However, not all of the conditions in the measurement of psychological ideal for all techniques applied to the statistics, such as:
* Many psychological aspects that are difficult to dikuantifisir scale let alone the ratio measurement interval (difficult to say, for example, attitudes towards alcohol someone / a group of people two times higher than the person / other groups);
* Difficult to assume that certain psychological variables terdistribusi to normal.
Therefore, statistical methods can be more nonparametrik considered the needs and conditions in accordance with science piskologi.
But before discussing methods of statistical nonparametrik, first need to understand what it is about parametrik.
Parametrik
Parametrik the understanding of the parameters, that is an indicator of the distribution of measurement results. Indicators of the distribution based on statistical measurements parametrik used to be the parameters of normal distribution.
Normal distribution, or also known by the term Gaussian Distribution, the two parameters, the mean (m) and varians (S2). Parameters provides the unique characteristics of a distribution based on "location" to its (central tendency), and as a statistical method, also base its calculations on both these parameters.
From this brief explanation appears that the use of statistical methods parametrik follow the principles of normal distribution (see wikipedia for further explanation).
The consequences of this understanding, the application of statistics parametrik need to pay attention to the following.
* Distribution of a sample made of objects of measurement derived from the distribution of the population that can be assumed terdistribusi to normal.
* A random sample is obtained, with the amount that can be considered to represent the population.
* Distribution is a normal part of the continuous probability distribution, so the scale pengukurannya also be continuous (interval or ratio) or a nominal scale which changed the proportion to be (can only be processed using the chi-square).
When the terms of this all met, then the new method can be applied statistics parametrik.
Nonparametrik
Nonparametrik term itself was first used by Wolfowitz, 1942. The other term frequently used, among other distribution-free statistics and assumption-free test. The terms of this, easily seen that the statistical method nonparametrik is a statistical method that can be used to dismiss the assumption that all melandasi methods parametrik statistics, particularly related to the normal distribution.
When used statistical methods nonparametrik? From the foregoing, it can be said with a simple test method is used when one of the statistical parameters parametrik not required (see flow chart below).
Statistical techniques nonparametrik
There are many techniques in the statistical methods nonparametrik.
To determine the appropriate statistical techniques, can be done simply with the following questions:
* What is the purpose of testing? [describe, test the difference, correlation, etc.]
* When to test the differences, there is a group of samples will be tested? [one, two, or more than two]
* When to test the differences, whether groups are from one population of the same (or a pair can be made), or groups independent of each other?
* What pengukurannya scale? [Ordinal or nominal]
These questions are basic questions to determine nonparametrik statistical techniques appropriate to the needs and condition data. However, to levels that more information needs to the power of each technique, tailored to the condition that more specific data.
Fill in the table below are some statistical techniques nonparametrik commonly used, with statistical techniques to compare with parametrik.
