CHAPTER 7: DATA PROCESSING AND STATISTICAL TREATMENT

Data processing involves:

• Input – it involves the responses from the research instrument by the subjects of the study.
• Throughput – it includes statistical procedures and techniques.
• Output – the results of the study which are presented in data matrix form.

Data processing
It consists of three basic steps:

• Categorization of data

It refers to the grouping of subjects under study according to the objectives or purposes of the study.
Five rules in categorizing research information by Kerlinger:
1.      Categories are set up according to the research problem.
2.      The categories are exhaustive.
3.      Each category is derived from one classification principle.
4.      The categories are mutually exclusive and independent.
5.      Any categorization scheme must be one level of discourse.
Coding of data – Information from the questionnaire, tests, interview schedules, rating scale and many others must be transformed into coded items to facilitate tabulation of data.
Tabulation of data – this is done by tallying and counting the raw data to arrive at a frequency distribution and to facilitate in organizing them in a systematic order in a table or several tables.
 Data matrix
-          Presentation of data in tabular form
3 Types of Data Matrices

• Univariate matrix – involves only one variable.
• Bivariate matrix – involves two variables.
• Multivariate matrix – has three or more variables in the table.

Dummy tables
-          Are helpful in preparing for the data matrix because they are used in planning, summarizing, organizing and analyzing the data on how the different variables differ with each other.
Statistical Treatment
-          It is a must that researchers diagnose the problem by using the appropriate statistical tool to arrive at accurate and definite interpretation of results.
Incorrect Statistical Tool
-          Percentage is incorrect or inappropriate statistical tool to scale options due to vague interpretation of results.
Univariate Statistical Treatment
-          The appropriate statistical tool for univariate problem is the weighted arithmetic mean and the like.

Bivariate Statistical Treatment in Experimental Research
-          The statistical tools for bivariate problem in experimental research are t-test and linear correlation.
Bivariate Statistical Treatment in Descriptive Research
-          The statistical tool used in bivariate descriptive research problems are z-test and linear correlation.
Z-test as Bivariate Statistical Tool in Descriptive Research
-          Z-test between percentages. The z-test is used to determine the significant difference between two percentages of related individuals in which the data are collected through survey.


Z =        P1 – P2

             PQ[1/N1 + 1/N2]
Multivariate Statistical Treatment
-          Two statistical tools used in multivariate experimental research problems with three or more variables are F-test or ANOVA, Kruskal-Wallis One-Way ANOVA and Friedman’s Two-way ANOVA.
F-test as Statistical Tool in Multivariate Experimental Research
-          F-test or two-way analysis of variance (ANOVA) involves three or more independent variables as bases for classification.
Friedman’s Two-Way ANOVA as Statistical Tool for Multivariate Experimental Research
-          Friedman’s two-way analysis of variance (ANOVA) is also a statistical tool used both in experimental and descriptive research problems. The formula is as follows:

Xr2 =        l2        ∑(ri2) – 3N(le + 1)
                                                                         Nk (k + 1)
Kruskal-Wallis One-Way ANOVA as Statistical Tool for Multivariate Experimental Research
Kruskal-Wallis one-way analysis of variance (ANOVA) by ranks is another statistical tool used in multivariate research problems both in experimental and descriptive researches. The formula is as follows:


H =     12       ∑[Ri2/n] – 3(N + 1)
                                                           N(N + 1)






Chi-Square (X2) as Statistical Tool for Multivariate Descriptive Research
Chi-Square (X2) 2 X 2 table. In chi-square (X2) 2 x 2 table or fourfold table, two discrete variables are involved to test if these variables are independent from each other.
Chi-Square (X2) 3 x 3 Table. It is also called nine-fold which involves three discrete variables to test if these variables are independent from each other.
Friedman’s Two-Way ANOVA by Ranks as Statistical Tool used in Multivariate Descriptive Research
-          It is used when the data from k related samples consist of at least an ordinal scale and have been drawn from the same set of observations to different population.
Kruskal-Wallis’ One-Way ANOVA (H) by Ranks a Statistical Tool in Multivariate Descriptive Research (Tied Observations)
The formula used is:

1 - ∑T
        N2 - N