Bootstrapping is also known as Resampling or Monte Carlo estimation. Bootstrapping and Resampling is used to establish the confidence intervals for any test statistics. It is not based on assumptions such as multivariate normal distribution. Instead it is based on the repeated samples from the researcher’s data.
Bootstrapping and Resampling is a nonparametric method of statistical inference. This means that Bootstrapping and Resampling does not use generic distribution tables (eg., normal distribution tables) to compute approximate p probability values. Bootstrapping and Resampling generates a unique sampling distribution based on the actual data at hand and uses experimental methods rather than the analytic ones. Bootstrapping and Resampling yields unbiased estimates, unlike approximation with generic distribution tables, because Bootstrapping and Resamplingis based on unbiased samples of all possible outcomes in the data being studied. In order to correlate Bootstrapping and Resamplingand Monte Carlo estimation, we should discuss them individually:
Bootstrapping uses repeated samples from the same original data sample to compute some test statistic. The distribution of this statistic is computed for a very large number of runs of the sampling process. It is used to estimate the variance of the statistic in the underlying population, thereby allowing the significance of the statistic to be estimated.
Resampling is the method used in bootstrapping to draw repeated samples from the original data sample. The general method is to select random cases with replacement from the original data sample, such that each drawn sample has the same number of cases as the original data sample. It is because of replacement that the drawn samples used by resampling methods include the repeated cases.
Monte Carlo estimation. Monte Carlo simulation methods can be used to obtain resampling results. In SPSS, the researcher enters the desired confidence level, the number of samples to be computed, and a random number seed. In order to duplicate Monte Carlo results, the same random seed is being entered each time.
There are certain assumptions made while conducting Bootstrapping and Resampling:
- Non-Parametric assumptions: Bootstrapping and Resampling is an experimental approach which ignores all parametric assumptions about the nature of the underlying data distribution. Thus, Bootstrapping and Resampling is based on non parametric assumptions.
- Sample size assumption: In Bootstrapping and Resampling, there is no specific sample size requirement. Therefore, the larger the sample, the more reliable the confidence intervals generated by Bootstrapping and Resampling.
In Bootstrapping and Resampling, there is an increased danger of overfitting to noise in the data. This problem can be solved by combining Bootstrapping and Resampling methods with cross-validation.
In SPSS, click “Analyze” from the menu, click “non-parametric tests” then select two independent sample tests and then click the “Exact” button. This button allows a choice of types of significance estimates. One of the choices is “Monte Carlo,” which is a Bootstrapping and Resampling method.
