Last comments YAY! :)






Psychology is an extremely broad research field and aims to fully understand everything about humans such as the way we think, the way we develop, the way we behave etc. Psychologists look at the psychological perspectives, the biological, the environmental- everything! Well at least we like to think everything is studied. The big question is, is it actually possible to measure every single variable?

On one hand one could argue that yes it is possible to measure every single variable. For example behaviourists believe that everything humans feel and think is expressed through an observable behaviour. If this perspective is taken then yes every variable can be measured. From this approach it is believed that all variables can be operationalized in that they can be clearly defined and accurately measured. This idea is often applied to internal variables known as hypothetical constructs and variables such as emotional states and intelligence are believed to be measurable through observable behaviours and performance on specific tasks. For example Bandura measured aggression through observing the behaviour of children when they were presented with and interacted with a bobo doll. Also Binet developed specific tasks in order to measure intelligence. Furthermore Freud used his tripartite model of the id, ego and superego and unconscious drives as a basis for his research. In Freud’s study of little Hans Freud measured the subject’s inner conflicts and phobia and cause of the phobia by analysing his dreams.

However the problems arise when psychologists try to measure these internal variables that cannot be directly observed. There is no definite way to ensure that the hypothetical construct is being accurately measured and there can be a very negative impact from the seemingly inaccurate measurements. For example Kaplan and Sacuzzo (2005) found a large number of African American children had been misdiagnosed with mental retardation due to the outcome of intelligence tests created by Binet. Research by Serpell (2000) found that one reason for this could be the fact that intelligence tests are biased to favour white, middle class children. Another issue is the conflicting research that appears when supposedly studying the same variable. For example Bowlby explained Little Hans’ phobia through an attachment theory that goes against Freud’s explanation. The fact that there is evidence for 2 different explanations of one variable implies there is not a valid or reliable way of measuring a hypothetical construct.

In conclusion, I don’t believe that every variable can truly be measured accurately. The research that attempts to operationalize and measure hypothetical constructs can be very helpful; however the theories should be applied with caution.

Comments for Nai :)





Difference between single subject design and case study One approach used by psychologists is the case study design method. This method involves researchers collecting in depth, descriptive data which is then interpreted, analysed and evaluated. Data can be collected from the participants past and present to ensure true context. Multiple methods are used to collect data such as interviews with the subject as well as their relatives and/or friends, observations, public records etc. For example the case study of ‘Little Hans’ conducted by Freud consisted of interviews with Hans, observations by Freud and his father as well as dream analysis. There is no cause and effect relationship established as the aim is not to manipulate variables to see specific outcomes. Treatments may be applied, but the standardised experimental procedure is not used. This method is often used for unusual and seemingly unique behaviours.

One advantage of the case study design is that it allows detailed and contextual data to be collected. Although cause and effect relationships are not established, the results found from these studies allow extended knowledge of theories that have already been established through previous research. As the design focuses on one subject, individual differences are accounted for.

One disadvantage of the case study method is that it is highly subjective. The data is mainly qualitative and descriptive and therefore needs to be interpreted by the researcher. This highly decreases the validity of the findings as researchers may be bias to look for evidence and behaviours that support their theory and may even ask leading questions when using the interview method. For example in the study of ‘Little Hans’ by Freud, Freud interpreted Hans’ dream with giraffes to be about Hans’ parents genitals! Although this made sense in Freud’s mind and supported his theory of the Oedipus complex, pretty much no one else would make this (rather disturbing) interpretation. Another disadvantage is that it is incredibly difficult to generalise or apply the findings from case studies to anyone else.

Unlike the case study design method, single subject designs aim to find a cause and effect relationship. The researcher has control over the situation and aims to show that the manipulation of an independent variable causes a change of the dependent variable. With this research design a baseline is first found by the researcher taking repeated measurements of the participant’s behaviour, which will be the dependent variable, before the intervention has taken place to ensure consistency. To do this the researcher must have control over the environment to make sure there are no extraneous or confounding variables affecting the results. Next in the intervention phase which involves the researcher applying the independent variable and then collecting more observational data. There may also be a reversal phase where the independent variable is then taken away and the dependent variable is measured again to see if the baseline is resumed.

One advantage of the single subject design is that the findings allow researchers to find the differences between individuals. One criticism of group experiments is that the significant differences between individuals is not accounted for or taken into consideration and single subject designs allow researchers to discover this difference. This research design also allows researchers to see a causal relationship for a specific therapy or treatment.

One disadvantage, like with case study designs, is that the results are not generalizable. The point of a single subjects design is to see the effect on individuals rather than groups.

Comments for TA :)





As psychology students we all know that correlation does not mean causation, however if a correlation is found, further research of the two variables could end up showing a cause and effect relationship.

So what is correlation, what is causation and what is the difference?

A correlation simply shows that two variables occur together and have some sort of relationship. These studies can conclude if there is a positive, negative or no relationship at all between two variables.  Correlation studies are usually conducted at the very beginning of a research area as preliminary research. For example a correlation could be found between children who watch more television and aggression levels. We would be able to conclude there is a relationship between hours of television watched and levels of aggression, however it would not be possible to conclude that watching more television causes aggression in children. As I will explain further in a minute, the correlation methodology is very different to an experiment looking for a cause and effect relationship. A correlation method design involves researchers simply observing two or more variables and a score will be obtained for each variable on different measurement scales. With the previous example mentioned researchers would get a value for how many hours of television each participant watches (maybe in a week?) and then would probably have a scale of aggression and would observe the participants for the same period of time (a week?) and give them a numerical value for aggression. In correlation studies there is no control over other variables, there is no manipulation of variables and different condition groups are not compared.

Causation occurs when one variable is found to directly cause the change in another variable and there is no other explanation for the change in the second variable. These experimental studies will usually have been based on the findings from a prior correlation study as the researchers would have needed to know if there was any relationship between the variable they were curious about. As I mentioned earlier, the methodology of experimental studies is very different to the correlation method. The major difference is the level of control. In order to establish a cause and effect relationship, the researcher must apply a high level of control over all other extraneous or confounding variables to ensure that there is no other explanation for their findings. For example, if a researcher was looking to find if classical music caused students to perform better in memory tests, the researcher would have to make sure the only variable changing was the music. The researcher would administer the exact same test to every participant, however may have a group that listens to the same classical music, a group listening to the same rock music, a group listening to the same R’n’B music and a group that does not listen to any music. The participants would all take the test in the same environment in that they would all complete the test in the room setting, but not the same room at the same time obviously. To ensure the highest level of control, researchers may ask all students to get 8 hours sleep the night before and not consume caffeine 24 hours prior to the study. The results from each group would then be compared to see if there is a significant difference. There are many other controls that could be applied to this study but I think you get the idea that there is a very high level of control, unlike with the correlational method. Another key difference is that the dependent variable will be measured using only one method, rather than having one for 2 or more variables.

So now we have a clear understanding of correlation and causation, it should be clear that correlation does not directly show causation. It implies that with further research and work, a cause and effect relationship could be found between the two or more variables that were found to have a correlation relationship. The lack of control from the design of the correlational method makes it impossible to conclude that there is a cause and effect relationship, although the media loves to take things out of context and blow them out of proportion by publishing headlines of causation when in fact only a correlation has been found. Correlation studies are not useless though as they are fairly easy and cheap to perform and the findings from these studies show the areas that need to be researched. If there was no correlation relationship, there would be no point in conducting a highly controlled experiment that could take a lot of time, effort and money. So correlation does not mean causation, however it shows there areas where researchers may be able to find a cause and effect relationship.