Comprehensive Notes on Frequency Distributions, Histograms, and Data Analysis

Class Limits, Boundaries, and Widths

• Class Limits:

  • Lower Class Limits: The smallest numbers that belong to the different classes.

  • Upper Class Limits: The largest numbers that belong to the different classes.

• Class Boundaries (Edges):

  • Class boundaries represent the exact numerical edges separating adjacent classes, preventing gaps between class limits.

  • Lower Edge Calculation: Calculated by determining the midpoint between the lower limit of a class and the upper limit of the preceding class. For a class with lower limit $125$, the lower boundary edge is $124.5$.

  • Upper Edge Calculation: For a class with upper limit $174$, the upper boundary edge is $174.5$.

• Class Width:

  • The difference between two consecutive lower class limits or two consecutive upper class limits.

  • Must remain constant across all classes in a frequency distribution.

  • Example: For lower limits $75$ and $125$, the class width is:         $\text{Class Width} = 125 - 75 = 50\,\text{seconds}$

• McDonald's Drive-Thru Service Wait Time Dataset ($N = 50$):

  • Class 1 ($75\,\text{s} - 124\,\text{s}$): Frequency = $11$

  • Class 2 ($125\,\text{s} - 174\,\text{s}$): Frequency = $24$

    • This interval contains the highest frequency. Most customer wait times fall in this class.

    • Converting $125\,\text{s} - 174\,\text{s}$ to minutes gives a typical customer wait time of approximately $2\,\text{to}\,3\,\text{minutes}$.

  • Class 3 ($175\,\text{s} - 224\,\text{s}$): Frequency = $10$

  • Class 4 ($225\,\text{s} - 274\,\text{s}$): Frequency = $3$

  • Class 5 ($275\,\text{s} - 324\,\text{s}$): Frequency = $2$

  • Total Frequency Sum:         $\text{Total Frequency} = 11 + 24 + 10 + 3 + 2 = 50$

Step-by-Step Construction of Frequency Distributions

• Class Selection Procedure:

  1. Select Lower Class Limit of First Class: Choose the smallest raw data value or a convenient value slightly lower than the minimum.

  2. Select Target Class Count: Decide on the number of classes to construct. Standard frequency distributions generally use between $5$ and $20$ classes.

  3. Compute Class Width:         $\text{Class Width} = \frac{\text{Maximum Value} - \text{Minimum Value}}{\text{Number of Classes}}$

  4. Round Up Class Width: Always round the calculated class width UP to ensure all raw data values fit into classes without omission or boundary errors.

  5. Construct Class Limits: List consecutive lower class limits by adding the class width sequentially. Construct upper class limits accordingly.

  6. Count Class Frequencies: Tally the raw data points belonging to each distinct class interval.

• Rules for Data Fitting:

  • No Overlap: Classes must be mutually exclusive. Every single raw data value must fit into exactly one class.

  • No Omissions: Classes must be exhaustive. No data values may be left out unless explicitly instructed.

Constructing Frequency Distributions in Excel

• Step-by-Step Excel Procedure:

  • Step 1: Enter all raw data observations into a single continuous column (e.g., A1:A51).

  • Step 2: Enter upper class limits (bin boundaries) into a separate adjacent column (e.g., C2:C6).

  • Step 3: Enter the array formula =FREQUENCY(data_array, bins_array).

    • Example formula: =FREQUENCY(A1:A51, C2:C6)

  • Step 4: Select the output cell range, complete the array formula, and label output columns clearly.

• Excel Function Definitions:

  • Data Array: The cell range containing all original quantitative raw data values.

  • Bin Array: The cell range containing the upper limits of each class.

Normal Distributions, Data Gaps, and Penny Weight Analysis

• Characteristics of Normal Distributions:

  • Displays an approximately bell-shaped symmetric curve.

  • Frequencies start low, increase to one or two high peak frequencies near the center, and symmetrically decrease back to low frequencies.

  • The left and right sides of the distribution form mirror images.

• Gaps in Frequency Distributions:

  • A gap occurs when one or more intermediate classes have a frequency equal to $0$.

  • Structural Meaning: Gaps strongly suggest that data points originate from two distinct sub-populations or groups (though multi-group data does not always produce gaps).

