MMPI-2, MMPI-RF and PAI

[EmotRegulation]

OK, ya'lls, I'd like to use your collective expertise here. I've searched the forums for information on preferences between the MMPI-2, MMPI-RF and PAI but it's a been a few years since the last thread of this nature, and I suspect the MMPI-RF is more widely used now than it was 3 years ago.

My questions are, particularly for those of you in academic medical centers or VAs, would you want incoming interns/new psychologists to have exposure to all of these tests? What would be the rank order preferences?

I ask because I'm teaching the graduate level personality assessment course, and we've always focused on the MMPI-2, but I'm thinking about moving the emphasis. The prevailing sentiment is that I *need* to still cover the MMPI-2 because it's "expected" for internships, but that perspective might not be accurate, and I personally prefer the PAI and the RF, so I'd rather cover those.

The other problem is that looking at psychometric information is difficult because the people publishing about these tests are, by and large, the people who created them, and thus are even more biased than your average researcher interested in promoting a particular perspective. Does anyone have any good resources--ideally those published by independent sources--reviewing the psychometrics and/or functional uses of these tests for general clinical practice (e.g., not forensic or police or another speciality function these tests are sometimes used for)?

---

Members don't see this ad.

 

[Therapist4Chnge]

1. MMPI-2RF
2. MMPI-2
3. PAI

Just my opinion.

 

[WisNeuro]

I second T4C's order. It's good to know the original scales of the MMPI since we get some old records of that, but in the VA, anything given in the past few years is pretty much all the RF.

 

[pavlove]

another +1 to T4C's post for general purposes. I have only seen a remaining strong preference for the MMPI-2 in forensic work.

That being said, I completed an assessment-heavy internship recently (2014-2015) and administered exclusively MMPI-2's (i.e. no RF) in a non-forensic setting.

 

[AcronymAllergy]

Agreed with all of the above.

Definitely keep the MMPI-2 training. It's still heavily used in some settings/by some clinicians, and will be found in reports for years to come.

For current administration, I predominantly use the RF (depending on the referral question, of course). It's shorter, and I find the interpretation to be more straightforward.

Get them at least a passing familiarity with the PAI. Same length as the RF, and the way the subscales are organized can be handy.

 

[futureapppsy2]

Do people use the PAI clinically very often? I've seen and used it in research but never as a standard part of a clinical assessment battery.

 

[cara susanna]

Do people use the PAI clinically very often? I've seen and used it in research but never as a standard part of a clinical assessment battery.

I actually used it a lot on internship.

 

[AcronymAllergy]

I actually used it a lot on internship.

And I used it very frequently in grad school.

 

[PsyDr]

I prefer the pai over the mmpi2. Lower reading level. putting the variability in the responses over increased questions compared to the mmpi2. Newer norms.

 

[AcronymAllergy]

I prefer the pai over the mmpi2. Lower reading level. putting the variability in the responses over increased questions compared to the mmpi2. Newer norms.

I remember attending one of Ben-Porath's talks at NAN a while back when he was discussing reading level of the MMPI-2 (RF in this case) vs. PAI. He'd mentioned that they were actually comparable in that regard, but came out ranked differently due to use of different programs to analyze them.

I also generally prefer the PAI to MMPI-2, but am becoming more and more of a fan of the RF.

 

[WisNeuro]

I think it mostly depends on the site and their history. Seems to be used much less in the VA. I've only seen it used a handful of times over the years at the 4 sites I have experience with. Pretty much all the MMPI-2 and RF in those places. In my patients, it's essentially another validity check.

 

[psychgirl77]

Do people use the PAI clinically very often? I've seen and used it in research but never as a standard part of a clinical assessment battery.

I am on internship and exclusively use the PAI. The only place I ever used the MMPI-2 was at a forensic practicum. Seems to be much more common in my area to use the PAI.

 

[Pragma]

Do people use the PAI clinically very often? I've seen and used it in research but never as a standard part of a clinical assessment battery.

I used it much more often than the mmpi at my amc sites.

 

[WisNeuro]

So help me FSM if anyone pulls out an MCMI!

 

[AcronymAllergy]

So help me FSM if anyone pulls out an MCMI!

