Tuesday, March 15, 2011

Disability Identity Questionnaire - Scales and Validity

Another one of my old surveys that I'm digging out and using SPSS on.

This one was inspired by an interesting book, Black and White Racial Identity: Theory, Research and Practice. In that book, they talk about stages in development of racial identity for both black and white people in US.

They also include a questionnaire that caught my eye, the Black Racial Identity Attitude Scale (RIAS-B). I decided to reword the questionnaire and give it to a bunch of (49 in total, though some left certain questions blank) disabled people to see whether they'd show similar patterns of responses.

There are four subscales on the RIAS-B, corresponding to four theorized stages: Preencounter, Encounter, Immersion and Internalization. Preencounter refers to a black racist, basically - someone who has internalized the prejudice aimed at them, or at the very least thinks it applies to 'other blacks' even if it doesn't apply to them personally. Encounter is a transition point when the Preencounter person suddenly realizes that their worldview isn't working for them, and they need to change something. Immersion is 'black supremacist/separatist', someone who rejects whites and views blacks as superior, and is seen as an initial solution to the problems that Encounter raises. Internalization is viewing both blacks and whites as equal, and wanting all diversity to be respected.

(Personally, I'd say those stages apply to me. Before I knew I was autistic, I viewed disability as interesting but never questioned the idea that it was a tragic thing. Then, I started thinking autistic was better than neurotypical, around the same time I was starting to realize I was autistic. At that time, also, I considered other disabilities to be totally different from autism. And then I started viewing all disabilities as part of diversity, and believing that no one kind of person was the best.)

So, here's the scale validity tests, first of all. (Note: Each statement is answered either yes or no, with a yes score counting for the relevant subscale.)

Preencounter

The Preencounter scale, with 9 questions, had a Cronbach's alpha of .578, which is poorer than my teacher's recommended .70, but similar to Ponterotto and Wise (1987)'s finding of .63 for the original scale.

I also analyzed correlations between the individual questions and the total score (p value of less than .05 is considered significant and italicized):
  • 'I believe that large numbers of disabled people are incompetent' - correlation .704, p < .001
  • 'I believe that nondisabled people look and/or express themselves better than disabled people' - correlation .513, p < .001
  • 'I feel very uncomfortable around disabled people' - correlation .191, p = .209
  • 'I believe that to be disabled is not necessarily good' - correlation .299, p = .046
  • 'I believe that disabled people should try to emulate nondisabled people and seek to cure or lessen their disability' - correlation .637, p < .001
  • 'I believe that disabled people are damaged in some way and are not meant to be' - correlation .692, p < .001
  • 'I feel guilty and/or anxious about some of the things I believe about disabled people' - correlation .403, p = .006
  • 'I feel that a disabled person's most effective way of solving problems is to become part of nondisabled society' - correlation .459, p = .002
  • 'I believe that nondisabled people are inherently more capable than disabled people' - correlation .645, p < .001
Furthermore, the two questions with the weakest correlations, 'uncomfortable around disabled people' and 'disabled not necessarily good' did not correlate with any other questions on this subscale, while the others all correlated with 1-6 other questions. Therefore, I decided to try dropping those two and recalculating the validity.

Preencounter2, with 7 questions, had a Cronbach's alpha of .714, which is above my teacher's cutoff. So, that's good.

Encounter

The Encounter scale, with 3 items, had a Cronbach's alpha of -.15, which is abysmally poor. In comparison, Ponterotto and Wise found an alpha of .37, which is still fairly poor.

Unsurprisingly, given the very small number of items, the correlation between each individual item and the total was high. However, none of the items showed any significant correlation with each other:
  • 'I feel unable to involve myself in nondisabled groups/activities, and am increasing my involvement in disabled groups/activities' - correlation -.226 with 'reading and thinking' and .05 with 'feeling guilty'
  • 'I find myself reading a lot of disabled literature and thinking about being disabled' - correlation -.226 with 'unable to involve' and .066 with 'feeling guilty'
  • 'I feel guilty and/or anxious about some of the things I believe about disabled people' - correlation .05 with 'unable to involve' and .066 with 'reading and thinking'
I decided to simply scrap this subscale. Two of the questions load on both Encounter and another scale, and both of those seem fairly well correlated with their respective scales. So those two should just be viewed as belonging to their scales.

Immersion

The Immersion subscale, with 7 items, had a Cronbach's alpha of .306, which is pretty poor. In contrast, Ponterotto and Wise found an aplha of .72 with the original version.

