Introduction

My study observed the relationship between classical music and IQ score, with a specific research question of “what is the relationship between classical music and IQ score in adults over the age of 60 currently residing on Nelson Island?” The population parameter of interest for this study was observing the IQ scores of individuals over the age of 60 currently residing in Nelson Island. When looking at research regarding these topics, there is a significant amount about how classical music affects performance in children, specifically in Chen et al. 2023, they observed the relationship between classical music and task performance in elementary school students. Another study, Bell et al. 2016, observed the relationship between classical music and spatial reasoning and mindfulness. With these studies in mind, I wanted to look at a older population as that seemed readily underresearched, and see if classical music had an effect on their IQ score. Before performing any data analysis, I suspected that those who listened to classical music would have a higher IQ score than those who did not.

Data Collection Methods

The observational units of my study were the mean IQ scores of individuals over the age of 60 on Nelson Island. The categorical variable of classical music was measured by either playing the islander 10 minutes of classical music before their IQ test or not. The quantitative variable (IQ score) was measured by administering an adult IQ test that took around 20 minutes. There was nothing that initially went wrong with this study as these islanders all were very willing to participate. I had a few individuals who opted out, but that did not have an impact on my data. One thing that was an issue for me was the selection process of the islanders. As opposed to using a random number generator to chose houses, I chose them somewhat arbitrarily as I scrolled to a random house and tested the individuals in the house if they were over the age of 60. For future research, an external tool like AI or a simulator should be used to ensure random sampling. This violates the random sampling aspect and may affect generalization to the larger public. The response rate was almost 100% as all the individuals were willing to participate, but I had around 2 individuals who opted out later on in the study. In order to ensure that the observational units remained initially selected, I had to visit after the classical music was administered (if they were in the classical music group), then after the IQ test was finished. This study did not take too much time, so it would be very easy to replicate. The most energy and effort was put towards finding the individuals that satisfied my parameters.

Descriptive Statistics

library(readr)
IQ_Data <- read_csv("~/Downloads/IQ_Data - Sheet2.csv")
head(IQ_Data, n=2)
bwplot(`classical music` ~ `iq result`,
horizontal = TRUE,
main="Side-by-side boxplots",
data = IQ_Data)

favstats(`iq result` ~ `classical music`, data = IQ_Data)

When looking at the side by side box plot, we can see that the means of the two groups are relatively similar. The standard deviations, or dotted lines, are also around the same width as the other, initially implying that there is no relationship. This is evident when we run the fav stats function: the mean IQ score of the group who did not listen to classical music is 106.16, and the mean IQ score of the group who listened to classical music was 104.94. These means were within 2 points of each other, and their standard deviations were very similar at about 17. So from the box plots and summarization of data in the fav stats table, I would infer that there is little to no association between the two variables.

Analysis of Results

My population for this study was adults residing in Nelson Island that are over the age of 60. The null hypothesis is that there is no relationship between classical music and IQ score. The alternative hypothesis is that there is a relationship between classical music and IQ score. These hypothesis can be defined in symbols (mu 1 is the iq score of individuals who did listen to classical music, while mu 2 is those who did not listen to classical music):

\[H_0:\mu_{1} = \mu_{2}\] \[H_A:\mu_{1} \neq \mu_{2}\] In this setting, there is a possibility for a type II error as there is enough evidence to fail to reject the null hypothesis when it could be true. Past research done on this topic has shown that there is a correlation between classical music and cognitive tests or performance, so maybe my population size or parameter leads me to accept the null, when in reality, there is a correlation. Since I have a population size larger than 20 and the distribution is relatively normal, I can generalize this study to the participants who opted to take it, specifically Nelson Islanders over the age of 60 who participated in my study. The only issue within the generalization aspect is the random sampling error I addressed earlier within this study. I did not use an external tool to ensure random sampling. When looking at the data collection to see if there is a correlation between these two variables, we can use a t-statistic as it is observing the two means of two independent groups. This study satisfies the corresponding validity conditions as there is independence between the variables, individuals were randomly assigned to each group (classical music or not), there were at least 30 people in each group, and the standard deviation was around the same for both groups. Below is the code for the t statistic calculation:

stat(t.test(`iq result` ~ `classical music`, data = IQ_Data))
##         t 
## 0.3664086

