NCU dissertations

Thought others might be curious. You can find Northcentral doctoral dissertations at

http://www.phddissertations.com/ .

They appear to be free of charge if you want to get pdf files in your mailbox. Otherwise, they charge for a copy on cd or hardcopy.

I wondered. Do NCU disserations make their way to UMI? Anyone know? I don't think that unaccredited schools do but I could be wrong.

 

Why not? NCU is regionally accredited. It is properly and legitimately accredited.
At present, UMI does not accept dissertations from unaccredited schools.

 

At UMI/ProQuest, I see 32 dissertations from NCU students in the last two years (the limit of one's ability to search).

The comment about unaccredited schools is a non sequitur for reasons Ike shared.

 

Rich,

Through our university library, I was able to search from 1997 forward and access 45 NCU dissertations dating back to 2003. What strikes me as somewhat unusual is a seemingly large number of very short dissertations. 18 of 45 NCU dissertations have less than 100 pages, and some of the most recent dissertations have as few as 47 pages. Of course, the length of a dissertation by itself is not a complete measure of substance. However, a large number of very short dissertations from one university in comparison to many others could possibly indicate a lower level of expectation and requirement from the school’s PhD candidates.

With this in mind, I briefly examined PhD dissertations from several random universities and grouped them by on-line, on-campus and mixed (on-line/on-campus) doctoral programs. Data were compiled surrounding the number of pages associated with 50 of the most recent dissertations published from each school. Smaller sample sizes were used for both Northcentral (45) and Touro (38), which represents all dissertations currently published by UMI/Proquest for these schools.

The following is a link to the data and general comparisons:

http://www.newknowledge.us/dissertation_compare/dissertation_comparisons.htm

Analysis was also performed to determine of what, if any, statistically significant differences might exist surrounding the average number of pages per dissertations, based on the school and respective PhD program participation type.

Analysis Overview:

GLM-univariate analysis variance was the primary method used to examine the between-subject factors, variances and subsequent main effects. Estimated marginal means were used for the multiple comparison tests. Post-hoc tests were not used because of the combination of multiple factors and unequal groups sample sizes. Levene’s test of equality of error variances was significant indicating violation of the homogeneity of variance assumption. Accordingly, log(x) transformations were performed on the dependant variable with successful results.

Univariate tests revealed that a statistically significant difference exist between the average number of dissertation pages among schools when all of the sampled institutions are included. ANOVA results show that the average number of pages for NCU’s dissertations is significantly less than essentially all other examined universities. Additionally, NCU represents the strongest underlying between-group effect and influence on the overall univariate results for both the school and type factors. Significant, but less frequent, differences also exist between Capella other schools. Occasional differences are also noted among several other schools. However, with NCU removed from the model, the remaining difference do not significantly influence the overall univariate test and therefore demonstrates that no overall significant difference exist among all remaining groups base on school or type of program. The latter is an important observation, because it offers initial evidence that, outside of NCU, there is little if any significant difference between the average length of dissertations (number of pages) among schools, regardless of whether the programs are delivered on-line, on-campus or by mixed methods.

Of course, this cursory examination does not comprehensively review a broad range of schools or offer an analysis of actual dissertation content; therefore its generalizability remains in question. Additionally, skeptics (I’m sure we have some here on DI) may raise objection sounding the usefulness or relevance of using the average page quantity as an indicator of dissertation substance or benchmark to examine levels of PhD program expectations. However, a thorough review of literature surrounding this subject and subsequent study may indeed provide some valuable knowledge.

Tom

 

I, too, noticed a number of short dissertations, but I wasn't compelled to analyze this effect.

I think page count can be an indicator, especially when mean page count is used. Any one or another dissertation might be short or long as the subject requires, but when we see piece after piece come in rather short, one has to wonder why. Why NCU's dissertations and not others?

 

but have you controlled for discipline?

It's possible that there is something significant here, but maybe not. One of the things that is not clear from your analysis is whether you compared dissertations from like disciplines. This is a critical factor in an analysis because difference in dissertation length varies greatly depending on the area of study. For example, it is not uncommon for dissertations in the humanities to be 300 pages or more; however, engineering dissertations are typically much shorter and often less than 100 pages.

 

Regardless of conclusions that are drawn, one thing is certain: An informal poll at DegreeInfo.com indicates that NCU is the most popular university for distance doctorates. Click here to see the poll.

 

Not necessarily. Because the poll does not involve sampling--because respondents choose whether or not they'll be part of the population--you can't really draw an inference about the population. It would be more accurate to say that respondents to the poll chose NCU more than any other school listed. We don't know what the non-respondents pursue (or will pursue), which might be entirely different.

 

LOL True!!!

 

Re: but have you controlled for discipline?

Good observation and point.

Data were collected from 50 of the most recent dissertations from each university, which based on the sample sizes should be a reasonable cross section of the published dissertations for each school. Since the purpose of my examinations was to compare the overall averages of all dissertations from each school, no effort was made to group the data directly by discipline or based on whether a dissertation was qualitative versus quantitative in nature. Therefore, additional research may indeed illuminate other possible confounding variables and perhaps the extent they may operate. Additionally, including data from a much larger number of schools would certainly add to the data’s value and generalizability of the results. However, for purposes of this cursory examination and intended scope, and based on the sample and effect sizes, I'm confident that the results as presented are reasonable and demonstrate something significant within the examined population.

Further to your point. The majority of NCU dissertations published by UMI/Proquest do appear to be social research, psychology or business based; few if any appear to be engineering or pure sciences related. Accordingly, I expect the addition of engineering or pure science dissertations to NCU’s sample would only further decrease their overall average page count and increase the observed effect. However, in my second model and analysis, I excluded NCU from the data, so this assumption is a mute point in terms of the subsequent results.

