Passive voice, agency, and survey format

I'm designing an opinion survey and one of the factors I am thinking of manipulating is the way the question is phrased. Specifically, I'm thinking of contrasting the following two item phrasings:

1) I think people with mental illnesses are frightening.

2) People with mental illnesses are frightening.

In both cases, the survey response options would be on a five-point scale ranging from strongly agree to strongly disagree. I think these wordings will impact responses, because in the first item you're asking someone their subjective opinion, but in the second item you're asking them to evaluate a statement presented as objective fact.

However, I'm having a hard time finding literature on this topic and I think part of the problem is that I'm not sure of the technical term for the linguistic element I'm manipulating.

Is this an example of active (sentence 1) vs passive (sentence 2) language? Or is this an issue of agentive (sentence 1) vs non-agentive (sentence 2) language?

Am I conflating multiple issues here (e.g., voice, case, agency, etc)?

I've pulled a lot of literature on active vs passive voice, and some on transitivity and agency, but most of the examples I've found are not in the first person, so I think that's also confusing me.

I would greatly appreciate guidance on how to describe the phenomena I'm trying to study, and any references that may be of use.

It sounds like you are talking about framing effect

The framing effect is an example of cognitive bias, in which people react to a particular choice in different ways depending on how it is presented

This is discussed extensively in the book Thinking Fast and Slow by Daniel Kahneman.

It also seems to relate to the broader areas of cognitive bias and response bias

all from Wikipedia