Blog comments-18/04

17 04 2012

http://psucfa.wordpress.com/2012/03/25/is-it-ethical-to-use-animals-as-subjects-when-conduting-research/#comment-72

http://kfh1991.wordpress.com/2012/03/25/can-correlation-show-causality/#comment-70 (I made a typo in this comment and corrected it in a reply!)

http://psuc0e.wordpress.com/2012/03/25/feeling-the-future-experimental-evidence-for-anomalous-retroactive-influences-on-cognition-and-affect-a-fancy-title-for-an-inaccurate-study/#comment-35

http://pipwinstone.wordpress.com/2012/03/25/correlation-does-not-imply-causation/#comment-53

Advertisements




Final blog?!?!

25 03 2012

Ok so for my big finale, I’m gonna write a blog on…………………………………………………………………………case studies.
Like I did with my last blog about sociometric tests, I’ll explain what a case study is, and go through the pros and cons of using them to collect data.

Case studies are basically the intense study of one individual. Their routines and life are deeply analysed so as to try and find an explanation for their behaviour. Freud used a lot of these while developing his work and theories. A famous example is the case of Little Hans, but I’ll talk about that later.

There are many different types of case study:

  • Explanatory-used to do causal investigations
  • Exploratory-a case study that is used as a prelude to more indepth research, which helps researchers gather more information to develop their research question
  • Descriptive-the case study is carried out after a research question has been proposed; data collected is compared to what the researchers expected to find
  • Intrinsic-a case study where the researcher has a personal interest in the case
  • Collective-a group of people are studied rather than one individual
  • Instrumental-the individual or group being studied allows the researchers to understand more than what can just be observed

Case studies can either be prospective or retrospective. Prospective case studies are when an individual or a group of people are observed over a period of time to determine outcomes, such as the progression of a disease. Retrospective case studies look at the history of an individual or group, to see what caused the disease for example.

There are many different ways to collect data in case studies. The most common method is direct observation of the participant(s), but they can also be interviewed, documents such as letters and newpaper articles and archival records such as census records can be looked at, as well as physical artifacts that the participant can be observed using. The researcher can also take part in participant observation, which is when the researcher is involved in serving as a participant in events and getting a true insight into what goes on.

One of the most famous case studies (and the favourite experiment that was taught at my college) is Freud’s of Little Hans. Little Hans had an irrational fear of horses, so his father wrote to Freud seeking help. Freud interviewed Little Hans, and received letters from his father describing his behaviour and dreams he was having. Freud explained that Little Hans was going through an Oedipus complex, and so feared his father, but because he didn’t WANT to fear his father, he displaced the fear onto horses, as they reminded him of his father.

Case studies can be very useful, as they can give us an insight into why an individual acts and behaves a certain way. Because we are able to only study one person, the data collected can be very indepth.

However, one of the aims of case studies is to generalise the results to other people, but this can be difficult when only studying a few, or even one person. It would not be possible to explain all phobias of horses as Oedipus/Electra complexes; someone may have been injured by a horse when they were younger etc.

The data collected is also very subjective, which could make it invalid. Freud’s data collection consisted of letters from Little Hans’ father, as well as a brief interview. It was said that Little Hans’ father was a huge fan of Freud, and so he would’ve been very excited to be in contact with him. This could also mean that he may have exaggerated or lied in his letters, to make his son’s case more interesting or impressive to Freud.

Overall, I think case studies are a very effective way to study people who are in a minor population, however I don’t think psychologists should be able to generalise their results; at the end of the day, everyone is different-individual differences.

http://psychology.about.com/od/cindex/g/casestudy.htm





Homework for TA-14/03/12

13 03 2012

http://prpnw.wordpress.com/2012/03/11/can-correlation-show-causality/#comment-38

http://laurencedown.wordpress.com/2012/03/04/week-twenty-onewhat-options-are-there-when-you-do-not-find-significance-despite-basing-your-work-on-a-solid-theory/#comment-60

http://psuc28.wordpress.com/2012/03/11/faking-it/#comment-94

http://psucd3.wordpress.com/2012/03/09/correlation-never-implies-causation/#comment-47





Sociometric Tests

10 03 2012

I thought I would write my blog on sociometry; including what it is, and what are the strengths and weaknesses of sociometric techniques.

Sociometry is a way of measuring relatedness between people. These sociometric tests can discover, describe and evaluate social status and structure, and can measure the acceptance or rejection felt between peers. Subjects within a group are usually asked to pick members that they like or prefer working with, depending on the context (i.e. if the subjects are school children or work colleagues).

The term was first coined by Jacob Moreno (1932-1938), when he was assigning residents from the New York State Training School for Girls to residential cottages. Sociometric tests were also used by Frederickson & Furnham (1998) to assess the social status of mainstreamed children with learning difficulties.

