Brian Hooker’s Findings Are Confirmed By CDC’s Results

ORtable

By Jake Crosby

Above is a table of omitted results from the original CDC study of age at MMR vaccination according to a video put out by the Autism Media Channel, and below is a table of results from Brian Hooker’s reanalysis of that study – since retracted by the publisher in breach of policies it claims to follow. In particular, note the boxed results of each showing risk from MMR vaccination before age 36 months in African-American children – both are significant, and the strength and precision of each are almost identical to one another.

hookerresults

This would eviscerate critics’ claims that Brian Hooker’s findings are invalid because his reanalysis did not employ the same statistical methods as the original CDC study. Within Dr. Hooker’s paper itself, it is also stated that his “results were also confirmed using a conditional logistic regression design similar to the DeStefano et al. [14] (CDC) study.” Another common criticism of Brian Hooker’s paper that it did not account for low birth weight children is easily refuted by another table of results showing a greater than two-fold risk for African-American boys even when low birth weight children are excluded.

Yet Dr. Hooker’s paper remains retracted in breach of the guidelines the publisher claims to follow when considering retractions. Even before the retraction, the publisher BioMed Central (BMC) had deleted the paper online in breach of its own policies on article removal. BMC has never offered any explanation concerning these issues in response to emails from Autism Investigated. Also yet to comment in response to Autism Investigated’s inquiries about the retraction is the Committee on Publication Ethics (COPE), whose guidelines BMC claims to follow when considering retractions and breached when it retracted Dr. Hooker’s study.

Meanwhile, CDC is pretending that its own study results were different from Dr. Hooker’s when they were clearly not. In a statement to ABC News insisting “There was no cover-up,” CDC said of Dr. Hooker’s findings, “it is hard to speculate why his results differed from CDC’s.”

Except they didn’t.

Jake Crosby is editor of Autism Investigated. He is a 2011 graduate of Brandeis University with a Bachelor of Arts in both History and Health: Science, Society and Policy and a 2013 graduate of The George Washington University School of Public Health and Health Services with a Master of Public Health in Epidemiology. He currently attends the University of Texas School of Public Health where he is studying for a Ph.D. in Epidemiology.

Share on FacebookTweet about this on TwitterShare on LinkedInShare on Google+Share on RedditPin on PinterestFlattr the authorDigg thisBuffer this pageShare on TumblrShare on StumbleUpon

143 Thoughts on “Brian Hooker’s Findings Are Confirmed By CDC’s Results

  1. both are significant, and the strength and precision of each are almost identical to one another

    Does this video also go to the trouble of describing the models that are prominently noted as column heads?

    • I think the table illustrates them well.

      • It illustrates the outputs. What the models are is completely unknown.

        Where’s the document? A still from a video doesn’t exactly do much for me. It is also unclear whether the AMC is now on the up-and-up, as opposed to the earlier betrayal.

        • Not completely – we know they used logistic regression both when combining racial categories and stratifying to specific racial categories. Certainly that’s enough to know that variables studied for the results in the Black model are the same as those studied for the results in Brian’s analysis – both results being virtually identical despite their being reached by different statistical methods.

          I don’t consider AMC on the up-and-up, but that is the original table of CDC results. That said, I wouldn’t trust AMC as to whether they had permission to release that table in its video.

          • Not completely – we know they used logistic regression both when combining racial categories and stratifying to specific racial categories. Certainly that’s enough to know that variables studied for the results in the Black model are the same as those studied for the results in Brian’s analysis – both results being virtually identical despite their being reached by different statistical methods.

            How do you know this? You have posted a heavily cut snippet of a table with no information regarding model used or even p-values. It appears as though it’s an unadjusted model. Furthermore Hooker’s relative risk of 2.30 translates to an OR of 1.30 and they are stratified differently so not the same numbers. If you look at the corresponding bin of 24-35 months you can see it spans 1 which nullifies the result. It isn’t omitted data you are presenting, just a preliminary pass-through.

            • We know that DeStefano et al. used logistic regression; Hooker’s relative risk of 2.30 would translate to an OR of 2.25 according to the table above, since both were stratified to African-American children irrespective of gender. The 24-35 month age range includes 1, but not the 12-15 month age range nor the 16-18 month one. It’s omitted data according to the whistleblower and coauthor of the study himself Dr. William Thompson. He would know better than you.

              • We know that DeStefano et al. used logistic regression; Hooker’s relative risk of 2.30 would translate to an OR of 2.25 according to the table above, since both were stratified to African-American children irrespective of gender.

                We know that he used conditional logistic regression for the final analysis. You have no idea if the table you show is an unadjusted logistic regression. There is no context, no p-values, nothing so your argument boils down to “because numbers”.

                The 24-35 month age range includes 1, but not the 12-15 month age range nor the 16-18 month one.

                Yes and? Do you even know what this means and that Hooker’s strata are different than that of the boxed stratum you show?

                It’s omitted data according to the whistleblower and coauthor of the study himself Dr. William Thompson. He would know better than you.

                “He said” is a pretty poor standard of evidence. Where is the protocol which explicitly stipulates this?

                • There is enough of a context to show that Brian Hooker’s RR is validated by DeStefano et al.’s OR. His stratum for race and age range for MMR vaccination are the same as that of the original study.

                  I would imagine it is among piles of evidence under review by Congress and possibly being withheld for legal purposes.

                  • His stratum for race and age range for MMR vaccination are the same as that of the original study.

                    No they aren’t; anyone looking at your red boxes can see that.

                    I would imagine it is among piles of evidence under review by Congress and possibly being withheld for legal purposes.

                    Um no. If Wakefield could post a snippet of a document, particularly one received by an FOIA then why not show the entire document? It looks very deceptive to me.

                    • No, anyone can see that it is stratified for the same age cut-off and race.

                      Wakefield is irrelevant.

                    • No, anyone can see that it is stratified for the same age cut-off and race.

                      The numbers in alleged CDC table that are boxed off are for the age stratification of 18-35 months whereas Hooker’s stratification that you are comparing is 24-35 months. There is a stratification for 24-35 months in the alleged CDC table, why aren’t you using that for comparison?

                      Wakefield is irrelevant.

                      If only but he is the one who supplied this heavily edited table with no context.

                    • You’re misreading Hooker’s tables. An age cut-off of 36 months simply means a comparison of autism diagnoses in children with MMR before 36 months compared to after. So I am simply comparing to the <36 months results in the CDC table, which Wakefield got from Hooker. So he is irrelevant.

                    • Just to clarify – it is Wakefield who is irrelevant, not Hooker.

                    • You’re misreading Hooker’s tables. An age cut-off of 36 months simply means a comparison of autism diagnoses in children with MMR before 36 months compared to after. So I am simply comparing to the <36 months results in the CDC table, which Wakefield got from Hooker.

                      The only way I’m misreading Hooker’s tables is if he compounded his age strata and if he did that then it makes his analysis even more wrong. You do realise that don’t you?

                    • But then that would make CDC’s analysis wrong, which was what Hooker was reanalyzing.

        • And FYI, I’ve just removed the hyperlink to the video. I don’t want to endorse it.

  2. Media Scholar on October 10, 2014 at 6:43 am said:

    CDC’s Thompson, the Autism-MMR whistle-blower himself, appears to have guided Hooker through the proper process of obtaining all the necessities required to “legally” re-analyze and correct the previously fraudulent results.

    In their efforts to disqualify what appears to be an impeccable paper full of re-analysis, the same old gang of Pharma pups are hanging out here and other places to boast about how clever they are and yet providing no proof they have had anything to do with the retraction at all, or to explain why they suddenly felt compelled to counter-attack coincidentally with the emergence of William Thompson, the CDC whistle-blower.

    A problem with peer-review??? Really???

    And the best these self-promoting publication pundits can come up with is to bully their way in with no consideration at all to the fact that they have conflicts of interests and are not revealing them. They have competing interests and are not revealing them. These pundits are not to be mistaken for peers. These pundits are far from neutral beings.

    Their lack of self-respect is pitiful.

  3. gruffalo on October 10, 2014 at 7:10 am said:

    Meanwhile, another at-risk group has surfaced: so-called “isolated autism”, unconfounded with other developmental illness/problems.

    http://vimeo.com/user5503203/review/108522744/186711a23b

    CDC manipulated the data, even reintroducing most of the confounders, but the association would not disappear. So they just left it out of the Pediatrics paper.

  4. White Rose on October 10, 2014 at 8:30 am said:

    How much do these interlopers get paid I wonder ? & so many of them here now .

    Herod – how much are you getting ? if the money is right – I’ll consider swapping sides – I dont come cheap mind . Please pass on the message to your handlers (or doesnt it work like that).
    How much are you paying the AoA lot ? We deserve at least double whatever Blaxman got (he is ydays news)

    How does it work ? Do you all meet up in a room to plan a coordinated attack ?
    I thought we were all just “crazzies” over here . But apparently Pharma Harma has to employ a team of
    mercenaries to counter mad loony conspiracy theorists .
    What is wrong ? Is Jake abit too good for you ?

    • Herod – how much are you getting ?

      If you’re referring to me, not a dime. I wouldn’t even take an offer such as you fantasize to exist were it to even exist….

      if the money is right – I’ll consider swapping sides –

      … which makes one of us.

      I dont come cheap mind .

      I’m afraid that you really do “come cheap mind,” from everything I’ve read.

      • If you’re referring to me, not a dime. I wouldn’t even take an offer such as you fantasize to exist were it to even exist….

        ****************

        Were it to even exist??? You’re kidding, right?

        http://www.indeed.com/cmp/Pharmaceutical-Company-Seeks-Social-Media-Specialists/jobs/Social-Media-Specialist-f74ee88e5efe3273

        The first two paragraphs of this JOB POSTING read as follows:

        “… Successful pharmaceutical company seeks a talented, motivated, and energetic social media specialist with excellent writing skills. The individual(s) must be able to commit their full attention to this position. We ask that only serious applicants submit their resumes, please.

