Our main prediction is that participants will believe ability to delay gratification will decrease. This will be analyzed using a chi square test. This will be done because we do not forsee collecting enough data for an ordered probit regression.
A secondary prediction, however, is that older respondents will be more likely to believe delay of gratification ability will decrease. For this we will run a flipped ANOVA, with participant age as the DV and the three response options as the IV. We will follow-up a significant omnibus test with LSD and Bonferroni post-hoc testing. The results from this analysis are qualitatively the same as an ordered logit regression, though the need for larger sample sizes can be better handled.
We will run epxloratory analyses on the rest of the data.