Jason Chaffetz is the Garo Yepremian of the U.S. House of Representatives, and I don’t mean that in a good way.

Mike Spagat and Paul Alper points us to this truly immoral bit of graphical manipulation, courtesy of U.S. Representative Jason Chaffetz.

Here’s the evil graph:

planned_parenthood_abortion_chart_bs

Here’s the correction:

planned_parenthood_corrected

From the news article by Zachary Roth:

As part of a contentious back-and-forth in which Chaffetz repeatedly cut off [Planned Parenthood president Cecile] Richards, the congressman displayed a slide with a graph that looked like this [top graph above]. When Richards said she’d never seen it before, Chaffetz replied: “It comes straight from your annual reports.”

Moments later, Richards shot back: “My lawyers just informed me that the source of this information is Americans United for Life, an anti-abortion group. I would check your source.”

But the source wasn’t the only problem. A cursory look at the graph, which comes from an Americans United For Life report about Planned Parenthood centers released in June, makes it seem like in 2006, Planned Parenthood performed far more cancer screening and prevention services than abortions, but that by around 2010 it performed an equal number of both, and by 2013 it performed far more abortion services than anti-cancer services.

The issue is important because as part of their effort to defund Planned Parenthood, Republicans have portrayed it as primarily an abortion provider, while the group’s defenders have said it mostly performs other women’s health services, like cancer screenings.

But look at the actual numbers in the graph. They show that in 2006, Planned Parenthood performed 2,007,371 anti-cancer services and 289,750 abortions. By 2013, the gap had closed slightly, but the group still performed many more anti-cancer services than abortions, 935,573 to 327,000.
Why does it seem otherwise? Because the “graph” has no y axis, which allows its creators to simply plot the results wherever they choose in order to create a compelling visual effect. That’s how 327,000 is made to look like a much larger number than 935,573.

What the slide actually shows, of course, is that the number of abortions performed by Planned Parenthood rose very modestly between 2006 and 2013, while the number of anti-cancer services it performed did indeed fall by more than half. But Richards said some of the services, like pap smears, dropped in frequency because of changing medical standards about who should be screened and how often. Displaying that information on an actual graph would show the line for abortions rising very slightly over the 7-year time period, and the line for anti-cancer services dropping, but always remaining far above the line for abortions.

I looked up Jason Chaffetz on wikipedia and found that this:

[Chaffetz] was the starting placekicker on the BYU football team in 1988 and 1989. He still holds the BYU individual records for most extra points attempted in a game, most extra points made in a game, and most consecutive extra points made in a game.

A placekicker of all people should understand the principle of division of labor. If you want to make a graph, get an expert to do it. Don’t use a double y-axis and, while you’re at it, don’t tell untruths about where you got it from.

Just kick the damn ball, and leave the passing to the quarterback, OK?

P.S. The bottom graph above is much better than the top graph but it’s still not perfect. The axis labels are too tiny to be readable, and there are way too many numbers on the y-axis. Tick marks at 0, 500,000, 1 million, 1.5 million, etc., would do just fine. Also those heavy black and red lines on the left, bottom, and right of the plot are bad news.

67 thoughts on “Jason Chaffetz is the Garo Yepremian of the U.S. House of Representatives, and I don’t mean that in a good way.

  1. The y-axis should measure/report totals costs (expenditures) on abortions vs rest, corrected for inflation. Then you have a clear idea of how much one costs compared to the other. That is the implicit comparison of interest. (Not relevant – but I believe that there should be more funding and the trade-offs much weaker).

    • Economist:

      It’s my impression that, from the anti-abortion side, it’s the number of abortions rather than the $ cost that is the key issue. In any case, I agree that there’s lots more information that can be conveyed. Rather than saying that a single graph should report this or that number, I’d prefer to recommend a display with multiple graphs, with one graph showing the number of incidents and another showing total dollars. Thus, my problem with Chaffetz’s graph is not that it’s numbers rather than dollars, but that it’s using the double-y-axis (which I just about always dislike) in a particularly misleading way.

      • Yes I have read your posts about the double y-axes. I actually use them in my work. But yes, everyone using them obviously has to worry about scales, intersection interpretation, and so on – a point brought out clearly by looking at the two graphs in this post.

