Monday, June 16, 2008

The Words We Use

Every one of us uses subjective words in our daily lives to express a level of confidence in the truth, accuracy or correctness of information about a fact or event. In genealogy reports, we often express that confidence in the correctness using words such as possible, probable, certain, unlikely, likely, impossible, perhaps, etc.

How do genealogists view these terms? What scale value from 0.0 to 1.0 do we assign to these subjective terms? This is the question Donn Devine asked recently on the TGF and APG mailing lists here. He is taking a survey to try to quantify them on the probability scale. Read his post to see the terms that he wishes to define on a probability scale.

Naturally, I sent him my list this morning before I went online to see if there are lists of scale values of probability for the subjective words that Donn is trying to define. I'll post my responses as a comment to this post, and I encourage you to do the same.

Are these terms really quantifiable? They are terms of probability and chance, and there is a scale from 0.0 (impossible) to 1.0 (certainty) as discussed in textbooks and this web page (among others).

I did find one article with a list of some of these terms on a scale, with plenty of discussion, with some value assigned to them - here.

It is important that we, as genealogy researchers and writers, have a common and familiar scale when we use these subjective terms. I encourage you to read Donn's post, and to send him your response with your own subjective scale (tell him I sent you!). He promised to post the results on the mailing lists when he compiles them.

1 comment:

Randy Seaver said...

Here are my responses to Donn Devine's question about those subjective words we all use...

Certainly - 99.5% (range 99% to 100%)
Probably -- 65% (range 50% to 80%)
Possibly -- 50% (range 35% to 65%
Likely -- 80% (range 65% to 95%)
Apparently -- 20% (range 15% to 35%)
Perhaps -- 35% (range 20% to 50%)

Other terms describing confidence levels

Almost certainly -- 95% (range 90% to 99%)
Undoubtedly -- 97% (range 95% to 99%)
Highly likely -- 90% (range 80% to 95%)
Very probably -- 80% (range 70% to 90%)
Unlikely -- 10% (range 5% to 15%)
Very unlikely -- 5% (range 2% to 10%)
Hardly possible -- 2% (range 0% to 5%)
Quite impossible -- 0% (range 0% to 0.5%)

I would have added Unlikely - 15% and Impossible - 0% to the first list so as to have a full range of terms.

So - my 9-point list after more thought about it would be:

0% -- Impossible
5% -- Highly Unlikely
20% -- Unlikely
35% -- Perhaps
50% -- Possibly
65% -- Probably
80% -- Likely
95% -- Very Likely
100% -- Certainly

A 7-point list might be:

0% -- Impossible
10% -- Unlikely
30% -- Perhaps
50% -- Possibly
70% -- Probably
90% -- Likely
100% -- Certainly

Certainly, it is probably impossible to assign likely values that could result in an unlikely scale of possibly subjective words.

I hope someone figures it out!