What is the difference between systematic observation and naturalistic observation? What are two possible disadvantages of using the survey method? What are the most important and most basic aspects of an experiment?

I suggest tackling the last three, because they are relatively easy to tie together. You can frame your discussion as a comparison of non-experimental versus experimental methods. To see how the topics connect, read on.
1. What is the difference between systematic observation and naturalistic observation?
Naturalistic observation involves collecting information that is offered spontaneously by the environment. Unlike an experimenter, the observer doesn't attempt to manipulate conditions. This can be a strength if you want to know how people behave in the real world. You don't have to worry so much that people are reacting in ways that are specific to the laboratory setting. Simple observation doesn't allow you to pin down causation, though, and it can lead to biases.
For example, how do you, the observer, decide if a person is sad? Your perceptions might be colored by your interests, preconceptions, cultural training, and so forth. To reduce observer bias, researchers often choose to use a more systematic approach. They start with rules for scoring what the observer sees. In this case, they would establish operational criteria for deciding if someone is sad. What specific behavior qualifies as evidence of sadness? How many times must a person adopt a particular facial expression? How long must the behavior go on?
Researchers engaging in systematic observation also typically look for evidence of interobserver reliability. They have multiple observers record the data, and then look to see if the observers agree. 
2. Describe two possible disadvantages of using the survey method.
A survey is another, non-experimental research method. People who design survey studies must take care to avoid obvious pitfalls. For instance, it's important to choose randomly from the population you wish to understand; otherwise, you could get a sample that isn't representative. Also, the wording of the survey is very important. You have to find language that doesn't influence the way survey respondents feel about the subject. That, too, could bias your results.
Let's assume you do that, though. You still have to deal with two key disadvantages. One is that a survey relies on self-reports: You merely learn what people are willing or capable of telling you. This introduces bias because people don't always provide correct information. They may lie intentionally, or they may be mistaken.
Another problem concerns causation. You are getting a snapshot of a population sample. It can reveal correlations, such as "respondents who say they were sad were more likely to suffer from insomnia." It can't tell you why the correlation exists, though. People might develop emotional problems as an effect of poor sleep, but they might also sleep poorly because they are sad. There could be a third variable that causes both sadness and poor sleep, like a medical condition. In short, surveys tell you something, but the information is necessarily biased by self-reporting, and it doesn't permit you to address causation.
3. What are the most important and most basic aspects of an experiment?
Experimentation differs from the other methods because it permits researchers to reach more definitive conclusions about causation. Researchers identify ahead of time what factors they think might be playing a causal role, and then think of a way to test their hypothesis. If their hypothesis is correct, what would we expect to see? They design experiments with these crucial features:

Manipulation. Researchers systematically change an independent variable (like sleep duration) to see if it results in changes to another dependent variable (like sadness).

Control. Researchers control the environment so other possible factors (like medical problems) remain constant across all versions of the experiment. Thus, if researchers test the effects of two different levels of sleep duration, they take pains to eliminate medical problems as a confounding factor. They could do this by using only healthy study subjects, or by making sure subjects suffered from the same medical problems under both conditions of the experiment.
To achieve these ends, researchers also make use of randomization. For example, researchers conducting a study on sleep and sadness might try to select volunteers at random from a given population. In practice, that's hard to do. The sort of people who volunteer may not be representative of the population as a whole. Even if that sort of selection isn't possible, researchers can apply another type of randomization: They randomly assign each volunteer to a particular experimental condition.
In summary, this is my overview of the topics you could choose for your essay. I believe you will find them easy to integrate, because you can unify the entire discussion in terms of the strengths and weaknesses of different approaches to research.