        About the costs etc. : by presenting two “trends”, the notion of a trade-off is at least implicit. My point is that if there is a trade-off it has to do with costs (and resources, and more generally binding constraints) and not with the number of people served. This is also closely related to the (factually incorrect) argument that Planned Parenthood funds are fungible. It is my impression that this graph is intended to speak to those issues.

        • There isn’t any such trade off since this is services that are generally reimbursed by insurance or are paid out of pocket by patients. There is no reason they couldn’t both have gone up or down (and those straight lines imply a two point chart). Saying there is a trade off is like saying that for a store the decline in customer purchase of oranges is a trade-off with an increase of customer purchases of milk. It might be a trade off for the customer but it’s a retail trade off. The trade-off for the store might be at a wholesale level where they are making predictions of customer demand that may or may not be correct, but that’s a completely different issue from these data from what I can tell.

  2. To be fair, there’s misdirection on both sides. PP also utilizes statistical legerdemain when they tout their 3% statistic to give the impression that abortions are a minuscule part of what they do. But that percentage normalizes abortions by service count.

    If you go to PP’s annual report and pull the numbers, they perform about 11 abortions for every 100 people (men and women) that they give service to, which is hardly insignificant. And while they don’t release the numbers, you can find the range of costs for an abortion, and calculate that abortions bring in about 25% of their revenue.

    • I don’t know the actual numbers, but I imagine it costs much more to perform an abortion than a pap smear or to give someone condoms. Is it surprising that abortions make up a larger chunk of revenue relative to their frequency? I guess the question is, what’s the right number, if there is one? If we were talking about dentists, would we focus on the relative number of routine cleanings to root canals (some high number, I assume) or the relative revenue of cleanings to root canals (some lower numbers since root canals cost much more)?

        • Yes, that’s what I said. “I don’t know the actual numbers, but I imagine it costs much more to perform an abortion than a pap smear or to give someone condoms. Is it surprising that abortions make up a larger chunk of revenue relative to their frequency?”

          Which part gets your question of ‘why?’

  3. The graph says in both words and pictures “abortions up-life saving procedures down”. Abortions are up and life saving procedures are down. That’s enough for it to qualify as “not evil”. Everything else need to make it “good” is at least an order of magnitude less important.

    • Anon:

      If you want to show, “abortions up-life saving procedures down,” you make 2 graphs, one showing abortions up, the other showing procedures down (incidentally, cancer screening does not necessarily save lives, as we’ve discussed on this blog occasionally, but that’s another story). What’s evil is sticking them on one graph while distorting the scale.

      One can do evil things for good ends. That’s another story.

      • I disagree with your “what’s evil comment”. See my comment below. I think it’s evil to scale “screenings” the same as “destroy fetus” thereby suggest some kind of equivalence. I could use that same trick to minimize any bad thing. Just graph the bad thing with something innocuous which has values so much larger it makes any changes in the bad thing seem small and unimportant.

        Bottom line: they correctly state in words, pictures and with the numbers provided (which were used to made the “improved” graph) that abortions are up and screenings are down. That’s not evil. Not by a long shot. Everything else is minor in comparison.

        • Anon:

          Chaffetz is the one who chose to display counts of screenings etc. and abortions on the same graph. Had he thought the comparison was a false equivalent, he could’ve displayed them on two graphs, which is what I’d recommend anyway as good statistical practice. What’s evil is to display the numbers misleadingly, to abuse the power of statistical graphics.

          I think we’d probably both be happy had Chaffetz displayed two time series side by side.

        • It isn’t misleading. The numbers are clearly there and it served the same practical purpose for me as two charts. Is two charts slightly better? Certainly if you have the room, but the difference between putting these on two or one isn’t the difference between “good” and “evil”.

          You’re being ridiculous for saying so, and even more ridiculous for toughing another chart which is doing its best to be just as misleading in the opposite direction (trying to make it look like there hasn’t been a significant increase in planned parenthood’s abortions).

        • Anon:

          No, once you are displaying numbers on a graph I think there’s some responsibility to get the axes correct. Also, you might feel like I’m being ridiculous but I think you’re also aware that I’m not the only one who thinks that graph was ridiculous and indeed reprehensible in its misleadingness.

