Investors are often told to consider their attitude to risk before choosing an investment. But what, you may be wondering, does "risk" actually mean?

"One definition is `the likelihood that the outcome you expect does not happen'," says Mr Graham Harrison, managing director of Asset Risk Consultants, which analyses the performance of investment funds.

This sums up the problem: risk means different things to different people because their expectations differ. For example, few investors worry about liquidity risk - the risk that you will not be able to cash in your investment when you want to. But in Malaysia, where many investors got stuck in 1998 when emerging stock markets crumbled, lack of liquidity is a real and present danger.

Then there is the risk of investing at the wrong time, the risk of under-performance and the risk of loss. There are so many types of risk that a nervous investor could be forgiven for converting everything into banknotes and putting it under the mattress.

Not taking enough risk is also risky. Mr Martin Spencer at Riskmetrics, an investment analysis company, says: "There are more people taking too little risk than too much. The risk of not taking enough risk is that you won't achieve your investment objectives."

This boils down to a reminder that risk equals return. The reason equities deliver higher investment returns than bonds or cash over time is because it is a riskier investment.

Yet there are ways for canny investors to lower the risk they are running without putting every last penny into bank deposits. Prof Paul Marsh, head of finance at the London Business School, splits investment risk into two categories: risk that can be lowered by diversification and risk that cannot. At its most basic level, diversification means "not putting all your eggs in one basket". But this means much more than just geographical diversification. It is also important to spread your investment between different fund managers and different asset classes - equities, bonds, cash and so on.

"By mixing asset classes you are far less likely to have a problem," says Mr Harrison. Mixing fund managers is also a good idea. "As investment houses become more disciplined and institutional, all their portfolios will reflect a house view," says Mr Harrison. "If that view is wrong, you have a problem."

Additionally, fund managers go in and out of fashion with as much regularity as investment styles. You have to mix firms with styles - "value" investors who seek out cheap shares or "growth" investors who emphasise fast-growing companies - as well as diversifying across continents or asset classes.

So how can you tell whether you are taking too much risk? There are, of course, a string of statistical measures of risk for an individual stock, based on past performance. But that, as any investor knows, is no guarantee of the future.

"Risk defies a lot of data crunching," says Mr Mark Rhodes of the British Financial Services Authority.

"People tend to look at one measure, whereas we prefer a range of measures. But, even then, risk doesn't tie up very well year-on- year. For example, even if you pick on the basis of consistency of performance, it is not certain that those stocks will continue to perform consistently in the future."

Most systems for measuring risk rely on measuring volatility - which is useful, but only part of the story.

"Hedge funds have a very low volatility, but the operational risk can be high if they are run by one guy in an attic," says Mr Harrison. "People tend to focus on volatility because it's easy to measure."

One crucial element in measuring risk is time. In simple terms, a 25-year-old investing for a pension might not be particularly reckless in putting money into Japanese smaller companies; for a 65-year-old with limited capital, it could be an act of supreme riskiness.

Viewed from a different angle, this means that the top 10 least risky (or least volatile) funds over one year will be very different to the top 10 over 10 years - a fact that investment advisers struggle to explain to clients.

"One problem is that people see daily market moves and find it very difficult to adopt a 10- or 25-year view. It's a temptation to bail out after three months' poor performance," says Mr Harrison.

Sadly, there are no simple answers when it comes to risk. But, without doing the calculations on your own situation very carefully indeed, you could end up carrying a lot more risk than you should.