Introductory Research Methods - Detailed Notes

Introductory Research Methods

Introduction to Data and Evidence

• Professor Hannah Keage

Overview of the Lecture

• What is statistical thinking?

• Why research and statistics?

• Understanding evidence

• Fundamentals of data:

  • Aggregation

  • Uncertainty

  • Sampling from a population

  • Causality

Statistical Thinking

• Definition: A method to simplify complex realities into understandable terms.

• Purpose: To summarize and capture essential data structure and function.

• Tools Provided: To navigate and comprehend uncertainty in knowledge.

Applications of Statistics

1. Describe: Simplifying complex situations for better understanding.

   • Example: Australia's population was 24.6 million in 2017.

2. Decide: Making informed choices based on data despite uncertainty.

   • Example: Projected increase of South Australia's population by 0.1%-0.9% per year.

3. Predict: Using past data to make projections about future scenarios.

   • Example: Australia's population estimate of 37.4-49.2 million by 2066.

Important Statistical Concepts

• Learning from Data: Utilizing earlier research to develop hypotheses and test against new data.

Aggregation of Data

• Definition: Combining individual data points into a collective summary.

• Significance: Helps in understanding broader trends from specific instances.

Understanding Uncertainty

• Risk Assessment (example of type 2 diabetes risk based on scoring):

  • Low risk (0-5 points): ~1 in 100 chance.

  • Moderate risk (6-8 points): ~1 in 50 chance.

  • High risk (9-11 points): ~1 in 30 chance.

  • Very high risk (20+ points): ~1 in 3 chance.

• Interpretation: Understanding variability in predictions (e.g., diabetes risk may vary if uncertainty is high vs. low).

Sampling from a Population

• Representative Sample: Accurately reflects the larger group.

• Biased Sample: Not reflective, leads to skewed data interpretation.

Causality and Statistics

• Correlation vs. Causation: Explore relationships without assuming direct cause.

  • Example: Correlation of cheese consumption to deaths from bed sheet accidents shows correlation but does not imply causation.

  • Caution: Always investigate causal links with experimental design; use terms like association when observing relationships.

Conclusion

• Statistical methods provide critical tools for decision-making, understanding populations, and interpreting data with an inherent uncertainty factor. Recognizing aggregation and sampling methods fosters better research practices. Proceed with caution when interpreting causal relationships in statistics.