        As the social media specialist, you will work with a team of fellow social media experts and writers to post and engage across several social media sites, such as: Facebook, Twitter, YouTube, Tumblr, and so on. Our goal is to expand the brand to every individual both domestically and internationally…”

        So although people like you enjoy treating us parents like idiots, most of us are well familiar with this ‘inter web” thingy.

        And were also pretty used to being lied to.

        • justthestats on October 13, 2014 at 7:58 pm said:

          Sounds like the job description of the guy who writes for the company’s social media account. You know, the guy who posts things to the company’s Facebook page and comes up with clever but harmless responses when someone tweets the company something funny. Pretty much every company has them nowadays.

    • I will readily concede, though, that “Herod” is by far the best attempt at ridicule-by-name that I’ve come across to date. It’s not quite rhyming slang, but the effortless amalgam with arguably frank ideas of reference suggest some marketable talent.

  5. Media Scholar on October 10, 2014 at 11:05 am said:

    Following the reanalysis link provided led to only notification of the published retraction.

    Apparently, this so called open access journal is anything, but open access.

    How does one, for example, raise questions about the clearly contentious , competing interests behind the cleverly devised retraction when their accusations and identities remain un-open access?

    Are the publishers so incredibly unaware of Hooker’s article? Are they unaware of the information provided to Brian Hooker by US Centers For Disease Control researcher turned whistle-blower, Dr. William Thompson, led to the reanalysis article? Is the publisher simply ignoring the likelihood that the whistle-blowing process involves, in this case as so many others, an enormous amount of unfair and inappropriate pressure being applied in the effort to intimidate to natural correlation between publisher, whistle-blower and reanalysis?

    The publisher has a moral and ethical obligation here to protect the printed word, the thoughts, and work exerted by a man obviously interested in avoiding reproach. The publisher has a professional obligation to maintain knowledge of the publishing profession. Instead of falling into the trap of being just the latest publication to inexplicably comply with the over-use of the remedy of retraction, particularly obvious by the certain sign of failure to provide open access to allegations and allowing for the testing of the purity of those making them.

    Otherwise, the journal Translational Neurodegeneration should be aptly renamed The Journal of Timid Turtles Which Retract When Frightened, or simply the Journal of Over-reacting Publishing Squirrels.

  6. This would eviscerate critics’ claims that Brian Hooker’s findings are invalid because his reanalysis did not employ the same statistical methods as the original CDC study.

    I’m afraid that this does nothing of the kind. I too would like to know what the column headings are and the statistical analysis applied as well as descriptive statistics. At best, it’s preliminary data that demonstrates why it was important to apply the appropriate statistics, i.e. conditional logistic regression controlling for numerous variables by collecting additional information from the hospital worksheets linked to birth certificate data. If this is what got Thompson’s knickers in such a knot then it’s just more evidence that he really doesn’t understand epidemiology any more than Hooker does.

    • “data that demonstrates why it was important to apply the appropriate statistics”

      -appropriate for hiding results you don’t like by diminishing your sample to dilute your findings

      According to Dr. Thompson, those results should have been published per study protocol. CDC breached that protocol.

      • -appropriate for hiding results you don’t like by diminishing your sample to dilute your findings

        Jake, have you ever even worked with complex datasets? What Hooker did is exactly what you are accusing DeStefano et al. of doing which is slicing and dicing the dataset in such a manner as to produce the desired result. Hooker didn’t even know to use an odds ratio instead of a relative risk, which conveniently gave a more “impressive” risk but was also wrong. There are numerous variables to control for in order to produce a reliable result, there is also still the matter of the temporal association of exposure with outcome which you have failed to address. Hooker treated this as a cohort study which is predicated upon the temporal association of exposure and outcome. I fail to see where you, in your own words have justified Hooker’s use of the innapropriate statistical model.

        • I have. Dr. Hooker was merely trying to see if he could find and quantify the associations CDC found but deliberately left out of its published results. He is open and honest about what he was doing; CDC is not. Your claim he used the wrong measurement is totally moot now that his relative risk is virtually the same as CDC’s odds ratio among African-American children. Sure 2.30 is technically larger than 2.25, but it’s a negligible difference. You keep harping on the need to control for additional variables without saying what those are when the whistleblower himself has said that the omission of results breached protocol. Dr. Hooker simply wanted to calculate relative risk for autism from MMR at less than 36 months, which as we know is virtually identical to CDC’s odds ratio and is therefore validated by it. He made no claims about temporality.

          • I have. Dr. Hooker was merely trying to see if he could find and quantify the associations CDC found but deliberately left out of its published results.

            And in doing so didn’t control for variables which are known risk factors for autism, again not to mention he doesn’t even know if exposure preceded outcome. You can’t say the CDC deliberately omitted this data; produce evidence.

            He is open and honest about what he was doing; CDC is not.

            So open and honest that Hooker would apply an erroneous statistical model that would give the desired result? So honest he would fail to include his conflicts of interest and those of his suggested peer reviewers?

            Your claim he used the wrong measurement is totally moot now that his relative risk is virtually the same as CDC’s odds ratio among African-American children. Sure 2.30 is technically larger than 2.25, but it’s a negligible difference.

            As I explained upthread, my claim that he used the wrong measurement is not only not wrong but is also a consensus amongst any epidemiologists and statisticians who cared to look at his methods. RR and OR are not the same measurements so I fail to see how you can claim that because they are similar, that somehow supports his results.

            You keep harping on the need to control for additional variables without saying what those are when the whistleblower himself has said that the omission of results breached protocol.

            I have said what some are and I should think that you would know also being in an epi programme. Gestational age, birth weight, parity and gender should have been controlled for. The “whistleblower” isn’t a whistleblower at this juncture, just a disgruntled gossip and “he said” isn’t sufficient. Evidence that this was the case please.

            Dr. Hooker simply wanted to calculate relative risk for autism from MMR at less than 36 months, which as we know is virtually identical to CDC’s odds ratio and is therefore validated by it.

            No it isn’t virtually identical nor validated because Hooker used a RR. He used the wrong statistical model to examine that dataset; it’s really that simple and no wonder the study was retracted. If I was a peer-reviewer I wouldn’t have accepted it at all.

            He made no claims about temporality.

            I know that, what a silly thing to say. Since he treated the data as a cohort study, temporal associations are at the crux. Hooker doesn’t know what he’s doing and he expects others reading his study to know even less.

            • Yes I can say the data CDC omitted this data, it is displayed in the table above and validates Dr. Hooker’s findings. The point of claiming a relative risk computed from a case-control study to be inaccurate is that it could differ from the odds ratio, but that is clearly not the case here as the two are the same. He said himself in his own paper that he used a statistical method similar to DeStefano et al. and got the same results. He also did disclose his involvement in litigation and the journal only stated there were undisclosed interests on his part, not on the part of any peer reviewers. Clearly, he was not trying to conceal any competing interests he may have.

              As I stated, he did exclude low birth weight children in one analysis and still got a very high risk – especially when stratifying for male gender among African-American children. I guess your argument rests on gestational age and parity – although other analyses done by CDC for thimerosal-containing vaccines for instance showed an increased risk for autism when both low birth weight and preterm children were excluded for the analysis. Seems like a tough sell to say it would be different for MMR. That leaves parity as the only remaining variable seemingly unaccounted for in either of these analyses – are you suggesting that after accounting for all these other variables failed to make a risk go away, that accounting that this one remaining variable would make all these findings just go away? That’s hard to believe.

              And Dr. Thompson is a whistleblower; documentation backs up what he said. An RR of 2.30 and OR of 2.25 seem “virtually identical” to me, especially when their precision estimates appear just as similar.

              Dr. Hooker never treated the data as if he could extrapolate temporal associations. On the contrary, he even said in his discussion section that some children counted as controls may have ended up developing autism since they were too young to be diagnosed at the time of the original study.

              Your repeated attacks on his paper are baseless. It is clear you’ve made up your mind about Dr. Hooker’s paper and are impervious to any information that might challenge your position. It’s getting to the point where you are no longer even arguing, just restating your position over and over which everyone here knows all too well now.

              • Btw, I notice other studies heavily relied on to back MMR’s safety such as the Madsen and Smeeth studies do not account for parity. Should we dismiss them too by your logic?

              • Yes I can say the data CDC omitted this data, it is displayed in the table above and validates Dr. Hooker’s findings. The point of claiming a relative risk computed from a case-control study to be inaccurate is that it could differ from the odds ratio, but that is clearly not the case here as the two are the same.

                Right because numbers. Hooker didn’t crunch the data as a case-control; he crunched it as a cohort study and without knowing temporality of MMR vaccination with ASD diagnosis, it’s dead in the water.

                He said himself in his own paper that he used a statistical method similar to DeStefano et al. and got the same results.

                “He said”, there’s that funny turn of phrase again which is meaningless since he doesn’t show his work, not even a table of descriptive statistics.

                He also did disclose his involvement in litigation and the journal only stated there were undisclosed interests on his part, not on the part of any peer reviewers. Clearly, he was not trying to conceal any competing interests he may have.

                Let’s see what he says, “Dr. Hooker has been involved in vaccine/biologic litigation.” That’s pretty ambiguous to an unitiated reader and you would be all over that if it was someone you didn’t agree with. He failed to include that he sits on the board of Focus Autism which funded the study and a no-no according to Biomed Central’s Competing Interest Policy: http://www.biomedcentral.com/about/editorialpolicies#CompetingInterests He failed to specify that he was a petitioner in an NVICP case. Biomed Central also has a peer-reviewer policy about COIs contrary to what you think: http://publicationethics.org/files/u7140/Peer%20review%20guidelines.pdf Of course we don’t know if that was an underlying factor for retraction or not.

                As I stated, he did exclude low birth weight children in one analysis and still got a very high risk – especially when stratifying for male gender among African-American children.

                Did he now? And what were the p-values of that analysis?

                I guess your argument rests on gestational age and parity – although other analyses done by CDC for thimerosal-containing vaccines for instance showed an increased risk for autism when both low birth weight and preterm children were excluded for the analysis. Seems like a tough sell to say it would be different for MMR. That leaves parity as the only remaining variable seemingly unaccounted for in either of these analyses – are you suggesting that after accounting for all these other variables failed to make a risk go away, that accounting that this one remaining variable would make all these findings just go away? That’s hard to believe.

