Ok, to let you know, I am an undergraduate student who has taken statistics, research methods and additional courses in psychology. I offered up some ideas earlier, thinking perhaps you were an undergraduate like me, not actually doing this research (which does sound very interesting!).
So I am in my last year, encountering for the first time just how involved this process is, through research projects in my courses. What I am learning is that just as everyone has said, what you are talking about cannot be done by one (specialist) person in a day, or in a week, even if they do have the extensive content knowledge required and are fluent in research methods and design. Even the selection of an existing measure, and showing how it might apply to your research, is time-and labour-intensive, even for an expert.
For example, you mention empathy. There are thousands of papers written on empathy. It has been defined and measured in various ways, looking at cognitive, emotional, behavioural and perceptual dimensions. So empathy, as a construct, has to be carefully considered in the light of existing research; just there, there are multiple streams. This is also true for mimicry and joint attention. Say that one or multiple forms of empathy (defined in one or more ways), or joint attention, or mimicry, are involved in attachment - how, and when? What evidence justifies your application of these concepts to your proposed method?
Further, it's one thing to talk about empathic processes between humans, but what additional variables would complicate extending (generalizing) that notion to robots? This should also be addressed. What about the nature of your stimulus materials? They have perceptual qualities that may affect the behaviour and subjective experience you observe and measure.
All of this involves assumptions, each of which must be substantiated by prior research in order to lead to valid interpretations. All of that work would go into selecting a questionnaire, and even more would go into creating one. (Also if language issues are involved - if you are translating English questionnaires to Japanese - that is another aspect of work that would likely involve pilot testing.)
(again, and i'm just a student - attachment seems to me to be an unsuitable construct for this proposal. in humans, attachment is thought to occur over repeated interactions with important people, in a number of contexts, over time, not one or five occasions; and it is definitely conceptualized as fundamentally driven by biological systems and needs. i sort of think limiting your measurements to the variables described in your second explanation might be more appropriate for what i can understand about your design, e.g., for cooperation, "how much did you feel you cooperated with the robot", etc. but i am REALLY not an expert and this is the sort of thing such a person could help you with.)
