[Interview]: Get to know Lazar

[Lazar]

do you ever have the eerie feeling that after you've thrown all these complicated modelling techniques to scores gathered through some sort of survey or psychological test... you might still not getting at anything?

i dunno, we both seem to come from the same tradition but sometimes i worry that after i've been doing all sorts of crazy factor analyses and SEMs, in the end, probably the test was badly designed to begin with and nothing that one comes up with even matters... :/

also... where did you interest with Bayesian statistics began? it's still relatively new among social scientists...

Oddly I think doing secondary data analysis has made me a better psychometrician because I HAVE to know what the data someone else collected can and cannot do. You still have to compromise a lot though and I often have to use less than perfect scales (i.e., one item self-concept scales). I do still have my doubts but where possible I try to have some concrete outcome because I think this provides more confidence. My biggest fear is not in the psychometrics but in the woefully inadequte causal models we end up having to use. Indeed I think psychology has become overly concerned with psychomtrics but has not really been much concerned with causal inference. Sociology is far in front on that score.

As for Bayes, I dont really remember. I can recall getting an interest and then working through Jackman's book but I am not sure where the initial interest came from.

[Dason]

Would you be willing to post a photo of yourself?

Even more importantly... would you be willing to post a few photos of your beagles? What are their names?

What is your favorite movie?

If you had an all expenses paid vacation to anywhere in the world - where would you go?

[Lazar]

Would you be willing to post a photo of yourself?

I will have to find one.

Even more importantly... would you be willing to post a few photos of your beagles? What are their names?

Yes I can do that. Their names are Pandora and Bella. Pandora is a cross with a cocker spaniel

What is your favorite movie?

Serenity is my favorite but I am not really a movie person though I much prefer TV series.

If you had an all expenses paid vacation to anywhere in the world - where would you go?

I would really like to go to Japan or South America.

[CowboyBear]

1. Of course the Aussies.

Are you relieved the Aussies are back ahead of Ireland in the T20 rankings?

More seriously: Where do you stand on the chi-square vs approximate fit indices debate in SEM?

[spunky]

More seriously: Where do you stand on the chi-square vs approximate fit indices debate in SEM?

do you REALLY wanna open that can of worms?

[CowboyBear]

do you REALLY wanna open that can of worms?

Haha! Yeah, if I say "approximate fit index" three times Les Hayduk might appear, and then we're all in trouble...

[Lazar]

Haha! Yeah, if I say "approximate fit index" three times Les Hayduk might appear, and then we're all in trouble...

See what CB is saying here SEMNET he is happy to betray the very principals of science!!!! [edit: This is my bad Les impersonation].

In all seriousness I am somewhere inbetween on this debate. I point to the Box quote that all models are approximations (heavily paraphrased) and thus for a suffciently sized sample any model will be rejected. This suggests a focus on statistical tests alone are unworkable. This issue is a realy problem for me as I have sample sizes in the 10 to 20,000 and beyond where the chi-square statistics is VERY unforgiving. Having said that reliance on fit indices too often leads to shovel misfit under the carpet and often there is no serious attempt to identify where misfit is coming from. This is a major problem. I try hard to a) always inspect MIs to see where the sources of fisfit are coming from and i make a point to know how the various fit indices are calculated and what biases they have. I am also a fan of using MC simulation to see what one should expect for a give sample and model.

p.s. note that I ignored the T20 question completely

[Dason]

What kind of music do you like?

Do you play any instruments?

Favorite Firefly character? Why?

[Lazar]

What kind of music do you like?

At the moment I am listening to metric's new cd and I really like all of their stuff. I also like Tegan and Sarah, Missy Higgins, Lady Hawk, Kimbra, La Roux, and Placebo. When in more fellow moods I listen To Martha or Rufus Wainwright. For older stuff I like Jonny Cash.

Do you play any instruments?

Yes I play the guitar, bass, and drums but have not touched any instrument in about six months. Actually I saw today that my guitar has a cover of mould on it which is not a good sign.

Favorite Firefly character? Why?

I did say Jayne earlier and he is my favourite particularly for comedy value in that vain I also like Wash "Its ok I'm a leaf on the wind"

[spunky]

i thought someone had asked this before but apparently no one has.... are you married? gf? children? (if there are no children, are you interested in having them in the future?)

[Lazar]

@Spunky

Yes I have been married for eight years (I got married at 20...gulp!). As for children I doubt it for a while my wife is just starting her PhD and I am trying to establish my career.

[trinker]

28? Wouldn't have guessed that. You've accomplished a lot early in life.

[Lazar]

Yes 28. PhD programs are slightly shorter here than in the US so I got out a little quicker than you guys will but yes I have been lucky to have had a good PhD supervisor and colleagues who got me a quality post-doc. All things that are very helpful in academia

[CowboyBear]

Continuing on the theme of opening cans of worms... What do you think of Joel Michell's critiques of "measurement" in psychology (or the lack thereof)? Do you have anything to do with him, being in the same city?

[Lazar]

Here is one of my dogs getting a present: