Summary of Comprehensive Guide to Business Statistics
Comprehensive Guide to Business Statistics for Students
Introduction
Descriptive statistics often begin by organising raw numerical observations so we can see patterns, shapes and summaries. Frequency distributions and grouped data are fundamental tools for summarising quantitative variables by counting how many observations fall into ranges called classes. This material explains how to build frequency tables, choose class widths and boundaries, and how to visualise grouped data with histograms and charts.
Definition: A frequency distribution is a table that groups raw data into classes and records the number of observations (frequency) that fall in each class.
Key concepts and step-by-step procedures
1. Range and number of classes
- Calculate the range of the data as: the maximum minus the minimum. In LaTeX: $$\text{Range} = \max(x_i) - \min(x_i)$$
- Choose a convenient number of classes; typically between 6 and 9 for many data sets. The exact choice depends on the sample size and the level of detail desired.
Tip: Fewer classes give a coarse summary; more classes give more detail but may make the table sparse.
2. Class size (class width)
- Compute the class size using: $$\text{Class size} = \frac{\text{Range}}{\text{Number of Classes}}$$
- Round the result up to the next whole number (always round up). For example, if you get $9.85$, choose $10$ as the class size.
- When possible, choose a class size that is a convenient round number (1, 5, 10, 20, 25, 50, ...).
3. Discrete vs continuous data
- Discrete data: values take separate whole-number values (counts). For discrete data you usually create classes whose limits do not touch; there is typically a gap of 1 unit between an upper limit and the next lower limit.
- Continuous data: measurements can take any value in an interval. For continuous data classes are contiguous: the upper boundary of one class is the lower boundary of the next.
Definition: Class limits are the smallest and largest values in the written class (e.g., for class “10–19”, 10 is the lower class limit and 19 is the upper class limit).
4. Class boundaries and continuity
- For continuous data you often express a class as "$a$ up to but less than $b$" so consecutive classes can share numeric endpoints without overlap problems (e.g., "$20$ but less than $30$" and "$30$ but less than $40$").
- Class boundaries (also called class delimiters) are used to avoid gaps when plotting histograms. For example, if class limits are 10–19 and 20–29, the class boundaries for continuous plotting might be 9.5–19.5 and 19.5–29.5.
Definition: A class boundary is the precise cut-off value between classes; it may be obtained by subtracting or adding half the unit of measurement to the class limits when switching between discrete and continuous representation.
5. Constructing a grouped frequency table (step by step)
- Sort the data (optional but helpful) and compute the range.
- Decide the number of classes $k$ (e.g., $k=7$).
- Compute class size $a$ with $$a = \left\lceil \frac{\text{Range}}{k} \right\rceil$$ where $\lceil,\cdot,\rceil$ denotes rounding up.
- Select the lower limit of the first class. It is convenient to start slightly below the minimum observation or at a round number.
- Create the classes by adding the class size repeatedly. For discrete data, ensure class limits differ by one unit where required; for continuous data use contiguous ranges or express as "but less than".
- Tally observations into classes and record the class frequencies $f_i$.
- Optionally compute cumulative frequencies and relative frequencies (percentages).
6. Example (continuous data outline)
Suppose you have distances: $117,119,125,128,133,134,134,135,135,136,138,138,138,138,142,142,142,144,144,144,144,145,146,146,147,147,148,148,148,152,152,156,158,161,163,165,168,173,176$ (40 values). Steps:
- Range: $$\text{Range} = 176 - 117 = 59$$
- Choose $k=7$ classes, compute class size: $$a = \left\lceil\frac{59}{7}\right\rceil = \le
Sign up for the full summaryAlready have an account? Sign in
Frequency & Grouped Data
Klíčové pojmy: Range equals max minus min, Choose 6–9 classes typically, Class size = Range/Number of classes and always round up, Use convenient round class widths when possible, Discrete classes usually have unit gaps; continuous classes are contiguous, For continuous data express classes as "a but less than b", When plotting histograms use class boundaries to avoid gaps, If class widths differ use frequency density = frequency/width