Research Methods: A Comprehensive Guide for Students and Researchers

Posted on Aug 2, 2024 in Psychology and Sociology

Research Methods: A Comprehensive Guide

Problem Statement

A well-articulated problem statement clearly defines the research problem and provides an argument that justifies the need for the study.

Purpose Statement

The purpose statement outlines the specific aim or goal of the study.

Variables

Independent Variables

  • Presumed cause or influence
  • Causes a change in the outcome of interest
  • May be an intervention

Dependent Variable

  • Presumed effect or outcome variable
  • Variable researchers are trying to understand, explain, or predict

Experimental Designs

Randomized Controlled Trial (RCT)

  • Best evidence that the intervention caused the outcome
  • Gold standard for establishing causation
  • Multiple RCTs can be combined in a meta-analysis

Limitations:

  • Many research variables cannot be manipulated ethically or realistically
  • Randomization may not be possible
  • Subject to the Hawthorne Effect

Quasi-Experimental Designs

  • Enhanced practicality
  • Some control
  • Participants may not always be willing to be randomized

Limitations:

  • Cause and effect may be difficult to establish
  • Non-equivalence between groups

Non-Experimental Designs

  • Study problems that cannot be investigated experimentally
  • Can gather information on relationships between multiple variables
  • Experimental studies often rely on descriptive correlational studies

Limitations:

  • No causation can be established
  • Self-selection bias

Manipulation, Observation, Randomization, and Control

  • Manipulation: Is there an intervention?
  • Observation: When and how often are the researchers collecting data?
  • Randomization: How many groups are being compared?
  • Control: What is done to the control group instead of the intervention?

Counterfactual

The counterfactual describes what is done to the control group instead of the intervention, such as:

  • No intervention
  • No treatment
  • Alternative treatment
  • Placebo
  • Standard care or usual care
  • Modified treatment (lower dose, partial intervention)
  • Delayed treatment (waitlisting control group)

Probability Sampling

Probability sampling ensures that each element of an accessible population has an equal chance of being selected for the sample.

  • Simple Random Sampling: Requires a sampling frame (list of all available members in the population)
  • Stratified Random Sampling: The sampling frame is divided into strata, and random sampling is completed from each stratum separately
  • Cluster Sampling: Successive random sampling at multiple stages
  • Systematic Sampling: Every kth case is selected

Non-Probability Sampling

Non-probability sampling methods do not ensure equal chances of selection for all elements in the population.

  • Convenience Sampling: Uses the most available individuals
  • Volunteer Sampling: Recruits participants through newspaper ads or posters
  • Snowball Sampling: Participants recruit other participants through word of mouth
  • Consecutive Sampling: Recruits all people from an accessible population over time
  • Quota Sampling: Ensures proportional representation of population strata
  • Purposive Sampling: Selects participants based on specific criteria

Sample Size

Power analysis can be used to determine the appropriate sample size for a study.

Informed Consent

Informed consent is obtained from participants before they participate in a study. This includes providing them with information about the study and their rights.

Confidentiality

Confidentiality procedures are used to protect the privacy of participants. This may include using anonymous numbers instead of names and securing information.

Statistical Tests

T-test

  • Used to test the statistical significance of a difference between means of two groups
  • Parametric (assumes a normal distribution)
  • Independent group t-test (means of different groups)
  • Dependent group t-test or paired t-test (means of the same group)

ANOVA (Analysis of Variance)

  • Measures significant differences in means of three or more groups
  • Parametric (assumes a normal distribution)
  • F-ratio
  • Multiple comparison procedures (post hoc tests) to identify which group caused the significant difference
  • Repeated measures ANOVA (when means are compared at different points in time)

Chi-squared Test (χ2)

  • Non-parametric test (does not require a normal distribution)
  • Compares differences in proportions in two or more groups
  • Nominal or ordinal data
  • 2 x 2 crosstab table

Correlation Coefficients (Pearson’s r)

  • Parametric
  • Measures the magnitude and direction of a relationship between two variables
  • Any correlation greater than r = 0.70 is considered strong

Multivariate Analysis

  • Deals with comparing three or more variables simultaneously
  • Multiple regression uses F tables when squared R2
  • ANCOVA (combination of ANOVA and multiple regression)
  • MANOVA (multivariate and ANOVA)