Of all the stages in a Kenyan postgraduate research project, data analysis is the one that students most consistently underestimate — and most frequently get wrong. The reason is not usually a lack of intelligence or effort. It is a combination of two problems that are far more common than they should be: students do not understand what data analysis is actually supposed to accomplish in the context of their specific study, and they do not know which analytical tools and techniques are appropriate for the data they have collected and the questions they are trying to answer.
The consequence of both errors is the same: a Chapter 4 that presents numbers, tables, or themes without genuine analytical depth — a chapter that a supervisor or panel reviewer will return for revision, often with the frustrating comment that the student needs to “go beyond description” or “link findings to the research questions.” By that point, the student has already spent weeks collecting data, cleaning it, and running tests in SPSS or STATA. The rework is painful and entirely preventable.
This guide will not teach you everything there is to know about statistics or qualitative analysis — that would take a shelf of textbooks. What it will do is give you the analytical framework, the practical decision-making tools, and the critical awareness you need to approach your data with genuine expert confidence: to choose the right methods, execute them correctly, interpret the results with scholarly rigour, and present them in a way that passes your panel on the first review.
At Tobit Research Consulting, we have provided data analysis support to over 5,000 postgraduate students across Kenyan universities — from Kenyatta University, the University of Nairobi, JKUAT, and Mount Kenya University to Strathmore, Egerton, Moi, Laikipia, and Kisii University. We have seen every data analysis mistake that can be made, and we have helped students correct them. This guide reflects that experience directly.
Table of Contents
- What Data Analysis Actually Is — and What It Is Not
- The Two Pillars: Descriptive vs. Inferential Statistics
- Qualitative Data Analysis: Thematic, Content, and Narrative Approaches
- Choosing Your Software: SPSS, STATA, EViews, and NVivo Compared
- How to Choose the Right Statistical Test — Every Time
- Writing Chapter 4: Structure, Presentation, and Interpretation
- The 10 Data Analysis Mistakes That Kenyan Panel Reviewers Flag Most
- Mixed-Methods Analysis: When You Have Both Numbers and Words
- The Data Analysis Pre-Submission Checklist
- How Tobit Research Consulting Supports Your Data Analysis
1. What Data Analysis Actually Is — and What It Is Not
Data analysis is the systematic process of inspecting, transforming, and modelling data with the goal of discovering useful information, drawing conclusions, and supporting decision-making — or, in an academic context, answering your research questions. In a postgraduate dissertation or thesis, data analysis is not merely the mechanical act of running tests in SPSS or applying codes in NVivo. It is a scholarly act of reasoning: taking evidence you have collected from the real world and using rigorous, transparent methods to determine what that evidence means in relation to the questions your study set out to answer.
This distinction matters enormously in practice. A student who “does data analysis” by running a regression in SPSS and pasting the output table into Chapter 4 has not conducted data analysis in the scholarly sense. They have generated numbers. A student who conducts data analysis understands why they chose regression, knows what the output means, interprets the coefficients in relation to their hypotheses, connects the findings to their theoretical framework, and discusses what the results imply for the field. The first student will be sent back for revision. The second will pass.
The fundamental rule of data analysis in postgraduate research: Every analytical step — from your choice of descriptive statistics to your selection of inferential tests — must be driven by your research questions, your research design, and the measurement level of your variables. Analysis that is not grounded in these three factors is not research. It is number-generation. Panellists know the difference, and supervisors catch it during review.
2. The Two Pillars: Descriptive vs. Inferential Statistics
Every quantitative data analysis in a Kenyan postgraduate dissertation works within two broad categories of statistics. Understanding the distinction between them — and knowing when and why to use each — is the first competency that separates a strong data analyst from a weak one.
Descriptive Statistics: Organising What You Found
Descriptive statistics do exactly what their name says: they describe the data you have collected. They summarise, organise, and present the key characteristics of your dataset in a way that provides a clear picture of what your respondents look like, how they responded, and what the basic patterns in your data are. Descriptive statistics do not draw conclusions about populations beyond your sample — they describe what you observed in the data you have.