 

Re: but have you controlled for discipline?

Good point--experimental design studies in the sciences do tend to produce much shorter dissertations than, say, qualitative studies in in the humanities and social sciences. However, NCU offers doctorates in business administration, education and psychology, rather than engineering or other "hard sciences".

I am hesitant to use dissertation length as a sole measure of quality; however, the unusually high number of short NCU dissertations is notable and could be used (in conjunction with other data, such as research design, type and amount of statistical measures used, and length of supporting documents in the appendix) to generate some interesting data about the level of quality expected by a recipient of a doctorate from Northcentral.

Some years ago, I did an informal study of the over 1500 dissertations in the library of Brigham Young University. the shortest--just over 100 pages--was an experimental study for a Ph.D. in physics. The longest--over 800 pages--was a qualitative comprehensive history of an Arizona Community College for a Ph.D. in history.

Thank you, tesch, for providing such fascinating data.

 

Interesting discussion....

I have a general stats comment here.... How would Type II errors be applied in this situation? For example, there have not been many disseratations at NCU in comparison to other measured schools. Also, perhaps the sample size is so small, that it is truly not representative of NCU in, say, 10 years when there are many more disseratations put out by them.

Finally, maybe their model helps make their dissertations much more succint and relavant. Who really knows... I am not trying to defend them but wanted to state that using statistics to prove a point may just confound it.

peACE,

Ace

 

Tesch, you did very well in your analysis. Your descriptive purpose that is supported by the data selection, analysis and conclusions for the cursory examination seems reasonable to me. The observations made by others are recognized, but, in my view, unneeded because as you stated, it a cursory examination. I did learnt something new about NCU dissertations today. Thank you for taking time to perform the analysis. Thank you.

 

Rich: "It would be more accurate to say that respondents to the poll chose NCU more than any other school listed."



Cehi: This is appropriately stated. I agree with you. Thank you.

 

Anthony Pina: "I am hesitant to use dissertation length as a sole measure of quality; however, the unusually high number of short NCU dissertations is notable and could be used (in conjunction with other data, such as research design, type and amount of statistical measures used, and length of supporting documents in the appendix) to generate some interesting data about the level of quality expected by a recipient of a doctorate from Northcentral."


Cehi: I agree. Thank you for your comments. This relate to my point that we did learnt something new about NCU dissertation lengths from Tesch's cursory examination. This new information could be used to generate more new information about NCU itself, depending on who has the time to provide more of the new information. Thank you.

 

Aceman,

You offer some interesting points and comments. However, let me address what I read as your questions/concerns.

A Type II error would mean that we missed the effect (failed to detect a significant difference when one does indeed exist). However, that is not the case here because our results identified that significant differences do exist between NCU and each of the other schools included in the dataset. The confidence intervals and effect sizes also demonstrate an adequate sample size.

Additionally, the sample size (45) collected for NCU is not dramatically different from the design goal of 50 collected from the other universities or the smaller sample size collected from Touro (38). Moreover, the data collected from NCU and Touro represented all dissertations published by UMI/Proquest to date, which is basically a census versus a sample, and is therefore an exact representation of the current population of dissertations published for both schools. To say that the current population of NCU dissertations today may not be representative of the next ten years is a fair statement. However, the current data analyzed does accurately represent the reality of today; we will have to wait to see if things change in the future.

In terms of a model to make dissertations more succinct and relevant, that is the intended model for basically all dissertations. Relevance, clarity and succinctness are key qualities for dissertations that most committees tend to strongly enforce. Accordingly, much pruning, trimming and rewriting is often required as part of an iterative and sometimes painful process. Therefore, I struggle with the notion of a unique NCU model that accomplishes the same level of substance, content and perhaps quality in significantly fewer pages as what could be accomplished with significantly more pages by all of the other university examined. Instead, I would venture to say that the bar, in terms of expectation at NCU, is somewhat lower—at least for the dissertations published to date.

To your last comment: statistics were used to further examine what appeared to fairly consistent average page counts among the majority of universities (my primary interest) and evaluate what seemed to be a much lower page average at NCU. Therefore, I do not think that applying statistical methods to examine possible significant variances is confounding. I would also argue that using statistical tools offers much more reliable information than conjecture or other forms of largely anecdotal evidence.

Just my opinion….

Tom

 

I applaud your statistical analysis but please don't lose sight of the fact that quality <> quantity.

I can’t speak for the quality of each dissertation, but since this thread got so wrapped up in equating dissertation length to quality, it’s at least worth a mention.

 

I agree with this principle as it applies to any single, individual dissertation. But as a group? It begs the question: what is it about NCU's process that produces dissertations that, taken as a group, are remarkably shorter than those found in the same fields from other schools?

Perhaps they've found some new way of getting it done in less space. Or, perhaps, they don't require the same level of exactness and completeness. I don't know.

 

Oh no! Not another RA school with inferior dissertations! :-(

 

Perhaps an even more accurate statement is that “quality does not [always] equal quantity”. This is certainly the case when considering dissertations within their individual context. However, when based on an overall group average from one university in comparison to that of many others, one could plausibly argue that length, as a matter of overall substance, may indeed be one of several valid indicators of quality. At least it is reasonable to ask why such differences do or do not exist.

Rich Douglas said it best, “Any one or another dissertation might be short or long as the subject requires, but when we see piece after piece come in rather short, one has to wonder why. Why NCU's dissertations and not others?”