🫁Intro to Biostatistics Unit 8 – Experimental Design

Key Concepts

• Experimental design involves planning and conducting a study to test a hypothesis while controlling for potential confounding variables
• Independent variable represents the factor being manipulated or changed by the researcher to observe its effect on the dependent variable
• Dependent variable measures the outcome or response that is being studied and is expected to change based on the manipulation of the independent variable
• Control group serves as a baseline for comparison, receiving no treatment or a standard treatment, while the experimental group receives the intervention being tested
• Confounding variables are extraneous factors that can influence the relationship between the independent and dependent variables, potentially leading to biased results if not properly controlled
• Randomization assigns subjects to different treatment groups by chance, ensuring that any differences between groups are due to the intervention rather than pre-existing differences
• Blinding conceals the treatment allocation from participants, researchers, or both (double-blinding) to minimize bias and placebo effects
• Statistical significance indicates whether the observed differences between groups are likely due to chance or the intervention, typically set at a p-value threshold of 0.05

Types of Experimental Designs

• Completely randomized design assigns subjects to treatment groups purely by chance, ensuring each subject has an equal probability of being in any group
• Randomized block design divides subjects into homogeneous subgroups (blocks) based on a specific characteristic before randomly assigning treatments within each block
• Factorial design investigates the effects of two or more independent variables simultaneously, allowing for the examination of main effects and interactions between variables
• Crossover design exposes each subject to all treatments in a random order, with a washout period between treatments to minimize carryover effects
• Matched pairs design pairs subjects with similar characteristics and randomly assigns one member of each pair to each treatment group
• Repeated measures design involves exposing each subject to all treatments over time, allowing for within-subject comparisons and reducing the influence of individual differences
• Latin square design arranges treatments in a grid to control for two sources of variability (rows and columns) while ensuring each treatment appears only once in each row and column
• Split-plot design combines elements of completely randomized and randomized block designs, with some factors randomized to larger plots and others to smaller subplots within each plot

Variables and Controls

• Independent variable is the factor manipulated by the researcher to observe its effect on the dependent variable (e.g., drug dosage, teaching method)
• Dependent variable is the outcome measured in response to changes in the independent variable (e.g., blood pressure, test scores)
  • Continuous dependent variables can take on any value within a range (e.g., weight, time)
  • Categorical dependent variables have distinct categories or levels (e.g., gender, disease status)
• Confounding variables are extraneous factors that can influence the relationship between the independent and dependent variables, leading to biased results if not controlled
  • Examples of confounding variables include age, gender, socioeconomic status, and pre-existing health conditions
• Control variables are kept constant throughout the experiment to minimize their influence on the dependent variable (e.g., room temperature, lighting conditions)
• Placebo control involves giving a treatment that appears identical to the active intervention but has no known effect, to account for psychological effects and biases
• Positive control is a treatment known to produce the desired effect, used to validate the experimental setup and ensure the study can detect a true effect
• Negative control is a treatment known to have no effect, used to establish a baseline and detect any confounding factors or biases in the experimental setup

Sampling Methods

• Simple random sampling selects subjects from a population purely by chance, giving each individual an equal probability of being chosen
• Stratified random sampling divides the population into homogeneous subgroups (strata) based on a specific characteristic, then randomly samples from each stratum in proportion to its size
• Cluster sampling involves dividing the population into naturally occurring groups (clusters), randomly selecting a subset of clusters, and including all individuals within those clusters in the sample
• Systematic sampling selects subjects from a population at regular intervals (e.g., every 10th individual) after randomly choosing a starting point
• Convenience sampling selects subjects based on their availability and willingness to participate, often used in pilot studies or when random sampling is not feasible
• Purposive sampling selects subjects based on specific characteristics or criteria relevant to the research question, ensuring the sample is representative of the population of interest
• Snowball sampling relies on initial subjects to recruit additional participants from their social networks, useful for studying hard-to-reach or hidden populations
• Quota sampling sets a predetermined number of subjects to be selected from each subgroup within the population, ensuring the sample reflects the population's composition

Randomization Techniques

• Simple randomization assigns subjects to treatment groups purely by chance, using methods like flipping a coin, drawing names from a hat, or using a random number generator
• Block randomization divides subjects into smaller, homogeneous subgroups (blocks) before randomly assigning treatments within each block to ensure balanced representation of key characteristics
• Stratified randomization combines stratified sampling and randomization, dividing the population into strata based on specific characteristics, then randomly assigning treatments within each stratum
• Covariate adaptive randomization uses information about subjects' baseline characteristics to assign treatments in a way that minimizes imbalances between groups
• Permuted block randomization creates blocks of a fixed size (e.g., 4 or 6) and randomly assigns treatments within each block, ensuring balanced treatment allocation throughout the study
• Biased coin randomization adjusts the probability of assignment to each treatment group based on the number of subjects already assigned, to maintain balance between groups
• Urn randomization involves drawing colored balls from an urn, with each color representing a treatment group, and replacing the drawn ball with one or more balls of the opposite color to maintain balance
• Minimization assigns subjects to treatment groups based on specific baseline characteristics, aiming to minimize the overall imbalance between groups across all factors