In a Kenyan postgraduate dissertation, descriptive statistics typically appear at the beginning of Chapter 4 and serve two purposes: they provide a profile of your respondents (the demographic and background information that tells the reader who answered your questionnaire), and they give initial, surface-level answers to your research questions before inferential analysis probes more deeply. The most common descriptive measures used in Kenyan university research are: frequencies and percentages (for categorical data), means and standard deviations (for continuous or ordinal data), and cross-tabulations (for examining relationships between categorical variables before inferential testing).
Inferential Statistics: Drawing Conclusions Beyond Your Data
Inferential statistics are where your analysis moves from description to argument. They allow you to use the data collected from your sample to draw conclusions — inferences — about the broader population from which that sample was drawn. Critically, inferential statistics enable you to test your hypotheses: to determine, with a defined level of statistical confidence, whether the relationships or differences you observe in your sample data are real (statistically significant) or could plausibly have occurred by chance.
At Kenyan universities, inferential statistics are the primary analytical engine for quantitative studies at Masters and PhD level. Panellists expect them. A quantitative dissertation that presents only frequencies and means — without any inferential testing — will almost always be returned with comments requesting analysis at a deeper level.
| Type | Purpose | Common Tests / Measures | When Used in a Kenyan Dissertation |
|---|---|---|---|
| Descriptive | Summarise and describe data characteristics | Frequencies, percentages, means, standard deviations, medians, mode, range, cross-tabulations | Always — at minimum for respondent profiling and preliminary findings in Chapter 4 |
| Inferential — Relationships | Test associations or correlations between variables | Pearson correlation, Spearman correlation, Pearson’s Chi-square, simple and multiple regression | When objectives examine the relationship or influence of one variable on another |
| Inferential — Differences | Test whether groups differ significantly on a variable | Independent samples t-test, paired t-test, one-way ANOVA, Mann-Whitney U, Kruskal-Wallis | When objectives compare two or more groups (e.g., gender, branch, region) on a measured outcome |
| Inferential — Predictive | Determine the extent to which independent variables predict or explain a dependent variable | Multiple linear regression, logistic regression, panel regression (STATA/EViews) | When objectives examine the influence of multiple independent variables on one outcome |
The p-value rule every Kenyan student must know: In most Kenyan university dissertations, statistical significance is set at p < 0.05 (the 95% confidence level). A p-value below 0.05 means the result is statistically significant — there is less than a 5% probability that the observed relationship occurred by chance. A p-value above 0.05 means the result is not statistically significant. Both outcomes are valid research findings. Do not omit non-significant results — report and interpret them honestly. Omitting non-significant results is a recognised data integrity error that reviewers flag.
3. Qualitative Data Analysis: Thematic, Content, and Narrative Approaches
Not all postgraduate research at Kenyan universities is quantitative. Qualitative studies — and the qualitative component of mixed-methods studies — require an entirely different analytical framework. Where quantitative analysis is about measuring variables and testing relationships with numbers, qualitative analysis is about identifying patterns of meaning in language, text, and lived experience. The goal is not to generalise to a population but to understand, in depth, the perspectives, processes, experiences, or contexts of the specific participants or documents you studied.
The most widely used qualitative data analysis approach at Kenyan universities is thematic analysis — a method for identifying, analysing, and reporting patterns (themes) within qualitative data. Thematic analysis does not require alignment with a particular theoretical framework, which is one reason it is so widely used across disciplines from education and social work to healthcare and business management.
The standard framework for thematic analysis in Kenyan academic research follows Braun and Clarke’s (2006) six-phase model: (1) Familiarisation — immersive reading and re-reading of transcripts or texts to develop deep familiarity with the data; (2) Generating initial codes — systematically identifying features of the data that are relevant to your research questions and labelling them; (3) Searching for themes — gathering codes into potential broader themes; (4) Reviewing themes — refining themes by checking them against the coded extracts and the full dataset; (5) Defining and naming themes — clearly articulating what each theme is and what it captures about the data; (6) Writing up — producing the analytical narrative that presents, illustrates with data extracts, and interprets each theme in relation to your research questions. In NVivo, phases 2–4 are facilitated by the software’s node-based coding system.
Content analysis systematically categorises and counts the occurrence of specific categories of meaning within text. It can be applied quantitatively (counting how frequently certain themes or words appear) or qualitatively (interpreting the meaning of content categories). It is widely used in Kenyan research studying policy documents, media content, institutional reports, and secondary data sources. Where thematic analysis is interpretive and concerned with the latent meaning within data, content analysis is more systematic and concerned with the manifest (visible, surface) content. Both can be supported by NVivo, though simple content analysis is also conducted manually.
Narrative analysis focuses on the stories people tell — how they structure their experiences, what events they emphasise, and what meaning they construct. It is used in qualitative studies at PhD level that focus on individual experiences, biographical accounts, or organisational histories. Case study analysis applies systematic cross-case or within-case comparison to examine the specific context, processes, and outcomes of one or more cases in depth. Both approaches require clear methodological justification in Chapter 3 and a rigorous, transparent analytical approach in Chapter 4.
4. Choosing Your Software: SPSS, STATA, EViews, and NVivo Compared
The choice of analytical software is not merely a technical preference — it should follow directly from your research design, your data type, and the analytical methods you stated in Chapter 3. Using a tool because your supervisor uses it, or because the university computer lab has it, without verifying that it is the most appropriate tool for your specific analysis, is one of the most common methodology errors in Kenyan postgraduate research. Here is what each of the four major platforms does, and when to choose it.
📊 SPSS (IBM SPSS Statistics)
SPSS is the most widely used quantitative analysis software at Kenyan universities, and for good reason: it is menu-driven, produces clearly labelled output, and handles the full range of tests that most Masters and PhD dissertations require — from frequencies and cross-tabulations through t-tests, ANOVA, correlation, and regression. SPSS is the default choice for survey-based quantitative studies in education, business, health, social sciences, and public administration.
Best for: Survey and questionnaire data; Likert-scale analysis; descriptive statistics; correlation and regression; t-tests and ANOVA; Cronbach’s Alpha reliability testing.
Descriptive Stats Regression ANOVA Reliability📈 STATA
STATA is more powerful than SPSS for advanced econometric and statistical modelling, particularly for studies involving panel data (longitudinal data tracking entities over time), instrumental variables, fixed and random effects models, survival analysis, and robust standard error corrections. It is the preferred tool for economics, finance, public health, and policy research at PhD level. While STATA has a steeper learning curve than SPSS (requiring command-line syntax for many operations), it offers greater flexibility and methodological precision for complex research designs.
Best for: Panel data; econometric modelling; fixed/random effects; advanced regression; public health and economics research.
Panel Data Econometrics Fixed Effects PhD Level📉 EViews
EViews is the specialist platform for time series analysis and forecasting, making it the standard tool for studies in economics, finance, and macroeconomics that analyse data tracked over time — interest rates, GDP, inflation, exchange rates, stock prices, or sectoral performance indicators. EViews excels at stationarity testing (unit root tests), cointegration analysis, Vector Autoregression (VAR), ARCH/GARCH models for volatility, and Granger causality testing. If your study uses secondary time-series data from sources such as the CBK, KNBS, NSE, or World Bank, EViews is likely your primary analytical tool.
Best for: Time-series data; macroeconomic and financial research; forecasting; ARCH/GARCH; cointegration and VAR models.
Time Series Forecasting VAR Finance & Economics🗣️ NVivo
NVivo is the leading software for qualitative data analysis. It allows researchers to import text data (interview transcripts, focus group transcripts, open-ended survey responses, policy documents, social media content), apply a systematic coding framework using nodes (themes and sub-themes), visualise relationships between codes through word trees and cluster analyses, and manage large volumes of qualitative data in an organised, auditable, and scholarly way. NVivo does not do the analytical thinking for you — it is a tool for organising and navigating your codes, not for generating themes automatically. The interpretation remains the researcher’s responsibility.
Best for: Interview and focus group transcripts; thematic analysis; content analysis; qualitative component of mixed-methods studies.
Thematic Analysis Coding Qualitative Mixed MethodsAt most Kenyan universities, the following discipline-software pairings are standard: Business Administration, Education, Social Work, Public Administration, Health Management — SPSS (with NVivo for qualitative components). Economics, Finance, Banking, Actuarial Science — STATA or EViews, depending on whether data is cross-sectional or time-series. Qualitative studies in Anthropology, Sociology, Community Development, Gender Studies — NVivo. Always confirm the expected software with your supervisor before committing to an analytical approach — and state your software choice and justification explicitly in Chapter 3.
5. How to Choose the Right Statistical Test — Every Time
Choosing the wrong statistical test is one of the most consequential errors you can make in quantitative research. It produces invalid results — conclusions that do not actually follow from your data — and it will be identified by any competent reviewer or panel member. The good news is that choosing the right test is not arbitrary or difficult once you understand the three questions that determine it.
| Research Question Type | Variable Measurement Level | Parametric Test | Non-Parametric Alternative |
|---|---|---|---|
| Relationship between two variables | Both continuous (interval/ratio) | Pearson Correlation (r) | Spearman Rank Correlation (ρ) |
| Association between two categorical variables | Both nominal or ordinal | Chi-square test of independence | Fisher’s Exact Test (small samples) |
| Difference between two independent groups | Continuous dependent variable | Independent Samples t-test | Mann-Whitney U test |
| Difference between two related/matched groups | Continuous dependent variable | Paired Samples t-test | Wilcoxon Signed-Rank test |
| Difference between three or more groups | Continuous dependent variable | One-way ANOVA (with post-hoc) | Kruskal-Wallis H test |
| Prediction of one outcome from one predictor | Both continuous | Simple Linear Regression | Spearman correlation as proxy |
| Prediction of one outcome from multiple predictors | Continuous outcome; mixed predictors | Multiple Linear Regression | — |
| Prediction of a binary (yes/no) outcome | Binary dependent variable | Binary Logistic Regression | — |
| Panel/longitudinal data (entities over time) | Continuous; panel structure | Fixed Effects / Random Effects (STATA) | — |
| Time-series relationships and causality | Time-series data | VAR, Granger Causality, Cointegration (EViews) | — |
Study A: “Influence of credit risk management practices on financial performance of commercial banks in Kenya (2018–2023).” The dependent variable is financial performance (measured as ROA — a ratio, continuous). The independent variables are credit appraisal, collateral management, and portfolio diversification (measured on Likert scales, treated as ordinal-to-interval composites). The study has one observation per bank per year — a panel structure. Correct tool: STATA with Fixed Effects or Random Effects regression (Hausman test to choose between them).
Study B: “Effect of teacher training programmes on student performance in public secondary schools in Nairobi County.” The dependent variable is student KCSE performance category (grades A–E — ordinal). The independent variable is whether a school’s teachers received the programme (yes/no — nominal). Correct test: Chi-square test of independence or, if controlling for multiple school-level factors, ordinal logistic regression.
6. Writing Chapter 4: Structure, Presentation, and Interpretation
Chapter 4 is where your research effort becomes visible as scholarship. It is the chapter that directly answers your research questions — and the chapter where the quality of your analysis, your command of your data, and your ability to reason from evidence are most on display. A well-structured Chapter 4 at a Kenyan university follows a consistent architecture, regardless of whether the study is quantitative, qualitative, or mixed-methods.
A brief paragraph explaining what the chapter covers, how it is organised, and how it connects to the research questions and objectives established in Chapter 1. Required at most Kenyan universities. Its absence signals a lack of scholarly organisation.
Report the number of questionnaires distributed, returned, and usable, expressed as both a number and a percentage. The response rate must be sufficient to justify the validity of your analysis. At most Kenyan universities, a response rate of 70% and above is considered satisfactory (Mugenda & Mugenda, 2003). If your response rate is below 70%, explain why and discuss its implications for the generalisability of your findings. Never proceed with analysis without reporting this — reviewers notice immediately.
Report the Cronbach’s Alpha coefficient for each scale or subscale of your instrument. A coefficient of ≥ 0.70 is the widely accepted threshold at Kenyan universities (Nunnally, 1978). Present results in a table showing the scale name, number of items, and Alpha coefficient. If any scale falls below 0.70, explain how you addressed it — either by removing weak items and retesting, or by acknowledging it as a limitation. Cronbach’s Alpha reports must appear before your main findings — they validate that the instrument was reliable before your results are presented.
Present the profile of your respondents using frequency tables and/or bar charts — gender, age, education level, years of experience, organisation type, or whatever demographic variables are relevant to your study. Every demographic table must be followed by an interpretive paragraph that discusses what the figures mean, not just repeats them. “Table 4.1 shows that 60% of respondents were female” is a repetition. “The majority of respondents (60%, n=180) were female, reflecting the gender composition of the health sector workforce in Nairobi County as documented by the Ministry of Health (2023)” is an interpretation.
For Likert-scale instruments, present means and standard deviations for each item or subscale, organised by objective or research question. Use a Likert interpretation guide to contextualise your means — for a 5-point scale, a mean of 1.00–1.49 typically indicates strong disagreement and a mean of 3.50–5.00 indicates agreement. State your interpretation key explicitly and apply it consistently. Standard deviations indicate the spread of responses: a low SD suggests consensus among respondents; a high SD suggests divergence that may warrant further discussion.
Organise your inferential analysis by objective or research question — not by test type. Each sub-section should begin by restating the specific objective and its corresponding null hypothesis (if applicable), present the relevant SPSS/STATA/EViews output in a cleanly formatted table, then provide a written interpretation that: states the test result (e.g., r = .542, p = .003); states whether the result is statistically significant (p < .05 or p > .05); states the decision regarding the hypothesis (reject H₀ or fail to reject H₀); and interprets the practical meaning of the result in relation to the objective. Never paste raw SPSS output tables into your chapter. Reproduce the relevant figures in a formatted APA-style table.
Present findings under thematic headings derived from your thematic analysis. Each theme section should: name and define the theme; present verbatim quotations from participants as evidence (indented, in quotation marks, attributed to a participant code such as “Respondent A, Interview 3”); and provide analytical commentary that explains what the quotes collectively reveal about the theme and how it relates to your research question. Quotes are evidence, not findings. The finding is your analytical interpretation of what the quote demonstrates.
A concise summary of the key findings from each objective, organised clearly so the reader can immediately identify what the study found in relation to each research question. The chapter summary must not include new analysis — it is a synthesis of what was presented above. It also serves as the bridge into Chapter 5, where findings will be discussed in relation to the existing literature and theoretical framework.
7. The 10 Data Analysis Mistakes That Kenyan Panel Reviewers Flag Most
Based on our work with thousands of Kenyan postgraduate students, these are the data analysis errors that supervisors and panel reviewers most consistently flag — and the ones that lead to the most painful and time-consuming revisions.
