General Statistics
Organizing and Decsribing Data
Introduction to Graphs

 
Introduction to Graphs

1. Know the advantages and disadvantages of frequency distributions and graphs compared to statistics to describe distributions.

Tables and graphs are good for quick, overview of distributions (frequency distributions) and serves as visual comparison of many distributions. They are especially useful to evaluate the shape of a distributions.

Statistics are better than graphs in providing specific parametric characteristics of distributions such as central tendency (e.g. Mean) and variability (e.g. Standard deviations).

2. Know the advantages and disadvantages of grouping and situations when grouping may be helpful.

The primary use of grouping data is for making graphs or table summaries.

Grouping (graphs or tables summaries) of data allows key characteristics of the data to be more easily interpreted or presented (a picture is worth a thousand words principle) than would be true if the raw data especially if large would be presented.

Grouping does not result in as much loss of detail as would describing groups with a single statistical parameter such as the mean.

Grouping or statistical parameters does results in loss of precision. 

3. Know the characteristics of common charts or graphs for summarizing data.
 
Frequency Polygon Bar Chart Histogram Pie Chart
Scatterplot Mixed Chart Stem and Leaf Box Plot

4. Know the basic principles for proper construction of charts and graphs.

The number of class intervals should be between 5 and 20

The class width can be determined by:

The ratio of the vertical axis to that of the horizontal axis should be such that the vertical axis (Y) is ¾ the length of the horizontal axis.

The location of the zero value for the vertical axis must be clear indicated (by a broken axis if necessary). Otherwise it is called a truncated chart. Figure 2b.1a shows a truncated chart where the zero point of the y-axis in not clearly shown. Truncated graphs often are misleading.
 
Figure 2b.1a Truncated chart (Height of men in feet between 20 to 35 years old - note that height starts at 4 instead of 0)

s

Figure 2b.1b Non-Truncated chart (Height of men in feet between 20 to 35 years old - note that height starts at 0)

5. Know how to compare the shapes of two distributions with different samples sizes on the same axis.

To compare the shapes of two distributions with different sample sizes on the same axis it is necessary to use percentage as the vertical axis.

Example: A survey of the opinions of 50 men and 80 women on an issue with Yes, No or None response is tabulated and graphed below for comparison
.
Table 2b.2 Response of Women and Men

Sample size for Men is 50 and Women is 80
 
 
Response Men

(# of)

Women (# of) Men

%

Women %
Yes 15 40 30 50
No 25 35 50 43.75
None 10 5 20 6.25

Percent is used for comparison

6. Know the shape of the following types of distributions, circumstances when each occur, and recognize examples of variables that would result in each shape:
 
Positively skewed Negatively skewed J-shaped Bimodal
U-Shaped Normal Rectangular Symmetrical

Figure 2b.3 Normal Distribution
 
Bell-shaped curve
(Asymmetrical shape)

Most random events or processes

Average Height
SAT scores

Figure 2b.4 Positively Skewed (skewed to the right)
 
Non-symmetrical curve

Longer right tail (few high values) 

Frequency bunched to left (low values)

Income, Salary Levels
Difficult tests

Figure 2b.5 Negatively Skewed (skewed to the left)
 
Non-symmetrical curve

Longer left tail (few low values) 

Frequency bunched to right (high values)

Fuel Consumption at various speeds
Easy tests

Figure 2b.6 J-shaped curves (Exponential)
 
Extremely skewed curves
(exponential curves)

Highest point at left or right extremes 

Fast growth or decay rates

Automobile depreciation
Investment growth
Unrestricted Population Growth

Figure 2b.7 Bimodal Distribution
 
Two high point separated by low points
(points not equally high)

Two different homogeneous characteristics mixed together
Or two very popular points of views grouped together.

Average Heights of women and men mixed together
Diameters of nuts and bolts mixed together

Figure 2b.8 U-shaped distribution
 
An extreme bimodal distribution where the two modes are at the extremes points (high and low values)

Extreme views or extremely polarized groups.

Attitudes toward abortions 
Failure rates of automobiles

Figure 2b.8 Rectangular (Uniform) Distribution
 
Equal Frequencies at each point

Each value randomly determined

Frequency of values on a fair die
Class size at each grade