Using examples describe and evaluate essential elements of a scientific research study.
Essential elements of a scientific research study include: research ethics, validity and reliability, statistics such as the average and range, experimental design, sampling, methods of data collection and hypothesis. There are obviously a lot of different elements of a study and those are just an example of them. There will be a few chosen to be discussed in more detail, the hypothesis, experimental design and validity. Different types of the elements as the different types can be very important into deciding the study.
A hypothesis is a stated expectation that is hoped to be proved in the research that is to be undertaken. There can be a one or two tailed hypothesis. A one tailed hypothesis is when the hypothesis is directional which means that the direction is stated. An example of this is in the study by Steptoe A., O'Donnell K., Marmot M. and Wardle J. (2008). In their study the hypothesis is that a positive effect is associated with longevity and favourable physiological function. This is one tailed because there is only one direction mentioned. A two tailed hypothesis is when the direction of the hypothesis isn't stated. A study by Brooks A. (2006) had the hypothesis that the participant's mood would have an effect on their food. As it can be seen, this is a two-tailed hypothesis because it does not state the direction in which the study is meant to go; it merely says that there will be an effect. Different studies use the two different hypotheses for a reason. A one tailed hypothesis can be used for a study to see whether the participants are affected by a particular condition in a certain way as it can be more specific. Whereas a two tailed hypothesis can be more about the fact that a factor will affect the participant but does not state how it will affect them. This would be useful if there hasn't been any conclusive research about the effect of the condition. Both types are both valid ways to state a hypothesis.
There are more different types of hypothesis. There is a causal hypothesis. This is a hypothesis that assumes that a condition will cause an effect on the participants. For a causal hypothesis, experimental research methods should be used. An example to shows this is to say: A causes B. If A causes B, then to change A, it should change B. Changes are noted to see if they are significant. In a study, the hypothesis was that nicotine dependence would affect the aggression of alcohol-dependent Turkish men who smoked cigarettes (Saatcioglu O. & Erim R. 2009). This causal hypothesis states that nicotine dependence has an effect on the participants' aggression. If they were to change their nicotine dependence it could be assumed that the participant's aggression would also change. The alternative to this type of hypothesis is the descriptive hypothesis. This hypothesis does not reference causation but assumes the different characteristics of behaviour. For example, a study that doesn't mention anything causing an effect but observes and investigates the relationship between: task interdependence, goal interdependence and individuals citizenship organization behaviour in group cohesion (Chen C. V., Tang Y., Wang S. 2009) is a descriptive hypothesis. From looking through journals, the descriptive hypothesis tends to be more common. There can't truly be any advantages and disadvantages to the different hypotheses. They truly depend on the research being carried out. If a study wanted to see if a condition affected a participant in a certain way and if changing the condition would change the effect, then it would be a causal hypothesis. Whereas if the researcher wanted to see how a condition affected a participant then a descriptive hypothesis would be used. However, depending on the type of research and hypothesis chosen, the research may have to be carried out differently.
In experimental design, there are two types: within subject design and between subject design. Between subject design - also known as independent design - is when participants are split into groups and the different groups each test one of the conditions. The results are then compared. In a study by Rousseau F. L. and McKelvie S. J. (2000), the participants were split into five separate groups where they did different instruction combinations. The study focussed on the effect of bogus feedback with the different conditions. Between subject design has flaws because more participants are needed - 196 participants were in the study above (Rousseau F. L. & McKelvie S. J. 2000) - and there can be variations in the way the participants are treated in the separate conditions. There are other considerations such as the fact that the different groups may have different types of people in them for example gender, age and personal history. However this can be controlled like in the Rousseau F. L. and McKelvie S. J. (2000) study were they made sure that the females were distributed equally throughout the different condition groups.
The other option is to have a within subject design group. This is a more constant option as all the participants experience all conditions. Within subject design means there can be less participants and no variance in the different conditions of the types of participants. There are still the problems of having a representative sample for the participants, but again actions can be taken to try and achieve it. A study showing an example of this, is a study by Todd J. et al (2009) in which all participants - apart from two due to fatigue - completed both experiments where they were tested on their motor response. One of the downsides to within subject design is that participants can get fatigue which happened in the above study (Todd J. et al 2009). Another downside is the order effect when participants remember what they had done previously and become trained. Order effects could be counterbalanced by the participants being split into groups and experiencing the conditions in different orders. Within subject design can have internal validity problems, these will be discussed presently.
There are two types of validity: internal validity and external validity. Validity makes sure that the condition that is wanted to be measured is being measured. Internal validity is when the researcher tries to control all the other variables that can corrupt the collected data. The variables are known as secondary variance. Internal validity can be compromised by variables such as: the difference in time of day, temperature, the difference in motivation, the strength of memory - if they are repeating a task - and the placebo effect where the participants expect the condition to have an effect. In the study by Brooks A. (2006), there were secondary variances that needed to be controlled. The experiment needed happy and sad events. To make sure that the events seemed evoked the necessary emotions, the participants had to provide as much detail as possible on sad and happy events. They had 6 happy events and 4 sad events. In the experiment after being told to imagine they were in the in the event mentioned they had to answer a questionnaire as to what they would eat after this event. To control the experiment even more, the researcher could have checked what food the participants actually liked. Also by repeating the same process ten times, there could be fatigue, also boredom as the questionnaire is repetitive. The study didn't mention what time of day the experiment was carried out or if the participants had eaten before hand as this could likely affect the data. External validity questions the degree to which the sample is representative. If a study is really controlled then the internal validity would be good. Alternatively, the external validity would be compromised as the resemblance to real life would deteriorate and it would be harder to generalise the results. The researcher would have to decide whether it would be more important to control extraneous factors or to have more ability to generalise with the results.
To conclude, there are many essential elements in a scientific research. Different types of hypothesis seem to really rely on the type of research being undertaken. In experimental design, within subject design seems to have the most advantages. However, there can be external validity issues as there tends to be fewer participants. If the researcher wanted to generalise, then it may be better to use between subject design as does tend to use a lot more participants. The problem if the researchers use between subject design and then generalise, the validity wouldn't be as strong so there results might not be accurate. On the other hand, if there are a lot of participants then the data might be less flawed because a pattern could be spotted easier. A causal hypothesis would be suited to a within subject design because it would be consistent with the participants soothe differences in effects could be more reliable. On the whole, the essential elements discussed depend on the type of research and the other essential elements.
Brooks, A. (2006) Changing Food Preference as a Function of Mood. Journal of Psychology, 140(4), 293-306.
Chen, C. V., Tang, Y. & Wang, S. (2009). Interdependence and the Organizational Citizenship Behaviour: Exploring the Mediating Effect of Group Cohesion in Multilevel Analysis. Journal of Psychology, 143(6), 625-640.
Hancock, K. (2008). Contact, configural coding and the other-race effect in face recognition. British Journal of Psychology, 99(1), 45-56.
Rousseau, F. L. & McKelvie, S. J. (2000). Effects of bogus feedback on intelligence test performance. Journal of Psychology, 134(1), 5-14.
Saatcioglu, O. & Erim, R. (2009). Aggression among Male Alcohol-Dependent Inpatients who Smoke Cigarettes. The Journal of Psychology: Interdisciplinary and Applied, 143(6), 615-624.
Steptoe, A., O'Donnell, K., Marmot, M. & Wardle, J. (2008). Positive affect and psychosocial processes related to health. British Journal of Psychology, 99(2), 211-227.
Todd, J. et al (2009). Slow motor responses to visual stimuli of low salience in autism. Journal of Motor Behaviour, 41(5), 419-426.