The one in Chandni Chowk, near the Red Fort in Delhi. It is a charity hospital. They take in sick birds on the condition that the birds once cured would be set free.
Technically speaking, any of these could be scientific or pseudo-scientific. It depends on several factors. We'll go through each one individually:
"Students gather to identify plants in an area"
If they had proper criteria for identifying the plants, and analysed a large enough area, then this would be perfectly valid.
If, however, they only analysed a gardener's custom-made flower garden and claimed these results were representative of the whole district, that would be pseudoscience.
"Students gather to record their beliefs about paranormal activity."
This is the most pseudoscientific one there is, but it's not complete garbage. If, for instance, they were conducting a study to see what percentage of young people believed in the paranormal, and what particular beliefs were most common, then that would be an entirely valid scientific experiment.
Simply writing down what their beliefs were without testing them would not be so valid.
"Students gather to weigh and tag a bird ready to be released by a bird hospital."
Probably fine, although for the results to mean anything they should probably have weighed the bird when it was admitted to the hospital as well. Also, they'd want to do this to several birds. A single bird is probably not representative of the entire population- it might not even be representative of birds released from the hospital.
"Students gather to record their results of water testing from a local stream."
Depends on the testing procedure and what they were trying to conclude from it. For instance, suppose they tested the stream after it had rained, found the lead concentration was high, and concluded that rain caused the lead concentration to increase. This would not be valid, because the lead concentration may have been high before the rain as well.
There is no generic type of experiment that's inherently valid or inherently pseudoscientific. It depends entirely on what you are trying to measure, what measurements you perform, and what conclusions you draw from those measurements. Some things to consider are:
* Whether or not the experiment yields useful information about what you are trying to study.
* Are there any other factors that could influence the results except for what you're tring to measure? If so, how do you compensate for them?
* Are your results statistically significant?
Here's a good example: suppose you had a cold, took some medicine, and got better. That proves that the medicine cures you, right? Well, not necessarily. You might have got better by yourself, and the medicine might not have done anything. There's no way of telling from this experiment.
Instead, what you could do would be to get a whole lot of sick people, and give the medicine to half of them. You then see if that half that has had the medicine is cured faster than the other half. If people who had been treated got better in a week, on average, whereas people who hadn't had it took an average of a month, you'd then conclude that the medicine did work.