Haha I've used it (and/or the MBMD) for organ transplant evals, but that's about it. And no, I've never seen one invalidated.

 

[WisNeuro]

Haha I've used it (and/or the MBMD) for organ transplant evals, but that's about it. And no, I've never seen one invalidated.

Mirrors my experience exactly.

 

[Pragma]

Haha I've used it (and/or the MBMD) for organ transplant evals, but that's about it. And no, I've never seen one invalidated.

Mirrors my experience exactly.

Ditto. I actually kind of like the MBMD.

 

[WisNeuro]

Ditto. I actually kind of like the MBMD.

I think it serves a very specific purpose. I just dislike the MCMI because it has no validity check whatsoever. You'll get a profile of complete PVT failure, completely invalid MMPI profile, and there's the MCMI-III, just smiling at you and saying "Interpret me in all my misleading glory!"

 

[cara susanna]

I've seen invalid MCMIs. But I've also used it in research so I've given a lot of them.

 

[WisNeuro]

I've seen invalid MCMIs. But I've also used it in research so I've given a lot of them.

You should write those rare occurrences up as case studies.

 

[Justanothergrad]

The RF is a neat tool in a lot of ways, but the cut-scores keep being lowered because of response style with endorsements now indicating that a T-score of 50 may be sufficient in some contexts. This makes it difficult to use the uniform norms as anything universally interpretable without delving into specialty comparison groups (not arguing against that, I'm an IPIP person so that's closer to what I think needs to be done...but it's not the traditional approach utilized by the RF (or personality assessment in general). The end-game clinical result is that I've seen providers interpreting relative peaks (i.e., within traditional normative ranges) of the RF as indications of significant pathology, and these are not folks that are in any way distant from the instrument's continued development.

The MCMI offers suggestions of personality dimensions that I've found more useful in terms of clinical conceptualization in a lot of ways than the PSY-5 of the RF/MMPI-2. This is mainly because the dimensions align more neatly to DSM/ICD and its not another interpretive step to say 'high disassociation equates to an indication of the following 3 things... now lets map out the PSY-5 and guess which PD best fits this pattern'. However, it suffers from difficulties with hardly ever being invalidated, as mentioned. It also continues to use corrected scores based on the validity indexes. Didn't we learn anything from K-correction?

I guess I just don't like our personality instruments, in general.

 

[Therapist4Chnge]

I think it mostly depends on the site and their history. Seems to be used much less in the VA. I've only seen it used a handful of times over the years at the 4 sites I have experience with. Pretty much all the MMPI-2 and RF in those places. In my patients, it's essentially another validity check.

I used the MMPI-2 and PAI during my time at the VA, as the MMPI-2RF had just come out (but wasn't available for use within our VISN yet). I use the BHI-2 for ~50% of my neuropsych cases (when I have more interest in adjustment following injury and there isn't a significant psych history) and the MMPI-2RF for the other 50%. The BHI-2 isn't a substitute, but it is pretty helpful w. certain neurologically compromised/post-injury populations.

 

[WisNeuro]

The RF is a neat tool in a lot of ways, but the cut-scores keep being lowered because of response style with endorsements now indicating that a T-score of 50 may be sufficient in some contexts. .

Man, we use the "VA correction" and pretty much wait til things are at >85 to consider it "high."

 

[Justanothergrad]

Man, we use the "VA correction" and pretty much wait til things are at >85 to consider it "high."

I suspect it all depends on the type of evaluation. VAs tend to produce flooded profiles while custody and fitness for duty evaluations under-report (those I mentioned). The issue is the same, there is no consistency in 'clinical level' and 'normal level' and there is far too much judgement being brought in, particulary for a test emphasizing Meehl's ideas of actuary assessment. I mean, after all, how does a test with this much interpretive looseness (no matter which direction) stay to that actuarial ideal?

I sometimes wonder if Meehl were to review our current use of the MMPI (both RF and 2), how he would describe it.

 

[PsyDr]

So help me FSM if anyone pulls out an MCMI!

I love the MCMI! It's sensitivity and specificity for detecting incompetent providers is outstanding!

 

[WisNeuro]

I love the MCMI! It's sensitivity and specificity for detecting incompetent providers is outstanding!

