Graham,
It’s not as ‘ad hoc’ as some of these references might suggest. It’s possible to calculate exactly how much of a sample you need to do but you need a statistician to work it out for you. See Wikipedia https://en.wikipedia.org/wiki/Statistical_power You just need to carefully define your variables and decide what your null hypothesis is going to be. I discuss sample number (or percentage) with researchers most months, and they all refer back to the power analysis undertaken by their collaborating statisticians.
Cheers,
Neil
From: Issues related to Historic Environment Records [mailto:[log in to unmask]] On Behalf Of Graham Lee
Sent: 07 June 2016 13:42
To: [log in to unmask]
Subject: % Sampling Strategies for site evaluation
Dear All – I have a question related to minimum acceptable percentage trenching sample to evaluate a greenfield site. The figure of 5% was used a lot in the past but is such a small sample still generally considered to be usefully valid? Can’t find anything in the CIfA or ALGAO libraries to update me so would be grateful for any help and useful (on-line) references.
Many thanks
Graham Lee
Senior Archaeological Conservation Officer
Please note that my usual working days are Tuesday to Friday.
North York Moors National Park Authority
The Old Vicarage
Bondgate, Helmsley,
York YO62 5BP.
Tel: 01439 772700