✅ What Passes Panel Review
- Reporting p-values accurately: p = .043 or p < .001 — never “p = 0.000”
- Presenting findings in formatted APA tables — not raw SPSS output screenshots
- Interpreting every table immediately after it appears in the text
- Testing and reporting normality and other test assumptions before applying parametric tests
- Reporting both significant and non-significant results honestly
- Linking every finding explicitly to the corresponding research question or objective
- Reporting Cronbach’s Alpha before presenting scale-based findings
- Using correct APA 7 table formatting: numbered, titled, no vertical lines, notes below
- Reporting effect sizes (R², Cohen’s d, η²) alongside significance values
- Explaining what the findings mean — not just what the numbers are
❌ What Gets Sent Back for Revision
- Writing “p = 0.000” instead of “p < .001” — a reporting error that signals statistical illiteracy
- Pasting raw SPSS output tables directly into the chapter without reformatting
- Presenting tables without any written interpretation below them
- Applying parametric tests without checking normality or homogeneity of variance
- Reporting only significant results and omitting non-significant ones
- Analysing data without connecting results back to stated objectives
- Missing or inadequate reliability analysis — presenting composite scores without testing Cronbach’s Alpha
- Reporting correlation coefficients without interpreting the direction and strength of the relationship
- Using R² alone without interpreting the adjusted R² for models with multiple predictors
- Treating Likert scale means as if they are ratio data without acknowledging the ordinal nature of the measurement
The “p = 0.000” error — why it matters: SPSS displays p-values rounded to three decimal places. When a p-value is very small, it appears as “.000” in the output. This does not mean the probability is literally zero — no probability in statistics is zero. The correct way to report this is p < .001. Reporting “p = 0.000” is a recognised statistical error that signals to reviewers that the student does not understand what a p-value is. It is one of the most common and most avoidable errors in Kenyan dissertation data analysis.
8. Mixed-Methods Analysis: When You Have Both Numbers and Words
Mixed-methods research is increasingly common at Kenyan universities, particularly in education, health sciences, public administration, and development studies. A mixed-methods study combines quantitative and qualitative data collection and analysis within a single study, and the power of the approach lies in how the two data types are integrated — not merely reported side by side.
At Kenyan universities, the most common mixed-methods design is the explanatory sequential design: quantitative data is collected and analysed first, and qualitative data is then collected and analysed to explain, contextualise, or elaborate on the quantitative findings. In Chapter 4, this means quantitative findings are presented first (with full statistical rigour), followed by qualitative themes, followed by an integration section that explicitly addresses where and how the two sets of findings converge, diverge, or complement each other. A mixed-methods Chapter 4 that presents quantitative and qualitative findings in separate sections without any integration fails the primary requirement of the design.
The integration question your panel will ask is: “How do your quantitative and qualitative findings speak to each other?” If you cannot answer that question with reference to specific findings from both components — if the two halves of your study exist independently with no analytical dialogue between them — you have not conducted a mixed-methods study. You have conducted two separate sub-studies and put them in the same document.
9. The Data Analysis Pre-Submission Checklist
Before submitting your Chapter 4 to your supervisor or panel, work through every item in this checklist. Each item represents a documented, recurring reason why Kenyan university panels return data analysis chapters for revision.
Data Preparation
- Data cleaning: Have you checked for missing values, duplicate entries, data entry errors, and out-of-range responses? Have you documented how you handled each?
- Variable labelling: Are all variables clearly labelled in SPSS/STATA with names that correspond to your instrument items? Have you assigned the correct measurement level (nominal, ordinal, scale) to each variable?
- Reverse scoring: If any items in your instrument are negatively worded (e.g., “I do not trust my SACCO’s financial management”), have you reverse-scored them before computing composite scores?
- Reliability testing: Have you run Cronbach’s Alpha for every scale and sub-scale, and reported the results in the chapter?
- Normality testing: Have you tested the distribution of your continuous variables before applying parametric tests, and reported the results?
Quantitative Analysis
- Correct test selection: Is every statistical test you used appropriate for the measurement level of your variables and the nature of your research question? Can you justify each choice if asked?
- Accurate p-value reporting: Have you reported all p-values correctly — never as “p = 0.000” but as “p < .001”? Have you stated significance thresholds explicitly?
- Complete results: Have you reported both significant and non-significant results, and interpreted both in the text?
- Effect sizes: Have you reported effect sizes (R², r, η²) alongside statistical significance values?
- Tables formatted correctly: Are all tables formatted in APA 7 style — no vertical lines, clear headers, table number and title above, notes below?
Qualitative Analysis
- Audit trail: Have you described your coding process clearly enough that another researcher could follow it? Have you mentioned the software used (NVivo or manual) and the number of initial codes generated?
- Themes grounded in data: Is every theme illustrated with specific, attributed verbatim quotations? Are quotations interpreted analytically rather than simply listed?
- Member checking / credibility: If required by your institution, have you addressed the credibility measures described in your Chapter 3 — member checking, peer debriefing, or thick description?
