So do you think Likert scales might actually be cardinal? I've heard people whisper about research findings hinting at that, but never heard anyone actually make the case.
0. I don't know. My heart says yes, they are approximately linear in the trait they measure, but I wouldn't bet my life on it.
1. We often find linear correlations between Likert scales and the variables we would expect to be related to happiness. Some people think this is very strong evidence that the Likert scales are linear in happiness, other people don't (e.g. it's possible that happiness is non-linear in Likert scales but the effect of happiness is non-linear on the dependent variable and these cancel out- whether this is plausible or ad-hoccery varies from person to person).
2. What I do think, almost enough that I would bet my life on it, is that the distribution of happiness, including between different groups of interest, is such that, and the the deviations from linearity in the Likert scale are small enough that, **where the numerical average happiness of one group on a Likert scale significantly exceeds the numerical average happiness of another group on a Likert scale, in almost all cases the real happiness of that group exceeds the real happiness of the other group.**. This is the truly important thing, as it allows us to use happiness
averages for many utilitarian purposes.
3. There are a variety of very cute methodologies that have been worked out to try and get at the real relationship between the Likert scale and the variable. While in still in development, these methodologies, if they worked, would allow us to correct Likert scales to linear, if they are not linear already.
4. I discuss these issues in chapter 6 of my thesis:
Haha that would certainly be in my interests! I should add that chapter 6 is still quite rough, also I just added in, in informal rough draft notes, a summary of this whole brouhaha to this chapter.
So do you think Likert scales might actually be cardinal? I've heard people whisper about research findings hinting at that, but never heard anyone actually make the case.
This is a great question.
0. I don't know. My heart says yes, they are approximately linear in the trait they measure, but I wouldn't bet my life on it.
1. We often find linear correlations between Likert scales and the variables we would expect to be related to happiness. Some people think this is very strong evidence that the Likert scales are linear in happiness, other people don't (e.g. it's possible that happiness is non-linear in Likert scales but the effect of happiness is non-linear on the dependent variable and these cancel out- whether this is plausible or ad-hoccery varies from person to person).
2. What I do think, almost enough that I would bet my life on it, is that the distribution of happiness, including between different groups of interest, is such that, and the the deviations from linearity in the Likert scale are small enough that, **where the numerical average happiness of one group on a Likert scale significantly exceeds the numerical average happiness of another group on a Likert scale, in almost all cases the real happiness of that group exceeds the real happiness of the other group.**. This is the truly important thing, as it allows us to use happiness
averages for many utilitarian purposes.
3. There are a variety of very cute methodologies that have been worked out to try and get at the real relationship between the Likert scale and the variable. While in still in development, these methodologies, if they worked, would allow us to correct Likert scales to linear, if they are not linear already.
4. I discuss these issues in chapter 6 of my thesis:
https://docs.google.com/document/d/17dw0_Ukp_98jmWBr8eyLkxSB1vXeLpi9YbodhQm3q2Y/edit#
Thanks for the response! I feel like I’m going to end up reading your entire thesis chapter by chapter at this point 😅
Haha that would certainly be in my interests! I should add that chapter 6 is still quite rough, also I just added in, in informal rough draft notes, a summary of this whole brouhaha to this chapter.