ONE-SAMPLE RUNS TEST FOR RANDOMNESS
ONE-SAMPLE RUNS TEST FOR RANDOMNESS
LEARNING OUTCOMEAfter study this topic, student will
be able to decide whether a sequence of events, items, or symbols is the result of a random process.
HOW DOES ONE KNOW IF THE DATA OBTAINED FROM A SAMPLE ARE TRULY RANDOM?
Procedures for investigating randomness are based on the number and nature of the runs present in the data of interest.
A run is a sequence of like events, items or symbols that is preceded and followed by an event, item or symbol of a different type, or by none at all.
We doubt the randomness of a series when there appear to be either too many or too few runs.
We can use runs test to determine the randomness.
It helps us to decide whether a sequence of events, items, or symbols is the result of a random process.
IS THIS RANDOM?
A researcher interviewing 10 people for a survey. Let the genders be denoted by M for male and F for female. Suppose the participants were chosen as follows :
Situation 1 :M M M M M F F F F F F
Situation 2 :F M F M F M F M F M
Situation 3 :F F F M M F M M F F
ASSUMPTIONS
The data available for analysis consist of a sequence of observations, recorded in the order of occurrence, which we can categorize into two mutually exclusive types.
We let n = the total sample sizen1 = the number of observation of one type
n2 = the number of observation of the other type
HYPOTHESES
A. TWO-SIDED H0 : The pattern of occurrences of the two types of
observation is determined by a random process. H1 : The pattern of occurrences is not random
B. ONE-SIDED H0 : The pattern of occurrences of the two types of
observation is determined by a random process. H1 : The pattern is not random (because there are too
few runs to be attributed to chances) C. ONE-SIDED
H0 : The pattern of occurrences of the two types of
observation is determined by a random process. H1 : The pattern is not random(because there are too
many runs to be attributed to chances)
TEST STATISTIC AND DECISION RULE
The test statistic is r , the total number of runs.
A. TWO-SIDED B. ONE-SIDED ( H1 has few
runs )
C. ONE-SIDED ( H1 has many
runs)
Reject H0 :r ≤ lower critical
value OR
r ≥ upper critical value
Reject H0 :
r ≤ lower critical value
Reject H0 :
r ≥ upper critical value
CRITICAL VALUE
Lower critical value can be determine from Table A.5 (Wayne W. Daniel, Applied nonparametric Statistics) with n1 and n2
Upper critical value can be determine from Table A.6 (Wayne W. Daniel, Applied nonparametric Statistics) with n1 and n2
OR Both Lower and Upper critical value can be
determine from Table M (Bluman, Elementary Statistics) with n1 and n2
EXAMPLE 1
On a commuter train, the conductor wishes to see whether the passengers enter the train at random. He observes the first 25 people, with the following sequence of males(M) and females(F).
F F F M M F F F F M F M M M F F F F M M F F F M M
Test for randomness at α = 0.05
SOLUTION
Step 1 : State the hypotheses and identify the claimH0 : The pattern of occurrences of males and females enter the train is determine by a random process.(claim)H1 : The pattern of occurrences of males and females enter the train is not random
Step 2 : Find the numbers of runs (test statistic).FFF MM FFFF M F MMM FFFF MM FFF MMTest statistic, r = 10 n1 = the numbers of female = 15
n2 = the numbers of male = 10
SOLUTION (CONT…)
Step 3 : Find the critical valuen1 = 15, n2 = 10
Lower critical value = 7.Upper critical value =18.
Step 4 : Make the decisionSince r = 10 which is between 7 and 18, do not reject H0.
Step 5 : Make a conclusionThere is not enough evidence to reject the claim that the pattern of occurrences of males and females enter the train is determined by a random process.
Table
EXAMPLE 2
20 people enrolled in a drug abuse program. Test the claim that the ages of the people, according to the order in which they enroll, occur at random, at α = 0.05. The data as follows :
18, 36, 19, 22, 25, 44, 23, 27, 27, 35, 19, 43, 37, 32, 28, 43, 46, 19, 20, 22
SOLUTION
Step 1 : State the hypotheses and identify the claimH0 : The pattern of occurrences of ages of the people enrolled in a drug abuse program is determined by a random process. (claim)H1 : The pattern of occurrences of ages of the people enrolled in a drug abuse program is not random.
