submit poster abstract using unvalidated measure?

[psych0000]

---

Members don't see this ad.

I'd like to submit an abstract very soon for a conference using data from a scale that has not yet been validated. The items are loosely based/adapted from a pre-existing scale. Is this a "no no" or will it fly for a poster abstract?

I am hoping the measure developer will get to completing validation for this adapted version in the future, for later use in a paper perhaps, but time doesn't allow for that with the deadline for this poster presentation, and I'd really like to use this set of items. Advice is appreciated ... I feel dumb asking the measure developer if this okay.

 

[deleted343839]

Depends on the conference, the quality of your study and write-up in other respects, significance of the research question, etc. There are "good" reasons to adapt existing instruments (eg, for use in a novel context for which the original version is not completely appropriate, etc.), and there are not so good reasons. People tend to make a disproportionate fuss out of minor deviations from validated instruments, IMO. However, the more you deviate from the original version, the more scrutiny is called for.

 

[Ollie123]

Depends on the conference. For most conferences, as long as your abstract contains actual words that have something at least vaguely to do with the conference topic and reflect something that reasonable audiences would construe as science (even crappy science), it will be accepted and you can present it. There are exceptions as some conferences are more competitive.

Even publications in top-tier journals often use modified versions of scales. Just depends on the context and whether or not you have justification for doing so. "We decided to expand the response scale from 5 to 9 to increase sensitivity to subtle changes" is very different from "Oops, we forgot to include a dozen items."

 

[Justanothergrad]

If you have a rationale and interesting study design, I doubt a conference reviewer would be a big problem with it. Do a little extra stats work on it to show evidence of validity and be sure to show the theoretical changes as a sentence in the methods section and it should be fine. Might consider an EFA or CFA in addition to a coefficient alpha to show strong evidence of validity. Don't give the full drawn-out model. Instead, just report the model fit statistics in the methods section of the poster. Shouldn't take more than a sentence for the validity evidence and one for the modification. You may want to also include a sentence comparing a sample item pre-post modification to give the reviewer an opportunity to see the types of changes.

This advice varies depending on how loosely-based the items are.