Bias in SONA

As we all know, SONA is the system by which much of Bangor’s Psychology research recruits its participants. However there are several problems that arise through the use of SONA and they are due to bias that occurs from the way SONA is designed. For example, an obvious thing to point out would be that SONA relies heavily on the student population of Bangor. Although the student population is an incredibly useful tool for research there are problems associated with using us, such as sampling bias. There is also a problem that exists in the was SONA is actually structured, the fact that we select which studies we participate in can lead to selection bias. Both of these bias’ could potentially have serious effects on the results of a study.

The fact that SONA relies on Psychology students causes a major amount of bias to enter the data, more specifically, sampling bias. Sampling bias occurs when the sample taken and researched is not random (http://www.medilexicon.com/medicaldictionary.php?t=10087). This means, for example, that if a researcher conducts a study and aims for the results to be generalised to a large population but only middle-aged women are tested, the results will be biased towards that group. It becomes apparent now that studies using SONA encounter a problem, they have a sampling bias towards students. This is a problem because students are generally WEIRD-er than a random sample from a population. Weird is an anagram for Western; students studied on are usually from Western European and North American cultures. Educated; due to the required qualifications to become a student, we are, as a group, highly educated. Industrialised; this refers to our desire to attain a highly paid job after we complete our time as students and are studying to attain this goal. Rich; comparative to much of the developing world, students are relatively well off. And Democratic; this to me is much the same as Western, as most of the countries included in that umbrella are Democratic and the population is raised as such. This idea was originally created by Henrich et. al. A summary of their paper can be found here (http://neuroanthropology.net/2010/07/10/we-agree-its-weird-but-is-it-weird-enough/). Coupling the implications of sampling bias with the assumptions of “WEIRD” students, it can be suggested that Psychological studies that rely on SONA suffer from a huge selection bias, this can affect the external validity of the study with extreme cases resulting in the study being no longer generalisable.

There is another type of bias that creeps into SONA studies is selection bias. Selection bias is an error that occurs due to the type of people who sign up for a study. For example, report rates of child abuse are susceptible to selection bias as differing cultures have differing views of what should be reported (http://www.ncbi.nlm.nih.gov/pubmed/9504213). Selection bias in studies means that people sign up for studies that they wish to participate in and this can affect the results (if you search wikipedia for participation bias there’s a funny example, but when I references wikipedia last time people laughed at me so I won’t link it here). With regards to SONA, when we are fulfilling our credits we are allowed to choose the specific study that we are interested in completing and often choose studies that we think would be easy or rewarding. If, for example you suffer from fatigue when viewing stereo 3D images then you would understandably be less likely to sign up for a study which was researching these images, this could be detrimental to the study if, for the purposes of the study the researchers needed to include participants who suffered from fatigue. Again this bias can affect the generalisability and the external validity of the study. This selection bias could really only be avoided in SONA if we were randomly assigned to studies as this would ensure we were not choosing only the studies that appealed to us (I also know this is wouldn’t work for various reasons, but it would deal with the bias).

Despite the great use SONA serves Psychological research there are problems that can arise through its use. The pool of participants it uses is comprised of students which means that results will be biased towards the student population. This can cause problems with external validity and generalisability, which ideally would not occur. Furthermore the ability to pick and choose which studies suit us the best can potentially cause selection bias which again can cause external validity issues.

About these ads
This entry was posted in Uncategorized. Bookmark the permalink.

12 Responses to Bias in SONA

  1. kfh1991 says:

    Yes, SONA does have limitations. But I think it’s a good way to introduce undergrads to the world of research, and give them a bit of an insight into how it works. Compulsory participation means that you have to sign up for something, and give it a go. Which may broaden your horizons when it come to choosing what kind of area of research you would like to do yourself for your own third year project. It also gives the budding psychologists the experience of being a participant themselves, so that when they come to setting up their own studies, they will be more sensitive to the needs of those taking part, rather than just focusing on what results they would like. SONA is a means to an end, and I think it serves it’s purpose well whilst also benefiting both those doing research, by giving them a readily available sample population, and those taking part, by allowing them an insight into new research. This study by Padilla-Walker et al. also found that when research participation by students was only voluntary for extra credit, those who did participate were scored higher on measures of academic performance than those who did not participate. (http://www.tandfonline.com/doi/abs/10.1207/s15328023top3203_2)