SOLUTION (CONT…)
STEP 2 : Find the number of runs (test statistic)Find the median of the data. First, arrange the data in ascending order.
18 19 19 19 20 22 22 23 25 27 27 28 32 35 36 37 43 43 44 46
Median is (27+27)/2 = 27. Compare the original data with the median.
Then, replace each number in the original sequence with an A if it is above the median and with B if it is below the median. Eliminate any numbers that are equal to the median.8 36 19 22 25 44 23 27 27 35 19 43 37 32 28 43 46 19 20 22
B A B B B A B A B A A A A A A B B B
Test statistic, r = 9 , n1 = # of B = 9, n2 = # of A = 9
SOLUTION (CONT…)
Step 3 : Find the critical valuen1 = 9, n2 = 9
Lower critical value = 5.Upper critical value =15.
Step 4 : Make the decisionSince r = 9 which is between 5 and 15, do not reject H0.
Step 5 : Make a conclusionThere is not enough evidence to reject the claim that the pattern of occurrences of ages of the people enrolled in a program is determined by a random process.
Table
EXAMPLE 3 (PG65)
Table in next page shows the departures from normal of daily temperatures recorded at Atlanta, Georgia, during November 1974*. We wish to know whether we may conclude that the pattern of departures above and below normal is the result of a nonrandom process.Departures from normal of daily temperatures recorded at Atlanta, Georgia, during November, 1974.
*Local Climatological Data, U.S. Department of Commerce, National Oceanci and Atmospheric Administration, Environmental Data Service, National Climatic Center, Federal Buliding, Asheville, Notrh Caronlina, November 1974.
DATA
Day 1 2 3 4 5 6 7 8 9 10
11
12
13
14
15
Departure from normal
12
13
12
11
5 2 -1 2 -1 3 2 -6 -7 -7 -12
Day 16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Departure from normal
-9 6 7 10
6 1 1 3 7 -2 -6 -6 -5 -2 -1
SOLUTION
Step 1 : State the hypotheses and identify the claimH0 : The pattern of occurrences of negative and positive departures from normal is determined by a random process.H1 : The pattern of occurrences of negative and positive departures from normal is not random.( claim )
Step 2 : Find the numbers of runs (test statistic).Replace the positive departures with A and the negative departures with B. So, the new data as follows :AAAAAA B A B AA BBBBB AAAAAAAA BBBBBBTest statistic, r = 8, n1 = # of A = 17, n2 = # of B =13
SOLUTION (CONT…)
Step 3 : Find the critical valuen1 = 17, n2 = 13
Lower critical value = 10.Upper critical value = 22.
Step 4 : Make the decisionSince r = 8 is less than lower critical value, reject H0.
Step 5 : Make a conclusionThere is enough evidence to support the claim that the pattern of occurrences of positive and negative departures from normal is not random.
Table
LARGE SAMPLE APPROXIMATION
Use when either n1 or n2 is greater than 20. Compare the computed z with Table A.2.
1
22
12
212
21
212121
21
21
nnnn
nnnnnn
nnnn
r
z
EXAMPLE 4
On a commuter train, the conductor wishes to see whether the passengers enter the train at random. He observes the first 30 people, with the following sequence of males (M) and females (F).
FFFF MM FFFF M F MM FFFFFF MM FFFFFF MM
Test for randomness at α = 0.05
SOLUTION
Step 1 : State the hypotheses and identify the claimH0 : The pattern of occurrences of males and females enter the train is determine by a random process.(claim)H1 : The pattern of occurrences of males and females enter the train is not random
SOLUTION (CONT…)
Step 2 : Find the numbers of runs and test value. r = 10 , n1 = 21, n2 = 9
6036.1
1921921
921)9)(21(2)9)(21(2
1921)9)(21(2
10
1
22
12
2
212
21
212121
21
21
z
z
nnnn
nnnnnn
nnnn
r
z
SOLUTION (CONT…) Step 3 : Find the critical value
α=0.05, z = 1.96, CV = ±1.96
-1.6036
SOLUTION (CONT…)
Step 4 : Make the decisionSince -1.6036 fall in the non-critical zone, do not reject H0.
Step 5 : Make a conclusionThere is not enough evidence to reject the claim that the pattern of occurrences of males and females is determine by a random process.
Example 1
Example 2
Example 3
Q2
Q1
Q3