Chapter Structure
- Response rate reported: Is the response rate stated as both a number and percentage, and evaluated against an acceptable threshold?
- Every objective addressed: Does each section of Chapter 4 explicitly address a stated research question or objective from Chapter 1?
- Every table interpreted: Does every table have a written interpretation immediately following it in the text?
- Chapter summary present: Does the chapter end with a summary that recaps findings by objective and bridges into Chapter 5?
10. How Tobit Research Consulting Supports Your Data Analysis
Data analysis is the technical and intellectual core of your research — and it is the point where the gap between a strong dissertation and a weak one is most visibly decided. At Tobit Research Consulting, we work with Masters and PhD students across Kenya at every stage of the data analysis process: from choosing the right analytical approach in Chapter 3 through to cleaning your data, running your tests, interpreting your results, and writing up a Chapter 4 that demonstrates genuine analytical expertise.
We do not conduct your analysis for you and hand you a chapter. We work alongside you — explaining what the outputs mean, helping you understand why you chose a particular test, and building the analytical competence you need to walk into a panel defence and answer any question about your data with confidence. Our approach is rooted in academic integrity and genuine student development — because the goal is not just to pass your panel, but to become the researcher your degree is meant to make you.
Expert Data Analysis Support for Kenyan Masters and PhD Students
Tobit Research Consulting provides comprehensive, integrity-focused data analysis support for students at KU, UoN, JKUAT, MKU, Strathmore, Egerton, Moi, Laikipia, Kisii University, and universities across Kenya. Our data analysis services include:
- SPSS data analysis — descriptive statistics, correlation, regression, t-tests, ANOVA, factor analysis, reliability analysis
- STATA data analysis — panel regression, fixed/random effects, robust SE, advanced econometrics
- EViews data analysis — time-series analysis, VAR, ARCH/GARCH, Granger causality, cointegration
- NVivo qualitative analysis — thematic coding, node structuring, theme development, output export
- Chapter 4 writing support — full findings write-up with correctly formatted APA tables and analytical interpretation
- Statistical test selection and justification — ensuring your methodology and analysis are consistent and defensible
- Cronbach’s Alpha and validity analysis — instrument reliability assessment and reporting
- Data cleaning and preparation — variable coding, missing value treatment, outlier analysis
- Mixed-methods integration — quantitative-qualitative integration for explanatory and exploratory designs
- Chapter 5 discussion support — linking findings to literature and theoretical framework
- Proposal development (Chapters 1–3) — background, problem statement, methodology, sampling
- Concept papers, conference papers, journal article preparation
- AI removal (Turnitin) and academic paraphrasing
- Full dissertation and thesis support from proposal to final submission
Whether you are starting your analysis for the first time or dealing with a panel revision that sent you back to your data, we are ready to help you understand your data, interpret it correctly, and present it with the rigour your degree demands.
Book a Free Consultation →📍 Bruce House, 4th Floor, Standard Street, Nairobi CBD, Kenya | Tel: +254 728 430 728 | tobitresearchconsulting.com
This guide is part of Tobit Research Consulting’s Postgraduate Research Skills Series. Key sources informing this guide include: Braun & Clarke (2006) — Using thematic analysis in psychology; Cronbach (1951) — Coefficient alpha and the internal structure of tests; Field, A. (2018) — Discovering Statistics Using IBM SPSS Statistics (5th ed.); Kothari, C. R. (2004) — Research Methodology: Methods and Techniques; Lincoln & Guba (1985) — Naturalistic Inquiry; Mugenda, O. M., & Mugenda, A. G. (2003) — Research Methods: Quantitative and Qualitative Approaches; Nunnally, J. C. (1978) — Psychometric Theory; Pallant, J. (2020) — SPSS Survival Manual; peer-reviewed literature on data analysis practices in East African postgraduate research; USIU-Africa Data Analytics Laboratory guidelines; and documented review patterns from Kenyan university panels at KU, UoN, JKUAT, and MKU.
