LEARN WITH A STUDY DESIGN

Animal-study examples

Explore independent groups, paired measurements, repeated visits and censored outcomes. Start with the animal and the question.

Free · Downloadable CSV · No account

AVAILABLE NOW

Try an analysis and customize its figure

Opening an example fills its data and figure labels. Review the design and confirm the assumptions before running it.

Inflammation

Inflammation at one endpoint

Runnable example · synthetic

Do mean paw-thickness measurements differ across three groups at a prespecified endpoint?

Experimental unit: One animal; one paw-thickness value per animal. Six different animals per group.

Analysis: Ordinary one-way ANOVA with all three Tukey comparisons; plot individual animals and mean ± sample SD.

Design, columns and interpretation

Independent groups, one continuous outcome, one factor. This toy design assumes treatment is assigned to individual animals and no unmodelled cage/litter clustering.

group = assigned group; unit = animal ID; value = paw thickness in mm.

Ordinary ANOVA needs suitable residual distributions and equal population variances. Several paws, sections or time points from one animal are not extra independent animals.

group,unit,value
Vehicle,group1-1,2.8
Vehicle,group1-2,3.1
Vehicle,group1-3,3.5
Vehicle,group1-4,3

First four records shown. The download includes the complete example.

Try Inflammation at one endpoint Download Inflammation at one endpoint CSV

Toxicology

A liver-enzyme endpoint

Runnable example · synthetic

Do mean enzyme activities differ among vehicle and two exposure groups?

Experimental unit: One animal; one serum ALT activity per animal. Six different animals per group.

Analysis: Ordinary one-way ANOVA and the full three-pair Tukey family; raw points and mean ± sample SD.

Design, columns and interpretation

Independent groups at one endpoint. The example assumes individual allocation and no cage/litter clustering.

group = assigned group; unit = animal ID; value = ALT activity in U/L.

These arbitrary values are not species reference ranges or evidence of toxicity. Real skewed/unequal-variance data may require another model. A nonsignificant test does not establish safety.

group,unit,value
Vehicle,group1-1,28
Vehicle,group1-2,34
Vehicle,group1-3,31
Vehicle,group1-4,39

First four records shown. The download includes the complete example.

Try A liver-enzyme endpoint Download A liver-enzyme endpoint CSV

Behaviour & neuroscience

An independent behavioural outcome

Runnable example · synthetic

Does mean distance travelled differ between two independently assigned groups?

Experimental unit: One animal; one total distance per animal during the same predefined observation window. Six animals per group.

Analysis: Two-sided Welch t-test, Treatment minus Vehicle; one prespecified comparison, with its difference interval.

Design, columns and interpretation

Independent treatment and vehicle groups, without repeated sessions or cage/litter clustering in the example.

group = assigned group; unit = animal ID; value = total distance in m.

Distance travelled alone does not establish anxiety or a mechanism. Multiple sessions from an animal need a repeated-measures design. Choose the outcome/window before inspecting results.

group,unit,value
Vehicle,group1-1,18.2
Vehicle,group1-2,22.6
Vehicle,group1-3,20.1
Vehicle,group1-4,25.3

First four records shown. The download includes the complete example.

Try An independent behavioural outcome Download An independent behavioural outcome CSV

Physiology

Grip strength before and after

Runnable example · synthetic

What is the mean within-animal change between two prespecified measurements?

Experimental unit: Six animals measured twice; n = 6 pairs, not 12 independent animals.

Analysis: Two-sided paired t-test of After minus Before; exact-ID connecting lines and a mean-difference interval.

Design, columns and interpretation

Exactly matched animal IDs in Before and After. Each value is a prespecified animal-level summary of any technical trials.

group = Before or After; unit = the same animal ID at both measurements; value = grip force in N.

The test assumes independent animals and suitable within-animal differences. An uncontrolled before/after change cannot isolate a treatment effect from time, training or other influences.

group,unit,value
Before,animal-1,1.1
Before,animal-2,1.25
Before,animal-3,1.08
Before,animal-4,1.32

First four records shown. The download includes the complete example.

Try Grip strength before and after Download Grip strength before and after CSV

PLAN YOUR DATA STRUCTURE

Templates for analyses not yet supported

Download and inspect these formats. They cannot be analysed in the current workspaces; longitudinal and survival tools still need implementation and independent validation.

Longitudinal physiology

Body weight over time

Template only · synthetic

Do body-weight trajectories differ between groups?

Experimental unit: Eight animals, four per group, each measured at three visits. Twenty-four rows are still eight animals.

Analysis: Requires a validated longitudinal model and planned treatment × time contrasts. Not available in the current workspaces.

Design, columns and interpretation

Repeated measurements with animal ID, treatment and numeric time.

group = assigned group; animal = stable animal ID; day = actual visit day; weight_g = body weight in g.

Do not import visits as independent animals or replace the longitudinal question with a series of unplanned t-tests. Missing visits and attrition must remain explicit.

group,animal,day,weight_g
Vehicle,group1-1,0,24.1
Vehicle,group1-1,7,24.9
Vehicle,group1-1,14,25.4
Vehicle,group1-2,0,25.3

First four records shown. The download includes the complete example.

Download Body weight over time CSV

Oncology

Tumour growth trajectories

Template only · synthetic

How do tumour-volume trajectories differ across treatment groups?

Experimental unit: Eight animals, four per group, measured at three visits; retain each animal’s identity.

Analysis: Requires a validated longitudinal model, explicit time/covariance assumptions and planned contrasts. Not available yet.

Design, columns and interpretation

Repeated tumour measurements. One trajectory per animal; the long-form rows are correlated.

group = assigned group; animal = stable animal ID; day = visit day; volume_mm3 = tumour volume in cubic mm.

Do not fit the concentration–response tool to growth over time. Tumour volume is not concentration. Transformations, missingness and humane-endpoint attrition need explicit handling.

group,animal,day,volume_mm3
Vehicle,group1-1,0,90
Vehicle,group1-1,7,143
Vehicle,group1-1,14,218
Vehicle,group1-2,0,105

First four records shown. The download includes the complete example.

Download Tumour growth trajectories CSV

Survival

Time to event with censoring

Template only · synthetic

How does time to a consistently defined, prespecified event differ between groups?

Experimental unit: One animal per row, with follow-up duration and event status. Six animals per group.

Analysis: Requires Kaplan–Meier estimates, censor marks, numbers at risk and a validated survival comparison. Not available yet.

Design, columns and interpretation

Right-censored time-to-event data with the same event definition and time origin in both groups.

group = assigned group; animal = unique ID; time_days = event or last follow-up time; event = 1 observed event, 0 right-censored.

event = 0 means no event was observed before that follow-up ended; it is not an event at that time or proof that the event never occurs. Do not analyse these times with an ordinary t-test.

group,animal,time_days,event
Vehicle,V1,8,1
Vehicle,V2,11,1
Vehicle,V3,15,1
Vehicle,V4,18,0

First four records shown. The download includes the complete example.

Download Time to event with censoring CSV