To get started, type one of these: helpwin, helpdesk, or demo.
  For product information, visit www.mathworks.com.
 
Bootstrap of variance values
Sample size:
Basic sample:

xx =

   -4.3256  -16.6558    1.2533    2.8768  -11.4647   11.9092   11.8916


n =

     7

Mean, sample standard deviation, and sample variance of sample:
 
   -0.6450   10.9373  119.6235

Basic sample sorted:

xxsorted =

  -16.6558  -11.4647   -4.3256    1.2533    2.8768   11.8916   11.9092

Five bootstrap resamples:
   11.8916  -16.6558  -11.4647    2.8768   11.8916   11.9092    2.8768
   -4.3256   11.9092    2.8768  -11.4647   11.9092   11.8916   11.9092
  -16.6558    1.2533   11.8916   11.8916    1.2533   11.8916   -4.3256
    1.2533   11.9092   -4.3256   -4.3256  -16.6558  -16.6558  -11.4647
  -16.6558  -16.6558   -4.3256   11.9092    2.8768   11.8916    2.8768

Sample variances of five bootstrap samples:
(The second column is the first column sorted.)
  137.4487   92.2090
   92.2090  105.9138
  113.5400  113.5400
  105.9138  137.4487
  144.0004  144.0004

 
Finding bootstrap CI for true variance:
Number of bootstrap replications (resamples):

nboot =

       10000

Finding sample variances of 10000 bootstrap replications:
Mean and Median of bootstrapped sample variances:
 
  102.9854  102.0089

Bootstrap median bias (proportion <= Observed value):
 
  6749/10000 = 0.6749  (Observed Variance: 119.624)

Smallest_7_of_Nboot =

    0.6275    0.6275    5.2542    5.9365    6.4933    6.4933    6.4933


Largest_7_of_Nboot =

  232.9883  232.9883  233.0359  233.0598  233.0598  233.1312  233.1312

Bootstrap 95% (percentile) CI bracketing Median:
 
   28.6504  102.0089  183.9559

This suggests that there might be a downwards bias in the CI.
HOWEVER zbmed is close to the original true variance.