                These are pretty rookie mistakes. Are you aware that tests exist to measure potential confounders before a final analysis is done? It’s also very dangerous and wrong to ASSume that an association present in one study will be present in another just because vaccines.

                And Dr. Thompson is a whistleblower; documentation backs up what he said. An RR of 2.30 and OR of 2.25 seem “virtually identical” to me, especially when their precision estimates appear just as similar.

                What documentation would this be? What “precision estimates” are there? All this drawn from a table with no context, nor p-values and stratified differently.

                Dr. Hooker never treated the data as if he could extrapolate temporal associations. On the contrary, he even said in his discussion section that some children counted as controls may have ended up developing autism since they were too young to be diagnosed at the time of the original study.

                Right, so you shouldn’t analyse the data as if it were based upon temporality which is what Hooker did.

                Your repeated attacks on his paper are baseless. It is clear you’ve made up your mind about Dr. Hooker’s paper and are impervious to any information that might challenge your position. It’s getting to the point where you are no longer even arguing, just restating your position over and over which everyone here knows all too well now.

                I’ve been offering reasoned, polite and supported discussion of Hooker’s highly flawed analysis so I take acception to your characterisation of baseless attacks. If by “impervious” you mean that I refuse to accept very poor statistical analyses and word-of-mouth as sufficient evidence then I guess I’ll wear that badge with honour.

                • “he crunched it as a cohort study” – No, he crunched it as cross-sectional.

                  “there’s that funny turn of phrase again which is meaningless” – Except it’s not uncommon for publications to refer to unpublished data. You can’t always publish everything due to space constraints, and in Brian’s case his findings are backed up by CDC’s anyway.

                  “That’s pretty ambiguous to an unitiated reader and you would be all over that if it was someone you didn’t agree with.”

                  -You don’t know what I’d do, and furthermore I think it would be pretty obvious that he was involved in litigation AGAINST the vaccine program.

                  “Did he now? And what were the p-values of that analysis?” Yes – in Table 4.

                  “These are pretty rookie mistakes” – So Madsen and Smeeth – two big studies used by the Institute of Medicine to exonerate MMR from causing autism made “rookie mistakes” by not accounting for parity? Good to know.

                  “It’s also very dangerous and wrong to ASSume that an association present in one study will be present in another just because vaccines.”

                  So an association exists for thimerosal, but not MMR? I would think if anything that gestational age would be more likely to confound associations with thimerosal exposure during the first month of life than for a vaccine not routinely given until after the first year of life.

                  “All this drawn from a table with no context, nor p-values and stratified differently.” – No actually, stratified to African-American children like in Brian’s paper. And the CIs of each findings are nearly the same.

                  “Right, so you shouldn’t analyse the data as if it were based upon temporality which is what Hooker did.” – He didn’t, that’s what I’m trying to tell you.

                  “I’ve been offering reasoned, polite and supported discussion of Hooker’s highly flawed analysis so I take acception to your characterisation of baseless attacks.” Well they are, and you’re just proving my point all the more now.

                  • “he crunched it as a cohort study” – No, he crunched it as cross-sectional.

                    Anyone looking at how he modelled the data can see that it was a cohort design and if that wasn’t enough, Hooker explicitely states that, “In this paper, we present the results of a cohort study using the same data from the Destefano et al. [14] analysis.” Taking a tumble down the hierarchy of study-design strength, particularly when the dataset available to him was sufficient to conduct a case-control is a bizarre strategy to salvage Hooker’s miscalculated results.

                    Except it’s not uncommon for publications to refer to unpublished data. You can’t always publish everything due to space constraints, and in Brian’s case his findings are backed up by CDC’s anyway.

                    You’re correct, however Hooker didn’t even publish a table of descriptive statistics. What could possibly be the excuse for that? Hooker’s findings aren’t backed up by the CDC since the table you posted has no context as far as modelling and the boxed strata are different. The comparable age bin is not similar to Hooker’s.

                    -You don’t know what I’d do, and furthermore I think it would be pretty obvious that he was involved in litigation AGAINST the vaccine program.

                    Actually I do but how is it obvious that Hooker was involved in litigation AGAINST the vaccine programme when all that was stated was, “Dr. Hooker has been involved in vaccine/biologic litigation.”? Was he a petitioner? Was he an expert witness either for a petitioner or for the DoJ? Furthermore, the syntax implies he had past involvement, not current. Why didn’t he just explicitly state that he was currently involved as a petitioner in the NVICP? Why didn’t he include his involvement as a board member for the very organisation that funded the study? Are you saying that it was acceptable to hold back such glaring COIs?

                    Yes – in Table 4.

                    And what were those p-values?

                    So Madsen and Smeeth – two big studies used by the Institute of Medicine to exonerate MMR from causing autism made “rookie mistakes” by not accounting for parity? Good to know.

                    Not at all. I should have stated that with more precision; testing for potential confounders exists and Hooker didn’t employ any. If parity was not a confounder in the Madsen and Smeeth studies, then no error was made. If you wish to hone in on a single confounding variable, then okay but the fact remains that Hooker didn’t adjust for any, save a crude and amateur analysis in table 4.

                    So an association exists for thimerosal, but not MMR? I would think if anything that gestational age would be more likely to confound associations with thimerosal exposure during the first month of life than for a vaccine not routinely given until after the first year of life.

                    No, actually an association does not exist for either if one follows the evidence. Why would gestational age be more likely for thimerosal exposure when it’s your own people who can’t seem to agree on causation to begin with? Given autism is a spectrum disorder, gestational age is a confounding factor. Period.

                    No actually, stratified to African-American children like in Brian’s paper. And the CIs of each findings are nearly the same.

                    No, the age strata that are boxed off in the alleged CDC table are not the same strata that Hooker has in his table. The fact that the numbers are the same raise more questions than answers I’m afraid.

                    “Right, so you shouldn’t analyse the data as if it were based upon temporality which is what Hooker did.” – He didn’t, that’s what I’m trying to tell you.

                    Then he should have done a case-control with conditional logistic regression instead of a cohort model with a chi-square (or Fisher’s exact test; it’s hard to tell which is another problem) and RR. It’s right there in his own words.

                    • He should have said cross-sectional in that sentence, but it doesn’t change the validity of his results. Relative risk would be more meaningful to the average person than odds ratios and this is an issue which effects everybody, so I would imagine that is why Brian Hooker conducted it that way. He said so himself it was a simple analysis, and his results are the same as for the study. He may not have provided summary statistics, but let’s not forget he was reanalyzing the same data from an already published study. The table’s overall age cut-off of <36 months and stratification to African-American race is apparent, just like in Brian Hooker's paper no matter how much you deny it. And you tell me what his p-values in table 4 are, you can see them too.

                      We don't know if parity was a confounder in Madsen or Smeeth because neither study attempted to control for it. You claim Brian Hooker didn't control for any, but he did account for some variables by stratifying by race and gender while providing results that both include and exclude low birth weight children. He was upfront about what variables he did not stratify or control for.

                      My point about gestational age was that since it is a birth outcome, chances are it would more likely confound associations with exposures closer to birth. Yet when these children were excluded along with low birth weight children, thimersoal exposure in the first month of life was still associated. MMR is not routinely given until much later, yet an association is still present in Brian Hooker's when low birth weight children were excluded.

                      He said so himself he did employ conditional logistic regression similar to DeStefano and got similar results, DeStefano's own results are virtually identical. He did not analyze the data as if it were a cohort study; that was a mistake in the sentence you quoted above.

                    • He should have said cross-sectional in that sentence, but it doesn’t change the validity of his results.

                      Again, that is a bizarre claim particularly since that isn’t the case. Hooker meant cohort and that’s what he published.

                      Relative risk would be more meaningful to the average person than odds ratios and this is an issue which effects everybody, so I would imagine that is why Brian Hooker conducted it that way.

                      Relative risk is easier for lay people to understand but that doesn’t mean it’s the correct measure of association or since it is roughly the same as an OR given the prevalence of the outcome, the presentation of it being a relative risk is what is problematic because Hooker didn’t have the information to present it as a relative risk.

                      He said so himself it was a simple analysis, and his results are the same as for the study.

                      I believe it was Narad who said that simple isn’t a virtue here and that is correct. A “simple” analysis doesn’t control for confounding variables and it shows here.

                      He may not have provided summary statistics, but let’s not forget he was reanalyzing the same data from an already published study.

                      Is that so? I thought he was more specifically examining an African American subgroup so then please tell me the breakdown of males, females, ASD, non-ASD, birth weight and age at MMR strata.

                      And you tell me what his p-values in table 4 are, you can see them too.

                      Statistically insignificant.

                      You claim Brian Hooker didn’t control for any, but he did account for some variables by stratifying by race and gender while providing results that both include and exclude low birth weight children. He was upfront about what variables he did not stratify or control for.

                      Stratifying does not control for confounding, that would be achieved with your actual statistical model and Chi-square isn’t adequate. He had a single result for excluding low birth-weight AA children and the association went away.

                      He said so himself he did employ conditional logistic regression similar to DeStefano and got similar results, DeStefano’s own results are virtually identical.

                      Then why would he opt to publish inferior statistical analyses if he had conditional logistic regression results?

                      He did not analyze the data as if it were a cohort study; that was a mistake in the sentence you quoted above.

                      A mistake that he repeated throughout the paper and also in a talk he gave about it?

                    • No, he’s just not as familiar with the names of specific study designs. Hence the incorrect reference in the paper and in his talk. There’s nothing bizarre about it, and it’s not “problematic” for relative risk to be used when referring to the prevalence ratio in a cross-sectional study. Risk is something that resonates more with the average person, which we agree on, so that would be why he reported results from the statistical methods that yielded those measures of association. Brian Hooker was very specific about what he was trying to find – a relationship between early MMR and autism in African-American children of different genders using the original study data. He found it; he even went further to exclude children on one analysis who could be a possible source of bias on his results, and the association stratified to the race and gender of particular interest held. However, he was still examining in the total sample too as well as in non-African-American children. He also clearly showed that some variables were not confounders but modified the effect since the association was different across race and sex groups, and he also showed low birth weight status did not have an effect because the association held for the primary at-risk group. The argument that his results are invalid because he did not study more variables is frankly misguided and could be used to dismiss any epidemiological study.

