To get started, type one of these: helpwin, helpdesk, or demo.
  For product information, visit www.mathworks.com.
 
Data in Table 3.1 page 39
EXPANDED DATA MATRIX:
Columns are Patient number, Before (X_i), After (Y_i),
  Z_i=Y_i-X_i, and 1 if Z_i>0, 0 if Z_i<0

dd =

    1.0000    1.8300    0.8780   -0.9520         0
    2.0000    0.5000    0.6470    0.1470    1.0000
    3.0000    1.6200    0.5980   -1.0220         0
    4.0000    2.4800    2.0500   -0.4300         0
    5.0000    1.6800    1.0600   -0.6200         0
    6.0000    1.8800    1.2900   -0.5900         0
    7.0000    1.5500    1.0600   -0.4900         0
    8.0000    3.0600    3.1400    0.0800    1.0000
    9.0000    1.3000    1.2900   -0.0100         0

SIGN-TEST ANALYSIS:
  
N=Sample Size, B=Sign Test Score, P=2-sided binomial P-value:

N_B_Pval =

    9.0000    2.0000    0.1797

Large-sample normal-approximation test for binomial probability:

Zval_Pval =

   -1.6667    0.0956

Note that the large-sample approximation for the sign test
  is not very accurate.
  
The HODGES-LEHMANN estimator for theta for After=Before+theta
  for the sign test is the median of the Z_i values:
Zmean is the mean of the Z=Y-X values.

Zmean_BMed =

   -0.4319   -0.4900

  
WILCOXON SIGNED RANK TEST:
  
Drop X and Y columns, add Abs(Z)
Columns are Ordinal, Z, Sgn(Z), and Abs(Z)

ddx =

    1.0000   -0.9520         0    0.9520
    2.0000    0.1470    1.0000    0.1470
    3.0000   -1.0220         0    1.0220
    4.0000   -0.4300         0    0.4300
    5.0000   -0.6200         0    0.6200
    6.0000   -0.5900         0    0.5900
    7.0000   -0.4900         0    0.4900
    8.0000    0.0800    1.0000    0.0800
    9.0000   -0.0100         0    0.0100

SORT the matrix by its FOURTH column = Abs(Z):
Then ADD a FIFTH column that are the signed ranks

ddx =

    9.0000   -0.0100         0    0.0100    1.0000
    8.0000    0.0800    1.0000    0.0800    2.0000
    2.0000    0.1470    1.0000    0.1470    3.0000
    4.0000   -0.4300         0    0.4300    4.0000
    7.0000   -0.4900         0    0.4900    5.0000
    6.0000   -0.5900         0    0.5900    6.0000
    5.0000   -0.6200         0    0.6200    7.0000
    1.0000   -0.9520         0    0.9520    8.0000
    3.0000   -1.0220         0    1.0220    9.0000

Data in ORIGINAL ORDER (by sorting on Pat.Number):
Columns are Pat.Number, Z, Sgn, Abs(Z), and Rank

ddx =

    1.0000   -0.9520         0    0.9520    8.0000
    2.0000    0.1470    1.0000    0.1470    3.0000
    3.0000   -1.0220         0    1.0220    9.0000
    4.0000   -0.4300         0    0.4300    4.0000
    5.0000   -0.6200         0    0.6200    7.0000
    6.0000   -0.5900         0    0.5900    6.0000
    7.0000   -0.4900         0    0.4900    5.0000
    8.0000    0.0800    1.0000    0.0800    2.0000
    9.0000   -0.0100         0    0.0100    1.0000


Tplus_Tmean_Tstd =

    5.0000   22.5000    8.4410

 
LARGE-SAMPLE NORMAL APPROXIMATION for two-sided Wilcoxon
  signed-rank P-value for H_0:theta=0:

Zval_Pval =

   -2.0732    0.0382

SIMULATING THE TRUE Wilcoxon Signed-Rank P-VALUE:
  
Using NSIMS random permutations:

Nsims =

       10000

INITIALIZING the random-number generator at

Wseed =

  2.1013e+005

FINISHED NSIMS simulations
 
SIMULATED value for true 2-sided P-value:
NLessEq=193  Nsims=10000   Pvalue=0.03860 (2-sided)
 
95%% Confidence Interval for true P-value (with Estimate in middle):

Plow_PEst_Phigh =

    0.0332    0.0386    0.0440

The HODGES-LEHMANN estimator for theta for After=Before+theta
  for the Wilcoxon signed-rank test is the median of
  the Walsh averages W_k=(Z_i+Z_j)/2:

Walsh averages (not in order; n=9, nwa=45, nrows=5):
 -0.952 -0.403 -0.987 -0.691 -0.786 -0.771 -0.721 -0.436 -0.481  0.147
 -0.438 -0.142 -0.236 -0.221 -0.171  0.114  0.069 -1.022 -0.726 -0.821
 -0.806 -0.756 -0.471 -0.516 -0.430 -0.525 -0.510 -0.460 -0.175 -0.220
 -0.620 -0.605 -0.555 -0.270 -0.315 -0.590 -0.540 -0.255 -0.300 -0.490
 -0.205 -0.250  0.080  0.035 -0.010

SORTED Walsh averages (n=9, nwa=45, nrows=5):
 ( 1) -1.022 -0.987 -0.952 -0.821 -0.806 -0.786 -0.771 -0.756 -0.726 -0.721
 (11) -0.691 -0.620 -0.605 -0.590 -0.555 -0.540 -0.525 -0.516 -0.510 -0.490
 (21) -0.481 -0.471 -0.460 -0.438 -0.436 -0.430 -0.403 -0.315 -0.300 -0.270
 (31) -0.255 -0.250 -0.236 -0.221 -0.220 -0.205 -0.175 -0.171 -0.142 -0.010
 (41)  0.035  0.069  0.080  0.114  0.147

Zmean_MedZ_MedWav =

   -0.4319   -0.4900   -0.4600

EXACT CONFIDENCE INTERVALS based on the Wilcoxon signed-rank statistic:

Estimator: -0.46

Table A4 for n=9  (n*(n+1)/2=45)  has:
   P(T+ ge 39)=0.027
   P(T+ ge 40)=0.020
   P(T+ ge 41)=0.014
   P(T+ ge 42)=0.010
   P(T+ ge 43)=0.006
 
Exact symmetric confidence intervals for theta:
  (-0.771, -0.142)  x=39  94.6% CI
  (-0.786, -0.010)  x=40  96.0% CI
  (-0.806,  0.035)  x=41  97.2% CI
  (-0.821,  0.069)  x=42  98.0% CI
  (-0.952,  0.080)  x=43  98.8% CI