Part 1.
1.1.a Null hypothesis: divorce either has no effect on or decreases anxiety in children
1.1.b We should fail to reject the null hypothesis (the groups did not differ)
1.1.c We should reject the alternative/scientific hypothesis
1.1.d We would make a TYPE II error (this is when there really IS an effect of the IV
on the DV but we could not find it)
1.2.a People are equally or less likely to comply with a difficult request when it is preceded by a less difficult one (or "foot in the door" technique has no effect on or reduces compliance)
1.2.b We reject the null hypothesis (people were more likely to comply)
1.2.c We fail to reject/find support for the alternative/scientific hypothesis
1.2.d We would make a TYPE I error (there really is no effect of the "foot in the door"
technique but we thought there was; the difference was likely due to chance)
1.3 Rejecting the null means we assume the IV did affect the DV - we assume the difference was NOT due to chance
3.1 IV is history of sexual/physical abuse (levels: abused & not abused). It is not ethically manipulable; it is a subject/participant variable
3.2 The DV is the emotional reaction to threatening stimuli; the researchers measure the change in expressiveness on the FAI
3.3 The null hypothesis is that abused children are more or equally emotionally reactive compared to non-abused children
3.4 The alternative hypothesis is that abused children are less emotionally reactive
3.6 The p-value of .03 is less than the alpha level of .05, thus suggesting we should reject the null. This indicates we observed a significant difference between groups in terms of emotional responses. However, the observed difference was the opposite of what we had predicted - indeed, abused children were MORE reactive than non-abused children. Therefore, the null hypothesis is still valid.
3.7 Based on the p-value alone, we should fail to reject the alternative hypothesis. However, since the observed difference between groups was the opposite of what was expected, we should revise our theory and reject the alternative hypothesis.
3.8.a If the abused group had a score of 4, the p-value would probably decrease. The decrease is driven by the fact that the observed difference between groups is larger (6 points) than it was before (3 points). The larger the difference, the less likely it is to occur due to chance alone, therefore the smaller the p-value. Remember, the p-value is the probability the observed difference was due to chance. So, a larger difference usually implies a smaller p.
3.8.b Now we can reject the null hypothesis. We observed a significant difference in terms of emotional reactivity between abused and non-abused children, and the difference was as predicted (i.e., abused children were less reactive).
3.8.c Now we would accept the alternative hypothesis - we predicted less emotional
responsiveness from abused children and that is what we found.
3.9 No, it means the difference is statistically significant (do not confuse "importance" with "significance").