Can I change the level of measurement?

Got it, thanks ?.

Parametric statistic: One-way ANOVA
Non-parametric alternative: Kruskal-Wallis’s test

I also want to compare three groups, instead of two groups. Which way I should adopt?

Thank you Dr. MEL and Colin!!

The assumptions are as follows:
• Normality: Data are normally distributed
• Homoscedasticity: Data from multiple groups have the same variance
• Linearity: Data have a linear relationship
• Independence: Data are independent

What are parametric assumptions?

Please consider using a non-parametric alternative, which is Wilcoxon sign test, instead of paired sample t-test. Many thanks.

Thx. But if my data don’t meet parametric assumptions, what should I do?

Please consider using paired samples t-test. Thanks.

I want to measure if there is a change in participants’ scores of subjective well-being from Wave 1 to Wave 2. Which statistical technique is required?

Yes, but you need to assess the benefits and disadvantages carefully when considering changing the level of measurement. To change a continuous variable to a nominal variable, you can collapse the continuous variable into a number of categories. To change a nominal variable to a continuous variable, a number of dummy variables can be created.