Monthly Archives: October 2011

homework for my TA

Standard

http://francesdevine.wordpress.com/2011/10/14/is-it-wrong-to-remove-outliers-from-data/#comment-10

http://roisin07m.wordpress.com/2011/10/12/week-3-wildcard-blog-is-it-dishonest-to-remove-outliers-from-data/#comment-20

http://abbydoesmath.wordpress.com/2011/10/14/declaration-of-helsinki/#comment-22

http://bar;ettdan.worpress.com/2011/10/14/an-example-of-statistics-in-the-real-world-and-its-effectiveness/#comment-14

 

Advertisements

Adoption rate decreases by 5%

Standard

For this week’s blog I am going to talk about adoption and the increase of the amount of children in care. This topic has recently been on the news and is quite a close topic to my heart as my parents have worked for social services for as long as I can remember. They used to do respite foster care which means the children come on regular visits; for example, on weekends, in order to support parents who need it. Adoption, in particular, is close to my heart as I do have four younger siblings that are adopted.
Last year, 2010, 3,050 children from care were adopted. This has in fact decreased from the previous year. The majority of the children adopted, 71% or 2,170 children, were aged between one and four years old. There are different reasons why children are adopted; out of all the adopted children last year 72% were because of abuse or neglect. It is absolutely disgraceful that people that are willing to abuse children even have them in the first place. When stories are on the news or I hear stories on this topic it literally makes my skin crawl. I get so angry that people would treat another human being in such a way.
The problem and main question is why is the adoption decreasing? The adoptive system is like sifting through mud with all the formality, testing and meetings. I completely agree and understand that each child needs a good and stable home to go to and of course the adoptive parents need to know what they are getting in to because I know it isn’t easy. But at the same time there are thousands of children who need a home to go to but spend months and months in care that do not need to. There are so many people out there who want to adopt but drop out of the system half way through because of the difficulty of it and frankly I think the system should be revised and concentrate more on the aim of the process and not nit picking on rules.72.4% of children last year, after the decision that they should be placed for adoption, were adopted within 12 months. I personally think that is a long time in care especially at a young age in the context of development and attachment.
There is more information on this topic on http://www.adoptionuk.org/information/103152/e_factsandfigs/ . As I said this is a topic close to my heart and do get very emotional when topics like this are on the news. It is interesting that there has been a 5% decrease in adoption over the past year and I would like to know your opinions on this and the amount of children that are in care.

Homework for my TA

Standard

I commented on these on Tuesday 11th October (I think):

http://mballen91.wordpress.com

http://hb90.wordpress.com

http://larabarker.wordpress.com

http://barlettdan.wordpress.com

http://statsblog2011.wordpress.com

http://roisin07m.wordpress.com

and on the 12th october:

http://lon03.wordpress.com

and on the 14th october:

http://abbydoesmath.wordpress.com

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

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

Standard

Well first of all, an outlier is an infrequent observation, they can very often be extremely different to the rest of the data. As an outlier isn’t consistent with the rest of the data meaning that they can completely distort the true result of the research.

Outliers can have massive effects on the mean which can sometimes cause validity issues. The reason why it can cause validity issues is if you have one outlier which pulls the entire data sets mean higher, the outcome may not be entirely valid if this ‘rogue’ data point is left in place. There is a graph which shows this very clearly on http://www.statsoft.com/textbook/basic-statistics/#correlationse. On this example, the majority of the data is clustered quite low except for the one extremely high outlier. I would say that it would be perfectly acceptable to remove this data point as it majorly affects the mean, which in turn affect the validity. However, if there were more than one outlier (especially in a small sample) it is questionable whether they should be removed or whether there is a reason why they are unexpectedly different.

Sometimes it is necessary to remove data when it falls outside of a certain boundary. That may seem as if the data is being manipulated, however, in research involving reaction times it may be necessary to exclude reaction times that are under 200ms. This is as it has been shown that this is the time it takes for the brain to process the information. On the other hand,you may not always want to do this; for example participants may get into a rhythm and anticipate their reactions. This could be an important thing to show in research, however, this does depend entirely on the hypothesis and what the researcher is looking for. It can be a tough choice when deciding whether to exclude data of whether it is significant and should be left well
alone.

In conclusion I do not think it is dishonest to remove outliers as they can cause validity issues and distort the true outcome. As for transforming the data, invalid data points and outliers can naturally do this, therefore I think it is perfectly fine to remove invalid data if it is obviously not valid in order to present the true outcome of the research and not a distorted version of it.

Do you need statistics to understand data?

Standard

I think the obvious and short answer is ‘of course’. Statistics make research and data understandable for everyone. Statistics can be daunting at times but are needed to interpret data and get some kind of meaning from it. If we did not use statistics to explain and represent the data then how would people be able to share the results it shows? People sometimes say ‘the data speaks for itself’ but I don’t believe that’s true. Without the mathematics (and yes there is maths involved, unfortunately) we would not be able to interpret any of this data. Even if you want to know the mean/average basic maths is involved. After these calculations and tests the data does tella story.

I did a quick search online on this topic and found this link http://www.emathzone.com/tutorials/basic-statistics/importance-of-statistics-in-different-fields.html, it helped me realise how many areas statistics are used in. As I mentioned in my last blogs that statistics are extremely beneficial and essential in many cases but I did not realise fully at that point at the extent of areas they are used in. Without statistics we would not be able to understand data. Statistics are used within many aspects of life, therefore without them general life would not be understood. The types of fields that use statistics in order to
understand the data are business, economics and of course sciences. Statistics in business are used to be able to understand the market and understand what the public want. And how would they know what they need to produce without conducting research and using statistics to understand data? Similarly within the field of economics it is essential to be able to understand the data to know the demands of the specific products. Also statistics within this field help to understand things like profit and inflation rate. Obviously statistics are essential in the world of science. Statistics enable people to understand what scientists are interested in and generally do provoke various attitudes toward their research. Statistics are also used to understand development in health and illnesses and help us as a society understand which illness’ are
becoming more widespread and help us to try and prevent future increase.

So as I said before I do believe statistics are needed to understand data and are beneficial in many areas of research and every day aspects such as economics.  Statistics are obviously beneficial in science, as I am sure you all know and have huge effects on the public’s opinions about certain topics, such as health changes.