Homework for TA

Approaching significance

 

Does “statistically significant” mean there is definitely an effect?

 

What options are there when you do not find significance, despite basing your work on a solid theory?

 

Internet Research: Yes or No?

‘All you need is… money!’

In this weeks blog I’ll shall be using a video to address the issue of money in research. The video is an interview with prof. ‘Brian Cox’ who is a current physicist and is fairly popular in the science world. The video outlines the context for which current spending in UK research is in relation to capital of a big corporation- Apple. Brian explains that in the last 25 years £3bn has been spent on research in the UK and that the money has been used efficiently. In this time things such as the worldwide web by CERN has been created. Also there has been advancements in aero-engineering and health care increasing life expectancy. Apple at the end of last year had £61.4bn in the bank. This just puts in context the enormity in the difference in spending and if that sort of money was spent on research how much further would we be? Would we even need half the products Apple sells?

 

Brian says that the reason for this imbalance in investment is due to the government. The government is the responsible for distributing the capital where it is most useful- this is also supported by other evidence (Webber & Wildavsky, 1986). However we elect a government based on there manifesto for which with either agree or disagree- we choose them to represent what we want as the people. Therefore this means that we drive where the funding for different things goes based on the feedback given in what we demand. The wealth of a nation and where it is distributed is done to ensure that most people are pleased.

 

The research done and things found are heavily influenced by how much is spent. This is shown by history and in times of wealth such as the victorian age many things are invented. For example in the victorian era the light bulb, steam engine and airplane were all invented which changed what we have today. However research is also limited by other factors and so throwing money at it won’t keep working there will be a optimum but currently funding is what limits a lot of research at the moment(Gulbrandsen & Smeby, 2005).

 

Brian goes on to say that investment in science will benefit humanity as a whole and maybe even more than apple is benefitting currently. However it has to be done in the right way. Brian expressed that if research is controlled and with a specific target this can also limit advancements. The research has to be done from a creative perspective if new and exciting things that create paradigm shifts that benefit humanity as a whole. Science need to be able to explore freely the edge of our knowledge to enable maximum discovery. Therefore, despite the share holders being put out, if apple gave a out of there wealth and put some money in to the right sort of research, humanity could be improved and a possible acceleration in discovery. This industrial funding can only happen if there is a change in perspective of research from the general public. A change so that there is a desire for research which would lead to an investment because of the want. Who knows, from the research something could be found that enables apple to bring out a new and exciting product that has never been seen.

 

 

Link for video: http://news.bbc.co.uk/1/hi/programmes/click_online/9707209.stm

 

Webber & Wildavsky, 1986: http://www.mendeley.com/research/history-taxation-expenditure-western-world/

 

Gulbrandsen & Smeby, 2005: http://www.sciencedirect.com/science/article/pii/S0048733305000958

Homework for TA

 

Should psychology be written for the layman or should science be exclusively for scientists?

 

Are theoretical research findings valuable?

 

Applied research findings are more valuable than theoretical findings

 

Is it ethically ok to use internet sources as data for qualitative studies?

‘Looking back to look forward’

History obviously has an effect on the present and the future through its influences, changes and themes. In light of this, history does and should have a lot of influence on the research that is carried out. However the past is often partially overlooked in not discussed in detail in research today.

 

The beginning of a research investigation obviously, and quite rightly, starts with a look at what hypothesis have been investigated in previous research. This is very important and is essential for the development of hypothesis in a scientific way. It enables proposed hypotheses to be justified and be able to be validated. This research is called logical positivism and is dominant in scientific research of today (Goodman & Kruger, 1988). However these are very specific and focus on on research area and almost always just consider ‘scientific’ research in the area.

 

Something that is often missing is a consideration of historical perspective. Historical perspective is different from just a look at history. A historical perspective takes something of today and looks through history to considers how and why it is as opposed to just looking at one part of history- helps understand the present not just that point in history (Lawrence, 1984). This is very useful in research because it is a greater understanding that may lead to better predictions on how things work. This does happen in research but the seems to be a lack of general historic influence and is more scientific research only focused. With psychology, which is a science of the mind, perhaps more discussion about general historical perspective in papers would help.

