Bias in Research

Psychologists and also the general public a lot of the time rely on research findings to make decisions on varying things such as theories and even their own health.  The worrying thing is that in some cases, these findings are either not entirely true or perhaps hiding the truth, which inevitably will mislead people and therefore cause decisions to be made which will be based on untrue ‘facts’.  One of the main causes for results being hidden or altered is bias in the research; the companies which fund the research a lot of the time would benefit from positive findings.  Therefore they might either purposefully pressure researchers or in some cases alter or hide results to make their product look better and therefore sell more.

One such example of bias in research is with a drug called paroxetine (or Paxil) which is an anti-anxiety medicine.  Results from four of the trials on this drug were suppressed, results which not only failed to show effectiveness of the drug among children and teens, but also which demonstrated a possible increased risk of suicidal tendencies.  The company which suppressed these results, GlaxoSmithKline, were forced to make a legal settlement which meant that they have now established an online registry which shows summaries of all the results of their sponsored studies which have happened after a certain date.  But the fact is that they suppressed results of trials in the first place, just to make their product look better and in the process not telling the public the possible risks.  This example shows possibly one of the worst dangers of bias in research, that people may die simply because results were suppressed.

So what can be done to combat this?  There are steps being taken: for example in June 2009, the FDA’s (Food and Drug Administration) Transparency Task Force was launched, which last year proposed draft plans that would make information about drugs and medical devices much more publicly available.
Although there is this progress, bias in research is still a real issue.  Without funding, there wouldn’t be nearly as much research actually completed, however if companies continue to fund research on their own products, bias will continue to occur.  It’s a sticky business with a lot of issues, so what are your views?


14 Comments (+add yours?)

  1. psychmja1
    Mar 13, 2012 @ 22:25:08

    An interesting topic this week 🙂 and one I had to go and look more up about. Did you know that Diederik Stapel, an ex-social psychologist produced some seriously biased (and extremely false) research right from his imagination it seems. One of the claims that was made was that “eating meat makes a person more selfish”… Erm ok? … And apparently “messy rooms make people more racist”. And people actually believed this because he was what we would call a reliable source. Yet, why do we think that researchers are trustworthy? Well apparently not, as you will see:
    John (2011)* conducted a survey of >2000 psychologists in America and found that 70% of them admitted cutting corners in reporting research findings, 1/3 said that when they had found something unexpected they had actually reported it as if they had hypothesised it from the start, and finally 1% admitted to completely falsifying the data. Well, this doesn’t do much for psychologies reputation does it?! It appears that psychologists can be very biased in research, making the research they present biased also!



  2. Trackback: Comments for my TA « psychmja1
  3. Ellie Simpson
    Mar 14, 2012 @ 01:18:10

    Yes I agree with you that the results of research affect many of the decision we all make daily; from what washing powder we buy to which school our children should attend. Some of these decisions are of little importance but others can be majorly important.

    The following articles emphasise the importance of accuracy in research, as individuals often make life changing decisions on the strength of research results.

    Healthy sisters have breasts and wombs removed to avoid cancer that killed their mother at 32
    Three healthy sisters have had their breasts removed – and two have had hysterectomies – to save their lives, after tests revealed that they had inherited the same faulty gene called BRCA1, and had a 50-60 per cent chance of developing ovarian cancer.

    Sisters both have double mastectomies after seeing their mother and grandmother die of breast cancer
    Two young sisters in perfect health have had their breasts removed to cheat the deadly disease that has claimed the lives of their mother, grandmother and two great aunts. They underwent a genetic test two years ago that revealed they were carrying the inherited cancer gene BRCA2. This gives them an 85 per cent risk of developing the disease.

    I find these two articles extremely thought provoking and absolutely terrifying. Such drastic measures based on the strength of the results of a statistical test, of the likelihood of possessing a certain gene, and the likelihood of actually developing the disease.

    Sorry I am unable to make the links work!


  4. psud46
    Mar 14, 2012 @ 16:29:46

    Another form of bias is one which we as experimenter’s are in danger of introducing to a piece of research. Sackett (1979) noted 56 varying type’s of experimenter’s bias that may influence results throughout the research process. One example of such is the observer-expectancy effect, in which a researcher may accidently convey their desired outcome of the study to a participant. However, since the research was conducted, it has become increasingly often for researchers to use individuals naïve to the nature of a study to gather data from participants.


