Introduction

I am interested in looking at caffeine and memory retention to see if drinking caffeine can help students retain information before a test. I am an avid caffeine drinker, so I am excited to investigate this study on the Islanders and see if they can score better on their memory test after drinking caffeine. My research question is: does caffeine intake affect a student’s memory during a memory test?

For this study, the variables that will accessed is the caffeinated drink, which is a categorical, binary variable and memory retention, which is a quantitative variable (a score out of 20 for all the correct words remembered), and this is also a proxy variable. I will be using a treatment vs. control group. The observational variables are the caffeinated (treatment group) or non-caffeinated beverages (control group). The total sample size is 54 subjects, with 54 total data sets collected for the study. This study is an experimental study since I am making my subjects consume a beverage. I participants came from the total count of the university; I chose the participants by scrolling through the given list and randomly stopping and choosing to ask for consent to participate. Those who consented were included in the study, otherwise another participant was randomly selected. I categorized into two groups for analysis: caffeine intake (coded as 0 for non-caffeinated and 1 for caffeinated beverages) and memory test (total correct answers out of 20). Prior to data collection, I assume that caffeine does not necessarily have an impact of memory, rather it enhances memory. I assume that memory scores will be relatively higher with caffeine than without caffeine due to the alertness of cognition; for example, studies have shown that “a 250 ml caffeinated beverage may also show an increase in positive effects on cognition at higher dosages (200–250 mg)” (Nehlig, 2010, p.3).

Note: I did not make repeat visits in order to obtain the observational units. This is because I made a mistake as a researcher and did not conduct my data based on my initial proposal. I accidentally conducted the study with the participants drinking the caffeinated beverage or no beverage at all, defeating the purpose of my study. This was also realized within the last day of my data collection, so with my time restraint, I was unable to do repeat visits

Data Collection Methods

For my data, the population are students at Hofn University in the the city Hofn in the Islands, where the Islands is a virtual “human” population. The Islands is a great place to conduct statistical studies where we are given assigned tasks to each subject (if they consent) and collecting data. During the experiment, participants were randomly assigned to either a 250 ml caffeinated or non-caffeinated beverage, which took no longer than 20 seconds. I randomized the assignment so that it reduces biases to have no confounding variable. Immediately afterward, participants completed a memory vocabulary test where they viewed a list of 20 words for one minute, waited for one minute, and then recalled as many words as possible in 30 seconds. Each participant’s total study time was roughly three minutes.

The total sample size is 53 subjects, 22 non-caffeinated subjects and 31 caffeinated subjects. I chose the participants by scrolling through the given list and randomly stopping and choosing to ask for consent to participate. Those who consented were included in the study, otherwise another participant was randomly selected. There is a total of 274 subjects at Hofn university. I categorized into two groups for analysis: caffeine intake (coded as 0 for non-caffeinated and 1 for caffeinated beverages) and memory test scores out of 20. The study is a theory based test, with met validity conditions (at least 10 “successes” and 10 “failures”). I collected 53 total data sets to ensure that my study meets the validity condition.

Descriptive Statistics

The null hypothesis states that there is no difference in memory scores between students who consumed caffeine and those who did not, \[H_0: \mu_{0} = \mu_{1}\]

The alternative hypothesis states that there is a difference in memory scores between the two groups.

\[H_A: \mu_{0} \neq \mu_{1}\]

A Type I error would represent concluding that there is a difference between drinking a 250 mL caffeinated beverage and doing better on a memory test when there is actually no association. A Type II error would represent failing to conclude that there is a significant association between drinking a 250 mL caffeinated beverage and doing better on a memory test when an difference exist. The null hypothesis states that there is no association that a student will do better on a memory test when drinking a 250 mL caffeinated beverage or not.

