FEM3002 Research Method: Topic 9

FEM3002 Research Method: Topic 9 Data Analysis (Part 2)

• Statistical Analysis involves descriptive and inferential statistics.
• Descriptive Statistics provides a way to describe data in manageable forms, including data collection, tabulation, summarization to create meaningful information.
• Descriptive statistics include quantitative descriptions of single variables, and the associations between variables.
• Data Reduction is a type of descriptive statistics, used to reduce unmanageable data and summarize it into manageable categories. For example, data on 100 different ages could be summarized into 3-4 age categories.
• Measures of Association show the nature and extent of the relationship between variables.
  • For nominal variables, it's Lambda.
  • For ordinal variables, it's Gamma.
  • For interval or ratio variables, it's Pearson's product-moment correlation (r).
• The Value of r (Pearson Correlation Coefficient):
  • 0.0: no linear relationship.
  • +1.0: strong positive linear relationship (as X increases, Y increases).
  • -1.0: strong inverse linear relationship (as X increases, Y decreases).
• Positive or Negative r: Graphs illustrate the direction of a relationship.
• Pearson Correlation Analysis: used in statistical software like SPSS, to find correlation between variables (e.g., age and income).
• Regression Analysis: represents relationships between variables via equations to predict values of a dependent variable based on independent variables.
• Simple Linear Regression: a basic regression equation (Y = a + bx + e) explains the relationship; where:
  • Y: estimated value of dependent variable
  • a: constant/intercept
  • b: slope, coefficient of independent variable (X)
  • e: error
• Significance Level: Researchers set a significance level for statistical tests using probability theory to assess if findings are real or due to chance.
• Multiple Regression: involves multiple independent variables (IV) to predict the dependent variable (DV).
• Inferential Statistics: helps draw conclusions from a sample to a population.
  • It's based on the assumption of normal distribution of population variables.
  • The area under a bell curve of the distribution is very important in inferential statistics
  • Inferential statistics also requires that the sample accurately represents the population from which it's drawn. Random selection is key in creating representative samples.

Software Usage

• SPSS: software used for performing statistical analyses like frequency analysis. SPSS has a user-interface to input data from a CSV file, and calculate values like correlations or regressions. A video on frequency analysis in SPSS is linked.