  2. danshephard says:

    Great blog
    University studies are basically all biased I agree and there is much evidence to support this. Smart, Reginald G, (1996) Subject selection bias in psychological research demonstrates this in a meta-analysis of The Journal of Experimental Psychology, the Journal of Abnormal and Social Psychology. The results showed that 73% of the studies in The Journal of Social psychology used college students where as the Journal of Experimental Psychology had 85.7% of studies using college students. In addition to this there was a gender bias with mainly male students enrolling in experiments. In addition to this only a few studies used samples that used a general adult population. So what does all of this result in? A bias in terms of age, social class, learning ability and other factors. As Rosenthal and Rosnow (1969) put it psychology research is the ‘psychology of psychology students and sophomores’.
    While it is clear that there is a selection bias and sampling bias some of the most ground breaking studies had these problems. An example of this would be Milgram’s (1974) obedience study, which used flyers to collect a sample. The self-selecting sample in this study produces a selection bias as the participants who are likely to sign up are probably going to be highly motivated individuals. Even the most respected psychologists such as Loftus and Palmer’s (1974) experiment on leading words, used a student population. Just like these studies most studies are likely to have sampling biases or selection biases. If it was possible to use a representative sample for every experiment that would be great as it would increase the generalizability of the results. However, this is just not plausible due to the amount of funding that would be needed for every experiment.

  3. statsjamps says:

    A very well thought out blog and interesting points. But have you thought that this might be the case in other research studies across the population and just not in SONA studies and that it is just a part of research studies. For example within in a study conducted by Barr and Hayan 96* they looked at deferred imitation and used 18 months olds only so therefore this can only be used to the 6 month old public like with students in the SONA studies.
    You suggest that within sona participation we have a choice when deciding what study we want to take part in. Well isn’t this the same for other studies conducted in other fields. Barr and Hayan 1996* also recruited through looking at birth times and areas and then phone the parents. Some researcher may advertise that the need participants and therefore the general population choose whether they want to take part; therefore recruiting in a similar ways to SONA recruiting.
    Also suggesting that coming to university we have a better understanding academically and therefore are biased with getting better results. This is true academically but myself coming to study psychology at university had little understanding of it as never been taught so when being a participant within the research is it fair to say that I would have the same or similar knowledge to a English university student who also has no prior knowledge in the field of psychology.
    Also you suggest that people may sign up for studies which interest them and avoid ones that they have no interest in. this may be true for a wide percentage, however many students who leave credits until the last minute may not have a choice if they want to gain the full credits for the due date.
    *The effect of Event Structure on imitation in infancy: practise makes perfect? Barr and Hayne 1996

  4. Pingback: Comments for this week! :) « statsjamps

  5. bangorlc25 says:

    A very interesting and thought-provoking blog. One part that really got me thinking was your point about the WEIRD sampling nature of using student participation in research. Because even though you have discussed it in a good amount of depth, i would just like to draw your attention to some of the interesting facts that have arisen from the so called WEIRD population. Students at the University of British Columbia published research citing the dangers of assuming that humanity closes matches that of university students. According to their research, between 2003 and 2007, 96% of all psychological samples had come from countries that made up only 12% of the world’s population and that undergrads made up 80% of study subjects in six top psychological journals. That to me is an astonishing figure. I think that you have really brought up such an important point here because, like you said, the problems relating to sampling bias stretch beyond SONA and it is indeed a wider problem for psychological research not only in this country but in the US as well. In my opinion, this has to be fixed because we can not assume results from student participation can be related to the wider population but unfortunately due to the accessible nature of students i doubt that will change any time soon.

