A: I do research on African American racial identity; in particular I'm interested in the meaning and significance that African Americans place in defining themselves in the context of race. I'm also interested in how that impacts other areas of their lives. We've looked at such things as academic outcomes, mental health outcomes, in the impact in the way in which African Americans experience and cope with racial discrimination, the process by which these attitudes are transmitted: race socialization processes between parents and adolescents.

A: We've developed a measure known as the MIBI or the Multi-Dimensional Inventory of Black Identity and we've refined that over the years. African Americans are very heterogeneous in the way in which they define themselves in the context of race, so we've developed a conceptual model in which we look at four different dimensions of racial identity. Two dimensions deal with the significance of race. The salience dimension looks at the extent to which race is a defining characteristic at any given moment in time. It's very situationally specific and it's driven by cues. The centrality dimension is the extent to which an individual normatively defines race. So salience is like a snapshot of what identity is prominent at any given moment in time, and centrality would be more like a movie.

The other two dimensions deal with meaning. What does it mean to be black? The first dimension, regard, is sub-divided into private regard and public regard. Private regard is the extent to which you feel positive about being black and other black people. Public regard is the extent to which you think others view blacks as positively or negatively. The last dimension is the ideology dimension. This is the dimension addressing what you think black people should do or should be like. We've identified 4 different ideological themes:

• Nationalist ideology emphasizes the uniqueness of the black experience and sees the world in terms of black versus others.


• Humanist ideology is on the exact opposite spectrum where you emphasize the commonalities of all people and you see individuals as individuals.


• Oppressed minority ideology emphasizes the commonalities between blacks and other oppressed groups.


• Assimilation ideology emphasizes the commonalities between African Americans and mainstream society.

With those 4 sub-dimensions of ideology, we don't define any one individual as one or the other but realize that people have different attitudes and beliefs and different domains in their lives. You might be nationalist in your political beliefs but humanist in terms of your interactions with individuals.

As you can see we're talking about a very complex model. We've also developed a specific measure for adolescents which we call MIBI-Teen, which is a 21 item measure. We are also in the future looking to develop coding schemes so that we can content code writing and do assessments from afar. With regards to salience, we're doing palm pilot approaches to measure the dynamic nature of salience and how it changes in different situations. We're very much interested in making sure that our instruments are valid and reliable measures of our underlying construct, but that our construct is more than just the instruments. By having multiple instruments, we're able to test whether or not the construct is really valid in different contexts and with different participants.

A: When I looked at the models of racial identity that were previously developed, the instruments never passed my "Uncle Robert Test". My Uncle Robert grew up in rural Alabama, and pretty much lived in rural Alabama until the day he died, with the exception of going off to fight in World War II. The instruments that were used to measure identity either didn't account for his experience at all, or they measured them in such a way that suggested he was either pathological or somehow flawed. This was completely inconsistent with most black people I knew, my uncle included.

Several years later, I had a group of wonderful graduate students who just got tired of hearing me gripe about the lack of decent measures, and said, "Put your money where your mouth is," and so we began to develop this model.

A: We ended up doing a lot of reading, going after many of the existing models, looking at every measure of identity we could find, and coming up first with a conceptual model. Before you can develop a measure, it's important that you're clear about what the construct is and (equally important) what the construct is not.

We were very clear that we were interested in attitudes and beliefs rather than behaviors, because behaviors are more likely to be influenced by contextual factors as opposed to what necessarily the individual brings to a situation. We got a pool of around 1,000 items from existing measures, and developed our conceptual model using the different dimensions. We tried to make sure that we defined these clearly, and that people were actually likely to differ on them, so that we would expect some variance. Then we thought about what types of beliefs might be associated more with a high score on centrality versus a low score, etc.

Then we wrote very specific definitions of each of the constructs (i.e. high centrality), included items from previous measures that were consistent with the construct. We also wrote items ourselves.

At that point we needed to test it out on students, which was a challenge in and of itself. When you're looking for African American undergraduates, the subject pool isn't as large as it might be for other folks. The measure had about 120 items starting off. After about 3 iterations, it came down to 56 item scale. Over the years, have piloted it out to community samples, adolescents, and other groups. We've done focus groups and other types of more qualitative approaches which helps us make sure we are measuring what we mean to measure.

A: We were very interested in making sure first and foremost that our items were face valid. For us the most important aspect really was face validity. The constructs themselves needed to be face valid. Before developing the instrument, we did focus groups and talked to various folks to make sure that our constructs actually made sense to African Americans. Were we on the right track? Or were we just being academics in an ivory tower not really measuring the folks that we wanted? Once we felt comfortable with the constructs, we wanted to make sure that our items were face valid, and represented the constructs. We weren't interested in fooling people per se. We wanted to make sure that the items under each construct made sense to people. Then we began to look to see whether or not the items had the appropriate construct validity. Did they hold together with regards to our factor analysis? Did they converge and diverge in ways that made sense?

We also needed to be sure that the measure was not a good predictor in any particular behavior, because again behavior is a function of both what the individual brings to the situation and situational context. For example, some measures use things like, "Do you have black art in your house?" as a measure of identification. Well, we argue that you could be highly identified as black, but if you live in Idaho chances are you can't find a lot of black art. On the other hand, you might not be very identified but live in southeast DC, and the store out front sells nothing but black art and so you'd have black art in your house. We wanted to make sure that the behavioral measures were related in ways that were consistent, but not too related.

We also checked to see if the constructs remained stable over time and also across different contexts. We developed a study where we put people in highly race-salient, highly gender-salient, and racially ambiguous situations to see whether or not that impacted individual scores on the MIBI; they didn't change the scores, so we felt that we were definitely on the right track. All of this has taken about 10 years.