To begin this semester blogs, I thought I would discuss the two basic research designs concerned with psychological research, those being between-subject designs and within- subject designs. Within this blog I am going to disregard the fact that the majority of the time we choose a design based on our hypothesis. I am going to look more into the pros and cons of both and decide despite my research hypothesis, which one in general appears to be the better design, resulting in more valid findings.
Firstly I shall discuss between-subjects designs; this is where different groups of individuals are compared. Therefore the researcher manipulates the independent variable, which creates different treatment conditions; separate groups of participants are then assigned to each of these different conditions. The dependent variable is then measured for each individual and differences are looked for between-groups. When given the basic description of a between-groups design, my initial thought is the scores are independent of each other and therefore I can be confident that they are clean results and have not be affected by practice effects, or fatigue, boredom or contrast effects. However in the past when thinking about conducting a between-groups experiment, I realise a lot more participants will be need than if I were to conduct a within-subjects design. As each participant is only giving one piece of data, if I were to conduct a study with 3 conditions needing 20 participants in each one, I would need 60 participants whereas within groups I would only need 20. This may prove an issue in situations where a population is hard to get hold of. Despite the advantages of different individuals being used in conditions, this is also a primary disadvantage, as it leaves the research open to the problem of individual differences. For instance if I were to conduct a study with my friend and boyfriend both participating, they differ in age, IQ, personality, interests but also the day before my experiment my friend may have had a sleepless night whereas my boyfriend slept soundly. Within my research the major concern for me would be these individual differences turning into confounding variables, basically is my finding actually because of the treatment or because of a confounding variable such as age. Sometimes we can just not tell.
Before choosing my winner, I shall discuss within- subjects designs, this is where a single group of participants are tested or observed in all of the different treatments being compared. Therefore the same group of individuals participate in all different levels of the independent variable. Again when given the basic description of within- groups design, my first thought is this is the complete opposite to between subjects in regards to confounding variables, at the very least they will be reduced. There are no individual differences to confound the results; the individuals in treatment 1 are exactly the same as treatment 2. You can more than likely conclude differences are due to the treatment rather than individual differences. Furthermore if individual differences are consistent across treatments, they can be measured and removed from the rest of the variance of the data. Despite its advantages, one major setback of within-subjects is that participants go through different treatments each administered at a different time, therefore time-related factors may come into play, such as fatigue and weather; also issues such as history, maturation, instrumentation, testing effects and regression hold disadvantages. Lastly such design is subject to participant attrition, where participants may simply just drop out, losing you valuable data.
After discussing both designs it is clear that both hold advantages and disadvantages, and it is apparent that the advantages of one design are essentially the same as the disadvantages for the other design. Therefore I have come to the conclusion that no design is better in general than the other in regards to validity; as when each design is faced with its disadvantages, each disadvantage is offered a solution. For instance between-groups main disadvantage is individual differences, this can be improved by establishing equivalent groups by random assignment or matching, also variance can be minimized by standardizing procedures and keeping the treatment setting stable. Lastly within-groups main disadvantage is order effects, dealing with such problem can be done through counterbalancing the conditions. No winner here guys 🙂