Quantitative Management & Analysis

Quantitative Management & Analysis

Learning Objectives

  • Describe concepts underpinning empirical (positivistic) research and questions answered using quantitative designs.

  • Outline and describe varying research designs associated with experimental and non-experimental (observational) research.

  • Define independent, dependent, and control variables.

  • Compare strengths and limitations of various types of quantitative research designs.

  • Differentiate between various probability and non-probability sampling methods.

  • Differentiate between descriptive and inferential statistics.

  • Describe four levels of measurement and means of reporting descriptive data, including central tendency and data distribution.

The Research Process

  1. Topic Selection

  2. Literature Review (understanding current knowledge, practices, and gaps)

  3. Formulating Research Questions

  4. Selecting Appropriate Research Design

  5. Data Collection (setting, sample, methods)

  6. Data Analysis

  7. Discussing findings in literature/practice

  8. Reporting findings

Quantitative Research

  • Definition: Quantitative research tests objective theories by examining relationships among variables or comparing groups.

  • Variables are measured, typically with instruments, to enable numerical data analysis through statistical procedures (Creswell & Creswell, 2024).

Quantitative Approaches

  • Reflect deterministic philosophy and positivist paradigm, employing the scientific method.

  • Examines causes, effects, interactions, and outcomes; reality is discovered probabilistically.

  • Deductive approach where concepts are reduced to variables and tested.

  • Findings result from careful observation, measurement, and interpretation of objective reality.

Study Designs in Quantitative Research

Categories of Research Designs

  1. Experimental Designs

    • Participants are subjected to an intervention (e.g., administering medication).

    • Main categories:

      • Randomized Control Trials (RCTs)

      • Quasi-experimental studies

  2. Observational Designs

    • No intervention is appliedā€”observes existing conditions or outcomes.

Experimental Designs

Overview of Experimental Research

  • Focus on doing something to participants (intervention).

  • Includes RCTs and quasi-experimental studies.

Key Concepts

  • Variables

    • Independent Variable: The manipulated predictor variable that predicts change.

    • Dependent Variable: Outcome or result affected by the independent variable.

    • Control Variable: Factors that remain constant during the study.

Essentials of Randomized Control Trials

  1. Control: Use of a control group to establish constants.

  2. Manipulation: Researcher intervenes with some participants.

  3. Randomization: Random allocation of participants to experimental/control groups.

Characteristics of Randomized Control Trials

  • Objective and systematic; considered the gold standard for reliable evidence.

  • Rigorous control of variables and random allocation (equal allocation chance).

  • Allocation concealment to ensure unbiased treatment assignments.

  • Blinding: Ensures participants do not know their allocation (can be single or double).

Example: Randomized Controlled Trial

  • Examined ultrasound-guided femoral nerve block (FNB) versus fascia iliaca compartment block (FICB).

  • Methods:

    • A double-blind RCT was conducted with both blocks given to the participants.

    • Results indicated no superiority in pain relief efficacy between FNB and FICB.

    • Reduction in pain scores: mean for FICB = 2.62, FNB = 2.30 (no significant difference, $P=0.408$).

Example: Pilot RCT

  • REFRESH Study: Investigated effectiveness of restricted fluid resuscitation in sepsis.

    • Design: Prospective, randomized, open-label.

    • Sample Size: 99 participants (50 restricted, 49 usual care).

    • Findings: Median volume administered in restricted group was significantly less ($p<0.001$).

Strengths of RCTs

  • Gold standard for assessing cause and effect.

  • Minimizes bias.

  • High internal validity.

Limitations of RCTs

  • Expensive and may have short follow-ups.

  • Volunteer bias and stringent conditions may not reflect real-world scenarios.

  • Some interventions are unsuitable for RCTs due to ethical or practical implications.

Quasi-Experimental Design

  • Non-randomized designs that may include a control group or none.

  • Classic example: Pre-test post-test designs and time series designs.

Strengths of Quasi-Experimental Design

  • Generally less resource-intensive than RCTs.

  • Ethical choice when RCT is not feasible.

  • Results may be more generalizable as they occur in real-world settings.

Limitations of Quasi-Experimental Design

  • Higher risk of confounding factors affecting results.

  • Caution required in causality interpretation compared to RCTs.

Observational Design

  • Non-experimental; can be descriptive or correlational.

  • Aims to observe, describe, and identify variables of interest and their relationships.

Descriptive Studies

  • Purpose: Observe and document situations as they naturally occur (Flanagan & Beck, 2025).

Types of Observational Studies

  • Longitudinal: Repeated measurements at multiple time points.

  • Cross-sectional: Snapshot of data at one point in time, helpful for understanding a population.

Strengths and Limitations of Descriptive Design

  • Strengths: Cost-effective, efficient, less time-consuming, helps determine issue scope.

  • Limitations: Cannot determine cause-effect relationships.

Correlational Studies

  • Examines relationships between non-manipulated variables.

  • Key takeaway: Correlation does not imply causation!

