I’m trying to build a set of measures for a set of latent variables regarding the identity functions of materials that I can apply to software developers and the things their teams create. After trying to write up a set, I tested their face validity by asking fellow PhD students to sort the proposed measures into “buckets” representing those latent variables. I explained the functions to my participants like this:
The following table contains definitions of the identity functions for you to use in your sort. There is a “NA” column if there is no good option, so please evaluate every statement.
| Effectiveness (EFF) | The thing allows me a sense of control and to show competence. |
| Emotional Regulation (EMR) | The thing is a means to generate, express, or regulating emotions (self-esteem, belongingness, continuity/stability). |
| Actual Identity (AID) | The thing represents who I am right now (my qualities and values, differentiating me from others). |
| Ideal Identity (IID) | The thing represents who I want to be (the qualities, values, and goals to which I aspire). |
| Personal History (PRH) | The thing provides a sense of continuity with my personal history and how I became the person I am today. |
| Symbolic Interrelatedness (SIR) | The thing represents a link to an individual (a friend, a family member, a colleague). |
| Social Identity (SID) | The thing represents a link to (or membership in) a group or subculture. |
| Not Applicable (NA) | There is no good relationship between a material identity function and this statement. |
As I told my committee:
Succinctly put, that first test was horrible. It did show the way to improvements, and a second test was much, much better! (Michelle has been extremely helpful, advising me on the creation of the instrument.) Now, I’ve launched a broader test of those measures, hoping to evaluate them with real software folk prior to rolling them out as part of the full survey.
The “sort” exercise is here: https://ischooluw.co1.qualtrics.com/jfe/form/SV_bK4t0yaPcxqn8Hz
I did decide to just stop trying to develop measures for symbolic interrelatedness. This is supposed to regard “relationships with specific others” and there are too many “specific others” involved in software development to include them all. Some research into symbolic interrelatedness is now “future work.”
With the set of latent variables reduced to six, I used Prolific.com to screen for software developers and ask them to respond to the set: https://wsu.co1.qualtrics.com/jfe/form/SV_ehe3FxInziFwNvM
Analysis of those results are in another post.
One Reply to “Early Measure Testing”