When looking at the p value, the probability of observing no relationship between the presence of classical music and IQ score is 1, or 100% probable, assuming that the null hypothesis is true. With a p-value of 1, we can conclude that there is no evidence against the null hypothesis, therefore we fail to reject the null hypothesis. In the context of the study, we can conclude that there is no relationship between the presence of classical music or not and changes in IQ score. The validity conditions for a theory based approach are satisfied (independence between variables, random assignment, and normality), so there is no need to use a simulation based approach.

p_value<-prop(~(diffmean = 1.22), data = IQ_Data)
cat("right-tailed p-value is", p_value)
## right-tailed p-value is 1
confint(t.test(`iq result` ~ `classical music`, data = IQ_Data))

For my interpretation of this confidence interval, I am 95% confident that the difference in mean IQ scores is between -5.39 and 7.83 points. Since 0 is included in this parameter, this further proves that there is no significant different between classical music and IQ score. This aligns with the p-value results that we found in question 5d.

Conclusion

Concluding the information stated above, there was no relationship between presence of classical music and change in IQ score within the population of adults over the age of 60 currently residing in Nelson Island (question 5c). From the observed statistics of the p value (P=1) and t statistic (t=0.366), along with my confidence interval (-5.38, 7.83), it is clear that there is no relationship between these two variables. As a whole, I learned a lot from performing this research study. I learned that research requires a lot of time and energy, so it is important to have your research planned out prior to performing the experiment, and ensure that you have all of the resources available. The data did not behave as I expected as I was hoping to see a relationship between the presence of classical music and IQ score. Since there were only 100 individuals (50 per group of listening/not listening to music) that participated in this study, we can generalize this study to the individuals over the age of 60 that decided to partake in this study. If I were to run this study again, I would try and obtain data values from other islands, or each town hall on each of the islands for the larger ability to generalize this study to a larger population. I would also try to perform this study at the same time every day to ensure there was no external factors or limited brain activity due to the time of day. I would also ensure that there was random sampling within what houses to choose in the island of Nelson. To build on my results, future researchers could investigate if the duration of listening to classical music affect IQ scores or performance. Another question would be to investigate if an increase of intelligence due to classical music listening is transgenerational.

Letter of Learning

The most meaningful part of this project for me was the ability to come up with and test our own research questions. I feel like college is a lot about analyzing other research and interpreting data from different sources, so I really enjoyed getting to formulate my own research question and experiment. The most interesting thing to me was the Islands. I thought it was so cool how a computer simulation could contain so many people with such different backgrounds and personalities. To be honest, it is a little bit creepy but works great when you are trying to conduct a study without an IRB. The most challenging part of this project was the data collection and analysis of results. The data collection was just very tedious as I had to return to the islands multiple times a day to check if the IQ testing was completed for all the participants. I struggled with obtaining the correct R code for my analysis, but found it helpful to reference past explorations and homeworks to help with that issue. One strategy I employed was using Google Sheets to track the scores of each of the individuals, along with the group that they fell under. This was super helpful as it helped me keep track of their names and what research group I put them under. Another strategy I used was starting this project early on. I knew it was going to take a decent chunk of time to collect and analyze my data, so I was sure to leave plenty of time to do so. Both of these strategies worked great and definitely helped the success of my project. One strategy that did not work for me was keeping all the islanders on the contacts page of the islands website. I initially tried to keep track of all my people in the contacts folder of the islands, but I was unable to organize each of the islanders into the groups they were assigned to. Other than that, everything went very well and I definitely learned a lot about how to collect and present data in a professional format!

Bibliography: references to literature mentioned in the introduction

Chen Lijun, Huang Meilin, Jang Xiaoliu, & Wang Xinjian. (2023). Does classical music make you smarter? A meta-analysis based on generalized Mozart effect. Advances in Psychological Science, 31(12), 2232. (https://doi.org/10.3724/SP.J.1042.2023.02232)

Bell, T. P., McIntyre, K. A., & Hadley, R. (2016). Listening to classical music results in a positive correlation between spatial reasoning and mindfulness. Psychomusicology: Music, Mind, and Brain, 26(3), 226–235. (https://doi.org/10.1037/pmu0000139)