Misleading Statistics

The other day, I got an email cartoon listing nineteen goods/categories whose prices have increased 10% or more over the past year. Nine were grocery products and one was men’s suits. The others were categories: gasoline, airline tickets, used cars, gas utilities, hotels, delivery costs, electricity, furniture, and cleaning products. The bottom line caption was: So how is inflation only 8.6%?

I checked the numbers against the latest CPI-U, and some were exaggerated. The email said used car prices were up 35%, but the CPI lists the annual inflation at 7.8% The email also listed gasoline increases at 49%, when the actual was 29%.

But the most misleading aspect is that all of the items listed by the email together comprise less than 30% of all the items that comprise the CPI-U. All food items comprise only 14%. Eggs [up 44%] only amount to 1/10 of one percent.

Overall food prices (comprising 14% of the CPI-U) increased 11.4%, but all commodity prices, excluding food and energy, only rose 6.3% over the last year.

Now neither 8.6% nor 6.3% is good, but citing huge price spikes in small segments of the economy is definitely misleading. It’s also politically effective. People don’t notice the prices that don’t rise or rise more slowly. They notice that gasoline and egg prices are way up, or that used cars are getting pricy.

Yet, I’d be willing to be that the majority of people who received that email or saw the original posting of the cartoon and its statistics will have the reaction that the government is grossly “cooking the statistics.”

And the government has been “readjusting” the statistics for years, but in little ways, such as reducing the impact of food and energy costs on the CPI, but its figures aren’t “readjusted” by the two and three-fold magnitudes suggested by the cartoon.

This statistical “discrepancy” also illustrates one of the biggest problems faced by a democratic high technology society, that fewer and fewer of the people who vote really understand either the technology or the economy underlying their society, and that lack of understanding becomes fertile ground for demagogues who offer falsified/incorrect facts, gross exaggerations, and beguiling simple (but unworkable) “solutions” to complex problems.

7 thoughts on “Misleading Statistics”

  1. Darcherd says:

    To quote (or possibly misattribute) Mark Twain, “There are three kinds of lies: Lies, Damned Lies, and Statistics.”

  2. Hanneke says:

    Statistics are really useful, but you have to be very careful about what you are looking at and comparing, and for what reason.
    For instance, the EU uses a “shopping basket” to compare inflation for a standard set of commodities across countries, which does not include rent because patterns of homeownership and renting, and heating needs, differ across countries.
    But rising rent has an important impact on family budgets, so our national inflation-determining “shopping basket” does include rent and heating costs. That meant that in the last few months, with rents and heating costs rising faster than most commodities, the national inflation percentage could be 2% higher than our nation’s EU inflation percentage. Neither is a lie, but they are meant to compare different things: your household’s past, present and future budgetting, versus how different EU countries are faring under the stresses of the pandemic, the war in Ukraine, Putin’s gas export stop and the OPEC’s decision to keep gas prices high and production low (much better for the future of the planet of course, but forcing this extremely fast switch to alternatives, preferably renewables, is hard for those who will be dealing with a lack of heating in the coming winter).

    1. R. Hamilton says:

      “much better for the future of the planet of course, but forcing this extremely fast switch to alternatives”

      I’ve long thought that there are limits to that, that too much force or too fast a switch might not even benefit the environment, insofar as rapidly maturing and deploying technologies requires a robust economy.

      Of course, I tend to think that any justification for great force of any sort (whether military or regulatory) needs to be both even greater than the problem itself, and needs to recognize that there will be resistance to sustaining extreme efforts and significant deprivation or losses over the long term. And needs to have a finite timeframe; expanding force of either sort significantly should not become a permanent condition, regardless of increasing complexity or population density or other factors; because there will always be future justifications for further expansions, so there needs to be about as much effort put into reducing force where alternatives are possible, as into increasing it where still or newly needed.

  3. R. Hamilton says:

    Readjusting statistics may well be legitimate, but given that blatant spin also exists, people might be forgiven for wondering if readjustments may well be influenced (in administrations of either party) by political considerations as well as by trying to keep the statistics a useful summary of the overall state, as much as a single number or small number of numbers can be.

    Except when trying to demonize the opposition (where it’s commonplace), there’s little incentive to overstate problems, and some to understate them, at least if the long-term solid evidence will be past one’s term limits. And long-term trends may be harder to judge even if legitimate changes to the CPI are made.

    There is room to believe both that the process tries to be legitimate, and that the outcome may be less legitimate.

    1. KTL says:

      The use of statistics doesn’t seem to me so much the problem, as the lack of an educated populace in the very basic foundations of statistics. The general population, and the media, are poor users of data in this country. Neither understand averages and distributions around that average. The media often publishes data that indicate a change (in a value) using only two numbers – current and one time point somewhere in the past (usually one year for annualized change)! Two data points are not a trend. This approach is both misleading and extraordinarily lazy. Many metrics can be understood much more readily when given the context of providing an historical graph over at least 1-2 decades (economic, crime, population, whatever. All are available covering many decades, if not nearly a century in this country). Paul Krugman of the NYT does the best job I have seen in using and illustrating data to make a point.

      Lower level education in the US would be much better served teaching basic statistics and risk assessment to students as a parallel set of courses to mathematics (or perhaps alternating with math concepts).

      1. Mayhem says:

        > The general population, and the media, are poor users of data in this country

        Don’t worry, they’re poor users of data in *every* country. Huff’s famous “How to lie with statistics” should be required reading everywhere at High School level, it’s nearly 70 years old today and still a perfect textbook for how everyone is manipulated in everyday life.
        That and the classic Yes Minister section on leading questions.
        https://www.youtube.com/watch?v=G0ZZJXw4MTA

  4. Anne Barton says:

    I have taken to browsing the USDA Food Plans as a more realistic source of info about how inflation is effecting the average person. These plans are based on a theoretical “shopping basket” of food based on budget. At times the price rises are moderate, and at other times, extreme. Between the winter months of 2021 and 2022, the expected expenditures for a thrifty for budget rose by over $100/ month for a family of four. So far, the summer month increases year-to-year have been much more moderate. Just an interesting resource to track food prices while avoid the cherry picking of huge increases or ignoring the rising price of staples while the price of luxuries stagnates due to decreased demand as more food dollars chase chicken thighs and rice and fewer chase filet mignon and organic blueberries. I can’t speak for other areas, but in my area peanut butter was sold out long before the recalls and rice remains sold out much of the time. Chicken legs are hit or miss. But I can buy as much bacon as I want to at $8/ lb.

    https://www.fns.usda.gov/cnpp/usda-food-plans-cost-food-reports-monthly-reports

Leave a Reply

Your email address will not be published.