M. MARIA ANTONY RAJ PhD Scholar (UGC-Junior Research Fellow) Department of Social Work Pondicherry University Email: antony.buddha@gmail.com Mobile: 9003348980 Scientific method and Logic of reasoning Induction Deduction Quantitative research and Qualitative research Steps in Research - Research Design - Sampling Techniques - Tool for data collection Methodological terms Population Collection of all individuals or objects or items under study and denoted by N Sample - A part of a population and denoted by n Variable - Characteristic of an individual or object. - Independent and Dependent variables - Qualitative and Quantitative variables Parameter Characteristic of the population Statistic Characteristic of the sample Hypothesis A tentative statement which can be put to test Karl Popper Conditions: 1. Minimum two variables 2. Variables should be definable and measurable 3. Already existing relationship between variables 4. Relationship should be tested through empirically and statistically Types of Hypothesis 1. Alternative Hypothesis a. One tail b. Two tail 2. Null Hypothesis Steps in Testing a Hypothesis 1. State the Null hypothesis 2. Select an Appropriate Test Statistic 3. Fix the Level of Significance () 4. Compute the Test statistic 5. Make Decision regarding Acceptance or Rejection of Null Hypothesis and in turn Research Hypothesis Rule of Acceptance or Rejection of Null Hypothesis P value < 0.05, Reject Null Hypothesis P value > 0.05, Accept Null Hypothesis Level of measurement The concepts are operationally defined to measure them. 1. Nominal - classified into different mutually exclusive categories Ex: Sex, religion 2. Ordinal - lowerness or higherness than other response-Rank order Ex: Education 3. Interval - values for the variable on an equally spaced continuum, -difference between 8 and 9 is the same as the difference between 76 and 77. Ex: IQ, level of awareness 4. Ratio - There is a rationale zero point for the scale, Ratios are equivalent, Ex: the ratio of 2 to 1 is the same as the ratio of 8 to 4. Permissible Arithmetic Operations Nominal Ordinal Interval Ratio Counting Greater than or less than operations. Addition and subtraction of scale values. Multiplicati on and division of scale values. Exercise: Identify the level of measurement for the following variables? 1. Age in Completed Years - 35, 54, 25, 5, 17, etc. 2. Sex - Male , Female 3. Socio-Economic Status - Low, Medium, High 4. Age - Young, Adult, Old 5. Time in Seconds - 5, 43, 52, 14, etc. 6. Size of Village - 400, 500, 431, 651, etc. 7. Religion-Hindus, Muslims, Christians, Jains, Sikhs etc. 8. Marks Obtained by Students - 35% , 48%, 78%, 25%, etc. 9. Days of a Week - Sunday, Tuesday, Saturday, etc. 10 Place of Birth - Rural, Urban, Semi Rural 11. State of Birth-Maharashtra, Bihar,U.P.,Gujarat, etc. 12. Field of Expertise - Science, Commerce, Arts, Engineering, Medicine, etc. 13. Position Held in a Organization-Manager, Supervisor, Driver, Peon, etc. 13 Systolic Blood Pressure - 130, 160, 110, 120, etc. 14. Intelligence Quotients - 65, 75, 89, etc. 15. Education level-Illiterate, Primary, Middle, HSC, etc. 16. Occupation-Agriculture, Labour, Service, Housework, Student, Aged, etc. Selecting appropriate statistical test Independent Dependent Parametric test Non Parametric Association Nominal/Ordinal Nominal/Ordinal Chi-square Relationship Interval/Ratio Interval/Ratio Correlation Karl Pearson r Spearman's Rank Correlation Significance Diff Nominal/Ordinal (2 category) Interval/Ratio Independent t test Mann-Whitney U Test or Wilcoxon Test Nominal/Ordinal (more than 2 cate) Interval/Ratio One way ANOVA Kruskal - Wallis Cause and Effect Interval/Ratio Interval/Ratio Regression Conditions for using Parametric Test 1. At least one variable should be interval/Ratio 2. Dependent variable should be normally distributed- Normality 3. Probability sampling Independent samples selected randomly Conditions for using Non Parametric Test 1. Samples compiled from different population 2. Nominal level of measurement 3. Very small samples (as small as 6 or 7) Testing Normality of the data Normality can be checked only for interval and ration scale If the data i.e., interval or ration variable under test is normally distributed then use Parametric tests Descriptive methods If the Mean and median are approximately same then the variable is normally distributed If Skewness and Kurtosis are approximately zero then the variable is normally distributed Diagram methods Histogram (the normality curve will be in positioned horizontally) Stem and leaf (the normality curve will be in positioned vertically) Box plot Test of Normality Kolmogorov Smirnov Test Note: steam & leaf and normal probability curve in diagram method and Kolmogorov Smirnov test in test of normality is strongly recommended to check the normality of variables. Reject Ho (significant level < 0.05) Absence of Normality Accept Ho (significant level > 0.05) Presence of Normality Types of Analysis 1. Descriptive 2. Associational 3. Inferential A. Summarisation of data 1. Frequency distribution 2. Graphical display 3. Measures of location (central tendency) 4. Measures of Dispersion 5. Measures of Skewness & Kurtosis B. Associational statistics 1. Measures of Association 2. Measures of correlation 3. Measures of Partial and Multiple Correlation(s) C. Tests of Hypothesis t-test, ANOVA, Chi Square, Proportion test, Significance of Correlation(s), Regression coefficients D. Multivariate Analysis Regression analysis, Logistic Regression, Factor, Cluster, Discriminant analysis Training on SPSS 1. Data Entry 2. Data Cleaning 3. Data Analysis a. Univariate Analysis b. Bivariate Analysis c. Multivariate Analysis References 1. Discovering Statistics Using SPSS by Andy Field THANK YOU