Six student projects, already defended.

The raw weekly records for 150 birds across six observation points had to be turned into a defensible answer to two questions: does the additive improve growth, feed efficiency and health, and is the effect dose-dependent? This required a complete statistical analysis robust enough for a dissertation defence.

I structured the data from a Completely Randomised Design (three doses: 0, 5, 10 g/kg; 50 birds each) and ran a full analysis pipeline in SPSS and Python: one-way ANOVA per outcome, Tukey HSD post-hoc, multiple linear (OLS) regression on body weight, a chi-square test on mortality and a Pearson correlation matrix, all at alpha = 0.05.

The client needed to test whether a simple, non-invasive body measurement (crown-rump length) could serve as a reliable field proxy for age, and to quantify that relationship while accounting for sex and breed.

Crown-rump length and dentition-estimated age were recorded for 200 zebu cattle across four breeds, then analysed with descriptive statistics, Pearson and Spearman correlation, simple linear regression, a Welch independent-samples t-test and one-way ANOVA at the 5% significance level.

Locality-specific prevalence data had been absent for over a decade, leaving farmers and veterinary services without an evidence base to target deworming and gauge which worm genera and host groups drive infection.

200 faecal samples were examined by saturated-salt floatation and sedimentation microscopy, then tabulated by parasite genus, sex and age, with chi-square tests of independence (SPSS) to assess associations between infection status and host factors.

The study needed to identify which diseases drive losses and to test rigorously whether disease outbreaks are statistically linked to reduced farm profitability, despite a modest sample and no laboratory confirmation.

Survey responses were summarised with frequency tables and 25 charts across demographics, disease burden, biosecurity, veterinary practice and socio-economics, then a 2x2 Chi-square test of independence (with Yates' continuity correction) related outbreak occurrence to perceived profitability.

The analyst needed to turn mixed-methods survey and interview data into defensible statistics — testing four hypotheses and ranking the relative weight of each challenge — while resolving a sample-size inconsistency (one summary table reported N=100 against N=107 analysed elsewhere).

In SPSS 26, each of the four challenges was tested against teacher burnout with Pearson correlation and simple linear regression at p < .05, then triangulated against thematically coded administrator interviews; descriptive item means were read against a 2.5 decision criterion on a four-point Likert scale.

Hand hygiene is the single most effective measure against healthcare-associated infections, yet adherence among healthcare workers is often suboptimal in low-resource settings — and a pile of raw questionnaire responses had to become a clear picture of knowledge, practice and the obstacles to compliance.

I tallied and analysed 100 questionnaire responses across socio-demographics, knowledge items, adherence and barriers, building frequency tables and charts (knowledge-correctness, professional mix, experience) and reading them against the WHO 'My Five Moments for Hand Hygiene' expectations.