATION BIAS Publication bias occurs when there is a selective publication of research findings based on their statistical significance or direction of results. It can lead to an overestimation of treatment effects or associations in the published literature. Publication bias can affect any study design if studies with statistically significant or positive results are more likely to be published than those with nonsignificant or negative results.

68. Systematic review and meta-analysis are two complementary methods used in research to synthesize and summarize evidence from multiple studies Defining the research question or objective. Conducting a comprehensive literature search. Applying predefined inclusion and exclusion criteria to select relevant studies. Extracting and synthesizing data from selected studies. Assessing the quality and risk of bias of included studies. Summarizing the findings and drawing conclusions based on the available evidence. Meta-analysis is a statistical technique used to quantitatively combine data from multiple studies included in a systematic review. It involves pooling the results of individual studies to derive an overall estimate of the treatment effect or association. Meta-analysis utilizes statistical methods to calculate summary effect measures, such as odds ratios, risk ratios, or weighted mean differences, along with their corresponding confidence intervals

69. Type 1 Error ( α error) A Type 1 error occurs when the null hypothesis is rejected even though it is true. In hypothesis testing, the null hypothesis represents the assumption of no effect or no difference between groups or variables. A Type 1 error is also known as a false positive or a false rejection of the null hypothesis. The probability of committing a Type 1 error is denoted by α ( alpha) and is typically set as the significance level, such as 0.05 (5%). Researchers want to control the risk of Type 1 errors to ensure that the evidence against the null hypothesis is strong enough before rejecting it.

70. Type 2 Error ( β error) A Type 2 error occurs when the null hypothesis is not rejected even though it is false. In other words, it is the failure to detect a true effect or difference between groups or variables. A Type 2 error is also known as a false negative or a failure to reject the null hypothesis when it should have been rejected. The probability of committing a Type 2 error is denoted by β ( beta). The complement of β is the statistical power of a test (1 - β). Power represents the probability of correctly rejecting the null hypothesis when it is false. Researchers aim to minimize the risk of Type 2 errors by maximizing statistical power, which requires an adequate sample size to detect the desired effect size

71. Sample size calculations take into account the desired level of significance ( α) and the desired power (1 - β) to determine the appropriate sample size. By considering both Type 1 and Type 2 errors, researchers aim to strike a balance between detecting true effects (minimizing Type 2 errors) while controlling the risk of false positives (Type 1 errors). The calculations help ensure that the study has a sufficient sample size to detect meaningful effects and achieve the desired level of statistical power, reducing the chances of making erroneous conclusions.

72. Methods of data collection

73. Types of data There are two types of data Primary Data— collected for the first time through direct interface with the persons. - Interview, Clinical Examination, Laboratory Investigations - Questionnaire, Schedule, and Proforma Secondary Data—those which have already been collected and analysed by someone else. There is no interface with the person concerned.

74. Secondary Data Medical Records of patients in clinics, nursing home, or hospital Diseases registries such as of Cancer and Thalassaemia Internet-based databases such as HIV maintained by US Bureau of Census for all countries of the world Reports of the survey (SRS, NFHS, DLHS, NACO etc.) Period reports of the government and nongovernmental organizations (National Health Profile and Family Welfare Statistics)

75. Methods of Primary Data Collection OBSERVATION METHOD : Commonly used in behavioural sciences It is the gathering of primary data by investigator’s own direct observation of relevant people, actions and situations without asking from the respondent. Types of Observation: 1. Structured – for descriptive research 2. Unstructured—for exploratory research 3. Participant Observation 4. Non- participant observation 5. Disguised observation Limitations: - feelings, beliefs and attitudes that motivate buying behaviour and infrequent behaviour cannot be observed. - expensive method

76. SURVEY METHOD : Approach most suited for gathering descriptive information. Structured Surveys : use formal lists of questions asked of all respondents in the same way. Unstructured Surveys : let the interviewer probe respondents and guide the interview according to their answers. Survey research may be Direct or Indirect. Direct Approach : The researcher asks direct questions about behaviours and thoughts. e.g. Why don’t you eat at MacDonalds ? Indirect Approach : The researcher might ask: “What kind of people eat at MacDonald’s?”

77. ADVANTAGES: -can be used to collect many different kinds of information -Quick and low cost as compared to observation and experimental method. LIMITATIONS: -Respondent’s reluctance to answer questions asked by unknown interviewers about things they consider private. -Busy people may not want to take the time -may try to help by giving pleasant answers -unable to answer because they cannot remember or never gave a thought to what they do and why -may answer in order to look smart or well informed.

78. Methods of Contact CONTACT METHODS : Information may be collected by Mail Telephone Personal interview

79. Mail Questionnaires : Advantages: -can be used to collect large amounts of information at a low cost per respondent. -respondents may give more honest answers to personal questions on a mail questionnaire -no interviewer is involved to bias the respondent’s answers. -convenient for respondent’s who can answer when they have time good way to reach people who often travel Limitations: -not flexible -take longer to complete than telephone or personal interview -response rate is often very low - researcher has no control over who answers.

80. TELEPHONE Interviewing : - quick method - more flexible as interviewer can explain questions not understood by the respondent - depending on respondent’s answer they can skip some Qs and probe more on others - allows greater sample control - response rate tends to be higher than mail Drawbacks: -Cost per respondent higher -Some people may not want to discuss personal Qs with interviewer -Interviewer’s manner of speaking may affect the respondent’s answers -Different interviewers may interpret and record response in a variety of ways -under time pressure ,data may be entered without actually interviewing

81. Personal Interviewing : It is very flexible and can be used to collect large amounts of information. Trained interviewers are can hold the respondent’s attention and are available to clarify difficult questions. They can guide interviews, explore issues, and probe as the situation requires. Personal interview can be used in any type of questionnaire and can be conducted fairly quickly. Interviewers can also show actual products, advertisements, packages and observe and record their reactions and behaviour.