                    • No, he’s just not as familiar with the names of specific study designs. Hence the incorrect reference in the paper and in his talk. There’s nothing bizarre about it, and it’s not “problematic” for relative risk to be used when referring to the prevalence ratio in a cross-sectional study. Risk is something that resonates more with the average person, which we agree on, so that would be why he reported results from the statistical methods that yielded those measures of association.

                      Just because something is more digestable for public consumption doesn’t mean it’s correct. Risk implies causation and Hooker did have a completeness of data to perform an RR (nor did DeStefano et al.).

                      Brian Hooker was very specific about what he was trying to find – a relationship between early MMR and autism in African-American children of different genders using the original study data.

                      Yes he was specific but he failed to analyse the data properly to do so. In order to examine a single variable’s effect upon an outcome, in this case MMR and an ASD then all other confounding variables must be controlled for. Since Hooker did not do this then all of those other confounding variables are still exerting their effects on the outcome and the results are unreliable.

                      He found it; he even went further to exclude children on one analysis who could be a possible source of bias on his results, and the association stratified to the race and gender of particular interest held.

                      No not exactly, anyone can see that statistical significance was lost. What he did wasn’t correct because:
                      a.) He still didn’t control for other confounders such as gestational age and SES (which is a surrogate). And,
                      b.) He reduced the sample power and that further eroded the reliability of his result.

                      However, he was still examining in the total sample too as well as in non-African-American children. He also clearly showed that some variables were not confounders but modified the effect since the association was different across race and sex groups, and he also showed low birth weight status did not have an effect because the association held for the primary at-risk group. The argument that his results are invalid because he did not study more variables is frankly misguided and could be used to dismiss any epidemiological study.

                      Please tell me how he demonstrated that confounding variables were rather effect modifiers. He didn’t show that low birth weight didn’t have an effect (see above). When you have numerous variables that are confounders then it is crucial to control for them and since he didn’t it is absolutely correct to say that that invalidates his results amongst numerous other things. I’m having difficulty reconciling how you can on one hand say that Hooker is so inexperienced that he can’t even correctly identify his own study design but that we should have confidence in his results which clearly weren’t even modelled correctly.

                    • He had a completeness of data to perform an RR for a cross-sectional study, which he mistakenly called a cohort study. It was simply an incorrect word choice, nothing that invalidates his findings.

                      Since you’ve previously only mentioned two variables – parity and gestational age – you have essentially your argument on the reliability of Hooker’s findings on those two variables. The former of the two was not controlled for in either of the other major studies used to defend MMR. Interestingly, you’ve now shifted your argument from parity and gestational age to SES and gestational age. Yet SES was not even controlled for by DeStefano et al, so you are essentially calling the original study unreliable. Either that, or you are really staking that argument on one variable. However, your very next assertion casts bigger doubt on the reliability of DeStefano than on Hooker:

                      “b.) He reduced the sample power and that further eroded the reliability of his result.”

                      The reduced sample power was due to the fact that it was restricted to the birth certificate group, which was how DeStefano et al. diluted the results to the null. Yet as Hooker has shown, the risk remains with a similar age cut-off within the same susceptible racial/gender subgroup with low birth weight children excluded. You tell me to “see above” in order to see how lbw did not have an effect, but you have yet to thoroughly demonstrate that. Meanwhile, gender and race clearly would be effect modifiers because the associations were different across different subgroups. This contrasts with a confounder, where the association is the same across variations.

                      All your frivolous efforts to pick apart and find fault with Brian Hooker’s paper is really getting redundant now. You also have yet to answer my previous comment as to how you would know Thompson’s status as a CDC employee no longer working with the other DeStefano coauthors when his position at NCBDDD would make Coleen Boyle his boss. I have yet to see anything in the public domain that would contradict that or which contradicts the fact that he’s a whistleblower for that matter.

                  • Actually, I was wrong: Hooker still crunched it as a case-control study. DeStefano’s study was population-based and extracted all cases from a defined population. He didn’t conduct a cohort study and should have referred to his measures as “odds” or “risk” instead of “incidence,” but that does not at all detract from his findings.

          • Jake, take a look at Hooker’s tables 2 and 4. Explain what you see.

            • I discuss them both in my article; I can’t read your mind, so why don’t you just share your thoughts?

              • You don’t have to read my mind, I would like to hear your explanation for the differences in Tables 2 and 4 and what they imply.

                • The latter was confined to the birth certificate sample and excluded low birth weight children.

                  • That’s it? No comment about what happened to the p-values? In the spirit of scientific discussion don’t you owe it to your readers what the difference was in those tables?

                    • It doesn’t matter, the confidence intervals are almost the same; the lower confidence limits of each are exactly the same.

                    • It doesn’t matter, the confidence intervals are almost the same; the lower confidence limits of each are exactly the same.

                      I beg to differ, it matters considerably. When he removed low birth-weight AA children, the association went away and the one p-value which is significant is questionable given the sample size. The lower CIs are not the same and I don’t know what that has to do with anything anyway.

                    • The association for all genders went away but the one for boys remained significant and strong. Table 4 was confined to the birth certificate sample so that would probably explain the diminished risk.

      • According to the Dr. Thompson, those results should have been published per study protocol. CDC breached that protocol.

        Thompson is neither an epidemiologist nor a statistician, he has to get others to do that work for him so how is he an authority on what should or shouldn’t be published regarding statistical results. Additionally, that wasn’t part of the study protocol, don’t know where you got that from.

        Why hasn’t Thompson removed his name from the original DeStefano study if he finds it so erroneous?

        • CDC describes him as an epidemiologist. I got the issue about the protocol from Dr. Thompson himself.

          Given that Dr. Thompson still works for CDC, I would imagine removing his name from the study at this point and time would put him in an even more precarious position.

          • CDC describes him as an epidemiologist.

            “Once described.” He has no formal pre-CDC epidemiological credentials. What does Atlanta require? The two-year EIS program? To what extent is this a routine requirement?

            I honestly don’t know, and I haven’t come across anything that helps with that situation.

          • CDC describes him as an epidemiologist. I got the issue about the protocol from Dr. Thompson himself.

            No, the CDC merely assigned him a group title; he’s a psychologist by education and training with perhaps some epi courses somewhere along the way. You “got the issue about the protocol” from Dr. Thompson himself but where is the documentation that racial strata were ever a subgroup analysis on the study protocol? You would demand the same level of proof of others making extraordinary claims.

            Given that Dr. Thompson still works for CDC, I would imagine removing his name from the study at this point and time would put him in an even more precarious position.

            That’s a cop out given he can actually claim whistleblower protection at any time and currently has legal representation.

            • It’s not so extraordinary given that Dr. Thompson is himself a coauthor and the existence of the omitted results above. Failure to report stratum-specific associations also goes against how epidemiologists are trained.

              Just because he has legal representation does not mean he is totally safe, and the fact that his identity was released against his wishes certainly didn’t help matters for him.

              • It’s not so extraordinary given that Dr. Thompson is himself a coauthor and the existence of the omitted results above. Failure to report stratum-specific associations also goes against how epidemiologists are trained.

                If Thompson is so wedded to the idea that the DeStefano study is fraudulent then it would behoove him to remove his name particularly if he has evidence that data was omitted. I don’t understand the excuses. As for the last sentence, “failure to report” would be predicated upon what was in the initial protocol, you have yet to establish that.

                Just because he has legal representation does not mean he is totally safe, and the fact that his identity was released against his wishes certainly didn’t help matters for him.

                Well you lay down with dogs…

                • Remove his name from a study byline alongside people he still has to work with? It’s not that simple, especially after having his identity leaked which was not his fault. If you want to deny what the whistleblower is himself saying, then go right ahead, but CDC has yet to give a coherent response if it ever does.

                  • Remove his name from a study byline alongside people he still has to work with?

                    I’d say he’s thrown his co-authors thoroughly under the bus and backed it up on them by now. Besides, he moved to a different group before this dust-up occurred so he isn’t working with them any more.

                    It’s not that simple, especially after having his identity leaked which was not his fault.

                    Oh but it is that simple if he stands by his allegation of fraud (his statement via his attorney says otherwise though) and need I remind you that Hooker and Lewis were just as complicit in this deception as Wakefield was?

                    If you want to deny what the whistleblower is himself saying, then go right ahead, but CDC has yet to give a coherent response if it ever does.

                    He isn’t a whistleblower and I don’t lend any weight to a highly-edited propaganda piece by people who deceived Thompson. All I have to go on is his own statement via his attorney. Frankly I’m not impressed with the CDC response either.

                    • “thrown his co-authors thoroughly under the bus and backed it up on them” – You mean blew the whistle on them? Also, how would you know about what his exact position is at CDC?

                      “…need I remind you that Hooker and Lewis were just as complicit in this deception as Wakefield was?”

                      No, it was Wakefield who outed him.

                      He is a whistleblower and has federal whistleblower status. I am referring to his statement.

                      “I’m not impressed with the CDC response” – Glad we agree there.

                    • Also, how would you know about what his exact position is at CDC?

                      It’s in the public domain.

                      No, it was Wakefield who outed him.

                      Hooker and Lewis participated in the recording and propaganda piece. Neither has admonished Wakefield for outing Thompson nor distanced themselves from him for doing so.

                      He is a whistleblower and has federal whistleblower status. I am referring to his statement.

                      No he isn’t; he has an attorney who specialises in qui tam cases that’s it.

                    • Where? How would you know all this?

                      I do find it troubling that neither Hooker and Lewis have spoken out, and I wish they would. However, that’s not to say that they were the ones who outed Thompson when it was specifically Wakefield.

                    • Where? How would you know all this?

                      Since you fancy yourself an investigator and all the information regarding Thompson’s change in departments is publicly available, you should have no trouble finding and reporting that.

