PSYC 3000 Lecture Notes  Lecture 3: Interquartile Range, Quartile, Box Plot
22 views1 pages
2 Aug 2016
School
Department
Course
Professor
PSYC 3000
Sept 15
Numeric tools
for normal distribution, we try and find the center of the data using the sample mean.
standard deviation is also used. It characterises dispersion.
dispersion means how observation are dispersed around the centre
often in research we don’t get normal distribution. Distribution are often skewed or oddly
shaped.
mean doesn’t work well in skewed or oddly shaped distribution.
in those cases the median is used
another solution is to use quartiles across the distribution. Separating the distribution by
quartiles allows us to establish landmarks and notice any outliers.
interquartile range is the difference between one quartile to another (q3q1)
Boxplots
a display of different landmarks on a distribution (quartiles)
line in the box is the median. It is not always in the middle of the box, it just needs to be inside
the box.
boxplots can be displayed vertically or horizontally. Most of the time it is shown vertically.
the size of the box is the interquartile range (q3q1)
width of the box does not matter, simply visual effect.
in the boxplot there is an invisible line. It’s called the upper fence
whisker (line on top of and bottom of box) always ends on a data point
the whisker does not go past fence and it stops at the last data point before the fence. If there are
data past the fence, they are outliers.
also an invisible line at the bottom called lower fence.
the fences do not show data point, the whiskers do.
if the median is close to one of the whiskers, it means the data is tightly clustered at one end and
very scattered in the other end.
boxplots can easily detect outliers. Anything beyond the top fence and bottom fence are
considered outliers.
stars= extreme outliers. These are often thrown out as data entry errors.
outliers only contain about 1% of the data.

find more resources at oneclass.com
find more resources at oneclass.com