Depends whether you have tits or not |
|||

Bebo |
Isn't the average around £23k now? |

Depends whether you have tits or not

Do tits affect your ability to calculate averages then?

Isn't the average around £23k now?

Depends whether you have tits or not

Do tits affect your ability to calculate averages then?

I have single source evidence that it could affect one's ability to do percentages.

You'd think it would be a figure they produce, but they don't.

No FOI request needed

http://www.statistics.gov.uk/cci/nugget.asp?id=285

Median salary is £26000pa

or http://www.statistics.gov.uk/downloads/theme_labour/ashe-2010/2010-all-employees.pdf for a more complete breakdown. (which interesting shows a lower median apparently for the same year)

Modal figures wouldn't really mean much because that could be thrown out by the breakdown of salary groups and the distribution of salaries is so skewed at high values (even if lots of people earn a high salary it's unlikely that a high salary would be the modal result unless you use unequal salary ranges to get your modal value)

And here was me thinking Modal was a type of cellulose textile!

Quick google later...

Ah- so presumably the most common salary...

So are you talking equivalent full time salary, or just actual salary independent of hours worked?

Dee

You'd think it would be a figure they produce, but they don't.

No FOI request needed

http://www.statistics.gov.uk/cci/nugget.asp?id=285

Median salary is £26000pa

or http://www.statistics.gov.uk/downloads/theme_labour/ashe-2010/2010-all-employees.pdf for a more complete breakdown. (which interesting shows a lower median apparently for the same year)

Modal figures wouldn't really mean much because that could be thrown out by the breakdown of salary groups and the distribution of salaries is so skewed at high values (even if lots of people earn a high salary it's unlikely that a high salary would be the modal result unless you use unequal salary ranges to get your modal value)

No no no - modal is the most frequently occurring salary, therefore the one nearest to which most of the population have to survive on.

The median is a completely different figure which really is skewed by very high and very low salaries.

And here was me thinking Modal was a type of cellulose textile!

Quick google later...

Ah- so presumably the most common salary...

So are you talking equivalent full time salary, or just actual salary independent of hours worked?

Dee

The most common salary is a very good way of putting it. I've asked for actual salary, so we'll see how they define it. If they ask which one I want, I'll ask for both

Isn't the average around £23k now?

Depends whether you have tits or not

Do tits affect your ability to calculate averages then?

I have single source evidence that it could affect one's ability to do percentages.

Isn't that going to be zero or will it only include people being paid a salary? In which case does that not exclude self employed people?

£13-14K is the 'modal salary' round here!

Isn't that going to be zero or will it only include people being paid a salary? In which case does that not exclude self employed people?

It's based on the figures that the ONS use for the average salary

You'd think it would be a figure they produce, but they don't.

No FOI request needed

http://www.statistics.gov.uk/cci/nugget.asp?id=285

Median salary is £26000pa

or http://www.statistics.gov.uk/downloads/theme_labour/ashe-2010/2010-all-employees.pdf for a more complete breakdown. (which interesting shows a lower median apparently for the same year)

Modal figures wouldn't really mean much because that could be thrown out by the breakdown of salary groups and the distribution of salaries is so skewed at high values (even if lots of people earn a high salary it's unlikely that a high salary would be the modal result unless you use unequal salary ranges to get your modal value)

No no no - modal is the most frequently occurring salary, therefore the one nearest to which most of the population have to survive on.

The median is a completely different figure which really is skewed by very high and very low salaries.

But because of the way the distribution is skewed the modal figure will vary according to the way you choose to group salaries. Salaries have to be grouped because salary is a continuous variable (1).

e.g. if a group of people have salaries of 10.2k, 10.5k 10.7k,12.3k, 35k, 40.3k 40.9k, 41.1k, 41.3k, 52k, 63k the the distribution is;

10-11k 3 people

11-12k 0 people

12-13k 1 person

...

40-41k 2 people

41-42k 2 people

so the modal salary is 10-11k. But group it differently and the distribution is;

10-12k 3 people

12-14k 1 person

...

40-42k 4 people

and the modal salary is 40-42k. Mathmatically both are right

Modal results are not really a suitable way to analyse continuous data, or data that has skewed grouping patterns. Because salaries are likely to have small intervals at the lower end of the salary range it is particularly prone to that (i.e. the same percentage difference at 10K can have two salaries in the same salary range whereas that percentage difference at 40k would put them in different salary ranges)

I see what you're trying to get at but there is no single correct 'modal' result for anything other than disrete data. Might it be better to use the percentile breakdowns which are available?

(1 - actually it's not but because it can vary by a single penny within a range which can cover hundreds of thousands of pounds it's needs to treated as one.)

Just had a response from them saying that the information is exempt as it is apparently something they can only do via their bespoke service.

I've asked for a quote

You'd think it would be a figure they produce, but they don't.