                      I do find it troubling that neither Hooker and Lewis have spoken out, and I wish they would. However, that’s not to say that they were the ones who outed Thompson when it was specifically Wakefield.

                      I reviewed the original video that outed Thompson and I’m afraid that Hooker is complicit which would explain why he isn’t speaking out against it. Hooker names Thompson in the video and even goes so far as to make a statement, “Dr. Thompson had appointed me his priest.” Now why would Hooker name Dr. Thompson in a video made PRIOR to outing Thompson if outing him wasn’t part of the plan? Furthermore, I understand Hooker is a religious fellow; betraying Thompson’s trust and confidentiality under that guise is even more reprehensible.

                    • Yes, but Brian Hooker never expected Wakefield would release the video without anyone’s permission.

  7. Lawrence on October 10, 2014 at 12:47 pm said:

    @Media Scholar – why hasn’t Hooker isn’t a statement? Perhaps because he realizes that he got caught with his hand in the cookie jar?

    I’m waiting for Jake to publish what he gets from Hooker – because Hooker would have received the official notice of retraction, which would articulate the exact reasons for the retraction.

  8. White Rose on October 10, 2014 at 12:49 pm said:

    Sullivan Carey lbrb – are you out there ? I see you are doing a lot of talking right now on ABC news .

    What is the rate of Autism amongst the Somalian community of the US & Sweden & Ldn ?
    Just answer that question . I think Brian Hooker can tell you (using the cdc’s own hidden figures)

    The mainstream press has finally picked up the story after weeks and weeks and weeks . ABC news .
    And are trying to do another Andrew Wakefield on it . Thanks ABC but read through it and I see another massive vaccine uptake scare is on the cards (maybe you should go back to the total news blackout) .
    Even by ABC’s own admission they have not covered this story for months now . Why ABC ? Why ?

    Instead of seeing my child fighting for his life after all the heinous vaccines now we can watch the cdc fighting for their very lives . Its payback time boys (and Gerberding et al).
    This is most enjoyable indeed . I smell prison time Colleen \ Marshalyn \ DeStefano !

  9. White Rose on October 10, 2014 at 3:51 pm said:

    Question for Larry et al:

    The CDC seems to be carefully monitoring the opt out rate which is occurring at a very exciting rate in the more affluent & educated areas of California .
    I wonder what the current rate of the genetic disease Autism is amongst those children whose parents have had the wonderful wisdom to opt out of all the vaccines ?
    The cdc must already be in possession of those figures and aere working on how to bend and skew that big fat zero % as we speak .

    Yes thats right everyone , in the prophetic words of Bob Marley ” No Vaccines . No Autism”

    • Lawrence on October 15, 2014 at 8:57 am said:

      @WhiteRose – since there are parents at AoA with unvaccinated autistic children, your statement is not factually correct (but then again, that’s nothing new).

  10. Yes thats right everyone , in the prophetic words of Bob Marley ” No Vaccines . No Autism”

    I will admit to being impressed by this allusion, although I’m not convinced that it was deliberate on WR’s part, rather than being just an idea-smoothie gone wrong:

    “Dr. Josef Issels first offered his alternative cancer therapies in a Nazi-fied atmosphere of ruthlessness and quackery. In the 1930s, chronic cancer patients consulted Dr. Issels and received his experimental ‘combination therapy,’ a regimen of diet, homeopathic remedies, vitamins, exercise and detoxification, among other holistic approaches. Today, his clinic offers training in cancer immunization vaccines, UV blood irradiation, oxygen and ozone therapy, ‘biological dentistry’ [tooth extraction], immunity elicitation by mixed bacterial vaccine, blood heating, and so on.”

  11. I’ve been mulling this over, and I don’t think the question is importune. Your blog, your rules, Jake, same as it ever was.

    But, given the insinuations/accusations below, I would like to straightforwardly ask whether you accept or reject the proposition that the dissenting or challenging commenters here are, in fact, bought and paid for.

  12. Media Scholar on October 11, 2014 at 8:16 am said:

    @Media Scholar – why hasn’t Hooker isn’t a statement? Perhaps because he realizes that he got caught with his hand in the cookie jar?
    ————————————————
    To suggest that Hooker has improprieties concerning his paper which shows the causal relationship between MMR and Autism is baseless, particularly since the basis of the allegations which led to the retraction are unknown except to those responsible for leveraging them against his publisher.

  13. Media Scholar on October 11, 2014 at 12:02 pm said:

    Why hasn’t Thompson removed his name from the original DeStefano study if he finds it so erroneous?
    ———————————
    Thompson could remove his name from the published study, but has correctly approached his participation in the appropriate manner. If he were to request his name be removed from the study, it would show his condemnation of the study is of small importance. Besides, what would he now be blowing the whistle about?

    The fact that Thompson is now poised as a whistle-blower brings significantly more weight in that he intends to take responsibility for his CDC group’s fraudulent research activities relatively common among US Centers for Disease Control’s stable of crooked nags.

    By now condemning the study yet remaining a co-author it’s a certain sign he values the importance of the Autism-vaccines causal relationship research endeavor and has been guided to the point of realization figures fudging in order to conceal the causal relationship between MMR vaccine and Autism does nothing to serve a greater good. It just makes more unsuspecting American moms and dads heart-broken. It certainly makes more innocent American babies contract Autism.

  14. Lawrence on October 12, 2014 at 2:11 pm said:

    @MediaScholar – that’s Orwellian in the extreme….so, by not retracting his name, he’s making a bigger statement? Yeah, he’s already made the statement that he supports vaccines 100% & would never encourage anyone not to get one, regardless of race.

    Doesn’t exactly sound like an individual who is positioned to “do anything” that would be favorable to the anti-vax Zealots.

    @Jake – so, what’s the hold-up with Hooker? Why hasn’t he responded to your questions about the retraction?

  15. Lawrence on October 12, 2014 at 10:11 pm said:

    @jake – not a “good time?” What kind of journalist are you?

    You make a number of claims that have zero back-up evidence….perhaps you’re giving Hooker a pass because of personal bias?

  16. White Rose on October 13, 2014 at 10:50 am said:

    Herod et al .

    We are simply warrior parents . The science isnt difficult to understand if you have a reason to do the work ie a vaccine damaged child .

    But like I always say to you “paid” guys , make sure you don’t listen to us , get your flu jab and your ebola jab and whatever other noxious nazi poison they try to pass off as a healthcare .

    Does anyone have the ingredient list to the Tekmira Ebola poison yet ? this will be interesting .

    signing off for now – in anticipation – of the staged fake Ebola outbreak in Northern Europe . Ldn & Paris ?
    But still yet they will allow the air travel to continue unabated . How else can you fake a pandemic ?
    If ebola truly is set to kill 1.4M by january you’d think the criminals at the CDC would try to quarantine the countries in West Africa , but it seems clear to me they want Ebola cases to appear in Texas (part of the plan.)

  17. White Rose on October 13, 2014 at 3:23 pm said:

    I have know the pain of Vaccine derived Autism for many years now .
    But I just viewed the latest revelations – “Isolated Autism – Your Child”
    and the hairs have just stood up on the back of my neck in incandescent rage , lets call it (to coin a new phrase) “CDC rage” .

    Colleen Boyle – where do you go from here . Can you be “reformed” ?
    I’m truly shocked at the absolute depths of the deep that you have operated at .
    Its hard to see how the establishment is going to make this go away .
    I’m shocked to my very core . This means that Colleen Boyle is a mass murderer .
    A bit of good news for Colleen I suppose – she has had 12 years of freedom she should never have enjoyed.
    But that liberty surely has to be all over now .

  18. Media Scholar on October 13, 2014 at 6:51 pm said:

    @MediaScholar – that’s Orwellian in the extreme….so, by not retracting his name, he’s making a bigger statement? Yeah, he’s already made the statement that he supports vaccines 100% & would never encourage anyone not to get one, regardless of race.
    ———————————————————-
    Not really. You just want to get him thrown off the study list of authors and save the crooked research generated by the skunks inside the CDC Autism research boiler room. You know, post a humble brag, pat yourself on the back, and go back to calling completely fabricated research fraud impeccably fudged “good science”.

  19. White Rose on October 14, 2014 at 2:08 pm said:

    ” he’s already made the statement that he supports vaccines 100% & would never encourage anyone not to get one, regardless of race.”

    That is not true , and you know it Lar .
    He has gone on record saying no pregnant woman should ever take the Hg Flu jab for a start .

    • Lawrence on October 14, 2014 at 7:19 pm said:

      “On record?” You mean those illegal recordings that Hooker made? That’s not exactly “on the record.”

      What we do have, for the record, is Thompson’s official statement through his attorneys which states, without equivocation that he supports vaccines and would never recommend against vaccination, regardless of race.

      • They were made legally.

        • Lawrence on October 14, 2014 at 11:04 pm said:

          @Jake – and the legal opinion you base that on is, what exactly?

          Again, Thompson’s formal statement is that he supports vaccines & would not recommend against them for anyone, regardless of race…..

          • That it was done in places where it was legal.

            • Lawrence on October 15, 2014 at 8:54 am said:

              @Jake – so please, do tell exactly where these recordings were made?

              And since Thompson himself has said that he was unaware that he was being recorded, Hooker has, at minimum, acted unethically….so, where is his statement on the retraction again?

              • Where it is legal.

                It’s not unethical to get what someone says on record in case for whatever reason they are unable to speak out later…

                I’m sure we’ll here it in due time…

                • Lawrence on October 15, 2014 at 1:21 pm said:

                  @Jake – that’s a mighty big “grey area” you’ve opened up….but then again, you seem to have no problem using / releasing private correspondence when it suits your own purposes, so perhaps asking you for a straightforward answer based on the facts is too much, considering your past efforts.

                  Under the “WhistleBlower” statutes, Dr. Thompson is free to say whatever he pleases & if the CDC takes action against him, they are liable and would be subjected to a major lawsuit….so at this point, nothing is preventing him from either confirming or denying what has been said.

                  One can assume that he is keeping his head down & hoping that all of this blows over….but his silence is deafening, as is Hooker’s at this point.