        • Anonymous is either being ingenuous or is truly challenged mentally. Stating the chart isn’t misleading is lying or willfully ignorant. I wonder if it isn’t a simple case playing devil’s advocate in order to generate an argument where none could possibly exist.

    • Huh? It’s totally evil. The chart does not reflect the numbers at all.

      The chart tells me of an approximate doubling in abortions & a halving in life saving procedures. Now tell me what the numbers say.

      “Abortions are up” but by how much? “Life saving procedures are down” but by how much? The relative numbers matter a huge amount in any rational decision making framework.

  4. “Evil” by the way is in the eye of the beholder. The charge could be at least as easily made at those scaling “screenings” at the same level as “destruction of a fetus” as if a count of one is in some way equivalent to a count of the other.

    What if some Iranian politician did the following: in order to deny the holocaust (which they seem to love to doing for some reason that escapes me) they made a graph of Jewish deaths throughout the 1940’s (~6,000,000), but superimposed a graph of world population (~2,000,000,000) drawing both to scale and used the smallness of the lower line to suggest those deaths weren’t a big deal? That’s basically what’s being done with that “good” graph.

    • I’ll go ahead and make the charge that your comparison of abortion services to the holocaust is offensive and misleading since one is the result of a lack of access to education/contraception/financial resources whereas the other one is the result of the ideologically motivated use of force.

      Come to think of it that dovetails nicely with the point of the corrected plot, which is that the number of screenings and the number of abortions changed for entirely different reasons. Apparently the number of screenings changed due to PP’s adherence to medical standards and we are left to wonder why the number of abortions changed.

      Maybe PP just really wanted to get more women to abort, but I’m going to just guess that it had something to do with demographics, education, and access to contraception.

      • I made no such charge. You’re being assigning. My point very clearly was you could make a graph like that “good” one which could hide ANY bad thing you wanted to hide or minimize. If you can’t understand that then your opinion counts for s**t. (see that’s me actually being offensive)

        • Writing “What if some Iranian politician did the following”, and then doing it, is thin cover for an opinion, especially when you troll the discussion by referring to abortion solely as “destruction of a fetus”.

  5. I completely agree with your premise that accurate truths should be conveyed and deliberately misleading tactics that use data graphics as a façade have no place in the public forum. However, your have chosen your words here very poorly.

    Your use of “truly immoral” and “evil graph” to describe graphical manipulation are insensitive given that the issue in question is in regard to abortion.

    It’s saying that an intentional patently bad graph is more morally reprehensible than the act of abortion itself. Perhaps that is exactly how you see it, but it is in poor taste to the many people who find abortion a more serious moral issue.

    I completely agree with your distaste for double-y-axis. If there is a story to be told here with data I feel it would be much more accurate and impactful told as a series of multiple graphs. I think Rep. Chaffetz has missed an opportunity in this specific regard that really could have helped his position.

    • Devlin:

      I agree, he missed an opportunity. And of course the “evil” line is hyperbole on my part, I thought that was obvious! Making a bad graph is nothing like as bad as mugging little old ladies on the street, or defrauding people with spam, hit-and-run driving, or all sorts of other crimes out there. I meant “evil” in context, in that this is not just a poor display, it’s actively misleading. It would be a challenge to put a communication crime on a commensurable scale of evil with evil actions that affect people directly.

      • Thank you for your comment. I feel we are in agreement. The juxtaposition of these two contexts was the impetus for my comment. I am reminded when I have used very similar language as you have here in my professional context (software) and that I should take care myself in my choice of language when it overlaps with other contexts.

    • It’s disingenuous to try and make this about how reprehensible abortion is to “some people”. If the concern was really the number of abortions Chaffetz might have focused on how important it was to get improved contraception (and STI protection) into the hands of people who need it.

      Chaffetz is clearly working hard to make it more difficult for low-to-middle income women to get medical care _and_ safe abortions. Chaffetz is doing this for political gain which makes ‘evil’ an entirely appropriate word to apply to him.

      But hey, congratulations on getting Andrew to rank this below spam e-mail.

    • Though this would depend on what types of real world and life altering decisions were made based on the intentionally distorted information and the harm done to those about whom the decisions were made.

  6. There are plenty of other things wrong with that graph. For example, the US population has increased about 6 percent from 2006 to 2013 so some increase in abortions are to be expected (as the revised graph shows). Also, Planned Parenthood provides millions of tests and treatments for STDs each year but somehow that was not included in the graph; neither is the fact that many other abortion providers have been closed down.
    http://www.bloombergview.com/quicktake/abortion-and-the-decline-of-clinics

    It should hardly come as a surprise that there has been a slight uptick in abortions provided by Planned Parenthood. Similarly, pap smear tests for cervical cancer are now recommended to only be done every 3 years instead of every year probably accounting for much of the decline in cancer screening numbers.
    http://well.blogs.nytimes.com/2012/03/14/new-guidelines-advise-less-frequent-pap-smears/?_r=0

  7. The details about the graph’s weirdness are sort of immaterial. Jason Chaffetz is, to use the popular phrase when it comes to climate change, “not a scientist.” He is also probably neither a mathematician nor a statistician. He is, however, committed to abolishing planned parenthood and so he latches on to any nugget, no matter how absurd, with energy and gusto. He fully expected that this was the crown jewel in his series of arguments–Richards annual salary, the (doctored) videos.
    Chaffetz and his aides, blinded by their beliefs, were sloppy and now have been caught on tape. We look forward to any “walking back” regarding the supposed source and the misleading graph itself. My guess, is that there may be some “doubling down” rather than any mea culpa. His side has lost the gay rights battle and are losing over Obama Care, thus the focus on Planned Parenthood.

  8. Another problem with both graphs: We don’t know what the in-between years look like, and therefore can’t appraise the year-to-year variation. Perhaps 2006 was chosen as the comparison to 2013 because it yielded the steepest slopes.

    • City:

      I thought that too at first, but then I looked carefully at the second graph and the lines are not quite straight, so I’m guessing that the second graph is actually displaying data from every year.

      • I dunno, Andrew. They look very straight to me. Plus, when do you see lines so perfectly fitting the points in real life? Also, is it likely that the folks who produced the second graph actually went back and found the data for all the other years so quickly?

  9. If the numbers are correct, now I am wondering why the number of abortions performed by PP increased while the number of abortions has been declining nationwide. Fewer abortion clinics?

  10. By only looking at mammographies, which have recently been recommended to be done less often and on a different population of women (in particular, not under age 50 except if there are special risk factors) the original Chafetz graph is further misleading. PP does a lot more than just mammographies. If you include other non-abortion services, the graph looks quite different. This is discussed here:

    http://www.dailykos.com/story/2015/09/29/1426029/-I-ve-corrected-the-chart-Rep-Chaffetz-presented-at-today-s-Planned-Parenthood-hearing

    Notice in particular the graph at the end of this article.

    • With apologies for going off on a tangent, confusion reigns regarding mammograms, and I would like to clarify Bill’s comment.

      Bill is presumably referring to the 2009 recommendations of the US Preventive Services Task Force (USPSTF). There are numerous other organizations, e.g. the American Cancer Society, the American Society of Breast Surgeons, and the American College of Radiology, which do not agree with the USPSTF. (Make what you will of the observation that the latter organizations all represents practitioners who have vested financial interests in more mammograms being done, whereas the USPSTF is comprised of independent experts in evidence evaluation who have no financial conflict of interest.) Moreover, evidence suggests that population usage of mammography has changed little since the USPSTF recommendation was issued. [Sorry, can’t find the reference for this right now.]

      Bill’s paraphrase “not under age 50 except if there are special risk factors” is not a correct restatement of the USPSTF 2009 recommendation. Their actual wording was “The decision to start regular, biennial screening mammography before the age of 50 years should be an individual one and take patient context into account, including the patient’s values regarding specific benefits and harms.” Note that the emphasis is on patient values, not risks. Risks are not even mentioned, though they are presumably part of the “context” referred to.

      The USPSTF has a draft 2015 recommendation (http://screeningforbreastcancer.org/?ds=1&s=breast%2520cancer%2520screening) which, regarding women under age 50, is even more clearly focused on patient values than the 2009 version.

      Disclosure: I served as a paid consultant to the USPSTF on the harms and benefits of mammograms in 2009 and 2014.

  11. Of course, part of the reason abortion revenues are as large a share of PP’s as they are is because in so many of the places they work, no one else offers abortions. I think they would be glad to have less of a “market share.” But the stigma and danger involved in providing abortions have meant an absence of alternative providers.