When I looked at individuals items and their correlation to the total, I found the following:
  • 'I feel unable to involve myself in nondisabled groups/activities, and am increasing my involvement in disabled groups/activities' - correlation .497, p < .001
  • 'I often find myself referring to nondisabled people in derogatory ways (eg 'stupid NTs')' - correlation .571, p < .001
  • 'I frequently confront the system and the people representing it' - correlation .605, p < .001
  • 'I believe that the world should be interpreted from a disabled person's perspective' - correlation .445, p = .002
  • 'I have changed my style of life to fit my beliefs about disabled people' - correlation .229, p = .125
  • 'I speak my mind regardless of the consequences, even fairly serious ones' - correlation .388, p = .008
  • 'I believe that everything about disabled people is good, and limit myself to disabled activities' - correlation .361, p = .014
The one uncorrelated item, 'changed life', actually correlated negatively with one other item on this scale ('derogatory towards nondisabled'). The weakest correlated item, 'limit self', did not correlate with any other item. So, I decided to try dropping those two items.

Immersion2 had a Cronbach's alpha of .417, still fairly poor. This scale probably needs to be either scrapped or redesigned, since it doesn't seem to apply very well to disabled people. It could be that too many questions involve activities that a disability might impact, such as being unable to involve oneself in nondisabled activities due to inaccessibility. Or, considering that 65% of my sample was autistic, it could be that they interpreted certain questions differently due to literalism or other cognitive differences. Or, perhaps this stage doesn't really occur in disability identity. Further study would be needed to determine this.

Internalization

The Internalization subscale, with 9 items, got a Cronbach's alpha of .695, a big contrast from Ponterotto and Wise's finding of .37. Rounded up, this would be exactly .70, so just barely meeting my teacher's cut-off.

For the individual items, I found:
  • 'I believe that being disabled is a positive experience' - correlation .748, p < .001
  • 'I know through experience what being disabled in our society means' - correlation .221, p = .160
  • 'I feel an overwhelming attachment to disabled people' - correlation .555, p < .001
  • 'I involve myself in causes that will help all oppressed people' - correlation .521, p < .001
  • 'I feel good about being disabled, but I do not limit myself to disabled activities' - correlation .717, p < .001
  • 'I feel excitement and joy when among disabled people' - correlation .641, p < .001
  • 'People, regardless of disability, have strengths and limitations' - correlation -.110, p = .487
  • 'I am determined to find my disabled identity' - correlation .522, p < .001
  • 'I believe that because I am disabled, I have many strengths' - correlation .708, p < .001
The question about people having strengths regardless of disability did not correlate with any other question (and only 1 of the 46 people answered 'no' to it). The question 'know what disabled means' was moderately (.379) correlated with 'strengths because of disability', but no other question. The remaining questions were correlated with 1-5 questions each (the one with only one correlation, 'attachment to disabled', was strongly correlated with 'excitement/joy with disabled'). So, I decided to drop the two uncorrelated questions.

Internalization2 had a Cronbach's alpha of .736, which is pretty good.

Average Scale Scores

I calculated scale scores by adding all the 'yes' items in a scale and then dividing by the number of items in the scale. So, 1 is the maximum and 0 is the minimum score on each scale.

For the original scales, I found the following:
  • Preencounter - mean .1827, SD .15741
  • Encounter - mean .4444, SD .24618
  • Immersion - mean .4441, SD .20254
  • Internalization - mean .7222, SD .22791
For my modified scales, I found:
  • Preencounter - mean .1094, SD .18889
  • Immersion - mean .5000, SD .26247
  • Internalization - mean .6739, SD .28676
Lastly, I took Preencounter2 - Internalization2 and divided the subjects into three subtypes based on the resulting score: Internalization (-.25 or less), Preencounter (.25 or more) and Unclassified (between -.25 and .25). The frequencies were 80% Internalization, 17% Unclassified and 2% (1 person) Preencounter. Which, by the way, means it would be a good idea to re-run this study in a group who have more Preencounter individuals.

1 Comments:

Blogger K said...

Hello, I'm a doctoral candidate currently working on my dissertation and a part of it is creating a disability identity scale. I was actually doing a literature search for scales when I came across your blog. It just so happens that I wanted to use Cross's racial identity theory to based my scale upon. The RIAS is based on Cross's theories and expanded upon by Helms. It looks like you're doing/have done the same thing. I'm wondering, is the study that you wrote about part of a thesis and has the scale been validated?

12:17 PM  

Post a Comment

<< Home