Descriptive Statistics

1
Descriptive Statistics
2
Exploratory Data Analysis (EDA)

• Searching for patterns in the data

3
Data Organization

• Coding sheets/scripts
• Stacked vs. Unstacked format
• Data entry problems - transcription errors
• Grouped vs. Individual Data

4
Graphing and Visualizing Data

• Bar graphs - best when the IV is categorical
  (nominal)
• Line graphs - best for illustrating functional
  relationships between variables

5
Line Graph Terminology

• Positive vs. negative acceleration
• Increasing vs. decreasing
• Monotonic vs. non-monotonic
• Floor vs. ceiling values

6
Scatterplots

• Best for two DV's when illustrating correlations
  or linearity

7
Pie Charts

• Best for proportions or percentages

8
Tables vs. Graphs

• Numerical precision lost in a graph, but
  relationships in the data are shown more clearly

9
Frequency Distribution of Data

• Histogram
• Columns should not be too narrow or too wide
• Positive vs. negative skew
• Normal distributions
• Outliers
• Stemplot
• Stem leaf
• All values are seen

10
Measures of Centers of Distributions

• Center
• Mode
• Median
• Mean
• Choosing a center
• Nominal - mode
• Ordinal - mode or median
• Interval or ratio - mean
• Skew - median
• Normal -mean

11
Measures of Spread of Distributions

• Range
• Problems sensitive to outliers scores in
  between not taken into account
• Interquartile range
• Variance
• Standard Deviation
• Choice of measure of spread
• If there are outliers or skew, choose
  interquartile

12
The 5-Number Summary

• min and max scores
• 1st quartile, median, and 3rd quartile scores
• Boxplot - visual representation of the 5-number
  summary

13
Statistical Measures of Description

• Used to describe data measures of association
  between variables

14
Measures of Association

• Pearson product-moment correlation coefficient
  (Pearson R)
• Used for two variables with continuous values
• Positive and negative relationships denoted
• Magnitude of the number tells you the degree of
  linear relationship between the two variables

15

• Factors that affect magnitude of the Pearson R
• Outliers
• Restricted range of scores
• Non-normal distribution of scores

16
Other Correlation Tests

• Point-biserial correlation
• Used when one variable is dichotomous
• Spearman rank order correlation (rho)
• Used to determine monotonicity
• Phi coefficient
• Used when both variables are dichotomous
• Linear regression analyses
• Used for predicting scores and values
• Multivariate Analysis