My statistical seven

Here are my seven rules for better understanding statistics.

  1. What is your emotional reaction?

How do you feel when you read about house prices, the rental market, housing affordability, new development, or the economy?

Many of us, if not most, accept or reject a statistical claim based on how we feel rather than on the actual facts.  Covid-related information is a perfect case in point.

So, my first rule – and often the hardest one to implement – is to notice your emotional reaction to a claim, rather than accepting or rejecting it because of how it makes you feel.

  1. What is the viewpoint?

Much of what we read about the property market (and also witness in a presentation, podcast, or video clip) is the ‘bird’s eye’ statistical perspective.

It is the big picture stuff, often summed up in one or two stats and covering a wide geographic area or product type.

I am guilty as charged when it comes to producing and sharing such information.

For example, according to Domain, the Brisbane housing market grew by 26% last year.

In contrast most of what we actually experience and hear from our friends, relatives and colleagues is a ‘worm’s eye’ view.  A ‘worm’s eye’ view comes from a personal experience.

Often the ‘bird’s eye’ and ‘worm’s eye’ view conflict.

“The Brisbane housing market is apparently booming, yet when my friend sold their home there last year, they really struggled to get their asking price.”

So, my second rule is that is often pays to combine the ‘bird’s eye’ statistical perspective with the ‘worm’s eye’ view.

  1. What is actually being said?

You should spend a few moments looking at the labels on a chart or the headings in a table and ask yourself if you understand what’s really being described.

Taking time to read the footnotes pays dividends too.

Many people over the years have told me that I said “x” when I said “y” and when I quiz them on why they think I said “x” more often than not that misunderstood the chart or table in one of my posts or during a presentation.

Now back to our Brisbane housing example.

What is meant by Brisbane?  Is it the Brisbane City Council area?  The Brisbane Statistical Division?  Or some other geographic area?

What is included in ‘housing’?  Detached houses only.  Attached dwellings too?  Only dwellings of a certain size or on land under a select square meterage?

What does last year actually mean?  Last 12 months to date?  Or a fiscal or calendar year time frame.

Is it a median or average price or some other measurement?

How old is the data and how big is the data set?  Is there a quorum to make such a statement like 26% price growth last year?

And has the data set been cleaned, removing agent’s advice or out of line sales, like family transfer’s or divorce settlements?

It pays to investigate such things and/or get your information from someone you can trust.  Hint, Hint!

  1. What is missing?

You should ask what is potentially missing from the data you’ve been shown, and whether your conclusions might differ if such exclusion/s were actually included.

Again, using our Brisbane price growth example.

Why pick just last year?  Best positive result?  What happens if we go back five years or better still ten?

What happens to the result if you break the region up into smaller areas?

Ditto when it comes to looking a resales only, especially without major renovations between sales.

  1. Who is saying it?

The cynic would place this rule at the top of the list and to be honest I often check out who is saying what before I invest too much time.

So, we all should look behind the statistics at where they came from.

We all have a bias, but some are more biased than others.

Also, for mine, experience counts.  Given my grey beard, bald head and increasing spread I trust an older head over a younger one.  But then again maybe I should revisit my first rule more often!

But then again older players in the statistical space sometimes have a reason to protect their past claims.  The same sometimes applies to official data sets as well.  Think the ABS and their labour force or CPI measurements.

I was told when I was a young lad – “Michael, forecast regularly but never let anyone see your track record”. Easier then, impossible now in the internet age.

Yet I have been known to delete the odd missive or two from my website!

But seriously folks in today’s post-truth world we seem to increasing want to maintain the narrative first and foremost regardless of the verifiable facts.

  1. Comparison and context.

We need to look for comparisons and context to help put any claim into perspective.

Let’s give the Brisbane housing example a rest of a comment and use some Covid-related intel to help illustrate this rule.

I keep reading about how deadly Covid is.  Yes, it has killed many, but it pays to understand the context.

Australia’s average death toll from all causes is about 160,000 per year.  That’s 440 deaths per day or just over 3,000 per week.

From the 1st of March 2020 to the 31st of January 2022, the average total deaths would be around 307,000.

The total number of Covid-related deaths in that period, whether “with” or “of” Covid, was 3,835 or 1.2% of all deaths.  

Sadly, and by way of comparison, in the 23 months to the 1st of February this year, Australia’s recorded number of suicides was 73% more than Covid deaths.  That’s 6,635 deaths.  Wow!

And it is also increasingly said that we best get used to hygiene practices like masking, deep cleaning, distancing, and isolation.  Yet the long-term normalisation of such practices reduces our exposure to microbes and thereby impedes the development of natural immunity.

Ironically vaccines may make the vaccinated more vulnerable to potential Covid virus variants.

Poppycock you say.

Well despite 76% and rising full vaccination since the 1st of November last year, more Australians have died “with’ or “of” Covid in the past three months than in the 20 months before.

Comparison and context matter.

  1. What is the time frame and relevance.

This is what I call “front page news” in my master class sessions.

When you are digesting information, I think is vital to ask if the data is really relevant and if so, then on what time frame does this relevance apply.

It is best maybe to think in terms of the ‘front page’ of the newspaper.  There are typically five or six stories on the front page, one of which leads with the biggest headline and usually the ink takes up about quarter of the front page space.  A second largish news account is next, followed by three or four smaller posts roughly of similar size.

So, what are the five or six things that really matter when understanding the topic in question?

Yet I think it is even more pertinent to view your ‘front page’ not through a daily prism but a longer time frame.  Weekly maybe, monthly hmmm, annually would be better.   But for many things a ten-year; generational or even lifetime horizon might provide more clarity.

In recent times any front page covering a weekly or even monthly time frame would feature Covid as either the lead or second story.

An annual and maybe even a ten-year front page would probably also hold a Covid column, but maybe not with a big headline.

A front page covering 20-odd years might also hold a Covid article but something small in the left hand corner.

And I do wonder if my lifetime front page (covering hopefully at least a 85 year time frame) would even hold the word Covid.

To close, let’s revisit our housing market growth example.

Any daily housing index is really just noise.  Ditto the monthly and even quarterly figures.  For mine, it is better to think on an annual basis and then best over a ten-year or longer time frame.

When it comes to the stuff that matters with regards to the housing market, may I suggest:

  • Market depth is important i.e., number of sales by price group and the underlying housing demand for housing it that area.
  • Supply, both new housing development and stock for sale, matter too and especially understanding the constraints (or otherwise) to that supply.
  • Timing also matters. Where is the market’s position in the property cycle and how has that marketplace behaved in the past?
  • What support is there? What is the relevant demographic, psychographic and economic trends?
  • Locational attributes. Is the locale unique and if so, why and will that distinctiveness mature?

And of course, I cover this material in the master class session too!

End note

So, these are my statistical rules of thumb or habits of mind that I’ve acquired, often the hard way, after nearly 30 years of doing this job.

Of course, who has time to apply these seven rules every time.  In fact, who can even remember them all.  I often can’t.

So maybe I could have saved you three or four minutes and me a couple of hours of my time and just written one rule, be curious.

I really do think we should all keep an open mind, asking on occasion if we might be mistaken or whether the verifiable facts have actually changed.

Oh, and remember most things you hear in the real estate space is either ‘recency and primacy’, in other words the most recent thing someone heard or was released and/or the tale that screamed the loudest.

Regrettably, many things released about the housing market don’t pass my seven rules or even a simple investigation by someone curious.

Copy link
Powered by Social Snap