### SELECTING THE STATISTICS

As he plans the analysis of the data the researcher should consider two sections of the research report in which statistics will be relevant. The first of these is the section in which the data producing sample is described and in which may also be compared to the selected sample and to the population. In describing the sample, the basic descriptive statistics of the summary frequency distribution and the appropriate measures of central tendency and variability serve to provide the reader with some insight into the nature of the respondents. Researches are interested in the usual demographic characteristics such as gender, age, occupation and educational level, but in addition anyone project will suggest other descriptive variables about which data should be collected.
Provided the data are available, the researcher should also employ inferential statistics such as Chi-square or the t-test to determine whether or not his data producing sample differs from his selected sample or population by selecting which analysis he will prefer at his early stage, the researcher structures the kinds of data he will need to produce about the population and can incorporate the search for these data into his data gathering plan.
The second section of the report in which statistical procedures plays role is in the reporting of research results. The selection of these procedures should be well structured by this point if the researcher has stated specific hypotheses and research questions. The necessity to test the hypothesis provides guidance to statistical procedures at the general level, with the decision as to the level of data available providing the key to which specific procedures are to employ. Thus, hypothesis which refers to the expected a relationship between two variables, immediately indicates the need for a correctional analysis. Once the researcher decides that the two variables will yield ordinal data, for example, he can move directly to the specification of the rank order correlation.
The specification of statistical analysis at this stage of the research also enables the researcher to estimate his data analysis cost in both time and money and make whatever arrangements are necessary to reserve time on data-processing facilities.
The elementary and special statistical techniques of analsysis are as follows:
1.     Elementary Statistical Techniques of Analysis
Most commonly used statistical techniques of analysis data are:
1. Calculating frequency of distribution in percentages of items under study.
2. Testing data for normality of distribution Skewness Kurtosis and mode.
3. Calculating percentiles and percentile ranks.
4. Calculating measures of central tendency-Mean, Median and Mode and establishing Norms.
5. Calculating measures of dispersion-Standard deviation, Mean deviation, quartile deviation and range.
6. Calculating measures of relationship-Coefficients of Correlation, Reliability by the Rank difference and Product moment method.
7. Graphical presentation of data-Frequency polygon curve, Histogram, Cumulative frequency polygon and Ogive, etc.
While analysis their data investigator usually makes use of as many of the above simple statistical devices as necessary for the purpose for their study. There are some other complicated devices of statistical analysis listed below which researcher use in particular experimental or complex casual comparative studies and investigations.
2.     Special Statistical Techniques of Analysis
The following are the special statistical techniques of analysis:
1. Test of students ‘t’ and analysis of variance for testing significance of differences between statistics especially between Means.
2. Chi-square test for testing null hypothesis.
3. Calculation of Biserial ‘r’ and Tetrachoric ‘r’ for finding out relationship between different phenomena in complex situations.
4. Calculation of partial and multiple correlation and of Bivariate and Multivariate Regression Equations for findings out casual relationship between various phenomena involved in a situation.
5. Factorial Analysis for the purpose of analysing the composition of certain complex phenomena.
6. Analysis of co-variance for estimating the true effect of the treatment after adjusting the initial effect.