  6. giggles20 says:

    Excellent blog, perfect subject to discuss with Bangor University students :D. Yes I agree that there are a few limitations with SONA as a experimental tool to use in research. For example SONA is open to the majority of Psychology students within the University but what about the remaining students who do not study Psychology? why do they not get a chance to participate? surely they would provide more valid data then the Psychology students who i am guessing know what the experimenter wants and knows the correct behaviour to portray. Therefore, may produce a set of bias results.

    On the other hand, I think SONA is a useful tool in terms of helping (psychologists to be) understand the full process of how to debrief participants, how to go about testing participants and the different experiments which can be done with the participants and most importantly how to receive informed consent!.

    Overall, I would like to think that the Psychology students in Bangor participate out of good will and not only to receive their credits! I think on the plus side of things, there are a lot of psychology students which means that random sampling is definitely on the cards. It may not be completely generalisable towards the public (due to the fact we are all students) but I think using SONA provides a good stepping stone/indication of the potential for future research into the many different areas of Psychology which are studied within Bangor University :)

  7. Pingback: comments for T.A 14/03/2012! | jessica0703

  8. psucab says:

    I hadn’t even considered that by us choosing the experiment this could effect the results!
    Some very good points. Despite all of Sonas faults I do believe that it is a great way to get students involved in the field of research. I would like to think I would participate in some experiments without the credit scheme…but in all honesty I don’t believe I would have participated in half as many.
    I suppose by volunteering (in a way) to sign up to these studies it eliminates investigator bias..however I believe that there are disadvantages this type of sampling (volunteer). This bias must be recognised when analysing ones results.

  9. Pingback: Homework for my TA | psucab

  10. psud87 says:

    yes sona could be bias,you talked about selection bias and sampling bias.but the truth is does it serve its research purpose in this varied study.i think it does and that takes care of sampling and selection bias .on the long run each research studies repeted over year can reduce error in getting a finding likened to each study.

  11. Pingback: home work for ta week 8 comments | psud87

  12. nat1990psych says:

    An interesting and insightful blog demonstrating the flaws in the use of SONA in recruiting participants, and I generally agree with all the points that you made, although there are some clear advantages of the system that I will note later.

    The first point that you noted is the fact that SONA is subject to selection bias, as students are free to choose which studies that they would like to participate in. As a result, people with similar attitudes towards a particular topic are likely to volunteer, thus resulting in the results being somewhat difficult to generalise to the population as a whole. Furthermore, students may choose topics that they have an academic interest in and therefore have a significant amount of knowledge into what is being investigated. Due to this fact, they may confound the results through demand characteristics, thus increasing the risks of a Type I or Type II error. However, the alternative is that students are forced to participate in studies that they would not ordinarily want to volunteer for meaning that they are somewhat reluctant to take part. An unwilling participant is not optimal for experimentation as they will be far more likely to withdraw early or respond without paying any attention to the task (Coulter, 1986). In this sense, it is far better to let participants to select what research they would like to be involved in as opposed to forcing it upon them.

    You also raised the fundamental issue of subject bias in the recruitment of participants via SONA. As you stated, students are often well-educated, of a particular age group, from Western civilisations, and of a higher socio-economic background (Wiederman, 1999). This means that it is often extremely difficult to generalise the findings from a study using SONA to the population as a whole, thus meaning the external validity of the research is reduced. However, the demographic characteristics of the student population is currrently in a period of change, with more mature students and individuals from lower socio-economic backgrounds being encouraged to pursue further education. This does not entirely diminish the problem of the lack of generalisability, but it does go some way in improving it.

    It would have been beneficial to also include the advantages of using SONA. One of these advantages is the fact that students are easily accessible as students have to compulsory partake in research in order to gain course credits. This means that researchers have a ready supply of participants so do not have to waste the time and money recruiting participants outside the university.

    Another advantage of using SONA to recruit participants are the benefits that the psychology students participating in the studies can gain from the experience. SONA gives students the chance to experience what actually happens in a scientific study. This experience will prove to be invaluable for students who will be able to gain a clearer understanding of how to conduct studies and information regarding psychological topics.

    Overall, there are clear advantages and disadvantages in the use of SONA however, it is currently the optimum way for researchers’ to investigate phenomena.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s