82. Determining the sample design

83. Designing the sample calls for three decisions: Who will be surveyed? ( The Sample) • The researcher must determine what type of information is needed and who is most likely to have it. H ow many people will be surveyed? ( Sample Size ) • Large samples give more reliable results than small samples. However it is not necessary to sample the entire target population. How should the sample be chosen? ( Sampling) • Sample members may be chosen at random from the entire population ( probability sample) • The researcher might select people who are easier to obtain information from ( nonprobability sample) The needs of the research project will determine which method is most effective

84. How many people will be surveyed? ( Sample Size ) Approach to Sample size calculation Precision of Estimation : Precision Analysis - a prevalence of 10% from a sample of size 20 would have 95% CI of (1%, 31%) – not informative precise a prevalence of 10% from a sample of size 400 would have 95% CI of (7%, 13%) – very informative and precise Hypothesis Testing of effects/relationship : Power Analysis - Sample size calculations are important to ensure that if an effect deemed to be clinically meaningful exists, then there is a high chance of it being detected i.e. that the analysis will be statistically significant. - If the sample is too small then even if large differences are observed, it will be impossible to show that these are due to anything more than sampling variation.

85. Factors that influence sample size calculation : a checklist

86. How should the sample be chosen? ( Sampling) 3 factors that influence sample representative-ness Sampling procedure Sample size Participation (response) When might you sample the entire population? When your population is very small When you have extensive resources When you don’t expect a very high response 96

87. Sample size estimation when selecting subjects from Hospital/Clinical Settings According to Taro Yamani formula   To determine the representative sample size for this study, the Taro Yamani’s formula was used. n = N÷ {1+N(e) 2 } , Where n = sample size N = Total average OPD visits/year or attending OPD   e = 1 - level of precision 95% Example: Say 1000 patients are expected/attended previous year n = 1000÷ {1+ 1000 (0.05) 2 } n = 285.71 n = sample size =286 finite population correction formula used when the sample is drawn from a finite population rather than an infinite population

88. Sample size estimation when estimating for Single proportion (Prevalence, Sensitivity etc.) p = proportion of diseased = 10% (prior Information) q = 1 – p = 90% Margin of error = 20% of 10%(p) = 2% n = (Z 1- α /2 ) 2 x p x(1-p)/(ME) 2 = (1.96) 2 x 10 x 90/ (2) 2 = 3.84 x 10 x 90/4 = 864

89. Sample size estimation when Estimating for the mean of a normal distribution Mean birth weight = 120 0z (prior Information) Variance σ 2 = 576 oz Margin of error (ME) = 5 oz Type-1 error = 0.05 n = (Z 1- α /2 ) 2 x σ 2 /(ME) 2 = (1.96) 2 x 576 / (5) 2 = 3.84 x 576/25 = 88 Z(1- α/2): The critical value from the standard normal distribution corresponding to the desired level of confidence (1- α). It represents the number of standard deviations from the mean, where α is the significance level or the probability of Type I error

90. Calculation of Z score Z = (X - μ) / σ Where: Z is the Z-score. X is the individual data point you want to standardize. μ is the mean of the distribution. σ is the standard deviation of the distribution. The Z-score allows you to determine the relative position of a data point within a distribution. A positive Z-score indicates that the data point is above the mean, while a negative Z-score indicates that it is below the mean. A Z-score of 0 means the data point is exactly at the mean.

91. Comparing Two Proportions Goal: When we Intend to compare two groups want to detect a difference between two population proportions Scenario: A controlled randomized trial proposes to assess the effectiveness of colony stimulating factors (CESS) in reducing sepsis in premature babies. A previous study has shown the underlying rate of sepsis to be about 50% in such infants around 2 weeks after birth, and a reduction of this rate to 34% would be of clinical importance.

92. Required Information - Primary Outcome Variable = Presence/absence of sepsis at 14 days after treatment (72 hours after birth) - Hence, a categorical variable (binary)=proportion - Difference of clinical importance=50%-34%=16% (0.16) - α = 0.05 (5% level of significance); Power =80%; Two-tailed test (Z 1- α /2 + Z 1- β ) 2 x [p1(1-p1) + p2(1-p2)] n = ---------------------------------------------------- = 146 where (p1 – p2) 2 n= sample size required in each group(146 x 2= 292) P1=first proportion=50%; p2=second proportion=34% Z α /2 = 1.96 for desired significance level; Z β = 0.84

93. Comparison of Two Means (Independent Samples) Scenario: When comparing the means of two independent groups (e.g., treatment vs. control). Formula: The sample size formula for comparing two means is given by: n = (2 * Z^2 * σ^2) / ( d^2) where: n = required sample size per group (total sample size will be twice this value) Z = Z-score corresponding to the desired level of confidence σ = pooled standard deviation (estimated based on prior knowledge or pilot data) d = desired effect size or difference in means

94. Comparison of Two Means (Paired Samples) Scenario: When comparing the means of paired or matched samples (e.g., before and after measurements within the same group). Formula: The sample size formula for comparing paired means is given by: n = (Z^2 * σ^2) / ( d^2) where: n = required sample size (number of pairs) Z = Z-score corresponding to the desired level of confidence σ = standard deviation of the differences between paired observations (estimated from pilot data or previous studies) d = desired effect size or difference in means