Hey i dont know who was asking about weather the test holds up in a criminal court, im a student of criminology and tecnically it does hold up as a 'reference' however it cannot be used to say point blank if a person is sane or insane because it doesnt work in that method. as with alot of personality tests it doesnt say wrong/right it merly states how many attributes of a certain personality type a person taking the test appears to have.

famously karla homolka took the test after her husband was arrested and the test was given by a Dr. Ardnt, who actually found that her test result didnt appear to be anything like her! 

However, when working on a case such as this (im sing this as an example, im sure there are other better examples of its criminal use) it was already apparent that she had some physchological damage and unusual behaviour. (look up karla homolka and paul bernardo if you wish to know more) The test was given to homolka and at the end of it was a good explination of her behaviour that was seen as 'put of character. 

As with most personality test they can be slightly vague and used to explain almost everything, but this one is extremly accurate. 

However as she was never put on trial these where never shown. 

I hope that helps, any other examples i would be very interested to find out about! 

Also would say that it is the type of test thats almost 'unbeatable' there are only two answers to each question, and alot of the questions influence each other, for example:

the difference in a person that 'reads articles about crime' harmless in itself, but when paired with 'finds it hard to control anger' becomes a slightly different statement.

nopolycop wrote on Jul 5^th, 2008 at 9:12pm:


I would argue that real world practice of instruments/knowledge that you do not fully understand the underlying theory of is of marginal benefit.  And a N=1 case study (your experience) does not carry as much generalizablity as the hundreds of published studies of the instrument. I think my last question would be what is your definition of "successfully manipulated"? Did you mean alter the profile only? Or did you mean alter the profile and have it not come across as invalidated or suspicious to a trained psychologist? What were your validity scale T scores? Who evaluated your results?  You would have had to have your profile evaluated by an independent outside psychologist to make a valid claim of successful manipulation. In other words, who did you fool?

Regardless, if the psychologists came back and told you that you seemed extraordinarily depressed, when in fact that you were not and this was the phenomenology you set out to fake, my money would be on that you had a bum psychologists.  However, in the case that they were fully competent, I would argue that you accomplished a feat that most others can not (according to the empirical literature to date) and therefore your accomplishment is far from "easy." Again I would go back to the argument that an N=1 case example carries less external validity for the general population than multitudes of studies involving thousands over the course of decades. If you can do it and the empirical literature demonstrates that most others can not, I do not think this qualifies as "easy", would you agree? Remeber, "easy" is a subjective experience.  What is easy for one person, may not be easy for the next person.

nopolycop wrote on Jul 5^th, 2008 at 9:12pm:



Ok. The short answer to your question would be no. So what evidence are you citing that shows the test is "easy" to manipulate? I have never seen a review, meta-analysis, or empirical study that supports this claim and could not find evidence of this doing a lit search just now.

I am glad to hear that everything worked out well for you in your eval for whatever position. However, I would give yourself more credit, and your method, alot less.  I would suggest that you actually accomplished very little in your method. Or at least you weren't really doing what you think you were doing. Let me explain:

You seem to be running on the premise that the questions are face vaild indicators of the construct they are measuring. They are not. The questions that seem like they belong to one type of disorder scale, often do not.  Assignment of questions to a particular scale (Depression, Schizophrenia, Hypochondriasis, etc) was done after the list of questions were normed on clinical samples during its original construction.  The "deviant response" is indicated by  individuals already known to have the disorder who endorse an item at a frequency that enables it to be a statistically reliable discriminator of that population from "normal" populations and other clinical populations (i.e.,empirical keying method), not by a clinical rationale of "this seems crazy" or "this seems like something someone with this disorder would endorse." It doesn't matter what the content of the question was, as long as the question happens to be statically discriminant among normal vs clinical populations. An example would be that, the question (regardless of its content) is endorsed by the depressed normative group so much more over any other group that it becomes a statistically reliable discriminator of that group, and hence, the construct of "depression." Assuming of course it wasn't also highly endorsed in the "normative" control sample. If it wasn't, it indicates this item is endorsed significantly more by the depressed group than both "normals" and those with other disorders. Therefore, it is viewed as a pure (nonoverlapping) indicator of depression. Endorsement of the item in the deviant direction identifies the test taker as responding in the same way as the depressed norm group did.  Item endorsement is aggregate, and after so many deviant endorsements consistent with that group, your T-score is raised on that scale. Remember, this is solely based on whether the question discriminates reliably between groups. It doesn't matter what the content of the question was. If it discriminates, it discriminates. Period. Certainly some of the questions are face valid, but the majority actually are not. Trust me, the questions that seem like they belong to one type of disorder scale, often do not, and the normative sample didn't always endorse or deny in the way you think they would have for particular questions. (i.e., questions do not always measure what you think they do). "I tortured animals as a child" was actually endorsed in only one particular disorder, and it it was NOT psychopaths.  It was found to be a discriminant question in another disorder and got coded on its scale instead. Hence, why I made the comment that an undergrad class in abnormal psych provides neither the time nor depth to fully understand the construction of the test and its subcales. I would not assume anything about the test until you you are formally educated on the method of construction and how exactly it does what it does. Psychologists are tricky fellows. One of the biggest reasons why the MMPI-2 is so loved by the field and has lasted so long (it was renormed in 1989 though) is there has been so much research done on the instrument on both its validity and its validity scales.  It's not an easy instrument to manipulate for the lay person, without it being picked up on. 

