Because animals cannot communicate, most of the human standards are not practical.  When writing on an exam, you should not use consent, right to withdraw, deception, anonymity, or debriefing to evaluate an animal study.  The only one that can be directly applied is "undue stress or harm."

A cost-benefit analysis should consider the number of humans that would benefit from the study, the effect on humans if the study is not conducted, the number of animals that are required for the study, and the harm that would be done to the animal.  It could also, potentially, not positive benefits for the animal (e.g. treatment for an animal's health condition). If the harm (costs) outweigh the potential benefits, according to both the APA and BPS, animal research should not be carried out.

The concept of "undue" stress or harm is not clearly defined and may be up to the researchers to interpret. In addition, one does not actually know the true value of an experiment until it is done.  If we actually knew what the outcome would be, we would not have to do the experiment. So, knowing what we hope will be the outcome may bias us in determining the potential benefit to humans. Finally, the value one puts on human vs animal life is not constant - it is also subjective. Different cultural norms may result in different attitudes toward the value of an animal's life.  This is why many are pushing for international standards in animal research.

By sharing information among researchers, we could reduce the number of animals used.  In addition, better techniques that do not lead to the death or impairment of an animal could reduce the numbers needed.  Better planning so that experiments are well controlled and have high validity from the beginning, rather than having to redo research because of mistakes that should have been accounted for.

By reducing the number of animals used, this will lower the sample size of studies, making the data potentially less reliable. This may lead to type I errors, rejecting a null hypothesis that should have been accepted.

Researchers can use less invasive techniques, limiting injections, ablations, and lesioning of the animals. Animals need to be provided with better living conditions that mirror what they would have under natural conditions. There should also be trained vets available to help with the care and treatment of animals, including painless euthanasia.

Laboratories may not be equipped to provide the requirements for these standards. This will increase the costs of research. This may result in less research being done in certain areas unless there is an increase in funding.

Having to have more members of the team may have the benefit of researcher triangulation.

When animals live in natural conditions, this will help to increase the ecological validity of the research and decrease confounding variables.  Animals raised in cages - either isolated or overcrowded - with minimal stimulation may actually not be representative of the animal.  Stressors induced by the conditions they are kept in may be a confounding variable in research.  This ethical standard may actually increase the quality of animal research.

We can use computer models, in vitro testing, and more clinical trials on humans.

Models are limited by our current understandings of biological systems. Using cell cultures is not the same as testing on the whole system - so this may not provide representative findings. There are also ethical concerns with carrying out more ethical testing on humans.