                • Hooker appears to work in both California (for Simpson University, a Bible college) and Washington (for ARES Corp., a consulting firm). Both California and Washington are “two-party consent” states, where it is illegal to record a conversation unless all parties consent to that recording. In Washington it’s a “gross misdemeanor,” but in California it may be a felony. I haven’t read the statute very closely, so I’m not sure about that–it might also be a misdemeanor.

                  I’m not very familiar with California or Washington law, but I’ve been involved in cases where the Illinois wiretapping law came up. (Illinois is also a two-party consent case.) Prosecutions based on such laws are not common, but they do happen. And regardless of whether or not charges are ever filed, it is in fact illegal to record a phone call without everyone’s consent in California and Washington.

                  Do you have reason to believe that Hooker made his recordings somewhere other than the states where he works and presumably resides?

  20. White Rose on October 14, 2014 at 9:18 pm said:

    Oh you really are a hoot Larry . Those recordings are illegal but the mass poisoning of millions of babies worldwide is somehow defendable ? yes that is what you are defending !
    The cdc has been caught red handed , lying . cheating , murdering and maiming the international masses and you are worried about the legality of a recorded conversation .

    Reminds me of the guards found in charge of the concentration camps at the end of the war who tried to continue on doing their murderous jobs regardless , despite the presence of allied troops & witnesses .

    Have a listen to this Lar :
    http://sfbayview.com/2014/10/cynthia-mckinney-on-autism-and-ferguson/

    I think you boys might be losing control of the situation . Looks like its spiralling out of control to me .
    Berlin Wall time . Its over Larry – give it up . The parallels to the holocaust are striking .
    How do you people hope to keep a lid on this ? I think its over for you . Everyone is talking – all at once .

  21. White Rose on October 15, 2014 at 8:14 am said:

    Msg for Dan Olmsted :

    So recently Dan you commended the cdc for its response to the Ebola outbreak in West Africa .
    I cannot get over that you would do smthg like that , particularly knowing all you do about the cdc.
    I already think AoA is a Pharma controlled opposition vehicle after the damage the Blaxman did to the Congress hearings and the 5000 parent Omnibus case . No I’m not going to stop bringing that up ever .

    However Dan , how you hope to ever get a compliment past everyone regarding the conduct of the murderous cdc is beyond me .

    I think the best way to think of Dan Olmsted’s comments , is anyone who would dare to overlook the despicable activities of Adolf Hitler during the 6 years of WW2 , and then compliment Adolf on the very fine car he designed (the VW beetle) . And the cdc has been getting away with this much longer than Adolf .

  22. Sam Hall on October 15, 2014 at 5:36 pm said:

    Colleen Boyle and her chums were seeking legal help back in 2002.
    I can only imagine being Colleen Boyle presently. I wonder where she will end up in the history books.
    A female scientist who knowingly falsified data which led to the murder of lord knows how many and the injury of millions of babies, black and white. That is quite an epitaph.

    How does that woman sleep at night ? I would be living in fear of my life had I been mixed up in all this illegal activity. There is no defence, just none.

    She will surely be one of the main civilian fall guys for this. What a fool ! what on earth did she ever have to gain from this. No money will ever have been worth it as her name and family name will be dirt in the books of (in)human(e) history. Almost as bad as having a serial killer in the family ……….oh wait …….

  23. Sam Hall on October 15, 2014 at 5:45 pm said:

    It is quite amusing here to see such a sad bunch desperately scrambling to make some sort of recovery from this situation. The numbers of parents who know the truth “hear this well” runs into many MILLIONS. A few old paid for crony’s on this and other websites isn’t going to change that. How long do you ‘paid for’ think you will be gainfully employed – the dam is leaking !

    Bottom line –
    The CDC falsified data and therefore research.
    That research concealed evidence proving the MMR caused Autism.
    That research was the basis for many other subsequent studies which means that they must be called into question now.
    The US government ultimately lied to its people and knowingly caused injury and death to millions of its citizens and used research to coerce other government’s around the world to do the same.

    How long do you think til the dam breaks ?

  24. A simple question for you, Jake, as an epidemiologist. What does this sentence from the Hooker paper mean and why is it important regarding the appropriateness of reporting unadjusted, unconditional risk ratios (or prevalence ratios if you call this a cross-sectional study)?

    “Control children were chosen from “regular” education programs and were within the same age group and schools of attendance or neighboring school as cases.”

    Surely you have been taught about individual and frequency matching. What is the appropriate method for handling the situation where individual matching was performed and artificial correlation was introduced (if, say, matched on age, sex, and school)? I’ll give you a hint, the answer is not to calculate an unadjusted, unconditional risk ratio (or prevalence ratio). If you should feel so inclined, I would recommend passing along to Dr. Hooker that for all future analyses of matched data, only methods conditioned on matched set are appropriate for individual matching and the matching factors must be included in the model for frequency matching.

    If the conditional results were the same, those should have been reported. A statistician, or yourself, might have been able to point this out before submission. As well as the importance of accounting for additional confounding factors, such as maternal characteristics.

    • You have a point, Hooker’s RR for the overall sample differ more from DeStefano et al.’s results than Hooker’s race-stratified RR, and it is weaker than DeStefano et al.’s too. However, Hooker was trying to replicate the race-stratified result and quantify it in terms of risk and how it differed across racial groups, which were virtually the same as ORs. That alone probably strained much of his publication space. So I can see why he would not have reported conditional results despite referencing them, even if reporting them would have arguably been the right thing to do.

  25. White Rose on October 16, 2014 at 11:43 am said:

    Ricky – are you (desperately) looking for that straw ?

    Keep on clutching boys .

    • @WR

      Maybe, just maybe, papers such as this one wouldn’t get dismantled (or rejected outright) so often if some attention was paid to appropriate methods that would pass statistical review. Remember this nugget from the retraction statement?

      “Furthermore, post-publication peer review raised concerns about the validity of the methods and statistical analysis, therefore the Editors no longer have confidence in the soundness of the findings.”

      But, that’s okay, this would only be important if you valued citable, scientifically valid evidence.

      • Lawrence on October 16, 2014 at 4:04 pm said:

        Since even Jake has said (on numerous occasions now) that Hooker was looking for a specific link in the original study, it is obvious that he only performed “just enough” statistical analysis to get the results he was looking for from the beginning…..this is a perfect example of crafting the process around finding the “right” result and not even attempting to appear impartial in the analysis.

        So, Dr. Thompson told Hooker what to look for – so Hooker only did enough to find exactly that, with no real attempt to discover if the facts were correct…..

        Real researchers don’t go into a study with the end-result already determined (they merely present a hypothesis & see if the data backed them up) – which is exactly what Hooker did himself, in this case.

      • The CDC knew 11 years ago, that there was a higher rate of autism among boys who received their MMR shot earlier than 36 months of age.

        They KNEW this…. and yet they made a conscious decision to keep that information from an unsuspecting public!!!

        And you think what’s important, is all this senseless semantics around how that truth is being suppressed …. again??

        I can guarantee you that statistics, no matter how they’ve been massaged, will mean NOTHING to the parents of MMR ravaged children, when they come to realize that those children could well have been perfectly healthy, if the truth about MMR hadn’t purposely been kept from them.

        • justthestats on October 17, 2014 at 8:20 pm said:

          The CDC knew 11 years ago, that there was a higher rate of autism among boys who received their MMR shot earlier than 36 months of age.

          …in one particular non-randomly-sampled dataset, which by its very non-randomly-sampled nature means that you can’t draw any influence about the prevalence in the population as a whole.

          Also note that they were measuring autism diagnoses, not autism itself. To do that, they’d have to assess every member of the sample themselves to make sure that the criteria for autism were applied consistently (reducing the rates of misdiagnoses and missed diagnoses) and they would have not sampled separately from a group that they knew was diagnosed with autism and one they knew was not diagnosed with and unlikely to have autism.

          Oh, by the way, you misread the table. Gender doesn’t show up in that table at the top at all. For the record, in the actual study, the adjusted OR for boys vaccinated before 36 months was a little high at 1.64, but not statistically significant, which means that there is a very good chance that the higher OR is a result of “luck of the draw”, not because there was anything real going on. And the “below 36 months” is cherry-picked there, because the OR at below 24 was 1.01 and the OR at below 18 was 0.94. So if you think those numbers mean that there is anything to a MMR/autism link, you should be advocating for people to get vaccinated before 24 months in order to slightly decrease the probability of getting autism.

          Anyway, I’m going to assume for the rest of this comment that you meant to type “blacks” instead of “boys”, because the table was about race. If that’s not the case you can ignore the rest.

          They KNEW this…. and yet they made a conscious decision to keep that information from an unsuspecting public!!!

          The table is conveniently missing any context that would say whether the numbers were adjusted for the non-random sampling and the multiple-comparisons effect. I suspect that neither adjustment was performed, and that means that the bolded numbers in the table are likely to be spurious associations. Seriously, at the beginning of a study you print out tons of tables like this with lots of spurious associations, and then you use the more powerful techniques to determine which are real. This kind of thing is just used to eliminate the obvious dead ends.

          But even assuming that the number are adjusted appropriately, that doesn’t mean it was wrong not to report them. Real associations don’t go away when you control for more variables — they typically get stronger. Since table 5 in the actual paper reports that the OR isn’t significant for race when controlling for birth weight, and maternal age and education, and the result in the table at the top of the page was just barely statistically significant to begin with, it’s pretty likely that it was a spurious result.

          Since you asked why they would choose not to report a likely spurious result, there are two possible explanations: 1) Why waste paper and readers’ time on irrelevant details? 2) In this particular case, they were afraid that people who weren’t familiar with how research is done and how common spurious effects are would misunderstand the data and think that it was some kind of real evidence that African-Americans should avoid getting vaccinated during certain ages. Given your reaction, that fear seems justified.

          And you think what’s important, is all this senseless semantics around how that truth is being suppressed …. again??