One advantage of sociometric testing is that because we are measuring group relations, this can help us come up with interventions to help improve these relations. For example, if we were carrying out sociometric tests on work colleagues, then this can help us pair workers together and reduce conflict and improve communication between colleagues. This has been shown in Hoffman et al’s (1992) study, where workers were asked to assess the effectiveness of an intervention that was designed from the findings of sociometric tests. They found that distrust/antagonism had been cut in helf and high trust had gone up by 19%.

However, because sociometric tests rely on self report, they suffer from all the weaknesses of this method, such as social desirability. This occurs when a subject wants to make themselves appear socially desirable to the researcher, so they may lie on personal questions to make themselves look better. Also, self report is completely subjective; a subject’s views may change depending on their mood-if they’re in a good mood they’ll typically be more positive and vice versa.

Also, there have been some ethical issues concerned with sociometric tests. This is mainly focused on negative views on fellow classmates in schools, and the possibility that children will compare their responses, resulting in negative social and emotional consequences for children who are not positively viewed by their peers.

Overall, I think the sociometric method is an extremely useful one, as its effectiveness has been shown in coming up with interventions to improve relationships between workers. However, researchers must be careful when using this, as there can be harsh consequences for the more negatively viewed subjects.

http://www.hoopandtree.org/sociometry.htm

http://www.ncbi.nlm.nih.gov/pubmed/9848265

http://www.globusz.com/ebooks/MarketingResearch/00000016.htm

http://www.vkmaheshwari.com/WP/?p=50





Homework for TA-22/02

21 02 2012

http://laurencedown.wordpress.com/2012/02/15/week-eighteen-why-the-file-drawer-problem-neednt-be-a-problem/#comment-49

http://prpsjj.wordpress.com/2012/02/16/why-is-the-file-drawer-problem-a-problem/#comment-39

http://roydeanschlipp.wordpress.com/2012/02/19/ethics/#comment-43

http://psucfa.wordpress.com/2012/02/19/the-truth-about-correlation-studies/#comment-26





The “File Drawer Problem”

17 02 2012

   When reasearch is conducted, the results and conclusion are often sent to journals to be published. If the research is interesting, and it results in a significant effect, then there is a fair chance it will get published. But what if your research showed no effect?

   The “File Drawer Problem” is when research with a true hypothesis, or contrary results to previously published papers is rejected by journals (Rosenthal, 1979). It has been found that research with significant results is three times more likely to be published than research with negative results (Dickersin et al, 1987).

   Significant results have been known to be published even if several papers beforehand have shown there to be no significant difference in the same area. This is thought to be because the editors of a journal won’t think readers would be interested in research that shows no effect; the most common cause for non-published papers is the investigators themselves not thinking people would be interested in null results (Easterbrook, 1991).

   John Ioannidis (2005) came up with 6 reasons why a lot of journals don’t get published:

  1. When the studies conducted in a field are small
  2. The effect sizes are small 
  3. There is a greater number and less preselection of tested relationships
  4. There’s more flexibility in design, definition, outcomes or analytical modes
  5. There’s greater financial and other interest and prejudice
  6. There are more teams involved in a field searching for statistical significance

   Ioannidis also came up with “remedies” of the file drawer problem:

  • Better powered studies (low bias meta analysis, large studies testing major concepts)
  • Enhanced research standards 
  • Considering what the chances are of a true or non-true result

   I believe that the file drawer problem is a serious matter, and that no research is pointless, even if it shows no effect. There could be something that we’ve believed for a very long time, and some research could disprove this thing!

   Publishing false true results can be a dangerous thing too. For example, the Thalidomide tragedy in 1960.This drug, known as thalidomide, was thought to alleviate morning sickness in pregnant women, and was sold over the counter. Many young women took this drug, and many would give birth to babies with terrible birth defects, such as badly deformed eyes, ears, nose or heart, and even phocomelia (all or parts of limbs are missing). The drug was tested on pregnant rats, and did not appear to harm them, and so the drug was deemed safe.

   I think there is no harm in publishing a null result; it could just be the case that there is no effect, it doesn’t have to mean the research is false, or even boring. Some people may want to carry out a study in an area, and if they see previous research got a null result then they can try and disprove it.





Homework for my TA – due 08/02/12

7 02 2012

http://psuc0e.wordpress.com/2012/02/05/research-designs/#comment-13

http://roydeanschlipp.wordpress.com/2012/02/05/outliars/#comment-30

http://katepsuc7d.wordpress.com/2012/02/05/differences-between-testing-of-within-subjects-design-and-between-subjects-design/#comment-35

http://amw1992.wordpress.com/2012/02/05/repeated-measures-t-test-advantages-and-disadvantages/#comment-32