A response to Janks Morton

Jank Morton’s response to my op-ed in the Baltimore Sun:

I reviewed your article, and I am at a point of concern as it relates to the symbiosis of the Justice Policy Institute and “Journalists” who continue to source this organization for this flawed comparison. I believe this point needs absolute clarification, and while you imply the truth lies between myself and the JPI, and you so diligently (and accurately) explain)the census data and its potential shortcomings, why did you not make one phone call to the NCES regarding the accuracy of the JPI data? If you would, please contact the NCES (202-502-7300) regarding the data for FY 2000 and you will find that NCES did not collect unduplicated 12 month headcount enrollment for African American males until FY 2001. The 603,032 number used by this organization, in its fraudulent inception was a mid-year forecast for fall enrollment ONLY (actual year end fall enrollment was 657,000) and their “study” did not include spring enrollment for 2000. And as we so readily refer constantly back to the JPI as “the source” for the cellblocks vs. classrooms data, and easily as my research and credibility is refuted by this source, why is that no one has “uncovered this story”. And your final assumption falls directly in line with fallacious postulate forwarded by this organization who massaged the data to justify an ends.

Finally your correct analysis of “the less representative the coverage of the survey, the less sure one can be of the accuracy of the estimated number. And – surprise – the Current Population Survey’s lowest coverage rate is among young black men” only helps support the argument that while the census coverage may be lacking as it relates to AA Males (a popular premise that I have not seen proved or disproved by any ones research) , the count still surpasses that which is incarcerated. So if they gather additional coverage is it not logical to assume the count would increase more? Please stop dumbing us down, do more fair and accurate research, before publishing and assuming the worst about us. – Janks Morton

My Response (sent to J. Morton):

My op-ed was a contribution to a discussion that Mr. Morton stated he hoped would be sparked by his film. Hopefully that discussion can continue without those with differing views slighting each other. I want to be absolutely clear that my intention of the op-ed was not to discredit Mr. Morton or his film. In fact, in the piece I mention that I am in agreement with him on two points: that there are more college-age black men in college than in jail and prison and that this age group comparison is the better one to make if one is to do so at all. But the bigger issue for me is not about the comparison between incarcerated black men vs. those in college. Rather, I am concerned about the overwhelming evidence that there are far too many black men incarcerated and the fact that they are so disproportionately represented behind bars. My analysis of the jail vs. college debate was to help us gain clarity on this issue in an effort to move on to the bigger questions.

There are currently over 830,000 black men in jail. Ultimately, whether that number is greater or less than the number in college doesn’t change this disturbing fact. Furthermore, I don’t consider stating this to be a problem which demands a solution to be assuming the worst of us or contributing to the negative image of black men. Nor does it assume that all black men are making no progress. Black men are making some progress; the increasing trend of black men attending college over the years is one example. It is also true that we are lagging in a number of areas—the incarceration rate is clearly one of them.

Since Mr. Morton expressed some concerns about the accuracy of what I presented I will lay out my rationale below and the facts behind them. However, as far as I am concerned you need not read further to get my point. Also, precisely for the reasons mentioned above, I do not find it productive to engage in a protracted debate about whether there are more black men in college than jail. But I do hope discussions about where we are, how we got here, where we should be headed, and how we should get there will continue.

Background on the two datasets

The National Center for Educational Statistics

In the article, I stated that the National Center of Educational Statistics (NCES) number is a more accurate number of black men in college. Here is some background of the dataset that may help further explain my reasoning. The methodology for the NCES number is pretty straightforward. It’s a headcount of all students in postsecondary schools from NCES’s Integrated Postsecondary Education Database (IPEDS). All schools that can disburse federal funds for financial aid (called Title IV schools) are mandated to complete a survey that includes questions about their student body. As a result, there is a very high response rate for this survey. For example, for fall 2005 enrollment, 99.9% of schools responded. There are some students from non-Title IV schools who are not included, but you can bet that number doesn’t change things much. No method of determining enrollment is perfect, but my position is that this approach is likely to be closest to capturing the real number.

The Current Population Survey (sometimes referred to as the “census”)

The reported Current Population Survey (CPS) number represents all Title IV and non-Title IV schools. However, the data collection method is very different from NCES. The CPS involves sampling the population and is more complicated in terms of how the number is derived. An important note is that this number is an estimate not a headcount like the NCES number. This means that it is based on the sample collected and then there is some number crunching involved to arrive at the 815,000 value for 2000 (and the 2005 estimate of 864,000). In basic terms, the estimate is determined by adjusting the sample number according to what is known about the entire population from which it was drawn.

When you have an estimated number like this that contains some sampling error, it is good practice to calculate a confidence interval if possible. The confidence interval takes into consideration sampling error and what it tells you is the probability of the actual number falling within a range of scores. I calculated this number for 2000 and found the 90th percentile confidence interval to be somewhere between 744,719 and 885,280. This means that, taking into consideration some error in calculating the estimated value, there is about a 90% chance that the real number falls somewhere within this range. As you can see, that’s a pretty wide range. If you are interested in seeing the details of how I calculated this you can email me.

Responses to Mr. Morton’s Comments

Below, I address the specific issues raised by Mr. Morton about my op-ed in a Q&A format:

Q: How do you know that the lowest coverage rate of the Current Population Survey is for black men? According to Mr. Morton there is no research demonstrating this.