          Using the correct statistics is the exact opposite of “senseless semantics”. If you try to use a fuel pump designed for a Ford F-150 as an oil filter for a Mini Cooper, you’re not going to get the results you want, even though they an F-150 and a Mini Cooper are both motor vehicles and oil and gasoline are both refined from the same goop. Likewise, if you use a statistical procedure designed for one circumstance in a different circumstance, you’ll get numbers, but the numbers won’t mean anything useful, and it’s just as justified for people to criticize you for that kind of inappropriate usage as for the oil filter mistake.

          In this case, Brian Hooker used statistics designed to be used with randomly-sampled data on a non-randomly-sampled dataset. So his results are only meaningful if they happened to non-randomly-sample a dataset that miraculously turned out to be just like a typical randomly-sampled dataset. Can you see why people who understand statistics are unimpressed?

          Furthermore, even if that choice was somehow justifiable, he didn’t control for other variables that may have had an effect, which is pretty hard to excuse when we know he had that data available to him. When you have the option to report statistics that are less likely to have spurious associations, and you report more potentially problematic ones instead, it looks very dubious.

          In summary, the only thing that was “suppressed” was a few spurious associations. I’ve seen enough spurious associations in academia and my day job that I don’t find them very interesting, and I doubt the editors of any journal would disagree.

          • Never mind Hooker’s RR is almost exactly the same as DeStefano’s OR.

            • justthestats on October 20, 2014 at 7:14 pm said:

              I’m not sure the fact that the RR numbers in Hooker’s table and the OR numbers in the purportedly DeStefano table are similar works in the direction you want it to. We know that Hooker used an inappropriate statistical method, didn’t correct for multiple comparisons, and didn’t adjust for the non-random sampling. The table at the top of the page is complete mystery-meat methodology.

              So there are three possibilities here:
              1) The mystery meat table uses the appropriate statistics, and the several problems with the Hooker numbers miraculously were either all had small effects or almost canceled each other out in every aspect.

              or

              2) They mystery meat table doesn’t use all the appropriate statistics to say what you’re trying to make it say, and so the numbers are more similar because they are methodologically similar. For example, it’s not unreasonable to infer that this table uses logistical regression because of the title. If it’s unconditional regression that isn’t adjusted for confounders or multiple comparisons, it will probably come up with similar numbers to a chi square or Fisher’s exact analysis.

              I know which option I find more likely, but without the methodology for the selectively-disclosed table at the top of the page, it’s impossible to say for sure.

              I have no doubt that you’re familiar with the topic, but for the benefit of your readers who are less familiar with this kind of research, I’ll note that the kind of statistics mentioned in 2) are useful for preliminary analysis , so it wouldn’t be surprising that they’d run such numbers. It’s not correct to interpret them as saying that there is an actual elevated RR for African-Americans receiving MMR at certain specific times but not other times. It is correct to interpret it as saying that race and age for first MMR are two variables that need to be included in the regression model. Given that that is “Table 5”, they ran at least four other selectively undisclosed analyses that would have been useful for identifying other relevant variables and possible relations between them.

              • According to the paper, conditional logistic regression similar to DeStefano was done, and the results were the same. In any case, the purpose of the reanalysis was to see if it could find the association originally found in CDC’s omitted results that should not have been omitted according to a CDC scientist, and it did.

                • justthestats on October 22, 2014 at 10:06 pm said:

                  Any thoughts on why there is nothing with 596 or 222 cases in either paper, but there is in the table? Also, any thoughts about why we should care about an association that goes away when you control for other variables?

                  • Probably due to the fact that they excluded any analysis stratified by race that wasn’t restricted to the birth certificate sample.

                    • justthestats on October 23, 2014 at 2:00 pm said:

                      I thought about that, but there are 571+384+65=1,024 controls and 199+137+19=355 cases in the birth certificate sample that specify race, so the numbers don’t match up.

                      The reason I’m asking about when we should worry about associations that disappear when you control for other variables is that accidental drownings are correlated with ice cream sales, but the correlation goes away when you control for the outside temperature, and even that correlation goes away when you control for attendance at public pools. It’s a pretty good assumption that if you could measure swimming pool usage in general that would be an even better fit. Does this mean that controlling for those better predictors is using statistics to hide the association in order to protect Big Dairy from their drowning-inducing products?

                    • There are only 355 cases in the birth certificate sample, as opposed to 596 in the table – nearly half as many cases. That would explain the depleted sample size and diluted findings that resulted from it.

                    • justthestats on October 23, 2014 at 8:12 pm said:

                      There are 333+230+40=603 cases in the total sample, so it doesn’t match up there either.

                    • You’re talking a difference of seven cases, so not a huge discrepancy. Maybe info on their racial status and/or vaccination history was missing?

                    • justthestats on October 24, 2014 at 4:34 pm said:

                      All 634 cases had vaccine data. 603 of them had race data.

                    • You mean all 624 cases?

                      I suppose it is possible that seven cases with available race data were excluded due to unavailability of other information such as whether or not they would belong in the “Isolated” subgroup. But again, we’re not talking a huge discrepancy.

  26. White Rose on October 16, 2014 at 4:05 pm said:

    This paper was retracted because the contents are “red hot toxic” for the whole political system .
    And you know it .

    But this isnt going away – you must know that – else why are you here ? We have concrete evidence of criminality at the cdc going back at least 13 years & I’d suggest much further .

    The media certainly knows it .They know how huge a story this is & how wide its ramifications would be – else why the complete and total international news blackout ?

    Meanwhile over at AoA , instead of rallying the parents of the “vaccine maimed” to fight and complain tooth and nail right now as one – they continue on with the T shirt competitions or crosswords or what ever else they do over there .
    AoA , a vaccine\autism pressure group , my foot , pah .

    • justthestats on October 17, 2014 at 8:25 pm said:

      The media certainly knows it .They know how huge a story this is & how wide its ramifications would be – else why the complete and total international news blackout ?

      I would think that a huge story with wide ramifications would be great for page views. Given how hard it’s been for the media to stay profitable lately, you would think someone would jump at the chance to break a huge story.

      Another possible explanation is that they’ve evaluated your evidence and the counter evidence and found the evidence for the “huge story” to be wanting.

  27. White Rose on October 18, 2014 at 2:03 pm said:

    “JustTheStats” I engaged a former senior managing editor from a major tabloid just days ago who point blank refused to even discuss the subject (now a government minister – nepotism? )
    I encouraged him to get his seasonal flu jab – as I do you .

    Bill Gates was seen in the offices of the Guardian just before all this story broke .
    No doubt putting them on a warning – publish and face closure no doubt .

    I understand the whole US media is tied up by just 5 companies – is this true ?

    We are censored all the time . The BBC endlessly put out nonsense misinformation about Autism ,
    older mothers , older fathers , people who drive green cars , people with blue cars . genetic this n that
    You know the nonsense I’m talking about , course you do , you probably lap it up .

    We have lots of examples now , people who formerly trusted government , fully vaccinated and lost children to the carnage , after a year or two in the vaccine abyss , pick themselves up , and go on to have more children , this time fully unvaccinated , and guess what no autism .

    Affluent large swathes of California now recording up 70% refusal rates – oh dear ! why ?

    • justthestats on October 20, 2014 at 11:05 pm said:

      I engaged a former senior managing editor from a major tabloid just days ago who point blank refused to even discuss the subject

      You might consider that either 1) you need to adjust your techniques for broaching the topic or 2) he believes that he’s already sufficiently familiar with the topic and isn’t interested in discussing it further. Just because you’re interested in something doesn’t mean he is.

      Since this acquaintance is not a journalist anymore, what were your goals for the engagement? Did it have anything to do with his current government position?

      (now a government minister – nepotism? )

      Without knowing whether this person is related to someone else in power, it’s impossible to say. You’ve certainly provided no information to judge either way.

      I encouraged him to get his seasonal flu jab – as I do you .

      I appreciate your concern for my health and the health of those I come in contact with. When I had a job that put me in frequent contact with people who were more at risk of serious complications, and well, more in contact with people in general, I did regularly. But now I probably won’t bother.

      From a rational economic perspective, it probably makes sense to get it even though the flu shot’s efficacy is lackluster, but there are the purely psychological factors of it being a hassle to do and the fact that there’s a non-negligible chance that either I wouldn’t have gotten the flu anyway or I’ll only come in contact with the wrong strain or that I’ll get the right strain but that the vaccine won’t be effective, and all of those are subjectively more annoying than the flu, which I’m at least used to. But if the flu were as awful for someone in my demographic as, say, polio, I’d get the seasonal shot in a heartbeat even with all the caveats I mentioned above.

      Your government acquaintance sounds like he’s the kind of person who really can’t afford a missed day at work, so your advice was especially good. You should evaluate your situation, including the number of people you come into contact with and their situations and decide how critical it is for you to get the shot, as well.

      Bill Gates was seen in the offices of the Guardian just before all this story broke .

      Was this the same Guardian that routinely thumbs its nose at the US and UK national security establishment, or some other Guardian?

      No doubt putting them on a warning – publish and face closure no doubt .

      See, this is what I mean by adjusting your technique. I’m a reasonable person who has changed his mind on a lot of things, including a few things that are very consequential in my life, based on reasonable arguments and good, solid evidence. But this sentence is the kind of thing that is not going to change my mind about anything.

      You see, Bill Gates is an absurdly wealthy person who made his money in an unrelated area and who is sometimes described as having signs of ASD himself. it’s hard to imagine how he could become involved in some kind of conspiracy to suppress evidence that vaccines were somehow causing autism, and how they would motivate him to toe the line.

      It’s even harder to imagine how he would become the Voice of Big Pharma delivering an ultimatum. It’s hard to imagine how such an ultimatum could be enforced. Either it wouldn’t be a very subtle, untraceable way of shutting them down, which would be hard to do quickly enough to matter, or it would be so fast and obviously related to the last things they published that the Streisand Effect would do the rest. It’s even harder to imagine such an amazingly effective conspiracy would be unaware of the Streisand Effect.

      So, you’ve jumped to the conclusion that Gates was delivering an ultimatum, but taking your report that he was seen at the Guardian to be correct, your listener can probably think of any number of more likely scenarios. For example, he might be being interviewed on the question of why his company has turned so much crappier since he left the CEO position, or what he thinks it can do to recover, or what the Bill and Melinda Gates Foundation has been up to recently. Any one of those would be more plausible. So your listener would probably think “If that’s the best he/she has, he’s got nothing.”