Lastly, you seem to not know or at least not understand the validity scales built into the instrument. Do you know what they are, their purpose, and the statics behind them? I doubt you would have discussed them in your undergrad class. Again, don't assume you understand how something works just because you have taken it. Obviously, we wouldn't want every person who takes it to be able to figure it out just by looking and thinking about it, right?

"When you did your lit search, did you discover any studies that attempted to prove or disprove the theory that the MMPI can be manipulated?"

Well it's not really "theory," its statistics. And yes there are hundreds of pubished studies on these issues dating back to the test's development. Meehl  (1948) was the first to report on the suppressor variable in psychological sophisticated malingerers, which lead to the development of the K scale. F and L were developed during initially norming. Actually the majority of the research on the MMPI-2 examines profile validity and manipulation issues. A Psychinfo search using Key words of "MMPI-2" and "malingering" yielded 188 published studies in in this area in just the past 3 years. A Pubmed search using the same key words reveals 166 studies. Because of my training and clinical practice I am already familiar with this literature. Overall, studies indicate the test has both high sensitivity and high specificity in identifying symptom exaggeration and/or manipulation in clinical samples, forensic samples, as well as those who were "coached" to simulate a particular disorder ("analogue" malingering studies).  Overall, coaching studies  demonstrate the ability of coached participants to suppress some the older standard validity scales fairly succesfully, but have typically been caught using the newer developed ones (eg.,FBS, Fp, Superlative Scale).

nopolycop wrote on Jul 9^th, 2008 at 2:02pm:


Yes, the research literature does indeed show that the MMPI-2 is a good measure for assessment of psychopathology. It is the most widely used personality/psychopathology screening in the Western world for this very reason. However, the test should not be interpreted blindly and the results should always be supplemented by other clinical information (such as an thorough clinical interview before hand) if diagnostic decisions are going to made. For example, If a cleint is in the midst of messy divorce and the partner is using every method possible to get child custody and take all of his financial assets, I would certainly expect some elevation of paranoia above the norm. In the context of this situational variable it is quite understandable, and I would explain this in the report so no one gets the wrong idea by just looking at the numerical score. 

I mentioned in my last post that both the sensitivity (proportion of actual positives which are correctly identified) and specificity (proportion of negatives which are correctly identified) of the individual subscales is very high. I do not have the sensitivity and specificity averages for each scale in front of me at the moment.  It varies by study and the population in question, but its in the very high 90% range. Perfectly acceptable for clinical use. I do not do these screens for enforcement agencies as part of my practice, but I see no reason why people should be paranoid about this issue.

Hi,
I was just reading the sample questions, and I have to take this test for a law suit I am very nervous, is there any way you can send me other questions, and what exactly they are looking for.  I had a bad fall on a cruise ship. and my neurologist wants me to take this test.  My lawyer says if I screw up on this test that could ruin my lawsuit.
Thank you

If you still have all of the questions.. I would love to get them.  Thanks. 

Does anybody know how many people receive negative results from taking the mmpi and do not get the jobs they had to take it for. After everything I have read, it hardly seems like the most effective tool for a psych evaluation.