Types of Data in Statistical Analysis

Introduction to Data

• Data is essential for statistical analysis.
• Collection of data is necessary to understand phenomena or processes.
• Each data unit collected is termed an observation.
  • Observations can be people, businesses, products, or time periods.
• Variables are the measurements we are interested in (e.g., age, sex, chocolate preference).
• In data representation:
  • Each row in a spreadsheet represents a single observation.
  • Each column corresponds to a variable.

Levels of Measurement

• The level of measurement determines the analysis and summary statistics that can be used.

1. Nominal Level (Categorical/Qualitative)

• Definition: The most basic level of measurement.
• Characteristics:
  • Descriptions or labels with no inherent order.
  • Examples: sex, preferred chocolate type, color.
• Data Representation:
  • Can be stored as text or numerical codes (numbers do not imply order).
  • Summary using frequency or percentage.
  • Cannot calculate mean or average.

2. Ordinal Level

• Definition: Measurement where order is meaningful, but intervals may not be equal.
• Characteristics:
  • Examples: rank, satisfaction levels, fanciness.
  • The gap between ranked values may vary (e.g., race positions, satisfaction levels).
• Data Representation:
  • Summarized using frequencies.
  • Some researchers calculate mean values, but it must be justifiable.

3. Interval Ratio Level

• Definition: The most precise measurement level.
• Characteristics:
  • Measurable data (e.g., number of customers, weight, age, size).
  • Can be discrete (whole numbers) or continuous (fractional numbers).
• Data Representation:
  • Mathematically versatile; common summary statistics include mean, median, and standard deviation.

Graphical Representation of Data

• Appropriate graph types depend on the level of measurement:
  • Nominal Data: Pie chart, column chart, bar chart.
  • Ordinal Data: Column chart, bar chart (not pie charts).
  • Interval Ratio Data: Bar chart, histogram.
  • Box Plots: Illustrate summary statistics neatly.
  • Time Series Data: Best displayed as a line chart.

Example Case: Helen's Choconutties Survey

• Helen surveys 50 customers about their preferences and satisfaction with Choconutties.
• Collected Data:
  • Nominal: Type of chocolate preferred (e.g., dark, milk, white).
    • Example Summary: 46% dark, 40% milk, 14% white.
  • Ordinal: Satisfaction levels (very satisfied, satisfied, etc.).
    • Example Summary: 32% very satisfied; 72% satisfied or very satisfied.
    • Average satisfaction score: 2.06 (interpretation debated).
  • Interval Ratio: Age, amount spent, number of chocolate bars.
    • Example Summary: Mean age is 38 years, mean groceries spent is $192, mean chocolate bars bought is 3.3.

Conclusion

• The type of analysis performed depends on the level of measurement of the dataset.
• Further insights can be gained from related resources such as the video "Choosing the Test."