Metainformation
| Tag | Value |
|---|---|
| file | Inferential_Statistics_uu-Standardized-coefficient-810-en_uu-Standardized-coefficient-810-en |
| name | uu-Standardized-coefficient-810-en |
| section | Inferential Statistics/Regression/Standardized coefficient |
| type | schoice |
| solution | FALSE, FALSE, FALSE, TRUE |
| Type | Interpretating output |
| Program | SPSS |
| Language | English |
| Level | Statistical Literacy |
Question
The influence of motivation (M) and intelligence (IQ) on study success (S) can be analyzed via multiple regression. A regression analysis with SPSS gives the output below.
In the regression model, which of the two predictors, M (Motivation) or I (Intelligence), is the most important predictor of S (Study Success) and why?
- FALSE: M (Motivation), because this predictor has the lowest test size (t) out of the two.
- FALSE: M (Motivation), because this predictor has the highest unstandardized regression coefficient (B) out of both.
- FALSE: IQ (Intelligence), because this predictor has the smallest standard error (Std. Error) of either.
- TRUE: IQ (Intelligence), because this predictor has the highest standardized regression coefficient (Beta) of either.
Solution
To determine which variable is the most significant predictor, 3 aspects can be looked at: - p-value: a more significant predictor is more significant than a less significant predictor. A smaller p-value means more significance (note that in this case the p-value is not represented accurately enough to make a statement based on the p-value. In fact, the p-values rounded to 3 decimal places are the same size). - t-value: a higher t-value means that the predictor is more important. (For this reason, answer option A is not correct, this is just the wrong way around). - Beta: standardized regression coefficients are given here. A higher beta means a more important predictor. This leads to the correct answer: D. Both the unstandardized regression coefficient (B) and the standard error (Std. Error) cannot be used to locate the most significant predictor without additional information about the scales.