Research Methodology II: Research Design; Primary and Secondary Data — Observation, Questionnaire, Interview, Library and Internet; Sampling and Sampling Techniques; Scales of Measurement; Analysis of Data and SPSS; Citation Patterns, Research Ethics, Bibliography and Report Writing

Primary data are collected first-hand by the researcher for the study at hand; secondary data already exist, collected by others for other purposes — the Census, National Sample Survey and National Family Health Survey reports, government annual reports, CAG and parliamentary committee reports, court judgments, archives, and published research. Secondary data are cheap and often large-scale, but may not fit the research question exactly and must be checked for reliability, suitability and adequacy.

The population (universe) is the whole set of units the study is about; a sample is a subset selected to represent it; the sampling frame is the list from which the sample is drawn (an electoral roll, a list of gram panchayats); the sampling unit is the unit selected. A census studies every unit. Sampling saves time and money and can even be more accurate than a census, because resources can be concentrated on quality. Sampling error is the difference between a sample statistic and the population value that arises by chance and falls as sample size rises; non-sampling errors — frame errors, non-response, measurement and processing errors — arise in censuses too and do not fall with sample size.

S. S. Stevens (1946) distinguished four levels of measurement, each permitting more mathematical operations than the last. The level determines which statistics can properly be used, so misclassifying a variable produces wrong analysis.

Attitude scales used in social research include the Likert scale (summated ratings of agreement, usually five points, from strongly agree to strongly disagree), the Thurstone scale (equal-appearing intervals judged by experts), the Guttman scale (cumulative items, where agreeing with a stronger item implies agreeing with weaker ones), Bogardus's social distance scale, and Osgood's semantic differential (bipolar adjectives such as efficient–inefficient). Analysis of data runs from editing and coding to tabulation, then descriptive statistics (frequencies, measures of central tendency and dispersion, cross-tabulation) and inferential statistics (tests of significance — chi-square for nominal data, t-tests and ANOVA for comparing means, correlation and regression for relationships). For a chi-square test on a contingency table, the degrees of freedom are (rows − 1) × (columns − 1). SPSS — originally the Statistical Package for the Social Sciences, developed in 1968 by Norman Nie, C. Hadlai Hull and Dale Bent and acquired by IBM in 2009 — is the most widely used software for such analysis, alongside Stata, R and, for qualitative data, packages such as NVivo and ATLAS.ti.

Citation acknowledges the sources of ideas, data and words, allows readers to verify them, and places the work in its scholarly conversation. The major styles follow two patterns: author–date (in-text parenthetical citations such as "(Riggs, 1964, p. 22)" with a reference list — the APA style, now in its seventh edition, and Chicago author–date) and notes–bibliography (footnotes or endnotes with a bibliography — Chicago notes style, common in history), with MLA (author–page) prevalent in the humanities. Citation analysis measures scholarly influence: Eugene Garfield's Science Citation Index and the journal impact factor, and Jorge Hirsch's h-index (2005) — a researcher has index h if h of his or her papers have each been cited at least h times.

The ethics of research protect participants and the integrity of knowledge. The core principles, set out in the US Belmont Report (1979), are respect for persons (informed, voluntary consent, and protection of those with diminished autonomy), beneficence (maximise benefits, minimise harm) and justice (fair distribution of the burdens and benefits of research). In practice they require informed consent, confidentiality and anonymity, avoidance of deception and harm, special care with vulnerable groups, review by an ethics committee, honest reporting without fabrication or falsification of data, and the avoidance of plagiarism — presenting another's work as one's own. In India the UGC (Promotion of Academic Integrity and Prevention of Plagiarism in Higher Educational Institutions) Regulations, 2018 grade plagiarism by the similarity found: up to 10 per cent is not penalised; Level 1 (above 10 up to 40 per cent), Level 2 (above 40 up to 60 per cent) and Level 3 (above 60 per cent) attract increasing penalties.

A bibliography lists all the sources consulted, cited or not, usually alphabetically by author; a reference list includes only works actually cited in the text; an annotated bibliography adds a short description or evaluation of each source. A research report conventionally has three parts. The preliminary pages: title page, declaration and certificate, acknowledgements, table of contents, lists of tables and figures, and an abstract. The main text: introduction (problem, objectives, significance), review of literature, research design and methodology, presentation and analysis of data, discussion of findings, and conclusions with recommendations and the study's limitations. The end matter: references or bibliography, appendices (questionnaires, schedules, additional tables) and index. Good report writing is clear, objective, consistent in its citation style, and honest about limitations.