  12. I agree that the graph shown is on a duplicitous arithmetic scale. My question is: why does Planned Parenthood play along with this?

    They could easily do a re-organisation. Think ala Google. It used to be Google Inc. owning a bunch of companies. Now it is (over-simplified) Alphabet Inc. owns Google Inc and Moonshots Inc.

    Planned Parenthood could reorganize to Parent Science Inc. that owns PP Abortions Inc. and PP General Reproductive Inc. Federal funding would then clearly and obviously only go to PP General Reproductive Inc. and never to PP Abortions Inc. (In fact the use of proceeds is already required to not go to abortions, but that seems like a technical detail to members of the general public (and politicians) who aren’t that financially sophisticated (and also have overcharged affect heuristic) and who can point to funds going to “Planned Parenthood” and all the abortions done by “Planned Parenthood”.)

      • I think that’s a rather thoughtful response.

        That said, taking a page from Negotiation 101, I think it is (most always) a huge mistake for people to focus on positions instead of interests. If PP’s interests are what they say they are, restructuring in a way to make it easier to get the exact same federal funding with the exact same use of proceeds restrictions, really does not cost them much at all, yet it alleviates sources of political gridlock, show inquiries in the House, etc. and should allow PP to focus its lobbying efforts elsewhere.

        As I recall, the mantra among the Left in the US is that abortion should be ‘legal and rare’. I haven’t heard such sentiments about condoms, the pill, the morning after pill etc. — I infer that is in recognition that abortion is a bit grizzlier than other types of methods for dealing with unwanted / unplanned pregnancies.

        I have a related concern that PP (and various grandees on the Left) haven’t been pushing IUDs heavily, but that’s a discussion for another day.

        (The following two articles do give a bit of color on this issue

        http://www.economist.com/news/united-states/21648692-family-planning-easier-if-you-dont-have-think-about-it-taking-bother-out-birth

        http://www.economist.com/news/leaders/21648637-america-has-high-rate-unplanned-pregnancies-greater-use-long-acting-contraceptives-would )

        At the end of the day, I’m trying to solve problems here. I suspect most of what goes on in DC (on both sides of the aisle) is bout grandstanding, but hopefully that isn’t a fair synopsis.

        • …really does not cost them much at all…

          Not sure how you reach that conclusion. Unless the two organizations are allowed to share resources, separating them into two arms-length organizations incurs a lot of deadweight loss. It would be very costly to have completely separate facilities for abortions and everything else, so cost savings can be realized if two organizations share physical premises. And it would be very costly to employ one set of doctors just for abortion procedures and another set for everything else… and so on. The more the two organizations share resources, the more they save and the less clear it is that funding to one organization does not benefit the other.

          Currently, abortion is priced according to the cost of the procedure and patients on Medicaid or Title X pay out of pocket, so government-provided health insurance doesn’t pay for abortions. As to direct grants, on the one hand, federal funds pay for large fixed costs that enable PP to keep clinics open and therefore to provide abortions; on the other hand, unless a reduction in funding is large enough to actually cause some clinics to close their doors, variations in funding would not have any effect on provision of abortion procedures. So does government funding actually pay for abortions? Sufficiently motivated individuals could argue it both ways.

        • You’re right to worry about scale and scope dis-economies. However there is a well trodden path for how these restructuring work in the corporate world. (For easily comparable case studies or general knowledge ask a lawyer about the way things are run when a large company does a ~19.9% carveout– or just spend a bit of time googling it). Very simplistically, there is no real change except there is a de minimis ongoing cost for entering in management service agreements between (a) Parent Science Inc. and PP Abortions Inc. and (b) Parent Science Inc. and PP General Reproductive Inc. There are also some one time legal costs associated with the friction of effecting the reorg. Since these are one-offs I tend not to spend much time on them, though the magnitude of these frictional costs can range from tiny to expensive depending on the existing complexity of PP’s internal legal arrangements (which as an outsider are out of view — in the corporate world if there is a deal breaker it tends to be adverse tax consequences but since PP is not a for profit company I don’t see how that would be the applicable).