I know, right? Anytime I see that in a neuropsych context where there is even a hint of forensics or malingering possibility, I know I'm dealing with someone who doesn't care about useful data.

 

[BuckeyeLove]

Seems to be much more common in my area to use the PAI.

I've grown weary of having to explain the item-overlap/cross loading and how it pertains to profile interpretation to randos, as have other forensic folks Ive worked with; consequently, PAI has largely been my go to.

 

[Ollie123]

My take on the literature (though admittedly, not my area of expertise) is that generally speaking; MMPI = Quantity of evidence; PAI = Quality of evidence. Meaning the MMPI has lots of research on it, but the PAI generally seems to perform better for most common purposes. If memory serves, its also cheaper (or at least was when I was looking into it several years back).

That said, I don't think this is a huge deal either way. Some internship sites will have a preference one way or the other, but I doubt there is any pattern to it. My program taught the PAI and it was part of our standard assessment battery in the department clinic, but also strongly encouraged us to give at least 1-2 MMPIs before applying for internship. Haven't heard of it being a major problem for anyone. Though I imagine the same would be true if they focused on the MMPI and recommended giving 1-2 PAIs instead.

 

[EmotRegulation]

The PAI is waaaaaaay cheaper (if given paper/pencil especially).

These are useful comments and perspectives, all! I'm also glad to see the diss on the MCMI. I want the test to be useful, and I guess I like that it maps onto the DSM, but I also hate the DSM conceptualization of PDs (I'm agnostic about the appendix version), so mapping onto the DSM is not really a great advantage in my book.

For what it's worth, I don't love personality assessment either, though I love the study of personality and (separately) the process of assessment. I just tend to find that people put too much stock in the personality measures when used to diagnose/describe individual people.

 

[Therapist4Chnge]

I utilize personality measures in a pretty narrow way (looking at how a person manages stress, resiliency, any red flags for suicidality, etc), as well as a different way to quantify when someone absolutely tanks a cog battery. It bugs me when people over-interpret and try to attribute all sorts of things to the person.

In "the real world" most of the time the referring provider just wants to know if the patient has any red flags or related concerns, and they care much less about the various scales and 5-10 pages of interpretation. I'm a 2 paragraph or less writer, which took some adjustment when I first started.

 

[WisNeuro]

The PAI is waaaaaaay cheaper (if given paper/pencil especially).

These are useful comments and perspectives, all! I'm also glad to see the diss on the MCMI. I want the test to be useful, and I guess I like that it maps onto the DSM, but I also hate the DSM conceptualization of PDs (I'm agnostic about the appendix version), so mapping onto the DSM is not really a great advantage in my book.

For what it's worth, I don't love personality assessment either, though I love the study of personality and (separately) the process of assessment. I just tend to find that people put too much stock in the personality measures when used to diagnose/describe individual people.

Yeah, I don't think we have any "great" personality tests, merely good or ok ones. The reason I dislike the MCMI so much is the lack of meaningful SVT measures, as well as not having seen any great refutation of the significant methodological issues in its validiational tests brought up by some (Retzlaff, Rogers et al).

That being said, I hate seeing diagnoses thrown in charts merely due to results from a personality measure, without a significant clinical interview being done. That's just lazy clinical work. Similar to T4C, I use it a small adjunct, or at times a validity check, it is never the basis for a diagnosis for me

 

[Rivi]

My personal preference is:

MMPI-2-RF
PAI
MMPI-2

I see the MMPI-2 getting used less and less, both within and outside the VA. The MMPI-2-RF seems to be getting used the most.

Wasn't the MCMI-IV recently released? Has the IV improved at all over the III?

 

[WisNeuro]

Wasn't the MCMI-IV recently released? Has the IV improved at all over the III?

Just got released this fall, I believe. No opinion either way as of yet, that's one that I have not reviewed any of the literature for.

 

[LETSGONYR]

I'm curious... anyone have opinions on these measures for use with kids/teens? In my (admittedly limited) experience, I've found the MMPI-A lacks sensitivity and kids with significant soc/emotional difficulties often have no elevations. I know the PAI age range extends lower, but I haven't used it.