A Feeding Trial With Confounded Results
An animal science class of 28 students is running a 42-day feeding trial with 18 crossbred steers housed in two pens. One ration is a donated high-energy mix; the other is the school's usual hay-based ration. Students weigh feed daily and cattle weekly on the school's only livestock scale, which is located at the far pen and has not been calibrated this year. The two pens differ: pen A is near the road and has a leaking waterer that keeps the bedding damp; pen B is shaded. Cattle were assigned to pens by convenience when they arrived from two sources. Students have already calculated average daily gain and found a 0.4 lb/day advantage for the high-energy ration, and they want to present it as proof at the county fair. The district veterinarian has not reviewed the protocol.
1. Identify the major threats to the validity of the students' conclusion and explain how each could produce the observed 0.4 lb/day difference.
Read the worked answer
Threats to validity. The 0.4 lb/day difference cannot be attributed to the ration because pen and ration are confounded. Pen A has a leaking waterer and road disturbance; wet bedding can reduce intake and increase health risk. Pen B is shaded, which can improve comfort and gain. Cattle came from two sources and were assigned by convenience, so source, breed, age, and initial weight may differ. The scale is uncalibrated and located at one pen, adding measurement error and possibly handling stress. Feed intake is measured by pen, not individual, so feed efficiency cannot be tied to cattle. There is no replication of pen, so pen is the experimental unit and the class has, in effect, two observations. Any one of these-especially initial weight and waterer status-could produce 0.4 lb/day. The defensible statement is an association within this trial, not proof.
2. Design a feasible redesign using the same 18 steers and facilities, or explain what claims the class can defensibly make without a redesign.
Read the worked answer
Feasible redesign. With 18 steers and two pens, perfect replication is impossible, but the class can improve inference. Fix the waterer and calibrate the scale before weighing. Record initial weight, source, and health score, then randomly assign within source and weight blocks. If the facility allows, use a crossover: feed ration A in both pens for 21 days, switch with a 7-day washout, and compare each animal's change; this helps separate ration from pen, though carryover and time effects remain. Keep feed and refusal records, sample feeds for analysis, and weigh at the same time of day. Treat pen as the experimental unit and acknowledge pseudo-replication. Without a redesign, students should present a descriptive case study with limitations and avoid causal claims. That is still rigorous if framed correctly.
3. What animal-care and communication steps should the teacher take before the fair presentation?
Read the worked answer
Animal care and communication. Before the fair, have the district veterinarian review the protocol, waterer repair, heat and road stress, and humane endpoints. Follow the school's animal-care approval process and document feed changes, health checks, and treatment. The leaking waterer is a welfare issue, not just a data problem; fix it immediately. For the presentation, students should show the data, the confounds, and what they would do differently. I would preview the display with the principal and avoid overclaiming. If the fair requires a conclusion, label it “preliminary” and include limitations. This models scientific integrity and protects the program from criticism that it used animals to prove a predetermined result.