 

Furthermore, more could be done with historiography in psychology. Historiography is the study of historical writing. Language and so writing expresses a lot about the culture and context of the time. Pieces of writing also say a lot about the way people thought. This therefore is a vast resource for which to gain historical perspective for current psychology theories (Danziger, 1994). Historiography is not new and there has been a lot learnt from it but I feel a lot more could be learnt if historiography was carried out from a psychologists perspective.

 

A look at history to further research is certainly not a new thing, however there can be a lot more could be done with it. By looking at history not only does it show what happened at the time but it reveals so much about what is happening in the world today. This can include psychology and with a vast resource that is historical writing it should be used more on its own and in general research papers.

 

References:

• Goodman & Kruger, 1988: http://www.jstor.org/discover/10.2307/258580?uid=18882&uid=3738032&uid=2134&uid=2&uid=70&uid=3&uid=5910784&uid=67&uid=18880&uid=62&sid=47698745277627
• Lawrence, 1984: http://www.jstor.org/discover/10.2307/258443?uid=18882&uid=3738032&uid=2129&uid=2&uid=70&uid=3&uid=5910784&uid=67&uid=18880&uid=62&sid=47698744797177
• Danziger, 1994: http://tap.sagepub.com/content/4/4/467.short

Homework for TA

The File Drawer Problem

 

Should psychology be written in the layman or should science be exclusively for scientists?

 

Is it ethically ok to use internet sources as data for qualitative studies?

 

SONA System – The best way to collect data for dissertations?

 

 

‘Research reaching people’

Imagine a father telling his young son that trains go at 200mph but no more information is given. The son is likely to think and therefore express that when a train is moving its going 200mph, which we all know is a fairly big exaggeration of the truth. It is only by the father telling the son that a train at top speed goes 200mph and that this can only happen on straight parts of the track and actually trains average about 95mph. As you can see by this short (and not 100% accurate) example that unless all the information is given small parts of the information can be heavily exaggerated. This is what is happening to the media. The scientists are being censored by governments and other interested parties to not reveal all about what research shows which ends up being counter productive as more scare stories come out due to exaggerated snippets of information and not the full story.

 

An example of this is a story that has recently been in the news about the Canadian government muzzling their scientists and not allowing them to be interviewed by reporters for certain stories. A protocol has been put in place by the government that states that a scientist working for the government must be first cleared by officials before they are allowed to be interviewed. This makes sense however this process can take several days and the request is often declined by which point the story has pasted. It is clear from the story that some very important research such as a hole in the ozone layer above the Arctic  is not being allowed to be reported not because the scientists do not want an interview but when posed to the government they refuse. As Professor Andrew Weaver, University of Victoria put it ‘The information is so tightly controlled that the public are left in the dark’. Full report at http://www.bbc.co.uk/news/science-environment-16861468

 

Another example a bit closer to home is the bird flu pandemic. It caused particular scare and confusion for a lot of the public about whether is was lethal; how infectious was it; and what should be done. A report by the New Scientist Magazine shows that researchers at the University of Rotterdam, Netherlands had created a mutated version of the bird flu which could go pandemic and be lethal to people. However they were not allowed to publish the complete research and it had to be censored. The reason for the censorship was to stop bioterrorism, which on the face of it seems like a legitimate reason and a sensible decision by the government. However the part in the research that was censored was that of the chemical make up of the virus which meant other scientist could not research it to find a vaccine. The full release of the research in this case would have had benefits that far outweighed any hypothetical bioterrorist threat. Full report at http://www.newscientist.com/article/dn21311-killer-flu-research-to-be-censored.html

 

I hope that you can see from the two examples above that withholding information when presenting research can be misleading and even dangerous but what are the implications for us as psychologists. Would studies such as Milgram’s and Zimbardo’ s even be allowed to be published today for concerns of terrorist mentally manipulating people (I know this is a little far fetched but you get my point). Lots of research is published and even presented in the media but without trying to sound paranoid is there groundbreaking research that is being hushed by the authorities and if so what are the ethical implications. Should there be such government and other authority controlled about what is presented to the public especially for science which is so close to home such as psychology. What is our obligation as psychologist. Is it to present everything or to do some censorship of ourselves to ensure that snippets of information from censored research is not presented by the media and exaggerated.

Homework for my TA

 

Should We Still Teach Psychoanalysis?

 

Is it Possible to Operationalize Every Variable?

