Quantitative Methods and Research

SESSION 15: Quantitative Methods

  • Focus on numbers and empirical data.

  • Use of close-ended questions in data collection.

  • Goal: Test objective theories and explore relationships between variables.

  • Involves measurement with instruments and employing statistical procedures.

  • Follows a deductive approach; rooted in a positivist worldview.

Mixed Methods

  • Combines qualitative and quantitative approaches for comprehensive analysis.

The Process of Deduction

  • Start from a theory.

  • Derive hypotheses.

  • Collect data to test hypotheses.

  • Interpret results, typical of quantitative research.

Epistemology and Worldview of the Researcher

  • Positivism:

    • Theories must be tested and refined.

    • Break down broader problems into smaller, testable hypotheses.

    • Emphasizes cause-effect relationships with measurable outcomes.

Key Concepts of Positivism
  1. Imperfect Evidence: Research evidence is always fallible.

    • Concept of falsification: cannot truly prove a hypothesis - only fail to reject it.

    • Karl Popper (1935):

      • Example: Observing white swans does not mean all swans are white; finding one black swan refutes that theory.

Relationships Between Quantitative Variables: Correlation

  • Types of Correlation:

    • Positive Correlation: As one variable increases, so does another.

    • Negative Correlation: As one variable increases, another decreases.

    • Zero Correlation: No observable relationship between variables.

    • Example: Time spent on social media vs. wellbeing.

Relationships Between Variables: Causation

  • Beyond correlation, researchers seek to understand causal relationships.

  • Other influencing factors include:

    • Time of year, parenting, family environment, and lurking variables.

Determining Causation
  • Difficult to assess causation through observational studies or surveys.

  • Easier in experimental studies that control for other variables through:

    • Randomization.

    • Balanced groups across conditions.

SESSION 16: Experimental Research

  • Flawed Study Example: Dr. Zimmo’s Sample Bias

    • Selected 25 students for feedback; only 10 responded positively.

    • Likely sampling bias due to non-random and incomplete sample.

Getting Primary Data

  1. Ask People: Interviews, surveys, focus groups.

  2. Watch People: Direct observation.

  3. Test People: Conduct experiments.

Sampling
  • Process of selecting units (people, organizations, etc.) from a population.

  • Goal: Generalize findings from the sample back to the broader population.

Sample vs. Population
  • A good sample is representative of the population.

  • Each participant should have an equal chance of selection to avoid favoritism.

Inferential Statistics
  • Allows predictions about a population based on sample data.

Randomization Techniques
  • Simple Random Sampling:

    • Similar to a lottery; involves number assignment and random selection.

  • Proportional Stratified Sampling:

    • Divides population into strata and samples accordingly to maintain proportional representation.

Qualities of a Good Sample

  • Representativeness and equal chances of being selected help reduce bias.

  • Bias can lead to inaccurate conclusions.

Examples of Biased Sampling
  • Voluntary Response Sample: May not represent the entire population due to self-selection.

  • Convenience Sample: Easy access to participants, but not representative.

Non-Response Bias Example

  • Example of survey on extramarital affairs where only 4.5% responded, potentially skewing results due to the behavior of non-responders.

Sampling Summary
  • Challenges in obtaining representative samples.

  • Random sampling is ideal but practically difficult due to biases:

    1. Selection bias from voluntary/convenience samples.

    2. Non-response bias.

Measurement in Research

  • Measurement: Assigning numbers to objects based on defined rules.

  • Variable: A measurable characteristic that varies among individuals (e.g., hours studied, screen time).

Types of Variables
  1. Quantitative Variables: Numeric values (e.g., counts, measurements).

  2. Categorical Variables: Qualitative differences that do not imply quantity.

Scales of Measurement
  1. Nominal Scale: Numbers as labels only; no order (e.g., category labels).

  2. Ordinal Scale: Numbers imply order without equal spacing (e.g., rankings).

  3. Interval Scale: Ordered with equal intervals; no true zero (e.g., temperature).

  4. Ratio Scale: Ordered, equal intervals, with a true zero point (e.g., weight).

Evaluating the Quality of Quantitative Research

  • Key Concepts:

    1. Objectivity: Findings should be independent of the researcher.

    2. Reliability: Consistency of measurements over time - described by the formula: X = T + E (observed score = true score + error).

    3. Validity: Measures how well an instrument measures what it intends to.

Validity & Reliability Combined Using Target Analogy
  1. Reliable but not valid.

  2. Valid on average but not reliable.

  3. Neither valid nor reliable.

  4. Both reliable and valid.