Problem 9 - An Introduction to Psychological Statistics

Question

You have become suspicious that the draft rankings of your fantasy football league have no predictive value for how teams place at the end of the season. You go back to historical league data and find rankings of teams after the draft and at the end of the season (below) to test for a statistically significant predictive relation. Assume \(S S M=2.65\) and \(S S E=337.35\)

Step-by-Step Solution

Answer

The draft rankings show no statistically significant predictive value.

Step 1: Define Hypotheses

Formulate the null and alternative hypotheses. The null hypothesis ( H_0 ) states that the draft rankings do not have predictive value, meaning no correlation between draft rankings and the end-of-season rankings. The alternative hypothesis ( H_1 ) posits that there is a predictive value, or a relationship exists.

Step 2: Determine Test Statistic

The formula for the test statistic, F, in the context of regression is given by \( F = \frac{SSM}{SSE} \). Here, \( SSM = 2.65 \) and \( SSE = 337.35 \). Substitute these values into the formula to get \( F = \frac{2.65}{337.35} \).

Step 3: Calculate the F-value

Perform the division: \( F = \frac{2.65}{337.35} = 0.00785 \). This F-value will be compared with the critical F-value from the F-distribution table to determine statistical significance.

Step 4: Conclusion and Interpretation

Compare the calculated F-value with a critical F-value from an F-distribution table at the chosen significance level (e.g., \( \alpha = 0.05 \)) with appropriate degrees of freedom. Since 0.00785 is likely less than the critical F-value, we fail to reject the null hypothesis, indicating no significant predictive value.

Key Concepts

Regression Analysis

The core of regression analysis is to understand relationships between variables. In the context of fantasy football rankings, this method allows you to see if early draft choices have any predictive power over end-of-season performance.
A regression analysis involves identifying and quantifying the relationship between a dependent variable (end-of-season ranking) and one or more independent variables (draft ranking). Here, the "predictive relation" is what analysts look to evaluate.
Regression helps in creating a mathematical model by minimizing the differences between observed and predicted values, often through a method called least squares. In this exercise, once the sum of squares due to the model (SSM) and the error (SSE) was computed, these figures were used to determine the test statistic, contributing to evaluating significance through hypothesis testing.
Understanding how well the model predicts outcomes is crucial. It distinguishes between strong, weak, or no relationships, guiding informed decisions based on statistical evaluations rather than mere assumptions.

Statistical Significance

Statistical significance is a way to test if the results of your analysis are likely to be true or occurred by random chance. In the exercise, we determine this by comparing the calculated F-value with a critical value from the F-distribution table.
A result is considered statistically significant if the calculated figure lies beyond a threshold defined by a significance level (often represented by alpha, such as 0.05). This criterion helps in deciding whether we reject or fail to reject our null hypothesis.
Statistical significance implies confidence in the findings. It helps determine if the data indeed supports a given claim or relationship, i.e., whether the draft rankings genuinely offer predictive power over end-season performance. However, this test does not guarantee practical significance, which refers to the actual relevance or impact of the results in real-world scenarios.

Null Hypothesis vs Alternative Hypothesis

Hypothesis testing begins with formulating a null hypothesis (\(H_0\)) and an alternative hypothesis (\(H_1\)). In this context, \(H_0\) claims there's no predictive value in draft rankings, implying no correlation. In contrast, \(H_1\) suggests a positive relation exists.
The null hypothesis is what researchers try to disprove or reject, motivated by trying to find evidence against a baseline assumption of no effect or relationship.
Typical hypothesis testing involves assuming the null hypothesis is true until statistical evidence suggests otherwise. As calculations are made, such as the F-value from the regression analysis, these values are compared against critical values to conclude whether they have statistical significance. If the test statistic falls into the rejection region determined by the critical value, \(H_0\) is rejected, pointing towards a significant relationship. However, in this case, the observed F-value was too low, leading to a failure to reject the null hypothesis, thus implying no significant predictive relationship was found.