Scribd
Upload
111 views24 pages

Summary - Non - Parametric Tests

The document provides information on several statistical hypothesis testing procedures including: - Z-test, t-test, and Wilcoxon signed-rank test for comparing means of single samples - T-test and Wilcoxon rank-sum test for comparing means of two independent samples - ANOVA and Kruskal-Wallis test for comparing means of more than two independent samples - Chi-square test for examining relationships between categorical variables in contingency tables Examples of hypotheses tests are given for comparing analgesic drug relief, ethnic group characteristics, and the relationship between bicycle helmet use and head injuries. Formulas and step-by-step procedures are outlined for conducting Wilcoxon signed-rank, Wilcoxon rank

Uploaded by

Survey_easy
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
111 views24 pages

Summary - Non - Parametric Tests

The document provides information on several statistical hypothesis testing procedures including: - Z-test, t-test, and Wilcoxon signed-rank test for comparing means of single samples - T-test and Wilcoxon rank-sum test for comparing means of two independent samples - ANOVA and Kruskal-Wallis test for comparing means of more than two independent samples - Chi-square test for examining relationships between categorical variables in contingency tables Examples of hypotheses tests are given for comparing analgesic drug relief, ethnic group characteristics, and the relationship between bicycle helmet use and head injuries. Formulas and step-by-step procedures are outlined for conducting Wilcoxon signed-rank, Wilcoxon rank

Uploaded by

Survey_easy
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 24

Mas S Mohktar

Email: [email protected]
Phone (office): 0379677681
One sample hypothesis testing procedure
 Normal distribution, known mean μ and standard deviation σ (big
number of sample)
 Z-Test

 Assumed normal distribution, unknown mean μ and standard


deviation σ (n between 10-30)
ҧ 0
𝑥−𝜇
 T-test 𝑡=
𝑠Τ 𝑛

 Not normally distributed (normally, n<10)


 Non-parametric
 Wilcoxon signed-rank
 A pre-post study
 Step 1
 State the null hypothesis - in this case it is that the median
difference, M, is equal to zero.
 H0: Median D=0

 Step 2
 Calculate each paired difference |D|.

 Step 3
 Rank the |D|, ignoring the signs (i.e. assign rank 1 to the smallest
|D|, rank 2 to the next etc.)
 Step 4
 Label each rank with its sign, according to the sign of |D|
 Step 5
 Calculate W+, the sum of the ranks of the positive |D|,
 and W−, the sum of the ranks of the negative |D|

 Step 6
 Choose W = min(W−,W+)

 Step 7
 Use tables of critical values for the Wilcoxon signed rank sum test to
find the probability of observing a value of W or more extreme
 If w > w = reject H0
 The table below shows the hours of relief provided by two
analgesic drugs in 12 patients suffering from arthritis. Is there
any evidence that one drug provides longer relief than the
other?
Two sample hypothesis testing procedure
 Normal distribution, known mean μ and standard deviation σ (big number of
sample)
 In actuality, two sample z-tests are rarely used, because the estimate for the SE for
difference used here is biased. Instead, statisticians use a two-sample t-test.

 Assumed normal distribution, unknown mean μ and standard deviation σ (n


between 10-30)
 T-test – paired sample
𝑑ത
 𝑡=
𝑠𝑑 Τ 𝑛

 T-test two sample


𝑥ҧ 1 −𝑥ҧ2
 𝑡=
𝑠𝑥ഥ 1 −ഥ𝑥2

 Not normally distributed (normally, n<10)


 Non-parametric
 Wilcoxon (Mann-Whitney) two-sample test
The Wilcoxon Rank-Sum Test
In a genetic inheritance study discussed by Margolin [1988],
samples of individuals from several ethnic groups were taken.
Blood samples were collected from each individual and several
variables measured. We shall compare the groups labeled
“Native American” and “Caucasian” with respect to the variable
MSCE (mean sister chromatid exchange). The data are as
follows:

Native American: 8.50, 9.48, 8.65, 8.16, 8.83, 7.76, 8.63


Caucasian: 8.27, 8.20, 8.25, 8.14, 9.00, 8.10, 7.20, 8.32, 7.70
 Step 1: State the hypotheses
 H0: NA=C or NA>C
 H1: NA ≠ C or NA<C

 Step 2:
 Set the criterion
 One-tail test or two-tail test? α=?
 Step 3

 Step 4
 Calculate: wNA = sum of the ranks for observations from NA (the
smallest sample among the two)
 wNA = 3+6+11+12+13+14+16 = 75

 Step 5
 Compare with value in the table
 If wNA > wTable = reject H0
 More than two sample hypothesis testing procedure
 Assumed Normal distribution
 ANOVA
 F = MS(B) / MS(W)

 Not normally distributed (normally, n<10)


 Kruskal-Wallis
 Refer to the Kruskal Wallis handout by Graham Hole Research
Skills Kruskal-Wallis
 Consider a 2 × 2 table
 The data consist of a random sample of 793 individuals who are
involved in bicycle accidents during a specified one-year
period
 To examine the effectiveness of bicycle safety helmets, we wish
to know whether these is an association between the incidence
of head injury and the use of helmets among individuals who
have been involved in accidents
To determine this, we test
 H0: The proportion of persons suffering head injuries among
the population of individuals wearing safety helmets at the
accident is equal to the proportion of persons sustaining head
injuries among those not wearing helmets
versus
 HA: The proportions of persons suffering head injuries are not
identical in the two populations

at the α = 0.05 level of significance


 The first step in carrying out the test is to calculate the
expected count for each cell of the contingency table, given
that H0 is true as the following:
 In general, if a 2 × 2 table of observed cell counts for a sample of
size n as follows,
 the corresponding table of expected counts is
 These marginal totals have been held fixed by design; we
calculate the cell entries that would have been expected given
these is no association between the row and column
classifications and that the number of individuals within
each group remains constant
 The chi-square test compares the observed frequencies in each
category of the contingency table (represented by O) with the
expected frequencies in each category of the contingency
table (represented by E) given the null hypothesis is true
 It is used to determine whether the deviations between the
observed and the expected counts, O−E, are too large to be
attributed to chance
 To perform the test for the counts in a contingency table with r
rows and c columns, we calculate

𝑟𝑐
(𝑂𝑖 − 𝐸𝑖 )2 𝑎 2
χ2 =෍ ~ χ (𝑟−1)(𝑐−1)
𝐸𝑖
𝑖=1

 where rc is the number of cells in the table


 the test statistic χ2 can also be express as

𝑛
𝑛[ 𝑎𝑑 − 𝑏𝑐 − ]2
χ2 = 2
(𝑎 + 𝑏)(𝑏 + 𝑑)(𝑎 + 𝑐)(𝑐 + 𝑑)
 Table in Appendix is a condensed table of areas for the chi-
square distributions with various degrees of freedom
 For a particular value of df, the entry in the table is the outcome
of χ2 df that cuts off the specified area in the upper tail of the
distribution
 the df = (c-1)(r-1)

You might also like