      Seriously, there is enough evidence here that you tend to draw unwarranted conclusions that you should consider running your arguments past someone else to make sure that they make sense.

      I understand the whole US media is tied up by just 5 companies – is this true ?

      I imagine that Google can answer that question better than I can. Having a few companies predominate isn’t great from an economic perspective, but it actually works against your implicit argument — larger companies with more clients are less beholden to any particular one. In this case, they don’t make a whole lot of money off of pharma advertisements, and they would stand to make a lot of money off of revealing some big conspiracy. After all, Bernstein and Woodward are two of the big reasons why the Washington Post is a household name, and the Washington Times is not.

      Plus, there are a huge number of smaller companies that would jump at the chance to become bigger companies. Revealing some huge conspiracy would be a pretty good way to do that. So strictly speaking, the answer to your question is “no.”

      We are censored all the time .

      I’d be fascinated to see some documentation backing that up. Are you, personally, being censored at this moment? Is there anything you can say about that without fear of serious repercussions? Can you comment on who is doing the censoring, and by what authority they are doing it?

      The BBC endlessly put out nonsense misinformation about Autism ,
      older mothers , older fathers , people who drive green cars , people with blue cars . genetic this n that

      I guess I have a couple of questions:
      1) How do you know that this is nonsense misinformation?
      2) How is it that the BBC doesn’t know that, or if they do, why?

      You know the nonsense I’m talking about , course you do , you probably lap it up .

      I could suppose that every qualified epidemiologist who studies autism is either incompetent or secretly doesn’t want people to understand autism. I don’t have any evidence for that, and it seems like that’s a whole lot of people in very diverse circumstances — I personally would expect someone somewhere on the planet to manage to be qualified, competent, honest, and not bizarrely malevolent.

      Do me a favor and take off the tinfoil glasses for a second. Use your imagination and imagine a world where autism really is a condition with complex etiology and multiple risk factors. Imagine that someone proposed that a certain factor was the cause, but the only research that was consistent with that theory had really small sample sizes and shoddy statistics, while several large studies from diverse groups all over the planet showed no association. If you were a competent epidemiologist truly trying to determine what the risk factors were for autism, would you spend all your energy on the proposed factor that didn’t pan out repeatedly, or would you try to find new risk factors and narrow down the already established ones more precisely? Food for thought.

      We have lots of examples now , people who formerly trusted government ,

      Ah, there’s your problem, right there. “Government” is way too big of an organization to fully trust or distrust. 😉

      fully vaccinated and lost children to the carnage ,

      I really don’t understand your use of the word “carnage” here. I’ve know parents who have had children die in traumatic accidents. I’ve known parents of autistic children. Their plights don’t seem the same at all to me. On a more personal note, I watched a relative pass away once. The experience was utterly unlike the realization that some of the more unusual behavior of my child could be explained pretty well by what they’re now calling “autism”. Utterly unlike it. There is no comparison.

      On the other hand, some of the things measles does to its survivors and non-survivors sound to me more like “carnage”, although I really associate the word more with things like being crushed, myself. Regardless, it’s not good rhetorical tactics to try to convince people that something causes something like carnage when the alternative causes something even more like carnage. It tends to distract instead of convince.

      after a year or two in the vaccine abyss , pick themselves up , and go on to have more children , this time fully unvaccinated , and guess what no autism .

      …except for the unvaccinated ones that do have autism… How do you explain those?

      All of the families I know with autistic children have more than child. None of them have more than one autistic children. I have not asked who has vaccinated and who has not, but since in one case the children were born long before the vaccine misassociation and in another case a parent is a doctor (and so unsurprisingly actually understands the issues involved), it seems that in at least two of the cases all the children were vaccinated. So it seems like the anecdata is against your theory, here, since subsequent children can be autistic or neurotypical independent of vaccination.

      I’ll note that the anecdata is not inconsistent with the possibility of genetic influence, however. As a simple example, imagine that autism is a simple Mendelian autosomal recessive trait. In that case, assuming asymptomatic carrier parents, the chance of an individual child being autistic is 25%. It’s not nearly that simple, of course, because if it was we would have figured it out by now, but having more genes involved than one doesn’t change the fact that parents can not show symptoms but have children that do, and that parents can have several children but not all of them show the disease.

      Anyway, since you asked my opinion, I think that parental age is probably a proxy for one or more real risk factors, and blue and green cars are correlated with autism for the same uninteresting reason that consumption of organic food is correlated with autism.

      The evidence for genes being involved is pretty strong, though. Even the causation theory you seem to favor pretty much requires genetic susceptibility, because not all vaccinated people are autistic.

      If you have the chance to talk to the parents of these unvaccinated children, I hope that you’ll show the same genuine concern for them that you did for me and your acquaintance and encourage them to vaccinate their children. After all, vaccines prevent diseases that are much worse than the flu, and it would be tragic for them to have a child killed or permanently disabled by a preventable disease.

      Affluent large swathes of California now recording up 70% refusal rates – oh dear ! why ?

      My guess: a potent combination of misinformation, faddish behavior, and the well-known tendency of people to try to do anything to protect themselves from the unknown, even if such behavior is ultimately not productive. Sprinkle a generous dose of Dunning-Kruger in a fairly small but supercritical number of cases.

      A believer in Big Pharma conspiracies might think that they’re secretly paid to not vaccinate their children so that they’ll get the horrible diseases that people in CA have largely forgotten about. It would be a good tactic, since it will ensure that every one else will continue to vaccinate after hearing about the children nearby disabled and killed by vaccine preventable diseases, while Big Pharma will make more money on the treatments for the children who get those diseases. A totally unethical win/win situation.

      A believer in astrology might suspect it has to do with minute tidal forces caused by distant chunks of dirt and balls of gas.

      A believer in Greek mythology might think Apollo has something to do with it.

      It is actually possible to find out why they are doing what they are doing. A competent sociologist would be your best bet, and if you truly are interested, you should try to convince one to do a study on it. What you should not do is try to do one yourself or convince someone who is sympathetic to your cause but not a sociologist to try to do it. The reason for this is simple: real sociologists train for years so that they learn how to recognize and avoid the kinds of mistakes that people who have no expertise in the field tend to make.

  28. White Rose on October 18, 2014 at 2:16 pm said:

    Msg for Ginger Taylor :

    Ginger , regarding these letters found here :
    http://focusautisminc.org/a-letter-recently-sent-that-clarifies-my-position-in-the-world-of-autism-barry-segal/

    I’m banned from AoA for continually asking the same questions about the Blaxman .
    Why are you defending the Blaxman ? I don’t get it .
    Has he been struck down deaf dumb and blind ? Why are you writing on his behalf ?
    Why doesn’t the Blaxman come out from whatever rock he is hiding under and answer the allegations ?
    Why are you & Dan Burns protecting him ? All he has to do is answer my questions ?

    Why do all these mercenaries camp out here on AI ?
    And not a one of them bothering with AoA ?

    Ginger – you and others better face up to some cold hard facts we are heavily infested with Pharma paid infiltraitors (sp.) – who knows you might be another one of them ?

    Dan Burns asked me previously – what have I ever done ? I have spent 7 years so far fighting .
    Doing my best , engaging talking to scientists , trying to talk to people of influence , complaining , whingeing , warning all my family and friends (even though many still defer to the lying media) .

    If we ever get that firm breakthrough (as I believe we will – and I think are very close to right now – despite the best efforts of AoA to prevent it – look at how AoA burys all the meaningful stories under a huge layer of nonsense all the time) – then we will hold the media to account for their role in all this .
    The media has colluded with the vaccine holocaust & the Hg-enocide in a way Hitler would be most proud .

    • Rebecca Fisher on October 20, 2014 at 7:08 am said:

      The reason there are more voices of reason here at AI is that, to his credit, Jake doesn’t filter anything like as heavily as AOA.

      Kind regards,

      Becky

  29. Doug Troutman on October 18, 2014 at 8:31 pm said:

    I guess I thought huge story was William Thompson admitted the CDC gins up its vaccine autism studies. Maybe some of you stat boys can locate the Verstraeten data sets so somebody independent could look at those.

  30. Pingback: Brian Hooker Rewards Wakefield For Betraying Whistleblower | Autism Investigated

  31. Pingback: Brian Hooker’s Findings Are Confirmed By CDC’s Results | Autism Investigated | Mommy's Knowledge

  32. Pingback: Brian Hooker's Findings Are Confirmed By CDC's Results | Autism Investigated

  33. Pingback: JAMA Study Implicates Early MMR Vaccine in Causing Autism - Autism Investigated

  34. Pingback: Review of Jain et al. JAMA 2015, and comments on MMR-autism vaccinepapers.org

  35. Pingback: Focus Autism Attorney: Wakefield "Hijacked The Story"

  36. Pingback: How Andrew Wakefield Alienated NAACP - Autism Investigated

  37. Pingback: Canary Party-Led Group Behind Failed SB277 Opposition - Autism Investigated

  38. Pingback: Referendum Is Wrong Because SB277 Is Unconstitutional

  39. Jake,
    Nice job bringing this forward as you did.

    The detractors will never accept what is presented quite clearly right in front of their face.

    I am now writing a book on “Genetic and Environmental Causes of Autism”. Please contact me offline re: that.

    In the process of reading 1300 studies in autism research, took a moment to summarize some of what the rest of the literature has to say about another adjuvant – aluminum.

    Have a look:

    http://jameslyonsweiler.com/2015/11/16/paging-dr-offit-your-aluminum-neurotoxicity-reading-assignments-are-ready/

    Keep up your own good works.

  40. Pingback: Neurodiversity Is Social Justice Cancer - Autism Investigated

  41. Pingback: CA SB277 Bill Co-Sponsor Called CDC Scientist a "Fraud"

  42. Pingback: Take It from an Anti-Vaxxer – There Is No CDC Whistleblower Scandal | Earth Med Research

Leave a Reply

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

Post Navigation