Correlational Study Design

  • Cohort Study: Prospective design starting with a presumed cause.

  • Case-Control Study: Retrospective design comparing cases and controls.

Strengths and Limitations of Observational Correlational Studies

  • Strengths: Efficient, collect real-world data.

  • Limitations: Cannot assume pre-existing similarities in compared groups; cautious interpretation needed.

Sampling in Quantitative Research

  • Sampling: Process of selecting units for study from a broader population.

  • Distinction between study population (entire group) vs. study sample (subgroup).

Sampling Methods

  1. Probability Sampling (Randomization): Each individual has an equal chance of selection.

    • Simple Random Sampling

    • Systematic Sampling

    • Stratified Sampling

    • Cluster Sampling

  2. Non-Probability Sampling: Does not give all individuals an equal chance of selection.

    • Convenience Sampling

    • Purposive Sampling

    • Quota Sampling

    • Snowball Sampling

Power Analysis

  • Purpose: Determines the minimum sample size required for a research study to detect effects when they exist.

  • Influenced by: Alpha level, effect size, sample size.

Example of Power-Sample Size Analysis

  • Example: Nurses in WA working night shift.

  • Population size = 28,240; confidence level = 95%; needed sample = 320, with a margin of error = 5%.

Quantitative Data Collection Methods

  • Empirical evidence is gathered through:

    • Physiological and biological measures

    • Observational measures

    • Psychological measurements

    • Questionnaires, surveys, scales

    • Quantitative interviews

Reliability and Validity of Measurements

Reliability

  • Consistency or dependability of an instrument with types including:

    • Internal consistency

    • Equivalence

    • Stability

    • Interrater reliability

Validity

  • Measures the extent to which an instrument accurately captures the concept it's intended to measure.

    • Types of validity include content, criterion, and construct validity.

Levels of Measurement

Level

Description

Examples

Nominal

Categorization without ranking

Gender, Eye colour, City of Birth

Ordinal

Categorization with ranking (unequal intervals)

Educational Levels, Customer Satisfaction, Ratings

Interval

Ranked with equal intervals, no true zero

Temperature, Calendar Dates

Ratio

Ranked with equal intervals, true zero point

Height, Weight, Age, Income

Data Analysis

Types of Data Analysis

  1. Descriptive Statistics: Summarizes the data (mean, median, mode).

    • Example: 65% of ED visits by individuals over 65.

  2. Inferential Statistics: Tests hypotheses about population using sample data.

    • Parametric tests (e.g., T-tests, ANOVA) require normally distributed data.

    • Non-parametric tests (e.g., Chi- Square) are not dependent on normality.

Level of Significance

  • P-value: Measure of the probability that an observed result occurred by chance.

  • Set before the study, often at $p< .05$ or $p< .01$.

Implication of the Null Hypothesis

  • Example Null Hypothesis: No difference in weight loss between two exercise groups.

  • A high p-value suggests accepting the null; a low p-value allows rejection of the null.

Bias in Research

Addressing Bias

  • Various biases include sampling, response, and confounding biases with corresponding methods to mitigate them:

    • Sampling Bias: Employ randomization and strict criteria.

    • Response Bias: Use blinding and standardized protocols.

    • Attrition Bias: Intention-to-treat approach.

Types of Bias

  • Hawthorne Effect: Influence on participants' behavior due to being observed.

  • Publication Bias: Effects on research visibility and interpretation.

Hierarchy of Evidence

  1. High Level: Systematic reviews, meta-analysis, RCTs.

  2. Level 2: Well-designed RCTs.

  3. Level 3-4: Cohort and case-control studies.

  4. Level 5: Systematic reviews of qualitative studies.

  5. Low Level: Expert opinions and panel reports.

Comparison of Quantitative and Qualitative Research

Qualitative

Quantitative

Semi-structured interviews, focus groups

Experiments, surveys, structured interviews

Subjective emphasis

Objective emphasis

Inductive reasoning

Deductive reasoning

Low structure

High structure

Longer time requirement

Less time required

Activities

Experimental Design Activity

  1. In an aged care facility, explore whether daily prune juice aids in constipation:

    • Experimental Design: Outline intervention and control groups

    • Non-Experimental Design: Describe observational approach

  2. Variables:

    • Independent: Daily prune juice

    • Dependent: Bowel movement regularity

    • Control: Other dietary factors

  3. Hypotheses:

    • Directional: Daily prune juice reduces constipation.

    • Null: No difference in bowel movements with/without prune juice.

  4. Randomization method for participant assignment.

  5. Description of blinding procedures.

  6. Causal link determination for prune juice impact.

Measurement Levels Activity

  • Identify measurement levels for items:

    • Number of pizzas eaten (discrete)

    • Color of cars (nominal)

    • Exam grade (ordinal)

    • Satisfaction level (ordinal/continuous)

    • Bank account amount (ratio)

    • GPA (ratio)

    • Political preference (nominal)

    • Happiness level (ordinal)

    • Average temperatures in Perth (ratio)