STATISTICAL ANALYSIS

Statistical Analysis



Statistical Analysis

Introduction

A statistical test is used in mathematical statistics to present observations on the basis of a reasoned decision on the validity and invalidity of a hypothesis to be made. Formally, a test is thus a mathematical function. Since the existing data realizations of random variables are, in most cases cannot say with certainty whether a hypothesis is true or not. An attempt is therefore, the probability for failure to control decisions about what a test at a specified significance level corresponds to. For this reason we speak of a hypothesis test or test of significance. When analyzing data measured by a continuous quantitative variable, statistical tests and contrast estimation frequently used are based on the assumption that we have obtained a random sample from a probability distribution of normal or Gaussian type. But in many cases this assumption is not valid, and others suspect that is not appropriate is not easy to ascertain, because they are small samples. In these cases we have two possible mechanisms:

The data can be transformed in such a way as to follow a normal distribution.

Or you can resort to statistical tests are not based on any assumptions about the probability distribution from which data were obtained, and are therefore called non-parametric test (distribution free), while the evidence assume a given probability distribution for the data are called parametric tests.

Discussion

In statistics, a hypothesis test is a process of rejecting or not rejecting (rarely accept) a statistical hypothesis, called the null hypothesis, based on a data set (sample). These inferential statistics: from calculations on the observed data, we issue conclusions on the population, their associated risk of error. Tests can be classified by purpose, type and number of variables of interest, the existence of a priori assumptions about data distributions, the means for providing the samples.

Testing by purpose

The purpose defines the objective of the test the hypotheses that we want to oppose, the information you want to extract data.

The compliance test - is to compare a calculated parameter of the sample at a pre-established value. This is called conformance testing to a standard. The best known are probably testing on the mean or proportion.

The fit test - is to check the compatibility of data with a distribution chosen a priori. The test most commonly used for this purpose is the test of fitness for the normal distribution.

The homogeneity test - (or comparison) is to verify that K (K> = 2) samples (groups) come from the same population or, it means the same thing, that the distribution of the variable of interest is the even in the K samples.

The test of association - (or independence) is to test the existence of a link between two variables. The techniques used depend on whether the qualitative variables are nominal, ordinal or quantitative.

Research Proposal 1

The Effects of scarce health care resources on Quality health care for Rural Populations

The research question in this case tells us that we have to find out the current healthcare access ...