University undergraduates are regularly used as research participants in Psychology; students are often willing to be involved for the benefits received from researchers or for course credits. Samples of university students can prove to be useful for researchers, with greatly varying interests and areas of expertise, as well as often being culturally diverse.
However, there are issues with student samples in research. Although such a diverse sample can often be achieved from universities, there may be far too many similarities to generalise findings to the population. University students are academics, and it should not be assumed that the behaviours of these can be generalised to the rest of the population. People who choose an academic career may think differently, and have very different philosophies than those who take a more vocational o approach. The vast majority of undergraduate university students are aged between 18-23, often considered to be an age where the individuals are rebellious and most likely to engage in risky behaviours
An example of the use of university undergraduates in research is the Clark and Hatfield study Gender Differences in Receptivity to Sexual Offers. The methodology involved a young, attractive stooge approaching students on a university campus and asking them one of three questions: ‘Would you like to go out tonight?’, ‘Would you like to come back to my apartment?’ or ‘Would you like to come to bed with me?’. They found that both sexes were equally willing to accept a date, although women were less likely to go back to the stooges apartment (6% of females agreed, 68% males agreed) and no women agreed to go to bed with the stooge, whilst 75% of the males agreed. Although this is an interesting study, the use of university participants when studying sexual behaviour lacks population validity. As mentioned before, university students are more likely to engage in risk behaviours than many other groups – and so to draw conclusions from this would be incorrect. Furthermore, the results of this study may unfairly misrepresent males. Between the ages of 18-21, males are considered to be at their sexual prime, and so more likely to engage in sexual activity, and so results should not be generalised to all males.
The accessibility of students for the purposes of research does make them an obvious choice of participant. However, it would be best to support any results with similar findings from different populations.
Ref.
Russel. D. Clark., Elaine. Hatfield., Gender Differences in Receptivity to Sexual Offers.
Retrieved from: http://www.elainehatfield.com/79.pdf
elliespsychblog 
Any piece of reasearch conducted is bound to hold a bias of some sort. As you pointed out, using university students may not always be the best participants to use as the results found may not be applicable to the wider population.
We might also want to consider that the increase in use of substances (some legal, some not so legal) cab also alter peoples judgements and they may make decisions which they otherwise wouldn’t have made, had they been at home just a few months before.
The timing of the study may also have been a factor in the students behaviours and answers as if it was in the first few weeks of starting term, they may be more susceptible to caving in on the social desirability bias, because they don’t want to be branded as a “wimp” (and other names, that one seemed most appropriate!) Overall, using university students are useful for a number of reasons, however, if we were to generalise the findings to the whole population, we should probably consider using a sample from another group of people in conjunction with the results of the university students.
Agreeing with elliespsychblog, it is best to confirm any results with data from a different population, as it can be biased to use students, if only a student population is used. Many students are academics, and are more likely to participate in studies related to their own subject choice. Perrin and Spencer (1980) recreated the conformity study by Asch (1955). They found that only 1 out of 396 trials conformed as opposed to the 36.8% found to conform in Asch’s original study. The students used by Perrin and Spencer were science students, who may have been more self-confident about judging line length, or have known about the original study, creating demand characteristics. Often students can know more about the research than the rest of population which can lead to difficulties in generalising results. Peterson (2001) would stress the importance of repeating studies upon non college based students before generalising any results found. However, this would not only apply to students, any research should be representative of the whole population when generalising results rather than focussing on one demographic.
References:
S., E. Asch (1955) Opinions and Social Pressure. Scientific American, 193, 31-35.
S. Perrin., & C., P. Spencer, (1980) The Asch effect: A child of its times? Bulletin of the British Psychological Society, 32, 405-406.
R., A. Peterson (2001) On the Use of College Students in Social Science Research: Insights from a Second Order Meta-analysis. Journal of Consumer Research, 28, 450-461.
The use of a sample of students in a study would only really seem generalisable to a wider population if it was simply just a study of students. Even this would have to be tested at a variety of universities though, because there is a wide range of students who go to different universities for different reasons. The use of students, as you said, is not generalisable because of the large difference in age, political views and general mindset compared to the rest of the population. The use of students could be favoured though because of how willing the students are for the rewards or benefits they receive.
It is highly difficult to avoid any bias at all within research, as the sample is always going to have at least one thing in common. However, this as mentioned, academics will have similarities with each other due to age, intellectual and also motivational levels. Even more so with credits, as most people choosing to take part in the research for this reason will also be studying the same area. However, due to the different locations students come from it takes away the bias of collecting people from the same area. The only real way of improving the bias in sample selection is by conducting it at locations of random gatherings of people from different areas and economical levels. This is done in the Piliavin study where the stooge pretended to collapse on a busy train. This means people from different areas could take part in the research, reducing the selection bias. However, even this has it problems as people have the similarity of being able to use the same train, where as other possible participants which could have affected the results in a different way was not able to take part as they have been busy.
Piliavin, I.M., Rodin, J.A. & Piliavin, J. (1969) Good Samaritanism: An underground phenomenon?
In studies, it is often the case that the best option is the easiest option that can get hold of as many people as possible quickly through universities. It is extrememly hard then to be able to commonise the results of a study to the whole population under scrutiny. Unless of course it is in our case, students studying psychology in Bangor. We can not get a full picture at all while trying to understand the working of the mind and behaviours. In the search of a sound hypothesis that has stood the test of falsifying, we need to use as many participants from as wide a variety from as many different backgrounds possible. By using university students we reduce this variety considerably.
One question that does arise about using university students is the ethics behind it. Students are forced to be part of researches as part of their studies and are given credits or printer credits for being part of them. Is ths ethical to do so? Students have a lot of heavy work to do already (I define this by the kind of work and the volume in each assignment and not according to the amount of assignments as such) so how should students be able to find time and resource to do extra studies as well?
With any group of students being used as participants, one must try and organise them so the reflection they give is as accurate as possible of the population they represent. A major factor that should be considered is the subject the students are taking and what year of their degree they are in, indeed what kind of degree they are taking. To illustrate the point further, for subjects such as Psychology, Biology and History (all subjects available at most sixth forms and colleges) the spread of socio-economic background for students varies greatly. However for students studying more highbrow and traditional academic subjects, such as Law, Classical Civilisations and English Literature the socio-economic average goes way up, where as with subjects such as Engineering, Business and Management Studies and Geography, the average goes down. Even the year of their degree must now be considered, as with Tuition fees rising, those at university after 2012 will tend towards the higher economic scale of income, as this now would be an indirect requirement for further education (BBC, 2010). All of these aspects, the course taken, the socio-economic background, the year in their degree can influence behaviour and personal opinions. Being such a large factor of university student life also, would you not expect this to be considered?
BBC, (2010). as found at http://www.bbc.co.uk/news/education-11677862