Chapter 2 - Using Graphs to Visualize Data - Part 2

Modern graphs

Used to illustrate average values, making them good alternatives to bar graphs

What are the two modern graphs

1. WHAT

2. WHAT

What are the two modern graphs

1. Violin plots/density plots

2. Jitter plots

Violin plot

Creates a graph of a continuous data distribution that looks like a stringed musical instrument 

In a violin plot the wider section have greater WHAT of response 

In a violin plot the wider section have greater FREQUENCIES of response 

Because the data are more densely clustered in places where the violin is WHAT the violin plot is considered a WHAT plot 

Because the data are more densely clustered in places where the violin is WIDER the violin plot is considered a DENSITY plot 

A violin plot shows the WHAT of data for a given variable 

A violin plot shows the SPREAD of data for a given variable 

Using a horizontal violin plot lets you see the amount of WHAT among ratings and the amount of WHAT for each rating 

Using a horizontal violin plot lets you see the amount of OVERLAP among ratings and the amount of VARIABILITY for each rating 

Jitter plot

Display individual dots on a graph to illustrate each and every participant’s score on one quantitative variable 

Dots on a scatterplot represents a WHAT participant’s score on two variables while the dots on a jitter plot correspond with only one WHAT measured 

Dots on a scatterplot represents a SINGLE participant’s score on two variables while the dots on a jitter plot correspond with only one VARIABLE measured 

Two common sources of untrustworthy or misleading data are:

1. WHAT

2. WHAT

Two common sources of untrustworthy or misleading data are:

1. Truncated Y-axis

2. Poor design of a data visualization 

Truncated y-axis

One that is shortened by cutting off one or both ends of it 

When the Y-axis is truncated it doesn’t show the full WHAT of all possible data values for the variable measures on the y-axis 

When the Y-axis is truncated it doesn’t show the full RANGE of all possible data values for the variable measures on the y-axis 

Use of a WHAT mark on the y-axis to emphasize the use of a truncated y-axis

Use of a DOUBLE HORIZONTAL SLASH mark on the y-axis to emphasize the use of a truncated y-axis

Poor design choice

Examples: having the shortest bar on the graph represent the largest % or placing days in the wrong order on the x-axis

When reading graphs, the two things to check for poor design choices are

1. Do the numbers appear to correspond WHAT

2. If the graph is representing data over time are the dates in the WHAT

When reading graphs, the two things to check for poor design choices are

1. Do the numbers appear to correspond VISUALLY

2. If the graph is representing data over time are the dates in the CORRECT ORDER

What are the four types of distributions for continuous data

1. WHAT

2. WHAT

3. WHAT

4. WHAT

What are the four types of distributions for continuous data

1. Normal

2. Skewed

3. Unimodal

4. Bimodal

The normal distribution is the most WHAT

The normal distribution is the most COMMON

Normal distribution or normal curve or bell curve

Shows data distributions in which half the data fall above the average score and half fall below the average score with most data clustered around the average and some data spread out in the tails

Skewed distribution

Data are not always distributed symmetrically

Outliers

are extreme relative to other data points for the same variable 

In skewed distribution, the data WHAT at one end of the distribution and WHAT at the other end 

In skewed distribution, the data PILE UP at one end of the distribution and TAPER OFF at the other end 

Positively skewed or right skewed

To the right or above the average 

Negatively skewed or left skewed

To the left or below the average 

Data always skew in the direction of the WHAT tail of more WHAT score in your dataset 

Data always skew in the direction of the LONGER tail of more EXTREME scores in your dataset 

The tail of a positively skewed dataset points to increasing WHAT number 

The tail of a positively skewed dataset points to increasing POSITIVE number 

The tail of a negatively skewed dataset points to increasing WHAT number 

The tail of a negatively skewed dataset points to increasing NEGATIVE number 

Unimodal distribution

when a distribution has only one mode or peak 

Bimodal distributions

Have two most frequent responses - sometimes they are equal and sometimes they differ