Australian Branch Stats?
by Marvin Shilmer 37 Replies latest jw friends

Petraglyph
@sir82 we're extrapolating the figures worldwide 
Marvin Shilmer
Something doesn't sound right  that's nearly half the current population of Australian JWs.
The extrapolated value is no more an no less than a result of the same ratio actually reported by the Australian Branch, which (based on the values included and cited above) is an annual average of 0.000401 per peak publisher for years 20052014.
Check the math if it doesn't sound right. That's why I've put this out there.

sir82
Aha  read too fast. Completely missed the "worldwide" there. I thought the figure was for Australia only.
Never mind  carry on.

Marvin Shilmer
'10+' could mean any number 10 and above, and 'Unclear' could mean any number 1 and above.
These were not being counted in the total.
I replaced all the '10+' with the number 10, and all the 'Unclear' with the number 1...
The total number of victims on this list is 1857.
This means the ratio is 1857 victims / 1006 abusers = 1.85/1 ratioOkay,
Where the RC states "10+" I'm going to assign 10 as the value.
Unclear means at least 1. For each of these instances I'm going to assign 1 as the value.
We'll see what that looks like.

millie210
Marvin I sent you a PM 
Marvin Shilmer
Of victims:
Based on the Australian Branch data, and subject to the same statistical base I describe above in my initial disclosure of BASIC STATISTIC,
For the 10year period of 20052014 I find victims at an annual average rate of 0.000645 per peak publisher. (As a hard number for the Australian Branch this is 412 victims for the period.)
Extrapolated for a worldwide total, for the 10year period of 20052014 the value is just over 46,000 (specifically in my statistical slice: 46,102).
PS: There were no "10+" counts in the data slice use in this 10year period that had to be adjusted to read "10". There were 3 occurrences of "Unclear" that were assigned a value of 1.

Petraglyph
What ratio did you identify from your statistical slice? It's evidently lower than for the whole data. 
Marvin Shilmer
What ratio did you identify from your statistical slice? It's evidently lower than for the whole data.
I'm not sure what you're asking me to calculate as a ratio.
Take a look at my second post in this discussion. There I express what I counted and did not count. Maybe that'll answer your question. If not please rephrase your question.

Petraglyph
If not please rephrase your question.
Ok.
For the whole dataset I got 1857 victims and 1006 abusers. This calculates to an ratio of 1.85 victims per abuser.
I'm curious as to this ratio for the 10 year data slice you have extracted, as it appears from the figures you've produced that the ratio will be lower for the latter 10 years than for the dataset as a whole  which would suggest there's a downward trend.
Also if you can separately determine the ratio for the period prior to 2005 we can absolutely confirm if there is a downward trend.

Marvin Shilmer
For the whole dataset I got 1857 victims and 1006 abusers. This calculates to an ratio of 1.85 victims per abuser.
I'm curious as to this ratio for the 10 year data slice you have extracted, as it appears from the figures you've produced that the ratio will be lower for the latter 10 years than for the dataset as a whole  which would suggest there's a downward trend.
Also if you can separately determine the ratio for the period prior to 2005 we can absolutely confirm if there is a downward trend.My data slice does take a look at each year and I noticed that 2005 has a significantly higher incidence rate than other years in the 10year period. (This is one reason why I presented an extrapolation based on a 10year average rather than any single year.) Before I finalize anything for wider use I'm thinking about removing the highest and lowest incident years to form a more conservative extrapolation. But before I do that I want to analyze the data more to see if i can identify why some years might have lower or higher numbers. There are instructions to elders over the last 25 years or so that might have driven some of what we find in the numbers. One reason why I selected the 10year period of 20052014 was to avoid some years (by a fairly wide margin) where numbers could easily be skewed either upward or downward for reasons of fluctuating Watchtower policy instructions to elders. What I've presented here is just a beginning.