 

Data, data, data!

 

Is psychology research making a differnece?

 

Is it dishonest to remove outliers and/or transform data?

 

Is it Ethical to use online sources in research?

 

The Helsinki Convention and the 5 basic ethical principles

Are hypotheses ever proven?

The answer principally no, but why when evidence supports a prediction.

 

A hypothesis is a prediction of a relationship between two variables (Gravetter & Forzano, 2009). These variables can be anything from ice cream flavours to types of bacteria. The relationship of the prediction can have direction such are when one thing increases so does another or more general such as there will a difference. The predicted relationship can be between anything such as shoe size and height. However such a relationship may be regarded as not plausible and so serve no useful progress in understanding. Therefore a hypothesis tends to, and for it to be scientifically robust then should, be based on evidence from previous hypothesis. Hypotheses are developed, improved and rejected based on supporting and non-supporting evidence but never proven or disproven completely. This is because hypothesis are supported based on probability and not certainty.

 

Tests, in most cases, cannot be carried out on all of the population because it is not possible to ask everyone in the world what their favourite colour is for example and so a sample is taken. This sample is used to represent the population and that the larger the sample size the more representative of the population it is- law of large numbers (Jakob Bernoulli, 1713). If the sample size is smaller a significant effect is harder to achieve thus showing that it is the probability of the sample being accurate to the population which determines whether a hypothesis is supported.

 

Ronald Fisher was the one who made it possible to test whether something was significant. He discovered many of the sample distributions of statistics along with the analysis of variance (http://www-groups.dcs.st-and.ac.uk/~history/Mathematicians/Fisher.html) which made it possible to see if two distributions are significantly different. These distributions were all based on how likely the difference of two samples from different populations would be accounted for by error. If greater than the estimated error then they are significantly different however it is dependent on how accurate the sample is to the population.

 

The probabilities which are used to calculate significance are based around normal distributions with certain levels of error. Therefore all supported hypotheses are just likely to be the case. Granted this probability is low and so the difference has to be fairly rare to be significant but there is a small chance that a supported hypothesis is due to just normal deviation in scores. This throws in to light a question that if nothing is absolutely proven what is the likelihood of a supported hypothesis having had a type 2 error or even an non-supported hypothesis having had a type 1 error?

Homework for TA

“Qualitative research isn’t as scientific as quantitative methods.”

 

Qualitative research isn’t as scientific as quantitative methods.

 

Qualitative research isn’t as scientific as quantitative methods

 

Qualitative vs. quantitative

‘Research Led or Data Driven?’

So you may be thinking is this even a question? Is there even a distinction? Surely all research is research led and data driven. But is it?

 

Firstly lets consider what research led and data driven means. Research led is fairly self explanatory but what does it look like. When research is research led is means that the origins of the hypothesis or question posed is from research in that field already. Research never pops up from no where an is more about development of theories than to completely start from scratch. This in itself is very objective because it means someone has considered a significant about of research before from which to base their current think. Data driven research however is also very objective because development of research is based upon what the measures find and the results. This sounds very close to research led because surely there is data in research?

 

The distinction lies then not in them as separate entities but the order in which they are considered and done when carrying out research. A lot of research is carried out by coming up with a fairly vague hypothesis that is chosen to try and be falsified. The research is done to see who has done it before and what method were used. These methods, or slightly adapted ones, are then used to gather results and see what is found. Conclusions are then drawn based on previous research but what research is chosen is completely down to the experimenter and limited by the experimenters knowledge. This is more data driven research. Other research is carried out by research around a general theme is done and hypothesis develop from interest in the reading. From this a experiment is developed and is a dominant part of the experiment as its also looked at in the conclusions of the study, this is more research led. The key difference is not the research considered but at what point and therefore the influence on the study it has. So which one of these is more objective? And which method of research will give robust results from which to develop further research?

 

I would say that research led style of research is a lot more objective and certainly is less likely to lead to experimenter bias. Furthermore this concept is not really considered by the APA because the regulations are upheld in both styles however they are fundamentally different which has an affect on the objectivity of the outcome. I feel this needs to be look in to further to keep science objective.

 

 

Resources used in this blog:

 

• Research methods for the behavioral sciences book- F. J. Gravetter & L. B. Forzano
• Research methods lectures
• Discussions with researchers