Tensions over Testing

Bear with me

The navigational solution to getting off Ben Nevis is a technique called a ‘dog-leg’. This is a technique to avoid a particular point for any reason and deviate around it. In this case, navigating off Ben Nevis entails making a ‘dog-leg’ around the top of Gardyloo Gully.”

The Summit is Optional Tip 57

This is excellent advice and in the olden days, before I could get Global Positioning System (GPS) fixes on a mobile phone, it is a technique I often used for navigating mountainous terrain in a mist. To operate any strategy that involves a dogleg, you have two navigational challenges. One has to do with bearings and the other with distance. Let us see what further advice the website gives.

“This dog-leg involves first heading on a Grid Bearing of 231 degrees for 150 metres. Then, you switch and follow a Grid Bearing of 282 degrees.”

?Provided you know how to use that essential piece of the hill-walker’s equipment, the magnetic compass, the bearing bit is fairly easy. Harder is to know when you have reached the kink in the dog-leg. In this case, you need to have ?walked 150m but how will you know that?

Once, twice, three times a maybe

I have used three techniques when hill-walking, I know that I use about 60 double paces to walk about 100m, so, were I to apply this to this case, 90 double paces should be about right. I know also that it would take me about 12-15 minutes to walk a kilometre so I should be able to manage 150m in about 2 minutes. (For difficult terrain I might take longer.) I could also use a completely different technique to try and locate the link in dog’s leg and that is to use an altimeter to tell me when I have reached the height of the bend in the leg. This works fairly well in steep terrain but not so well if the area is flat.

I have used each of these at one time or another and sometimes all three together. Of course, there are some problems First, if I use more than one method there will always be some disagreement between them, which may be negligible but perhaps not always so. Then I have to resort to judgement to resolve any conflict and perhaps also pause to consider if I can work out why they disagree. Second, since the methods don’t always agree, if I have only used one, it might not be right. Thirdly, even, if the methods all seem to agree, they might be wrong (although I must confess, I have never doubted the result when this happens).

Of course, a modern GPS system should in theory trump them all. I say, in theory, because the systems seem to be usually very accurate but occasionally and alarmingly I have found them to be quite wrong. In any case, If your phone battery has died and you have forgotten to pack your power bank, it is as well to be able to use such old-fashioned systems.

Never mind "where's the point?", what is it and why?

The reader may be excused for thinking that this blog is a bit of a dog’s leg and perhaps even a shaggy dog story. What is the point I am trying to make? The point is this. In making decisions in ‘every day life’ or for that matter in science generally, it seems to be reasonable to have and use a number of 'instruments' at your disposal and nobody thinks any the worse of you for having more than one way of testing a hypothesis (for example, ?a hypothesis about the position you are at).

However, in statistics, our instruments are statistical tests or estimation procedures. Testing the same question more than once is something that usually carries a warning. In subsequent blogs I shall try and address this issue and consider to what extent the navigational analogy applies.

However, I am currently feeling my way cautiously in an intellectual fog, so it may take me some time to see clearly.

Watch this space...

...if you can find it again.