关于嵌套回归模型层级测试的疑问：F值计算与显著性判断

Hey there, let's break down your questions about hierarchical regression step by step—since you already know multiple regression, this should click pretty quickly, especially focusing on your anxiety-related analysis.

Key Clarifications on F Change vs. F Ratio in Hierarchical Regression #

Hierarchical (nested) regression centers on testing the marginal contribution of adding groups of variables one at a time. The F Change value you're looking at isn't the same as the overall model's F Ratio—each F Change corresponds to a specific step where you introduce new predictors (like anxiety-linked variables) to the existing model.

1. Interpreting the first F Change (2.783) and its significance

• This 2.783 is the F-statistic that tests whether the first group of new variables you added (e.g., trait anxiety scales, initial stress predictors) significantly improves the model's ability to explain variance in your anxiety outcome, compared to the baseline model (e.g., just control variables like age or gender).
• How to test significance:
  • The simplest way is to check the Sig. F Change value in your output table. If this p-value is less than your chosen alpha level (usually 0.05), the added variable group makes a statistically significant contribution to explaining anxiety.
  • If you don't have the p-value handy, use the F-distribution: calculate the degrees of freedom for this test (number of variables added in the step, followed by the residual degrees of freedom of the new model). Compare your F-statistic (2.783) to the critical F-value for those df—if your value is larger, the change is significant.

2. The second F Change (2.775): No, you don't add it to the first one!

Your assumption here is a common misconception—this second 2.775 is a standalone test, not a sum of the first F Change. It tests whether the second group of new variables (e.g., coping strategies, social support factors tied to anxiety) adds significant explanatory power on top of the model that already includes the baseline variables + the first group you added.

• Significance check works the same way:
  • Look at the Sig. F Change for this step. A p-value < 0.05 means this second variable group adds unique, significant insight into your anxiety outcome, beyond what the previous model already captured.
  • Again, you can use the F-distribution with df equal to the number of variables added in this step and the residual df of this latest model to compare against the critical value.

Quick tip for your anxiety-focused analysis

Hierarchical regression is ideal for your work because it lets you isolate the impact of different sets of predictors on anxiety. For example:

• Start with a baseline model of demographic controls.
• Add a group of variables like past trauma history.
• Add another group like current lifestyle factors.
  Each F Change tells you if that specific group is making a meaningful difference in explaining anxiety, over and above what you already accounted for.

内容的提问来源于stack exchange，提问作者user196525