          Under the most basic setup: for the management services agreements, existing (and all future) buildings, equipment and labor are all contracted at arms length rates from Parent Science Inc. to its subs. This kind of thing is done all the time in the corporate world (and yes there are definitely disconcerting practices with respect to transfer pricing to dodge taxes by some companies, but the complexity of tracing between different boxes and associated legal agreements, legitimate or not, tends to blunt affect heuristic and associated outrage, methinks.)

          The reasons for spinoff, carve-outs, and restructurings are many, but a classic reason is to give more clarity to the public — which shows up in the public (financial and possibly tax) reporting done by said companies.

          The benefits of a restructuring may ultimately prove elusive for PP, but the costs are very much limited.

        • Seems reasonable as far as it goes, but as you say, the benefits of a restructuring may ultimately prove elusive — it may seem to be simply a technical detail to members of the general public (and politicians) who aren’t that financially sophisticated (and also have overcharged affect heuristic).

          It occurs to me that given the general opposition to abortion, splitting might not be a tactically sound move — carving out the abortion provider makes it easier to attack…

  13. Two words: “dimensionless ratio”

    A plot of N(t)/N(0) for both quantities could be put on the same axis, and would clearly show the percentage decline in one and the percentage rise in the other. No absolute quantities and their associated relative importance need be involved. Also, assuming price hasn’t changed dramatically, if you only have revenue data, you’d get a similar result for Dollars/Dollars as N/N (though if unit price has been changing… not so much).

    Of course the purpose of the linked graph is obviously not accurate portrayal of information.

    • Daniel:

      I love dimensionlessness but I often have a problem with these sorts of baseline graphs in that they privilege whatever was happening at the time point that is labeled as t=0.

      • Daniel, exactly right. We need to see the numbers for the years in-between to understand the variation. For instance, we would expect the number of PAP smears to decrease quickly after the recommendations changed and then level off. A constant rate of decrease over the years shown doesn’t make sense, given a change in medical practice in the middle of the time period.

      • It’s not so much “hides the fact” as emphasizes different information. I agree with Andrew that t=0 gets special status, and that can be especially bad when t=0 has noisy data. There are other alternatives, like using the average N over the timespan of interest, but that will make both series go through 1 somewhere in the middle of the interval, which is more or less arbitrary. Another option would be to scale both series by some other value that is of interest, such as N(abortions) / N(total_patients) and N(screening)/N(total_patients).

        As pointed out above, comparing 1 pap smear with 1 abortion is pretty callous to the people who are vehemently against abortion, so if the thing of interest is how have things changed in time, eliminating the scale can be a reasonable thing to do.

  14. As good as andrewgelman.com is, its effectiveness is limited. Chaffetz told Blitzer of CNN,
    “I stand by the numbers. I can understand where people would say the arrows went different directions, but the numbers are accurate. And that’s what we were trying to portray.”

    • I think in every case where someone might want a double-y axis they are better off creating two (or more) appropriate dimensionless ratios and plotting the two data sets with a single dimensionless y axis. So I agree with the idea that it should never be necessary to use double y axes.

      • They are useful for “at-a-glance” displays because they allow packing more information in the same space (eg multiple time series). You need to know what you are looking at first though, they are not so useful for communication.

        • Anon:

          If you want to display two time series, you can either put them on the same scale (as in the bottom graph above) or simply make two graphs, side by side. These would be legible and would easily fit in the space occupied by either of the graphs above, in addition the multiple-graphs strategy generalizes directly to displaying 3 or more trends.

        • Andrew,

          I use them to fit as much info as possible onto a single display. I would not use them to compare one series vs the other. Actually, one exception is to comparing some cyclic phenomena. In the former case the scale does not matter, in the latter using the same scale may obscure the relevant detail if the values are very different.

  15. There is a chart here showing that if you graph all of PP’s services, the “pro-active health” services (of which cancer screening is only one) stayed quite constant, as did (if you take them on the same scale) their abortion services:
    http://www.vox.com/2015/9/29/9417845/planned-parenthood-terrible-chart
    Basically, the decline in cancer screening services is replaced by STD screening services and treatment. Chaffetz seems to have been trying to show that PP’s claims to being primarily a health services provider, so including the broader range of data falsifies his claim.

Leave a Reply to Paul Alper Cancel reply

Your email address will not be published. Required fields are marked *