A: I assure you that I did not make this up. The Census Bureau itself provides this information about the Current Population Survey. For example, in 2004, the lowest coverage rate among all combinations of age and sex groups was for black men aged 20-24 where the coverage was 61%. So for this group, undercoverage was 39%. The overall estimated undercoverage rate of the survey was 11%. But don’t take my word for it–see for yourself (table 1): http://www.census.gov/population/www/socdemo/school/2004SchoolS&A.pdf

Q: Doesn’t the fact that there was a low coverage rate for black men mean that the Current Population Survey number is an underestimate of the real number of black men in college? Or put another way, doesn’t it mean that if coverage were greater, the estimated number of black men in college would be higher?

A: Not necessarily. It all depends on who is in the sample. Think about it this way: if the sample contains a disproportionate number of college-going men in the sample in comparison to the real proportion of this group in the nation, there is the possibility of the estimated number actually being too high (an overestimate). On the other hand, it could be an underestimate if non-college-going men were disproportionately sampled. The point is, since we don’t know the “real” total number of black men in college we don’t know if we’ve oversampled or undersampled. That’s one of the reasons confidence intervals are useful—to give us a range of numbers that account for what is referred to as “error” in the sampling. But of course, if you can just count everyone you don’t have this problem.

Q: How about Mr. Morton’s comment regarding 12-month unduplicated data not being collected for black males from NCES until 2001 and the number you presented only representing fall enrollment?

 

A: I’m not entirely clear on what his argument is here, but it is true that the NCES number I mentioned is a count for the fall (and that they did not collect 12-month unduplicated data until 2001), but so is the CPS estimate Mr. Morton places more faith in (both surveys are conducted in around October). So they are comparable in that regard—they both represent the enrollment for about half the year (that is, if you think of the school year divided into 2 semesters). Likewise, the BJS jail and prison number that the NCES and CPS enrollment numbers are compared to represents one time-point at the middle of the calendar year (June)—not an entire calendar year. In terms of time frame, all of these numbers are roughly comparable. Furthermore, if one were to take the 2000 NCES number Mr. Morton claims to be more accurate (657,000) it’s still lower than the total number of black men in jail and prison in 2000 according to BJS (791,600).

Calculating the confidence interval for 2000 Current Population Survey (CPS)

I calculated the confidence interval based on the equation on the Census Bureau’s website found here. This is the equation:

for02.gifIn the equation, “s” equals the standard error, “b” is a parameter provide by CPS, “T” is the total number of college-aged black men in the country between ages of 16 and 24 (11,529,241, I got this number from the 2000 decennial census), and “x” is the estimated value (815,000). Then the lower limit of the confidence interval is x – (1.645*s) and the upper limit is x + (1.645*s). So after doing the calculations based on these numbers, the 90th percentile confidence interval is between 744,719 and 885,280. Again, this means that, taking into consideration some error in calculating the estimated value, there is about a 90% probability that the real number falls somewhere between these two numbers.

Advertisements

4 Responses to “A response to Janks Morton”


  1. 1 SL November 14, 2007 at 12:32 am

    While I find Mr. Morton’s attempt to challenge myths and stereotypes about African American men praise worthy, I found some aspects of his documentary seriously disturbing. In particular the implication, however subtle, that black women are responsible for perpetuating negative stereotypes about black men.
    I understand that the documentary is about “what black men think,” but is there any way to empower black men without devaluing the important contributions of black women to black men and the community in general?

  2. 2 JJ December 17, 2007 at 1:18 pm

    “As a black man, I am in total agreement with SL. Janks not only devalues black women, but black men as well–esp. if they think differently from him. As indicated in his remarks to Mike’s op-ed in the Baltimore Sun, he ends it with: “Please stop dumbing us down, do more fair and accurate research, before assuming the worst in us.”

    His recent comment to Mike’s blog article “Tech Check: Highschool Graduation Rate” he said, “I can’t seem to find 50% or 75% or anything in between.”

    Maybe Janks himself might consider ‘more fair and accurate research?’ Maybe stop ‘assuming the worst in us?’ One thing for sure, from his recent comment, Janks needs to learn how to determine dropout/graduation rates. With only the enrollment chart he used, God couldn’t find ‘50% or 75% or anything in between’, either. Now, is Janks being fair and accurate here?

    Finally, Janks might consider the effects of ill-willed comments, which often boomerang on the person who sends them out. For instance, “Please stop dumbing us down” might be seen as Janks “dumbing himself down.”

    Nothing good ever comes by putting someone else down, unless you possess the shock jock skills of a Jason Whitlock. And I only know of one Jason Whitlock.

  3. 3 KM January 7, 2008 at 11:50 pm

    It’s fairly easy to get on here and assail Morton’s documentary. Won’t cramp my fingers up doing that. I agree with the author’s (Dr. M. Strambler) response to the far more important underlying issues that are driving this debate: in short, there are more college-age black men in college than in jail and prison and this age group comparison is the better one to make if one is to do so at all. The more important issue however is not about the comparison between incarcerated black men vs. those in college. Rather, it is the overwhelming evidence that there are far too many black men incarcerated and the fact that they are so disproportionately represented behind bars. Whether are slightly more or less than 830,000 AA males in jail and the exact number (doesnt that change each day?) is not the issue. There is a bigger issue here that Morton is simply trying to avert.

    As the creator of a rather biased and slanted documentary, you should be open to more constructive debate and opposing view as they are presented to you.

  4. 4 lawn mower repair edmonton July 27, 2013 at 7:15 pm

    This paragraph will help the internet users for setting up new weblog or
    even a weblog from start to end.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s





%d bloggers like this: