Experimental Design, Graphical Analysis, and Statistical Methods in Biology

Administrative Guidelines and Action Items

• Current Week Responsibilities:

  • Watch Slide Set 1 (Course Policies Recorded Lecture).
  • Complete Homework 1 (Course Policies) by Sunday at $11:55\,\text{pm}$. Work must be completed on a plugged-in computer rather than a mobile device. Assignments may be edited prior to submission, but once submitted, cannot be reopened or revised.
  • Complete Weekly Quiz 1 covering Slide Set 2 material. Quiz opens Friday at noon ($12:00\,\text{pm}$) and closes Sunday at $11:55\,\text{pm}$. Located in the Weekly Quiz module. Must be taken on a plugged-in computer. Time limit is $10\,\text{minutes}$ for $5\,\text{questions}$.

• Upcoming Week Responsibilities:

  • Read textbook sections $1.1$, $1.2$, $1.3$, and $1.4$, as well as pages $6\text{--}9$ of the Interpreting Graphs handout.
  • Complete Textbook Reading Assignment 1 by Wednesday at $11:55\,\text{pm}$ (accessible via the Textbook Reading Assignments module).
  • Complete Homework 2 ("How College is Different from High School") by next Sunday at $11:55\,\text{pm}$.
  • Complete Weekly Quiz 2 covering Slide Sets 3 and 4. Opens next Friday at noon ($12:00\,\text{pm}$) and closes next Sunday at $11:55\,\text{pm}$.

• Recitation Requirements:

  • Recitation sessions begin next week.
  • Review the Recitation Syllabus in the Canvas Administration module prior to attending.
  • Read pages $6\text{--}9$ of the Recitation 1: Interpreting Graphs handout prior to recitation.
  • Bring a copy (physical or downloaded on a phone) of the Recitation 1: Interpreting Graphs handout to recitation.

Core Learning Objectives

• Master the foundational principles of scientific experimental design.
• Correctly identify and label structural components of real-world biological experiments.
• Analyze experimental setups to identify procedural errors, confounding variables, and logical flaws.
• Formulate testable hypotheses and design controlled experiments to test specific predictions.
• Interpret quantitative graphical data and summarize relationships in precise narrative language.
• Assess basic statistical metrics (including measures of variance and significance) when drawing conclusions from data.

The Scientific Method

• Definition & Overview: The scientific method is an iterative, structured process utilized by researchers to observe natural phenomena, form mechanistic explanations, test hypotheses through controlled experimentation, and refine scientific understanding.

• Sequential Steps of the Scientific Method:

  1. Observe a pattern in nature: Identify an unexplained phenomenon or consistent empirical trend.
  2. Ask a question: Formulate a targeted question regarding the underlying causes or mechanisms of the observation.
  3. Form a hypothesis: Propose a testable, mechanistic explanation that answers the question.
  4. Make a prediction and conduct an experiment: Predict specific outcomes based on the hypothesis, then design and execute an experiment to test the prediction.
  5. Analyze the results: Evaluate experimental data to determine whether actual results match the predicted outcomes.
  6. Report the results: Document conclusions and disseminate findings to the scientific community.

• Iterative Experimental Decisions:

  • If experimental results DO NOT match predictions: The hypothesis must be rejected or revised. Formulate a new hypothesis and initiate new experimental testing.
  • If experimental results DO match predictions: The hypothesis is supported, but not permanently proven. Test the hypothesis using new experimental systems, different environmental conditions, or alternative methodologies.

• Case Example: Workplace Teaspoon Attrition:

  • Observation: Teaspoons regularly disappear from shared breakroom facilities.
  • Study Reference: "The case of the disappearing teaspoons: longitudinal cohort study of the displacement of teaspoons in an Australian research institute" by Megan S C Lim, Margaret E Hellard, and Campbell K Aitken.
  • Setting: The Macfarlane Burnet Institute for Medical Research and Public Health (Burnet Institute) in Melbourne, employing approximately $140\,\text{people}$.
  • Study Objectives: Quantify the rate of spoon loss and determine if attrition correlates with relative spoon value or room placement.

Core Components of Experimental Design

• Variables and Setup:

  • Independent Variable: The single experimental factor intentionally manipulated or altered by the investigator across treatment groups.
  • Experimental Group: Treatment conditions in which the independent variable is manipulated.
  • Control Group: The unmanipulated baseline condition used for comparison against experimental treatments.
  • Standard Conditions: All procedural, biological, and environmental factors held strictly constant across every treatment condition to prevent confounding.
  • Dependent Variable: The outcome factor or biological response measured to quantify the effect of the independent variable.

• Five Common Errors in Experimental Design:

  1. Inadequate Sample Size: Testing too few subjects, increasing susceptibility to random error and reducing statistical power.
  2. Lack of Standard Conditions: Failing to keep non-experimental conditions uniform across treatment groups, introducing uncontrolled confounding variables.
  3. Lack of Replication: Failing to repeat trials or perform experiments across independent replicates.
  4. Lack of Appropriate Controls: Omitting baseline unmanipulated groups or using invalid controls, making result interpretation ambiguous.
  5. Investigator or Subject Bias: Subjective bias influencing data collection, observation, or outcome assessment.

Applied Analysis of Experimental Flaws: Commute Case Studies

• Scenario Objective: Determine the fastest daily commuting method between Jersey City and Rutgers-Newark under the hypothesis that public transit is faster than driving.

• Case Study 1: Seasonal Traffic Variations:

  • Design: Driving every day during the holiday week between Christmas and New Year's Day versus taking public transit during the fourth week of January.
  • Conclusion: Driving is faster.
  • Design Flaw: Lack of Standard Conditions. Driving traffic patterns during major holiday weeks are drastically lighter than typical late-January traffic conditions.

• Case Study 2: Insufficient Sample Size:

  • Design: Driving on a single Monday of an average week versus taking transit on a single Monday of a similar week ($n = 1$ per treatment).
  • Conclusion: Public transit is faster.
  • Design Flaw: Sample Size Too Small / Lack of Replication. Single trials cannot account for daily transit delays, single-car accidents, or anomalous weather events.

• Case Study 3: Subjective Expert Bias:

  • Design: Consulting an auto mechanic of $20\,\text{years}$ for an opinion rather than measuring actual commute times.
  • Conclusion: Driving is faster.
  • Design Flaw: Investigator / Subject Bias. An automotive repair specialist maintains an inherent professional bias toward vehicle usage.

Experimental Design Application: Poison Ivy and Remedy Testing

• Biological Mechanics:
  • Skin contact with poison ivy (Toxicodendron radicans) causes severe contact dermatitis.
  • The primary toxic allergen responsible for the reaction is a class of lipophilic catechol derivatives known as urushiols.
  • Jewelweed (Impatiens capensis) is traditionally utilized as an herbal remedy for poison ivy dermatitis.

• Designing a Controlled Experiment:
  • Study Population: Define subject selection criteria and ensure randomized assignment to treatment groups.
  • Urushiol Application: Apply standardized concentrations of pure urushiol oil to specified skin sites across subjects.
  • Remedy Formulation: Specify the exact plant preparation tested (e.g., pure leaf extract, stem mash, whole-plant homogenate, or jewelweed-infused soap).
  • Control Treatments: Establish negative controls (e.g., distilled water) and positive controls (e.g., commercial dish soaps or proven active ingredients).
  • Standardization: Maintain uniform wash duration, water temperature, skin location, and ambient humidity.
  • Data Collection: Record quantitative skin dermatitis scores on a standardized scale (e.g., $0\text{--}14$) at set post-exposure time points.

Graphical Representation and Data Presentation

• Selecting Appropriate Graph Types:

  • Bar Graphs: Ideal for displaying discrete, categorical groups (e.g., comparing mean systolic blood pressure between high-stress and low-stress cohorts).
  • Line Graphs / Continuous Curves: Essential for displaying continuous functional relationships or time-series changes (e.g., tracking change in body temperature over time).

• Table Formatting Conventions:

  • Independent variables belong in the leftmost column.
  • Dependent variables belong in subsequent rightward columns.
  • Lizard Body Temperature Dataset:
  • Time $t = 0\,\text{min}$, Temperature $T = 25\,^\circ\text{C}$
  • Time $t = 30\,\text{min}$, Temperature $T = 27\,^\circ\text{C}$
  • Time $t = 60\,\text{min}$, Temperature $T = 29\,^\circ\text{C}$
  • Time $t = 90\,\text{min}$, Temperature $T = 31\,^\circ\text{C}$
  • Time $t = 120\,\text{min}$, Temperature $T = 32\,^\circ\text{C}$
  • Time $t = 150\,\text{min}$, Temperature $T = 32\,^\circ\text{C}$

• Graph Axis Assignments and Scaling:

  • X-axis: Assigned to the independent variable (e.g., Time in minutes).
  • Y-axis: Assigned to the dependent variable (e.g., Temperature in $^\circ\text{C}$).
  • Scale Choice Effects: Altering Y-axis boundaries impacts visual slope interpretation. Truncating axis bounds exaggerates subtle variations, whereas expanded ranges dampen visual changes. Axes must remain consistent and clearly marked.

Statistical Evaluation in Data Analysis

• Error Bars: Graphical representations displaying measurement uncertainty or sample dispersion around a calculated mean. Larger error bars signify greater variation within the dataset.

• Standard Deviation (SD):

  • Quantifies dispersion of individual data points around the sample mean.
  • Under a standard normal distribution, approximately $68\%$ of individual sample observations fall within $\pm 1\text{ SD}$ of the mean.

• Standard Error of the Mean (SEM):

  • Quantifies how precisely the sample mean estimates the true population mean.
  • Calculated as standard deviation divided by the square root of sample size: $\text{SEM} = \frac{\text{SD}}{\sqrt{n}}$.

• Statistical Significance and p-Values:

  • Statistically Significant: Indicates an observed difference is highly unlikely to be the result of random chance.
  • p-Value: The probability that an observed outcome (or a more extreme result) occurred purely by random variation.
  • Standard Threshold: $p \le 0.05$ is widely accepted as the threshold for statistical significance.
  • Research Application: In a study measuring student figure annotations before and after pedagogical intervention, marginal notations per figure panel increased significantly from Week 1 ($n = 14$) to Week 15 ($n = 16$) ($p < 0.001$ by paired t-test; two subjects omitted from Week 1 due to incomplete assessment protocols).

Systematic Frameworks for Graph Analysis

• Step 1: Describing a Graph (EASUT Framework):

  • E (Experimental Units / Subjects): Identify biological or experimental test subjects.
  • A (Axes): Identify variables assigned to X-axis, primary Y-axis, and secondary Y-axis.
  • S (Scale): State numeric range boundaries and tick intervals on each axis.
  • U (Units): Identify exact units of measurement (e.g., $\mu\text{g/g}$ of plant material, score $0\text{--}14$).
  • T (Treatments): Detail all experimental, positive control, and negative control treatment groups.

• Step 2: Interpreting a Graph (PVC Framework):

  • P (Pattern): Identify overall trends, relative bar heights, and rate changes.
  • V (Variation & Statistics): Examine error bar magnitudes, overlap between error bars, and notation symbols indicating significant differences (e.g., asterisks).
  • C (Conclusions): State evidence-based conclusions directly supported by the statistical data.

• Applied Study Example: Botanical Remedies for Dermatitis:

  • Source Citation: V. Abrams Motz et al., Journal of Ethnopharmacology 143 (2012) 314–318.
  • Experimental Variables: Dual Y-axes comparing rash score ($0\text{--}14 \pm \text{SE}$, left axis, bar height) against lawsone concentration ($\mu\text{g/g}$ plant material, right axis, line plot).
  • Evaluated Treatments: Distilled water, Impatiens capensis extract, Lawsone solution, Impatiens balsamina extract, I. capensis mash, I. balsamina mash, I. capensis soap, I. balsamina soap, and Dawn dish soap.
  • Analytical Result: Treatments marked with an asterisk ($*$) demonstrate statistically significant diminution of rash development relative to negative control treatments.