The Impact of League Tables on University Application Rates

In the UK, the popular use of league tables in assessing the comparative standing of universities has escalated to such an extent that most UK national newspapers now publish them. A review of literature to date suggests there has been little academic research undertaken to establish the value of such tables or to assess their impact upon student application rates. This paper explores the relationship between league position and application rates in the UK, using rankings from the Sunday Times as an example, with important lessons for higher education across the globe. Results show that in the period when league tables were newly introduced, there was a significant negative association between league table position (ranked from highest to lowest) and the change in application rates (that is, the higher the league position, the faster the growth in applications) an association that subsequently weakened and may have even have reversed. The paper goes on to explore the notion of various divisions within the overall league table, suggesting that there are effectively two 'selecting' divisions, typically choosing from a surplus of applicants, and two 'recruiting' divisions, typically working to avoid an applications shortfall against target numbers.     

Incorporating Uncertainty Into Parallel Analysis for Choosing the Number of Factors via Bayesian Methods

A number of psychometricians have suggested that parallel analysis (PA) tends to yield more accurate results in determining the number of factors in comparison with other statistical methods. Nevertheless, all too often PA can suggest an incorrect number of factors, particularly in statistically unfavorable conditions (e.g., small sample sizes and low factor loadings). Because of this, researchers have recommended using multiple methods to make judgments about the number of factors to extract. Implicit in this recommendation is that, when the number of factors is chosen based on PA, uncertainty nevertheless exists. We propose a Bayesian parallel analysis (B-PA) method to incorporate the uncertainty with decisions about the number of factors. B-PA yields a probability distribution for the various possible numbers of factors. We implement and compare B-PA with a frequentist approach, revised parallel analysis (R-PA), in the contexts of real and simulated data. Results show that B-PA provides relevant information regarding the uncertainty in determining the number of factors, particularly under conditions with small sample sizes, low factor loadings, and less distinguishable factors. Even if the indicated number of factors with the highest probability is incorrect, B-PA can show a sizable probability of retaining the correct number of factors. Interestingly, when the mode of the distribution of the probabilities associated with different numbers of factors was treated as the number of factors to retain, B-PA was somewhat more accurate than R-PA in a majority of the conditions.     

Loglinear analysis of cross-classified ordinal data: applications in developmental research

This article provides an introduction to loglinear analysis of cross-classification tables, including tables with nominal and ordinal variables. Loglinear models offer several advantages over the more commonly used chi-square test of independence, including the ability to analyze 3-, 4-, and higher-way interactions, the ability to determine whether the association between variables is linear or nonlinear, and the ability to interpret scale scores assigned to categories of an ordinal variable. After a review of the advantages of loglinear modeling, the chi-square test of independence is compared with the loglinear model of independence. This comparison serves to introduce the notation and terminology of loglinear modeling. The overall strategy of loglinear modeling is introduced next; then special loglinear models for ordinal data are reviewed. Each model discussed in the article is applied to data from the developmental literature.     

Time and response matching with topographically different responses

Four pigeons were exposed to several nonindependent concurrent variable-interval schedules of reinforcement. One schedule component required a keypecking response; the other component required a treadlepressing response. The birds matched the ratio of their behavior (as measured by responses and time) between the two topographically different responses to the ratio of reinforcement in those two components. When additional foods not contingent on a keypeck or treadle-press were then added, the birds matched time spent in the components to total rates of food delivered in those components; response matching was somewhat disrupted. The matching law, developed under concurrent variable-interval schedules requiring similar responses, can thus account for choice behavior involving topographically different responses.     

The development of a hypertext-based tutor for Geographical Information Systems

Experience gained in the construction of the world's first Geographical Information Systems tutor, ARCDEMO, emphasised the importance of accommodating different learning strategies. ARCDEMO, while highly successful, suffered from static graphical displays, a single pre-determined access path, and an overall design which restricted its scope in use. The Geographical Information System Tutor (GIST) was developed to address these problems using Apple's object-orientated HyperCard software. GIST incorporates a 'point-and-click' interface with graphical cues to initiate operations which include animated demonstrations, step-by-step illustrative graphics and graphical displays capable of user modification. The system was designed with two alternative access modes, serial or browsing , to allow different classes of user to work through the material in their preferred fashion. The use of GIST in undergraduate and short course teaching has shown how users respond well to the interface and find graphical explanations easy to follow.