Class 5A-2
Page 1: Class Information
Class Title: PH 45 Class 5A
Page 2: Agenda
Overview of Midterm
Population Sampling
Power and Sample Size
Updates and Reminders
Page 3: Midterm Overview
In-depth discussion of the upcoming midterm exam.
Page 4: Midterm Exam Overview
Format: 45 questions, 100 points
Question Breakdown:
True/False (13 questions, 2 pts each)
Topics: Research concepts, study design, validity, ethics
Multiple Choice (17 questions, 2 pts each)
Topics: Study types, bias, statistical principles, research protocol
Short Answer (5 questions, 3 pts each)
Key terms: health research, blinding, ecological fallacy
Research Project (10 questions, varied points)
Apply methods: exposure/outcome variables, hypothesis, statistical test
Page 5: Reminder – Categorical versus Continuous Data
Categorical Data: Represents groups or categories, not numerical
Nominal (no inherent order): Examples include Gender, Race/Ethnicity, Yes/No
Ordinal (ordered but uneven intervals): Examples include Education Level, Pain Scale
Examples: Blood type (A, B, AB, O), Smoking status (Current, Former, Never)
Continuous Data: Represents measurable numerical values
Interval (no true zero): Temperature (Celsius, Fahrenheit)
Ratio (true zero exists): Age, Weight, Blood Pressure, Income
Examples: Hours of sleep per night, Body Mass Index (BMI), Cholesterol levels
Page 6: Population Sampling
Introduction to population sampling in research studies.
Page 7: Types of Research Populations
Four types of populations to consider when collecting data:
Target population: Broad population for study results applicability
Source population: Defined subset for potential participants
Sample population: Individuals invited to participate
Study population: Eligible members who consent and participate
Page 8: Types of Populations
Flow:
Target population: General population aimed for understanding
Source population: Specific individuals for a representative sample
Sample population: Individuals asked to participate
Study population: Eligible participants
Page 9: Target and Source Populations
Importance of a well-defined study question to identify target population
Target: Results applicable to the general public
Source: Narrowed down participants based on the study goal
Page 10: Sample Populations
Considerations for sample populations:
A source population larger than the required sample may need subset selection.
Sampling Bias: Non-representative samples due to systematic differences. Example: Study on smoking habits using participants from a gym.
Non-random sampling bias: Equal chance of selection is essential.
Page 11: Sample Populations (Cont.)
Probability-based Sampling Methods: Ensure representation of source population
Simple random sampling: Each individual has an equal chance
Systematic sampling: Selecting at regular intervals
Stratified sampling: Dividing into subgroups and sampling randomly
Cluster sampling: Whole clusters are randomly chosen
Page 12: Examples of Probability Sampling
Distinct methods of probability sampling:
Simple random sampling: Every person has equal selection chance
Systematic sampling: Every nth person post random start
Stratified sampling: Random samples from various strata
Cluster sampling: Geographically defined groups selected
Page 13: Sample Populations (Cont.)
Convenience Population: Non-probability sampling for accessibility
Be cautious as these populations may differ systematically from intended representative populations
Page 14: Sample Populations (Cont.)
Aim: Achieve a sample population representative of the source
Bias mitigation through careful planning and appropriate sample sizes:
Issues such as Berkson's bias and Healthy Worker bias can arise
Page 15: Sample Populations (Cont.)
Additional biases:
Exclusion bias: Differential eligibility criteria for cases and controls
Page 16: Study Populations
Participation Rate: Percentage of sample population in the study
A high participation rate reduces selection bias.
Low response rates can lead to nonresponse bias
Page 17: Populations for Case-Control Studies
Steps in case-control studies:
Identify source of cases
Select a valid control group meeting the same criteria
Page 18: Population Example of Case-Control Studies
Example:
Target population: American women aged 70-79
Source population: Women with hip fractures at St. Luke's Hospital
Sample population: Women with fractures & controls from friends
Study Approach: Interview participants for data collection
Page 19: Populations for Cross-Sectional Surveys
Ensuring that the study population represents the target population
Random sampling ensures a representative sample
Convenience populations can lead to bias and are unsuitable for cross-sectional studies
Page 20: Example of Cross-Sectional Study
Example:
Target population: High school students in North County
Source population: All students in North County
Sample population: Random students from selected homerooms
Study Approach: Participants complete anonymous questionnaires
Page 21: Populations for Cohort Studies
Sampling methods must align with cohort study design
Identifying source and sample populations mirrors case-control studies for prospective compares
Page 22: Population Example of a Cohort Study
Example:
Target population: Children with cystic fibrosis in Canada
Source: Patients aged 2-12 at participating hospitals
Sample population: Random sampling of 25% invited for participation
Page 23: Populations for Experimental Studies
Sampling methods prioritize validity and safety
Define strict criteria to minimize risks
Page 24: Population Example of an Experimental Study
Study Question: Effectiveness of nutritional counseling in preventing weight gain among first-year college students
Targeted sampling from first-year students in seminars
Data collected on nutritional assessments following structured guidelines
Page 25: Sample Size and Power
Overview of concepts related to determining adequate sample sizes.
Page 26: Importance of Sample Size
Balancing participant recruitment is critical: too few compromises validity, too many wastes resources.
Page 27: Sample Size and Certainty Levels
Sample size defined as the number of observations and critical for achieving desired certainty levels in quantitative studies
Page 28: Sample Size and Means
Relationship of different sample sizes to the means and variability illustrated through examples.
Page 29: Sample Size and Confidence Intervals
Explanation of confidence intervals as a statistical estimation and their importance in data analysis
Page 30: Sample Size and Confidence Intervals
Visualized confidence intervals based on varying sample sizes
Page 31: Sample Size Estimation
Use of sample size calculators to determine appropriate participant numbers to improve research validity
Page 32: Type 1 and 2 Errors
Definition of statistical errors and their implications on study results
Type 1 Error (α): False positive conclusions
Type 2 Error (β): False negative conclusions
Page 33: Type 1 and Type 2 Errors
Graphical representation of statistical errors in relation to study populations
Page 34: Power Estimation
Importance of power in detecting significant differences
A high number of participants increases the likelihood of identifying true effects
Page 35: Power and Error Simplified
Simplified definitions of Power, Type 1 & 2 Errors for understanding research implications
Page 36: Why Run a Power Analysis?
The advantages of conducting a power analysis for study design efficiency and validity
Page 37: Refining the Study Approach
Adjustments needed based on participant availability for maintaining statistical power
Page 38: Updates and Reminders
General announcements for upcoming assignments and deadlines
Page 39: Updates and Reminders
Reminders:
Final research question and hypothesis due
Midterm on Thursday, February 6th
Study guide available online