No FOI request needed

http://www.statistics.gov.uk/cci/nugget.asp?id=285

Median salary is £26000pa

or http://www.statistics.gov.uk/downloads/theme_labour/ashe-2010/2010-all-employees.pdf for a more complete breakdown. (which interesting shows a lower median apparently for the same year)

Modal figures wouldn't really mean much because that could be thrown out by the breakdown of salary groups and the distribution of salaries is so skewed at high values (even if lots of people earn a high salary it's unlikely that a high salary would be the modal result unless you use unequal salary ranges to get your modal value)

No no no - modal is the most frequently occurring salary, therefore the one nearest to which most of the population have to survive on.

The median is a completely different figure which really is skewed by very high and very low salaries.

But because of the way the distribution is skewed the modal figure will vary according to the way you choose to group salaries. Salaries have to be grouped because salary is a continuous variable (1).

e.g. if a group of people have salaries of 10.2k, 10.5k 10.7k,12.3k, 35k, 40.3k 40.9k, 41.1k, 41.3k, 52k, 63k the the distribution is;

10-11k 3 people

11-12k 0 people

12-13k 1 person

...

40-41k 2 people

41-42k 2 people

so the modal salary is 10-11k. But group it differently and the distribution is;

10-12k 3 people

12-14k 1 person

...

40-42k 4 people

and the modal salary is 40-42k. Mathmatically both are right

Modal results are not really a suitable way to analyse continuous data, or data that has skewed grouping patterns. Because salaries are likely to have small intervals at the lower end of the salary range it is particularly prone to that (i.e. the same percentage difference at 10K can have two salaries in the same salary range whereas that percentage difference at 40k would put them in different salary ranges)

1 - actually it's not but because it can vary by a single penny within a range which can cover hundreds of thousands of pounds it's practical to treat it as one.

I see what you're trying to get at but there is no single correct 'modal' result for anything other than disrete data. Might it be better to use the percentile breakdowns which are available?

All that works beautifully if you have one or two values, as your example shows, but when you have many millions of figures, and do it to the nearest £500 or £1000, which is what I've asked for.

The same effect applies if you have millions of data points, in fact it's more likely to occur then as the skewing of the data is more obvious and consistent, for example (making up numbers to illustrate the point)

10-12k 1 million

12k-14k 1.5 million

14-16k 2.5 million

...

40-42k 0.9 million

42-44k 0.9 million

44-46k 0.9 million

46-48k 0.9 million

48-50k 0.9 million

Modal result is 14-16k, but most people are on 40-50k. The long tail on high salaries means that they will never appear in the modal data if you use equal salary groups. Use unequal salary groups to address that and you might as well use the percentile breakdown.

(this is where being able to post a quick sketch would help)

What we need is a bell curve, but the chance of getting that are pretty remote.

I do understand what you're saying - I just don't think the figures will come out like that.

What we need is a bell curve, but the chance of getting that are pretty remote.

I do understand what you're saying - I just don't think the figures will come out like that.

It does come out like that. e.g. from the figures at http://www.statistics.gov.uk/downloads/theme_labour/ashe-2010/2010-all-employees.pdf the percentile breakdowns are

10% £6480

25% £11076

75% £32665

90% £46428

so 15% of the population have a salary in a £4500 wide bracket and 15% of the population have a salary in a £14000 wide bracket. Use equal salary groups to determine the mode and you are guaranteed to get the lower result even though there are the same numbers of people those groups.

Those are not actual figures, they are estimates based on a sample of about 24 thousand jobs aren't they?

Actually, just checked. It's based on 181,000 jobs. What a load of twaddle.

No no no - modal is the most frequently occurring salary, therefore the one nearest to which most of the population have to survive on.

If only things were that simple. However, am I correct in thinking that many people aren't in full time work and there are many who have "portfolio" employment, i.e. more than 1 part time job. Then there are us pensioners (millions of us dragging the economy down). Regarding the impact of bosoms on earnings, I understand that a disproportionate number of part-timers are ladies.

At a guess, the largest group salary-wise are those of us without a salary.

Those are not actual figures, they are estimates based on a sample of about 24 thousand jobs aren't they?

A random sample of 1% of PAYE records. So a sample set of about a quarter of a million, so it should be pretty accurate. It will have missed the black economy and those who for some legitimate reason are outside of the PAYE system, not sure how you could meaningfully measure those though.

http://en.wikipedia.org/wiki/Income_in_the_United_Kingdom#Percentile_points_for_income_of_individuals_before_tax

That of any interest?

http://en.wikipedia.org/wiki/File:British_household_income.jpg

Now that's the sort of thing

But up to date.

Actually, just checked. It's based on 181,000 jobs. What a load of twaddle.

Twaddle? Why do you say twaddle? Even if it's not representative for one reason or another, such as the black economy or undeclared offshore earnings, you're not going to get a different result from a FOI to the same department with the same underlying data.

http://en.wikipedia.org/wiki/File:British_household_income.jpg

Now that's the sort of thing

But up to date.

DINKS?

Quote: |

Post Tax Household Income (Britain, households with two adults, no children, 2006) |