this is gonna be a long answer and the info found is found from this web site http://www.14-19nw.org.uk/mod/page/view.php?id=8875
Experiments involve manipulating the independent variable in order to measure its effect on the dependent variable. External variables will be controlled so that the casual relationship between the independent and dependent variable is seen; a cause and effect relationship. This means that any difference measured is caused by the independent variable, or because of differences between participants. In a strict experiment, it is possible to minimise the chance of one group of participants differentiating naturally against the other group by pre-testing and matching abilities between groups, or by getting a large sample divided at random into groups.
As a laboratory experiment increases control and accurate measurements of variables, this increases objectivity and replicability.
If there is tight control over variables as in the case of a laboratory experiment, then a study can be repeated in order to verify the findings as a laboratory experiment is standardised.
If experiments are conducted in the participant's natural environment (as is done in a field experiment), this makes the experiment more natural with the possibility of having more ecological validity and less bias from sampling, and fewer demand characteristics.
In laboratory experiments, it is argued that not all variables can be completely controlled. In methods such as a field experiment, it is more likely that an external variable will cause bias as not all aspects of an experiment outside of a laboratory situation can be controlled. It is also difficult to determine the cause and effect in a natural experiment due to loss of control over external variables and no direct control over the independent variable (s).
In some cases such as a laboratory experiment, experiments can be artificial in terms of their method and reduce ecological validity (we cannot generalise the findings to reality).
Such an artificial situation can cause participants to respond in a different way to normal and create effects such as the researcher effect and demand characteristics. But, trying to overcome the researcher effect and demand characteristics can cause the researcher to break ethical guidelines. A field experiment reduces the researcher effect and demand characteristics, but there are disadvantages of using this method such as lack of control over variables, participants and lack of informed consent. Such experiments can also be expensive and are difficult to repeat.
The investigation of correlation is convenient when the ability to control variables is difficult, impossible or unethical as behaviour does not have to be controlled. Correlation can tell us about the strength and direction of a relationship and can identify a pattern between variables in order to inform further experiments. It can also quantify data from observations and give a correlation coefficient as a result.
Cause and effect can not be established as two variables can be related because they are affected by a third variable; only linear relationships can be measured through simple correlations. The relationship can be described better when a third variable affects both variables (Confounding variable)
Interpreting correlations can be difficult as we must accept that there is a possibility the relationship between any two variables will not be meaningful as both can be affected by a variable which hasn't been considered or measured.
We often see that the natural observation method is free from the researcher effect and demand characteristics and has high ecological validity, especially if the observation is covert. A natural observation can produce ideas or validate findings from experimental studies. Sometimes, we see that natural observation is the only ethical or practical method.
In a controlled observation, there is more control over the environment which leads to more accurate observations.
It is thought, in a participant observation, that the natural behaviour of individuals is observed without the need to be covert. The method also overcomes the ethical problems of a covert observation and collects detailed and very in-depth information and data.
When an individual is not aware that he or she is being observed, as happens with a covert observation method, ethical issues arise from lack of informed consent. If the participants are aware that they are being watched, they won't behave normally, which is also a weakness as there will be bias in the data.
Cause and effect can not be established in this method either. An observer may have a bias, confounding variables can not be controlled and repetition is a problem, especially in the natural observation method as well as the controlled observation method.
Methods of questioning individuals.
In general, interviews produce large amounts of detailed data particularly about beliefs or inner thoughts. They are easily quantified, are reliable, with the ability to be repeated and the findings can be generalised. Some types of interviews such as semi-structured interviews are quite flexible, sensitive and can be easy to analyse. On the other hand, unstructured interviews collect detailed and very valid data and are natural.
Interviews depend on self-report data which may be untrue. Cause and effect can not be proved. The fact that semi-structured interviews are flexible can be a weakness because phrasing and timing can reduce reliability. Answers to open questions are more complex to analyse.
Structured interviews have less validity and can disregard data because answers are closed and restricted.
Questionnaires gather large amounts of standard data which are quite quick and convenient to collect. They are very repetitive for the researchers and are easy to score apart from open-ended answers. Tests such as IQ tests are easy to score.
Questionnaires lack flexibility and are based on self-report data which has been biased by motivation levels. Producing a test which has high reliability and validity can prove difficult.
Questionnaires can be biased when individuals give answers that are socially desirable and respond in a particular way.
Data with high ecological validity can be collected as case studies gather very detailed and in-depth data which is sometimes missed by superficial methods. Often, case studies are the only suitable method of studying behaviour and is often the only possible method due to rare behaviour e.g. Genie's deprivation (Curtis, 1997).
As case studies gather in-depth data about an individual, the data can not be generalised as the sample is not representative of the entire population. This method is expensive and time consuming which means it is impossible to repeat.
Cause and effect can not be proved 100% and the method has low reliability as a number of case studies involve recalling past events and can be open to memory distortion. The researcher's lack of objectivity can also reduce the reliability of the results.