a. Is there a main effect of note-taking type? Explain your reasoning.
There is a main effect of note-taking type as the line indicating scores (since z-scores indicate comparison to mean with regards to standard deviation) for students taking longhand notes is always above the line for students taking laptop notes. In other words, the students who took longhand notes always had higher scores, regardless of types of questions, than the students who took notes on their laptop.
b. Is there a main effect of type of question? Explain your reasoning.
There is no main effect of the type of question as one end of the line is, in one case, less than the other end going from left to right, and in another case, the other end is higher than the first end. This indicates that students with a laptop note-taking style did worse than before on new questions while students with a handwritten note-taking style did better than before. Therefore, a type of test was not seen as having an effect of being more difficult across both groups.
c. Is there an interaction effect? Explain your reasoning.
There is an interaction effect as in order for an interaction effect to occur, the lines need to not be parallel. As we can see from the the graph, these lines are not parallel and there is an effect of differences on the variables of questions for the groups.
d. Interpret these results in the context of the study.
Since z-scores indicate the placement of a student's score among the test results, a display of all z-scores would be a good estimate of test score distribution. We have seen that there was a main effect of note-taking type, indicated by continuously higher scores for students with a laptop note-taking style. We have also seen that there was no main effect of the type of question as the conceptual questions led students with laptops to do worse while leading students with longhand notes to do better, meaning the type of questions were not easier or more difficult for all. Finally, we saw that there was an interaction effect of both line being non-parallel. This final note meant that there was a difference in scoring based on the type of group a student was in. Additionally, if both the lines were parallel, then there would be no or a minimal difference on the scoring based on the type of group a student was in.
#TODO- Defend correlation
2. Given that “correlation isn’t causation”, why do we even care about correlation? Describe two reasons.
We still care for correlation as it can offer an accurate comparison between the linear relationship of two data sets. One reason that correlation is useful is that it can be used to offer information about future trends that are likely to happen, such as bird migration south in North America and weather patterns becoming increasingly colder. Another reason to implement correlation is to find a negative association with two groups, such as a preventative treatment/practice and rates of cancer in a population, which can prove very useful in prevention rather than treatment for a general population that is susceptible.
#TODO- Propose hypothetical scenarios
The first reason may be that both there is an outside element that is having an interaction effect on both the salaries of professors and the price of alcohol, such as bouts of inflation or deflation of the economy at that point in time, indicating that the value of money was more/less than it was before and therefore even though professors were paid more/less and alcohol costed more/less, the value given was relatively not significantly more/less than before.
Another reason could be that either alcohol prices are having an effect on the salaries of professors, or the salaries of professors are having an effect on alcohol sales. This can be for a number of reasons, such as a far-fetched idea that as professors are paid less, they assign more homework and fail more students, making the sale & rise of alcohol to adult student consumers to rise as they cope with their failing grades.
As both groups do not seem to be otherwise correlated in reality, a final reason could be that there is no such correlation and that the findings are due to an error in comparing groups that do not have a linear relationship, or excessive testing in different towns over many different time periods to get a p-value that is significant.
The third reason (error) seems the most plausible as there does not seem to be a logical connection between salaries of professors and alcohol prices since professors present a marginal population that do not seem to otherwise consume more alcohol than the general population. The first reason (outside influence) seems the second most plausible for the reasons described above in there not being an influence from one to the other. This first reason may be plausible under the effects of an outside force as both professors and consumable goods form a part of the larger society that can be influenced through laws or the economy. The second reason (one affects the other) is the least probable as professor salaries do not have a logical effect on alcohol sales, or vice versa, that can be readily explained.