• Penny Weight Case Study ($N = 72$ Pennies):

  • Class $2.50\,\text{g} - 2.59\,\text{g}$: Frequency = $18$ (or $19$) pennies

  • Class $2.60\,\text{g} - 2.69\,\text{g}$: Frequency = $0$ pennies (GAP)

  • Class $2.70\,\text{g} - 2.79\,\text{g}$: Frequency = $0$ pennies (GAP)

  • Class $2.80\,\text{g} - 2.89\,\text{g}$: Frequency = $0$ pennies (GAP)

  • Class $2.90\,\text{g} - 2.99\,\text{g}$: Frequency = $2$ pennies

  • Class $3.00\,\text{g} - 3.09\,\text{g}$: Frequency = $25$ pennies (Highest Frequency Peak)

  • Class $3.10\,\text{g} - 3.19\,\text{g}$: Frequency = $8$ pennies

  • Total Sum Calculation:         $18 + 19 + 0 + 0 + 0 + 2 + 25 + 8 = 72\,\text{pennies}$

  • Analytical Findings:

    • The typical weight cluster centers around $3.00\,\text{g}$, where the frequency peaks at $25$

    • A multi-class gap occurs between $2.60\,\text{g}$ and $2.89\,\text{g}$

    • Because of the prominent gap and dual clustering, this distribution is not normal.

Relative Frequency Distributions and Percentage Conversions

• Relative Frequency Formula:     $\text{Relative Frequency} = \frac{\text{Class Frequency}}{\text{Total Frequency (Sample Size)}}$

• McDonald's Wait Time Relative Frequency Table ($N = 50$):

  • Class $75\,\text{s} - 124\,\text{s}$:         $\text{Relative Frequency} = \frac{11}{50} = 0.22 \implies 22\%$

  • Class $125\,\text{s} - 174\,\text{s}$:         $\text{Relative Frequency} = \frac{24}{50} = 0.48 \implies 48\%$

  • Class $175\,\text{s} - 224\,\text{s}$:         $\text{Relative Frequency} = \frac{10}{50} = 0.20 \implies 20\%$

  • Class $225\,\text{s} - 274\,\text{s}$:         $\text{Relative Frequency} = \frac{3}{50} = 0.06 \implies 6\%$

  • Class $275\,\text{s} - 324\,\text{s}$:         $\text{Relative Frequency} = \frac{2}{50} = 0.04 \implies 4\%$

• Rules for Relative Frequencies:

  • Percentage Conversion: Convert relative frequency decimals to percentages by multiplying by $100\%$ (e.g., $0.48 \times 100\% = 48\%$).

  • Reporting Standards: Always explicitly output percentage values (48\%$) when questions request percentages, rather than stopping at decimal form (0.48).\n * **Sum Constraint**: The sum of all relative frequencies in a distribution must equal 1.00$or$100\%.\n\n# Cumulative Frequency and Running Totals\n\n* **Definition**:\n * Cumulative frequency (running total) sums the frequency of a given class with the frequencies of all preceding classes.\n\n* **McDonald's Wait Time Cumulative Frequency Construction (N = 50)**:\n * **Less than 125\,\text{seconds}$**: Cumulative Frequency =$11\n * **Less than 175\,\text{seconds}$**:$11 + 24 = 35\n * **Less than 225\,\text{seconds}$**:$35 + 10 = 45\n * **Less than 275\,\text{seconds}$**:$45 + 3 = 48\n * **Less than 325\,\text{seconds}$**:$48 + 2 = 50\n\n* **Cumulative Property**: The cumulative frequency of the final class must equal the total sample size (N = 50).\n\n# Histograms and Graphical Data Presentation\n\n* **Definition & Purpose**:\n * A histogram is a graphical representation of quantitative data using adjacent bars derived from a frequency distribution.\n * Used to analyze distribution shape, estimate central tendency and spread, and identify outliers.\n\n* **Structural Elements of a Histogram**:\n * **Horizontal Axis (x-axis)**: Represents quantitative classes (class limits or boundaries).\n * **Vertical Axis (y-axis)**: Represents frequencies or relative frequencies.\n * **Bar Heights**: Equal to class frequencies or relative frequencies.\n * **Bar Spacing Rule**: Adjacent bars must touch without gaps, unless an actual gap exists in the underlying data (a class frequency of 0$$).

• Software Chart Editing (Excel):

  • Default bar charts created by spreadsheet software place space between bars.

  • To turn default bar graphs into accurate histograms, chart settings must be edited to remove or minimize gap width between adjacent bars so that they touch.