For the current study, the question below was chosen from a suggested list.
Does drinking caffeine affect students’ performance on tests?
The choice of this question was not accidental but instead has a strong justification. Living in a college community often involves severe stress and health risks. Academic responsibility for one’s own learning and the desire to perform at a high level encourages young students to use radical ways to keep their bodies functioning. These include the consumption of coffee and caffeinated coffee drinks in order to maintain alertness and clarity of judgment at unnatural times for a healthy biorhythm. Such phenomena are particularly acute during exams and tests when students are required to demonstrate the knowledge and competencies they have learned on their own.
It is true that exams and final tests are a stressor for the student. In an attempt to perform at their best, students spend the last days before the test repeating theoretical basics and learning new knowledge. Such a large amount of information cannot be absorbed systematically, so students often forgo healthy sleep and devote all their waking hours to studying. In this sense, coffee is a good tool for maintaining such a state when one does not want to sleep.
Nevertheless, it is easy to assume that such activities do not turn out to be absolutely harmless for the body. A large amount of caffeine consumed overstresses the nervous system, as a result of which the student experiences severe emotional stress, hallucinations, and memory loss. Obviously, the consequences of such sleepless nights before exams can be precisely the opposite of the expected effect, when due to lack of sleep and large amounts of coffee, reflection ceases to be clear, and the theoretical base is forgotten. Thus, the critical interest in this experiment is to determine the relationship between coffee consumption before exams and students’ own performance on those exams.
Description of Hypotheses
The initial research question for the entire experiment is to determine the above relationship. In doing so, the working hypothesis is that consuming caffeine (in coffee or coffee drinks) before exams significantly reduces performance on those exams. This question allows us to create an alternative hypothesis, from which the null hypothesis is also formulated:
From the mathematical formulations of the above two hypotheses, we can see that two samples (populations) are used, for which the mean is created. It is worth saying that the focus is set on the scores on the exams, with drinking coffee before the exams as an intervention. It is seen that the alternative hypothesis states that the average for the first population in which students drink coffee should be lower than for the population without the intervention.
Description of Information Collection and Analysis
A description of the procedures for collecting and analyzing the information collected forms the center of the design of the entire experiment. In order for the results to be reliable and for the conclusions to be extrapolated to the general population, the impact of errors and limitations must be minimized. For this reason, some of the steps described below may seem redundant at first glance, but in fact, they carry the goal of reducing the impact of threats and improving the quality of the results.
The primary issue is to determine the mechanism for selecting participants. As follows from the hypotheses stated, it is necessary to collect two samples, one of which will be exposed to the intervention and the other (control) will not. It should be taken into account that selecting students from only one group may not be correct since potential faculty bias cannot be ignored. The same is true for a single university, as there may be differences in teaching tactics and grading methods for exams between different institutions. For this reason, the best solution is to collect students from different universities within the same city.
Each sample should contain about 75 participants who have volunteered to contribute to the study. Both groups should also have an equal or close balance of incoming females and males, as well as a roughly similar average age of the entire sample.
The first group is the one exposed to the intervention throughout the experiment. Three days before the exam, they take four cups of coffee or 400 mg of caffeine per day. The second group, accordingly, does not consume any coffee, energy drinks, or coffee. Two important points must be discussed beforehand. First, the university exam is not a real exam because if the hypotheses turn out to work, the researcher assumes responsibility for the low results in the students. To delineate these responsibilities and prevent respondents’ scores from falling, the exam takes place in a mock format, with the test, assignments, and procedural parts being kept real.
Students will not be alerted to the falsity of the experiment because it may affect their determination to prepare for it. Thus, the entire test acquires blind status, which in the limit increases the reliability of the results. Second, students are not allowed to drink more than four standard cups of coffee a day, as this could negatively impact health than expected.
On the day of the test, students write an exam in one discipline (e.g., math) and formally stop participating in the experiment when they finish writing the test. Each student’s results are collected and sent to a follow-up stage using statistical tools. However, to form a more exciting and meaningful study, it would be a good idea to ask students after the exam is over about their emotions and stress over the past few days. This could be a fully structured questionnaire that represents a different line of inquiry and is not related to the test results. That said, randomly discovered data may prove valuable in elucidating the nature of the relationship between coffee, exam preparation, and students’ stress levels.
Statistical processing of the results can be conventionally divided into two stages: descriptive statistics and hypothesis testing. Descriptive statistics includes determining the demographic characteristics of each sample, calculating the mean score, median, range, standard deviation, and IQR. These results will provide an initial indication of trends. Such data alone, however, are not all-encompassing to the objectives of the experiment.
Therefore, it is necessary to perform hypothesis testing using the test of the mean of two independent samples with the hypotheses stated in Step 2. The standard Z-test is used, which is calculated based on the mean, the number of observations, and the standard deviation. If the final p-value result is above the critical level of significance (0.05), then there is no substantial evidence to reject the null hypothesis and vice versa. Thus, with this test, it will be possible to statistically reliably indicate whether there is an effect of coffee consumption on exam results.