Download as rtf, pdf, or txt
Download as rtf, pdf, or txt
You are on page 1of 14

Resistant indicators &

percentiles
1. Exploratory data
analysis
•0exploratory data analysis
used prior to inferential
statistics
•1alternatives to traditional
measures of central
tendency and variability
•2strengths of more
traditional methods can
also become limitations
•3 uses every score in
distribution
•4 but extreme values can
pull the measure
1. Exploratory data
analysis
•5 resistant indicators,
however, change
relatively little if portion
of data replaced by new
numbers that may be
very different from
originals
•6 less vulnerable to
atypical scores
1. Exploratory data
analysis
•7 1.1 Resistant
indicators of central
tendency
•8 median is more
resistant indicator of
central tendency than
the mean
•9not responsive to
every score in
distribution

1.1 Resistant indicators


of central tendency
•10 Why
does the median not
change?
•11consider the conceptual
definition
•12 finding the median
•13do not get hung up on the
formula
•14 show you a more
generic one later
•15odd or even number of
scores
•16 easy with odd as long
no repeats
•17 my example of 17
scores
1.2 Resistant indicators
of variability
•18 vulnerability
of variance
and range to extreme
scores
•19 fourths or hinges
•20 how to calculate them
•21 follow steps in text
•22will not have to do it for
exam, but will need to do for
homework
1.2 Resistant indicators
of variability
•23 two fourths or hinges
•24 upper and lower fourths
•25one above and one below
the median
•26 fourth spread
•27difference between the
fourths
•28 75th – 25th percentile
•29 not really true, but close
enough
1.2 Resistant indicators
of variability
•30 outliers [beyond 1.5 ×
fourth spread]
•31below and above the
fourths
•32 extreme scores
[highest score within
1.5 × fourth spread]
•33 graphic representation

•34 boxplot
1.2 Resistant indicators
of variability

1.2 Resistant indicators


of variability
•35 interpretation of the
boxplot
•36 location of median
•37 symmetry or not
•38length of whiskers [line]
compared to box
•39 box is short & line is
long [more leptokurtic]
•40 box is long vs. short
line [more platykurtic]
•41 SPSS /R
•42 ability to interpret is key
Applied stats
1.2 Resistant indicators
of variability
•43 Why are these
measures resistant
indicators?
based on median
•44
which represents central
half of distribution
median not influenced
•45
by deviant scores
•46identifies those more
deviant scores through
use of outliers and
extreme scores
2. Percentile point
•47 definition of median
•48 Mdnnot a score, but a
point on the scale
•49 common sense approach
•50 formula will write itself
•51 no need to “memorize”
•52 cumulativefrequency
distribution very helpful
•53 What
is 50% of 15
scores?
•54 7.5 scores

2. Percentile point
•55 medianis special type of
percentile
•56 the 50th percentile
•57 calculate any other
percentile
•58 12th percentile, 94th
percentile
th
•59 try
finding 80 percentile
on your own

th
L6-q3. The 80 percentile is
X f cf
29 1 15
28 5 14
27 4 9
26 1 5
25 2 4
24 1 2
23 1 1
The 80th percentile is
X f cf
29 1 15
28 5 14
27 4 9
26 1 5
25 2 4
24 1 2
23 1 1
2. Percentile point
text formula for
•60
percentile point
2. Percentile point
•61 P = percentile rank as decimal
fraction
•62 N = number of scores in
distribution
•63 L = lower real limit of score value
containing the required percentile
point
•64 nb = # of cases below L
•65 nw = # of cases falling within
score value containing the required
percentile point
•66 i = size of score interval

2. Percentile point
•67 grouped data, same basic
approach
•68 McCall

•69not all texts or computer


programs use the
interpolation method
•70 assumption

•71values are evenly


distributed within the
category
•72 may not be true but…

You might also like