Diary of a Disordered Mind

an online scrapbook 
Tags


Filed under

statistics

 
March 11, 2009

Understanding Risk

via news.bbc.co.uk magazine

This is taken from a great article from the magazine section of news.bbc.co.uk.

Looking at those statistics cold we're led to believe that bacon is the route to an early grave. Should I be cutting all delicious meaty food-stuffs from my diet? Well, as this articles shows, I shouldn't be overly concerned. What these types of statistic often fail to show is the initial risk of said fatality from occurring. As it happens, around 5 in every 100 people get colorectal cancer. This of course takes no lifestyle choice into consideration so if you're already guzzling a whole hog every day, your risk is probably higher. Anyway, as the article describes, an increased in risk by 20% would equate to 6 in every 100 people suffering from colorectal cancer. So although risk is greatly increased, your overall statistic chance of suffering the disease is only slightly increased. Make sense?

Similarly, although drinking too much increases breast cancer risk by 12%, it correlates to an increase of only 1 in every 100 (from 10 to 11%) people.

What they're trying to show is, it all depends on initial risk.

Filed under  //   science   statistics  
Loading mentions Retweet

Comments [0]

March 8, 2009

Quote

There are three kinds of lies: lies, damned lies, and statistics.
via en.wikipedia.org

Filed under  //   quote   statistics  
Loading mentions Retweet

Comments [0]

March 8, 2009

Tesco's 'Real Baskets' - Figure Manipulation?

via tesco.com

Asda recently ran an advertising campaign stating that, according to 'independent price checker' mysupermarket.com, Asda have a greater number of cheaper products compared to their rivals. Fair enough, that's easy to understand; when you directly compare products, Asda appear cheaper. That makes a lot of sense to me.

Tesco, of course, are having none of it! What (as far as I can make out) they think is that, although Asda are demonstrably cheaper on the majority of produces they stock when compared with Tesco, these products are NOT products the majority of people buy. So what if Fray Bentos pies are cheaper at Asda if Britain is populated by vegetarians? So Tesco have set about proving that it is they who are cheaper if you look at all the products people buy in an entire weekly shop. This is easily done as we all have to go through the checkout when we buy our shopping, surely Tesco will use this data to in it's calculation?

Err, no.

In fact Tesco decided to only use purchases made by clubcard holders. I'm assuming they did so because it's easier to get data from their clubcard database than from raw checkout data rather than because clubcard holders buy the cheapest products. So we now have data for over 2 million weekly shops, we can analyse these and workout whether Asda is cheaper or more expensive.

Well, in fact, no.

What Tesco decided to do was take a 'random sample' consisting of 10 percent of these weekly shops and work out from these whether the shopper would've been better off going to Asda. They then round back up to 100 percent. Why? Is it really necessary dividing and multiplying by 10 to work out the figure?

A sceptic would suggest that Tesco's method is a way of fudging the results in their favor. I suppose it all depends on how 'random' the selection really is. But I'm positive that a more accurate result could be calculated using ALL of the available data rather than a 10 percent of it.

To be honest, I don't care who's cheaper, I would much rather go to Sainsbury's.

Filed under  //   shopping   statistics  
Loading mentions Retweet

Comments [0]