Data Collection Strategies

• Surveys and questionnaires gather self-reported data from subjects using a standardized set of questions, which can be administered in person, by mail, phone, or online
• Interviews involve a researcher asking subjects open-ended or structured questions to collect detailed, qualitative data on their experiences, opinions, or behaviors
• Observations involve researchers directly watching and recording subjects' behaviors or events of interest, either in a natural setting or a controlled laboratory environment
• Physiological measurements collect data on subjects' bodily functions and processes, such as heart rate, blood pressure, or brain activity, using specialized equipment (e.g., ECG, EEG)
• Behavioral tests assess subjects' performance on specific tasks or activities, such as cognitive tests, physical assessments, or simulated scenarios
• Archival data involves using existing records or datasets, such as medical records, government statistics, or historical documents, to answer research questions
• Ecological momentary assessment (EMA) collects real-time data on subjects' experiences, behaviors, and moods in their natural environment using methods like diaries or mobile apps
• Biomarkers are measurable indicators of biological processes, such as blood tests or genetic analyses, used to assess health status, disease risk, or treatment response

Statistical Analysis Approaches

• Descriptive statistics summarize and describe the main features of a dataset, such as measures of central tendency (mean, median, mode) and dispersion (range, standard deviation)
• Inferential statistics use sample data to make generalizations or predictions about a larger population, testing hypotheses and estimating parameters
• t-tests compare the means of two groups to determine if they are significantly different, assuming the data follows a normal distribution and the groups have equal variances
• Analysis of variance (ANOVA) tests for differences in means among three or more groups, partitioning the total variation into between-group and within-group components
• Correlation analysis assesses the strength and direction of the linear relationship between two continuous variables, using measures like Pearson's or Spearman's correlation coefficients
• Regression analysis models the relationship between a dependent variable and one or more independent variables, allowing for prediction and estimation of the dependent variable's value
• Chi-square tests examine the association between two categorical variables, comparing observed frequencies to expected frequencies under the null hypothesis of independence
• Non-parametric tests, such as Mann-Whitney U, Wilcoxon signed-rank, or Kruskal-Wallis, analyze data that does not follow a normal distribution or has ordinal or ranked variables

Ethical Considerations

• Informed consent ensures that subjects understand the purpose, procedures, risks, and benefits of the study and voluntarily agree to participate
• Confidentiality protects subjects' personal information and data from unauthorized access or disclosure, using methods like anonymization or secure storage
• Beneficence requires researchers to maximize the potential benefits of the study while minimizing any harm or risks to subjects
• Justice ensures that the benefits and burdens of research are distributed fairly among different groups, avoiding exploitation or discrimination
• Respect for persons recognizes subjects' autonomy and right to make their own decisions, providing them with sufficient information and the freedom to withdraw from the study at any time
• Scientific integrity involves conducting research honestly, objectively, and transparently, avoiding fabrication, falsification, or plagiarism of data or results
• Social responsibility considers the broader societal implications and consequences of research, ensuring that studies address important issues and contribute to the public good
• Institutional review boards (IRBs) review and approve research proposals to ensure they meet ethical standards and protect the rights and welfare of human subjects

Real-World Applications

• Clinical trials test the safety and efficacy of new drugs, medical devices, or interventions in human subjects, following a carefully designed protocol and regulatory guidelines
• Educational research evaluates the effectiveness of different teaching methods, curricula, or educational policies on student learning outcomes and achievement
• Environmental studies investigate the impact of human activities or natural processes on ecosystems, biodiversity, or environmental health, informing conservation and management strategies
• Market research assesses consumer preferences, behaviors, and opinions to inform product development, pricing, or advertising strategies for businesses
• Psychological research explores the underlying mechanisms and factors influencing human cognition, emotion, and behavior, contributing to the understanding and treatment of mental health issues
• Public health research examines the distribution and determinants of health and disease in populations, informing the development of interventions and policies to improve health outcomes
• Social science research investigates social phenomena, such as social interactions, cultural practices, or political attitudes, to understand and address societal issues and challenges
• Sports science research applies scientific principles to enhance athletic performance, prevent injuries, and optimize training and recovery strategies for athletes and coaches