Skip main navigation

New offer! Get 30% off your first 2 months of Unlimited Monthly. Start your subscription for just £35.99 £24.99. New subscribers only T&Cs apply

Find out more

Measuring Personal Income

In this video Professor Melanie Luhrmann explains to Professor Michael Spagat the pitfalls of measuring personal income as well as possible solutions.
Hello, everybody. I’m here with Dr. Melanie Luhrmann, who is in my own economics department at Royal Holloway. Melanie, you’ve done a lot of work on economic inequality. So let me just ask you a question that a lot of people have on their minds, is inequality growing? Shrinking? Staying about the same? Well, that’s a difficult question. It obviously depends on where in the world you look. It also depends on what you think inequality is. Inequality really could be measured in lots of different dimensions. The most typical would be income inequality. But if you think about it when we’re thinking of inequalities, we’re thinking of how are the poor faring relative to the rich? Is there a massive gap?
And then the question we need to answer first, is who’s poor, and who’s rich? And how do we measure that? OK, so measurement is really important. So what are what are the issues that we face when we try to measure, let’s say, income? Income is the most frequently measure used. Obviously, we need the income data to measure the question. What happens in service is the same as what happens on the Sunday lunch table, when you’re sitting with friends and family members. And you ask them, Mike, how much do you make a year? You might actually frown, and say, I don’t really want to tell everybody what my income is. So people might give you an evasive answer.
The same happens when you ask open-ended questions– what is your annual income before taxes– in surveys. When you ask open-ended income questions in surveys, you get a negligible fraction of people who refuse to respond. So an open-ended question will be simply, what is your income, and then you fill in the blank? Exactly. And so you get a lot of just, no response? Exactly. And the second problem that you’re catching when you’re asking open-ended questions is that basically, you have the interview on the telephone you don’t quite remember exactly what your income is so you’re rounding. You know the order of magnitude so you’re rounding. You’re saying oh, my income is 25,000 pounds a year.
But reporting income relatively precisely might ask for some questions, too. Think– for example– of student loans in the UK. Student loans in the UK only need to be repaid after you’ve studied if you earn above a certain income threshold. This income threshold has recently been raised from 25,000 to 25,750 per year. An now if you were interested in knowing whether this rising of the threshold has had impacts on the fraction of people who have to repay their loans, then you need to know not just that it’s around 25,000, but where exactly it is. Is it at 19,750? Is it at 25,999? We call that rounding, when people round to the next 1,000 pounds or the next 500 pounds.
That can become a problem for certain questions that you may want to ask, too. I see. But for some purposes, you might want to just accept very rough answers. You just gave an example where you actually want a pretty precise answer, but sometimes, we might be content just to know if it’s between 10, 20, 20 and 30, et cetera– like that. Yeah, that’s very practical because it’s much easier in service to ask people– Mike, are you making between 20,000 and 30,000 a year, or between 30,000 and 40,000, or between 40,000 and 50,000? We call these bracket questions. And this bracket format seems much less intrusive to people. And so, you get a much higher response rate.
So you don’t face a problem that you have asked income, you want to now look at income inequality. But 20% of the respondents of your survey have given you no indication of what their income is. So can we take that as a fact that when you just go with wide categories, more people give an answer? Yes, yes. That’s a known survey technique. So really, how you want to ask about income always depends on what you want to do with the data afterwards. And are the problems different at the high end of the spectrum compared to the low end? You’ve been giving examples of incomes between 20,000-30,000, 30,000-40,000, but we might be particularly interested in extremely high incomes.
Do we think that those are more difficult to ferret out and to get accurate information on? So the main problem that arises if you’re interested in incomes at the very top– so rich people like Bill Gates, also– then if you had Bill Gates in your sample and you asked him what he’s making, the problem that arises is his income is so far out there that simply by looking at this data point, you would be likely to know that it’s Bill Gates you’ve been asking. So then we get into the whole area of data privacy and sensitivity concerns, that’s why usually, in surveys, income is top-coded. Top-coded means there is a maximum income beyond which incomes are not reported.
So basically, if somebody earns 2 million pounds a year, they might be recorded in a 1 million or more category. OK, Melanie, thank you very much for coming. It’s been a very interesting interview. It was a pleasure. Thank you very much.

Professor Melanie Luhrmann (promoted from Dr. Luhrmann after we filmed the clip!) introduces yet another important idea into our measurement discussion.

Melanie is a Professor at Royal Holloway, University of London and a Research Fellow at the institute for Fiscal Studies. Her research interests are in applied microeconomics and public economics, particularly the analysis of education, health and family economics.

People may refuse to disclose sensitive information, such as their incomes, to survey interviewers. Melanie also points to a proven, if partial, solution: ask people to just place their incomes within wide income ranges instead of demanding precise income numbers. This method of dividing possible responses into broad ranges has the additional advantage of not requiring more precision than respondents actually know.


Can you think of ways to apply this insight to the measurement of other things besides income?

This article is from the free online

Survival Statistics: Secrets for Demystifying Numbers

Created by
FutureLearn - Learning For Life

Reach your personal and professional goals

Unlock access to hundreds of expert online courses and degrees from top universities and educators to gain accredited qualifications and professional CV-building certificates.

Join over 18 million learners to launch, switch or build upon your career, all at your own pace, across a wide range of topic areas.

Start Learning now