Shulou(Shulou.com)05/31 Report--

Today, Xiaobian will share with you the relevant knowledge points on how to look at the bootstrap mediation effect. The content is detailed and the logic is clear. I believe most people still know too much about this knowledge, so share this article for your reference. I hope you will gain something after reading this article. Let's find out together.

The test for mediating effects in bootstrap does not need to look at p values; the test for mediating effects in Bootstrap method is judged according to whether the interval "BootLLCI, BootULCI" contains 0. If it does not contain 0, the mediating effect is significant, and if it contains 0, the mediating effect is not significant.

Operating environment of this tutorial: Windows 7 system, DELL G3 computer

Bootstrap mediating effect depends on p-value

The Bootstrap method is used to test the mediating effect, not by P value, but by whether the interval (BootLLCI, BootULCI) contains 0. The mediating effect is significant without 0, and not significant with 0.

In the case given, the value of the mediating effect (indirect effect) is 0.1969, which is significant, and the total effect of the independent variable on the dependent variable is 0.9373, which means that the mediating variable mediates 21% of the effect (0.1969/0.9373), which is an incomplete mediating.

Note also that the above figures are non-normalized effect values. SPSS only gives normalized values for intermediate effects, but not for total and direct effects.

In fact, using normalized or non-normalized effect values to calculate the proportion of mediating effects yields similar results.

Bootstrap is most widely used in tests for mediating effects.

Other methods of mediating effects include:

·Most commonly used: Stepwise test regression coefficients (stepwise method)(Baron & Kenny, 1986)

Step one: check if X→Y, which is c, is significant. If it's not significant, don't do it.

Step 2: Test whether X →M, M →Y are significant, both must be significant, even if one is not significant, do not do it.

Step 3: If all of these are significant and c'is smaller than c then it's partial mediation. If c'is not significant, then it is completely mediated. This is rare.

The Sobel Method:

The test power is higher than the stepwise test, but assuming that a*b obeys a normal distribution, even if a and b are both normal distributions, the product is usually not normal.

So Sobel has limitations.

Bootstrap advantage: no normal distribution required, higher sensitivity (easier to produce significant results)

Bootstrao test for mediating effects, taking Process plug-in in SPSS as an example:

Step 2: Set parameters.

Select dependent, independent, and control variables from Variables, which will form your regression equation.

Model number Select 4, which is the model number for the mediation analysis. If you choose any other model number, you will report an error.

Number of bootstrap samples, is the bootstrap sample size mentioned above, the default is 5000, generally between 1000 and 5000, usually 1000 or 5000. Bootstrap samples vary in size, resulting in slightly different data.

Also check Save bootstrap estimates and bootstrap reference for model coefficients.

Click [Options] in the upper right corner, check show total effect model, and click Continue

Finally, click [OK] to get the calculation result.

Result of calculation:

1. Regression Results with Mediating Variables as Outcome Variables

2. The dependent variable is the outcome variable. At this time, we can get the influence of MV on DV and the direct influence of IV on DV.

3. Total effect model: This is the total effect of the independent variable IV on the dependent variable DV without the mediator MV; that is, all the effects of the independent variable on the dependent variable before mediation, and the effects of the independent variable on the dependent variable after mediation will be divided into two parts: direct effects on the dependent variable (direct effects) and indirect effects on the dependent variable (indirect effects).

Here's the point!! The most direct mediation effect is here!!!

4. Results of the mediating effect test.

That's all for "bootstrap mediation effect". Thank you for reading it! I believe everyone has a great harvest after reading this article. Xiaobian will update different knowledge for everyone every day. If you want to learn more knowledge, please pay attention to the industry information channel.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.