For my quantitative response variable and binary categorical explanatory variable, I conducted my first part of analysis by creating a side-by-side box plot and a numerical summary statistic to see if there appears to be an association between the two variables. Here is the beginning of my statistics:

Importing my .csv file

Memory_and_Caffeine <- 
  read_csv("~/Documents/math247/Memory_and_Caffeine_Data.csv")
head(Memory_and_Caffeine, n=2)

Side-by-side boxplot

bwplot(caffeine ~ memory_score, 
       horizontal = TRUE, 
       main="Side-by-side boxplots",
       data = Memory_and_Caffeine)

Summary of Statistics

favstats(memory_score ~ caffeine, data = Memory_and_Caffeine)

The side-by-side box plot below compares the two variables, caffeine and memory score. We can see that the jagged edges on the left of the box represents the “lowest” memory score out of 20 and the jagged edges on the right represents the highest score out of 20. The x-axis represents the caffeine intake, where it should be encoded as a 0 and 1 instead of a 1 and 2, where 1 represents 0 (i.e., no caffeine) and 2 represents 1 (i.e., caffeine). There is a single outlier, that is a person who did not drink a caffeinated beverage scored 8/20. This can be due to a confounding amount of reasons, but it doesn’t mean much. The summary of statistics show 0 for no caffeine and 1 for caffeine. We can see that the means are roughly the same score, median is 1 off and the SD is pretty similar.

Analysis of Results

In this study, I will be using a t-statistic to find if there is a comparison between the memory scores of subjects who consumed caffeine and those who did not. Considering the relatively small sample size, this statistic should access whether there are any observed differences in memory scores are statistically significant, thus helping to determine the effectiveness of caffeine on memory retention in this specific population:

t Statistic of Interest

stat(t.test(memory_score ~ caffeine, data = Memory_and_Caffeine))
##         t 
## 0.2199557

This t statistic suggests that there is weak evidence against the null hypothesis of no difference in memory score and caffeine consumption. However, just finding the p-value does not provide enough statistical significant evidence to reject the null hypothesis. Thus, I will need to find the p-value, in this case, I will need to find the two-sided p-value since we are accessing that there is no difference.

Finding the p-value

two.sided.p.value<-pval(t.test(memory_score ~ caffeine, data = Memory_and_Caffeine))
cat("the two-sided p-value is",two.sided.p.value)
## the two-sided p-value is 0.8268788

This p-value represents the probability that our null hypothesis is true. In this case, the value 0.82 suggests that we do not have strong evidence to reject our null hypothesis because our p-value is quite large, 0.05 is significant.

Therefore, with a t-statistic of 0.219 and a two-sided p-value of 0.82 we cannot rejected our null hypothesis, in “support” of our alternative hypothesis. We have proven that our theory-based test meets the validity condition that both sample sizes are large (generally considered to be at least 20 with no strong skewness0.

This study can be generalized to population at Hofn University, but not across the other islands. This is because the sample size is repetitively low and we have proven that there is not strong evidence to reject our null hypothesis. Additionally, the data does not provide evidence that caffeine makes memory retention better. This study suggests that caffeine does not have effect on memory retention that the subjects participants score was the score if they had not drinken caffeine.

Finding the 95% confidence interval

confint(t.test(memory_score ~ caffeine, data = Memory_and_Caffeine))

The 95% confidence interval suggests that we can be 95% confident that the true difference in memory retention scores between students who took caffeine and those who did not is somewhere between -1.29 and 1.60. In other words, we cannot reject the null hypothesis that there is no difference in memory retention between the two groups. This supports our conclusions in both statistical analysis.

Conclusion

This study accessed if there is a difference in memory retention for students who drink caffeine or not, specifically students at Hofn University. While there were very very slight differences in means (0.16), statistical tests (t-statistic and confidence interval) revealed no significant impact of caffeine on memory scores.

Therefore, memory retetion appeared similar regardless of caffeine intake. However, the limited sample size restricts generalization to the entire university population. Future research with a larger sample could provide more definitive results, potentially exploring the influence of caffeine dosage and individual variations in caffeine processing on memory.

Bibliography: references to literature mentioned in the introduction

The article below is a study that looks at college students (n=83) where they drank a caffeinated beverage before starting a memory test. It was found that memory and predictions of memory can be influenced in different ways if they drink caffeine before they study or take a test.

Kelemen, William L., and Catherine E. Creeley. “State-Dependent Memory Effects Using Caffeine and Placebo Do Not Extend to Metamemory.” The Journal of General Psychology, vol. 130, no. 1, 2003, pp. 70–86,
https://doi.org/10.1080/00221300309601276.

The article below is similar to the other study above, but focuses more on overall cognition rather than just memory. It found that there is indirect action of caffeine on arousal, mood and concentration contributes in large part to its cognitive enhancing properties.

Nehlig, A. (2010). Is Caffeine a Cognitive Enhancer? Journal of Alzheimer’s Disease, 20(s1), S85–S94. https://doi.org/10.3233/JAD-2010-091315