The Music and Memory Coursework
Ms Borg teaches a Year 11 psychology class of 24 students. For her coursework she proposes to test whether background music impairs memory for words. She plans to split her own class into two groups of 12: Group A learns a list of 20 common nouns while a Maltese pop song with Maltese lyrics plays through a speaker; Group B learns an equivalent list in silence. Each group gets two minutes to study, then writes down as many words as possible. To save time she runs Group A during the first lesson on Monday and Group B during the last lesson on Friday. She plans to tell students the study is about listening skills so they do not guess the hypothesis, and she does not intend to ask for parental consent because it is only a classroom activity. She will read the results out in assembly so other classes can learn from them.
1. Identify two methodological weaknesses in Ms Borg's design and explain, for each, exactly how it could distort her conclusions.
Read the worked answer
The clearest weakness is a confound with time of day. Group A studies in the first lesson on Monday; Group B in the last lesson on Friday. Attention, fatigue, hunger and anticipation of the weekend differ systematically between those slots, and all of them affect memory. So if Group A recalls more words, Ms Borg cannot tell whether silence caused it or whether the morning slot did: the independent variable is entangled with an extraneous variable.
A second weakness is that she uses her own class, so participants are not independent in any meaningful sense. They talk between lessons, the two lists can be shared, and demand characteristics are high because they know her and want to please her.
A third is that the lists are described only as equivalent. They must be matched for word frequency, concreteness, imagery and syllable count, otherwise list difficulty rather than music could drive the difference.
A fourth is allocation: keeping the class seated is not random assignment. Randomly assigning the 24 students, or testing everyone twice with counterbalancing (music first for half, silence first for half, different but matched lists each time), would remove the time-of-day confound and allow a repeated-measures comparison.
2. Explain why the cover story and the missing consent are ethically problematic here, and propose how each issue could be resolved.
Read the worked answer
The cover story is deception. Deception is defensible only when the research cannot be done another way and the deception is harmless, and it must be followed by a debrief. Here the aim is to stop students guessing the hypothesis, which is a legitimate reason, but her students are themselves studying psychology and will learn from this that researchers conceal things. The remedy is not necessarily to abandon the cover story: she could be honest about the broad area (a study of learning and background sound), tell them afterwards exactly what was withheld and why, and let them withdraw their data at that point.
The consent issue is more serious. The participants are 15-16 and therefore minors, the study is coursework, and the findings will be made public within the school. Informed consent should come from parents or guardians as well as from the students, and students should be told they may stop at any time without penalty.
Reading results in assembly is public disclosure. Even without names, a class of 24 is small enough that individuals may be identifiable, and low scores may be misread as low ability. This breaches confidentiality and adds social risk without scientific gain. A written report seen only by the teacher and examiner would be sufficient.
3. Explain what a matched-pairs design would involve here and whether it would fix the problems you identified.
Read the worked answer
Matched pairs would mean pairing students on something known to affect verbal memory, such as their current coursework grade or a baseline recall score from a short preliminary test, then placing one member of each pair in the music condition and the other in silence. This addresses individual differences in memory ability directly by design, and it avoids the order and practice effects that a repeated-measures design can bring.
It does not, however, touch the central problem. If the pairs are still split across Monday morning and Friday afternoon, condition remains confounded with time. Matching balances the people, not the testing conditions. To fix the confound she would need randomised time slots or a counterbalanced repeated-measures design in addition to matching.
Matching is also only as good as the matching variable. If she matches on a grade that is unreliable, or on a baseline test that does not predict the task, the matching achieves very little and may create an illusion of control. So matched pairs is a genuine improvement in one respect, not a cure for the design.