901170
| Question | Answer |
| Linear vs. everything else | linear is a straight line, everything else is curvilinear |
| Why are curvilinear ignored in the literature? | People generally don't think in terms of polynomial predictions, theories. They don't know how to compute them. They don't know how to create a figure for them. They struggle interpreting them. |
| The most famous curvilinear? | Yerkes-Dodson Law. Performance x Levels of Arousal As levels of arousal increase, performance increases descreasingly until it hits an asymptote and then performance increases level off |
| Compute heirarchical regression | Multiply the IV by itself, do nothing with the DV, Enter terms in ascending order ones step at a time, look at the R squared change values |
| R squared change values | Any values above .015 should be taken seriously. If below .015 but still registering as significant then perhaps look at it more closely. |
| Slope on a linear? | Constant |
| slope on a curvilinear | is not constant, it changes across the x variable. |
| Quadratic has | one distinct curve |
| cubic has | two distinct curves |
| In the macro the x values are how many SDs? | -2SD - M - 2SD Mean is the centre point. This can mean sometimes raw data is left out |
| Any interesting results should be instantly? | replicated |
| If there is a significant linear relationship then perhaps there may be some going on in the data. | So take a look at the quadratic and cubic etc. |
| polynomial relationship qualifies the linear relationship. | It tells you more about how the relationship is working. That the Pearsons r is not constant across the whole range of the IV but changes |
| Shapes of quadratic slopes | U-shaped, inverse U-shaped Upward, downward, level |
| Finding polynomials | Begin with null hypothesis, ora weaker than expected significant correlation, or don't know |
| Non-sginificant linear is related to quadratic how? | It is not. They are all independent. A non-significant linear relationship does not mean that there is not a sig. quadratic. They don't affect the presence of each other |
| A non-sig quadratic looks like what? | A flat line |
| Does the direction of the analysis/graph, i.e. from x to y or y to x, make a difference? | Yes it does. In some instances it has been shown that x can predict y but y does not predict x. x squared and y squared are not mathematically equivalent. |
Want to create your own Flashcards for free with GoConqr? Learn more.