Ms. McLain's Stretch Day Presentation (ANOVA & Tukey Test)
Visualizing Your Data (Statistics & Probability Website_Brown University)
Independent variable (manipulated by the experimenter) on x-axis
Dependent variable (measured or observed by experimenter) on y-axis
Title should be of sufficient length to describe in detail what the graph displays (anyone looking only at the graph should know what is shown)
Both axes should have label and units (with uncertainties if applicable)
A Legend is needed when multiple data sets are graphed on one chart
Error bars (typically showing standard deviation) should be added when appropriate, and discussed in the DEC analysis
Each title below is linked to further instructions on how to create the graph (e.g., LINE GRAPHS:)
used to represent a quantity varying with time (rate) or repeated measurements
used when D.V. shows continuous change
fractions/decimals are possible
slope of "line of best fit" is the rate of change (tangent line in trigonometry; first derivative in the calculus)
useful for finding trends or correlations
multiple lines on one graph possible when scaled the same (shows comparisons)
used when there are distinct categories (spaces between the bars)
used when D.V. is measured in whole numbers
can be horizontal, vertical, grouped, range, or stacked (composite) as shown below respectively
good for comparisons (discontinuous data sets)
show fraction of a whole (100%)
used to compare parts of something studied
data at one point in time (snapshot)
vertical bar chart that shows statistical distributions
used to compare measurements against specifications (values)
not used to show varying time (snapshot)
no gaps between bars (continuous data sets)
bars have two important qualities: height & width (width is the same for all bars)
individual data points are not shown
y-axis shows frequency of values for the interval (width)
use two sets of I.V. to find a third (D.V.)
n-axes (1 per variable used)
used in conjunction with a straightedge (ruler)
used when approximate answer is useful (slide-rule)
limited by precision of worst measurements
used to show change in frequency & distribution of species (ecology)
used along an environmental gradient (x-axis)
number or biomass on y-axis
drawn symmetrically above and below 0 line
one kite for each species
used to show how two numerical variables relate (correlation)
each point represents one relationship
slope of "line of best fit" represents variables' correlation
Range: -1 < r < 1
weak --> strong
negative --> positive
none; no correlation
used to (dis)prove cause-and-effect relationships
does not prove one variable causes the other (causality)