  5. psud0b
    Mar 14, 2012 @ 16:48:49

    Although it is possible that companies funding research into their own products are obviously likely to try to change/give misleading results, this isn’t the only way in which results can be biased. They can be biased by the experimenter themselves – if the experimenter expects certain results, they may be more likely to see what they expect rather than what is happening (for example if they believe that boys are more aggressive than girls, they will probably pay more attention to boys’ aggressive behaviour and ignore girls’, in order to disprove their null hypothesis). Rosenthal & Fode (pg 178 in Research Methods for the Behavioural Sciences) investigated rats learning mazes – some people were told that their rats were good at learning a maze, other were told their rats were bad (all rats had the same capabilities), but the participants who thought their rats were ‘maze bright’ treated the rats differently, which then led to the rats actually performing better than the other group’s rats (who were believed to be bad at learning mazes). I thought this was a really good example of how easily particpants/experimenters can accidentally effect the results of a study – the participants weren’t actively trying to make their rats better at learning mazes, but it happened anyway. This is obviously very dangerous in research, because it will completely distort results and lead us to believe something that may not necessarily be true! A good way to avoid such problems is using double-blind research studies, which means that researchers and participants aren’t aware of the predicted outcome of the study.


  6. groundblogday20
    Mar 14, 2012 @ 19:26:09

    I remember hearing about this at A level and being horrified that this was allowed to happen and that GSK even tried to cover it up and this blog prompted me, like other people who have commented, to look it up once again. Aside from the medical side effects which are fairly numerous there were also allegations of ghostwriting taking place (Roehr, 2011), meaning that the researcher did not (or in this case was not allowed to) write up their own study. Whether or not this is true it did make me think about whether not ghostwriting should be allowed. David Healy (a psychiatrist at Cardiff University) estimated that at least 50% of papers on drug trials are ghostwritten so obviously it is a common phenomena and many of the issues involving Paxil have since been atrributed to the ghostwriter rather than the researcher suggesting that it can be a serious problem. Yet researchers have shown little movement towards eradicating ghostwriting or at least making it more transparent who has contributed what to an article (Collier, 2009).

    This is unrelated to the rest of my comment but there’s more info on the Paxil story at this link if anyone’s interested:


  7. exactestimates
    Mar 14, 2012 @ 20:41:57

    You make several good points here, especially that money should not dictate whether research is conducted or stopping results being published for one’s own gain or reputation. I believe that for an article or study to be scientifically published, they should be conducted by third party establishments, that would not care one way or another about the results as it would not affect them. As this could be prone to bribes etc, it should also be conducted as double blind studies (Tsui et al, 1986) so that the researchers do not know what conditions they are testing so no bias can occur. Any that are not published in this way should just be ignored, from the layman to experts in the specific field.
    Awesome blog though! Loved it 🙂


  8. Trackback: Comments for Blog 3 (14th March) « psud0b
  9. Trackback: Blog Comments for Shanti, Week 8 | jameezio
  10. Trackback: Comments for my TA – 14.03.12 « groundblogday20
  11. ellislee15
    Mar 14, 2012 @ 22:00:19

    I think people just need to be more wary about what they are reading and who is funding the research behind these so called ‘facts’. A big example of this is Andrew Wakefield, he wrote a paper saying that the MMR vaccine caused children to develop autism, but he broke ethical guidelines and has major biases in his work. There is actually no evidence at all that the MMR vaccine can cause autism, but because of this one guy many parents decided not to get there children immunised, which is just stupid. Which brings to the point of how easily research can manipulate your views as you believe it to be scientific and true. If you want to read more about this there is a whole wikipedia article on it –


  12. stach22
    Mar 17, 2012 @ 11:46:28

    Funding Bias is definitely a tricky one- with the fact that for studies and research to happen they have to be funded by an outside source the majority of the time. It a bias that unfortunately occurs more than we know. When researching this i found an article in the washington post ( In many cases with publication bias it is not to do with researches changing the results of a study but rather manipulating the methods or testing against a condition that they know does not work effectively, or a treatment that is long and time consuming. An example of this is when pfizer conducted research on therapy for depression or drug treatment. They found that they’re drug worked better than therapy but on closer inspection it was found that those that were assigned to the therapy condition dropped out and affected the outcome.


  13. kmusial
    Mar 27, 2012 @ 08:25:32

    Media often uses statistics to advertise products. However there many cases in which the results of their experiments are not honest. This is because the media researchers interpret their data to get the results they want, rather than to get the true, accurate, and unbiased results. This is the word play which is not really fair. According to Mark Suster’s research, almost 75% of statistics in media are modified to some extent, or even made up. What puzzles me further is that 79% of people agreed with the statement “Statistics can be trusted to give an accurate description of the facts”.


  14. Trackback: (FINAL) Comments « kmusial

Leave a Reply

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

You are commenting using your 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

Enter your email address to follow this blog and receive notifications of new posts by email.

Join 18 other followers

%d bloggers like this: