Factor analysis by generalized least squares

Aitken's generalized least squares (GLS) principle, with the inverse of the observed variance-covariance matrix as a weight matrix, is applied to estimate the factor analysis model in the exploratory (unrestricted) case. It is shown that the GLS estimates are seale free and asymptotically efficient. The estimates are computed by a rapidly converging Newton-Raphson procedure. A new technique is used to deal with Heywood cases effectively.The work on this project was done when the first author was Research Statistician at Educational Testing Service, Princeton, N. J. The second author was in part supported by a grant from the Research Committee of the University of Wisconsin Graduate School. The authors wish to thank Michael Browne for many helpful comments and Marielle van Thillo for valuable assistance in the numerical computations.     

INTERMITTENT REGRESSIONS IN A SERIAL AFTERIMAGE EXPERIMENT AS SIGNS OF SCHIZOPHRENIA

S mith G. J. W., R uuth E., F ranzén G. & S jöholm L. Intermittent regressions in a serial afterimage experiment as signs of schizophrenia. Scand. J. Psychol ., 1972 , 13 , 27 33.—A group of 35 young patients suspected of schizophrenia were tested in a serial afterimage (AI) experiment and were also rated on the Rockland-Pollin scale. Subjects with high psychotic scores showed one or more regressions to immature (size-constant, positive, physiognomic) images in their Al serials, images typical of children who cannot clearly distinguish their Ails as subjective phenomena distinct from the outside world. The Al test thus indicates schizophrenics to be characterized by intermittent withdrawal into a world of experiencing the self and non-self as not clearly distinguished. Regressions were rare in two control groups.     

Search processes for detecting repeated items In a visual display

Evidence for successive serial exhaustive searches was found in a task requiring detection of repeated letters in visually presented lists of varying length. The time per comparison was 25 msec. Important differences in processing appeared between instances of adjacent repeated letters and instances of nonadjacent repeated letters.     

Orientation effects on line afterimages

Lines of various orientations were viewed as prolonged afterimages in six experiments. The duration of unitary appearance was not influenced by line orientation for monocular afterimages, but there was a marginal effect for afterimages generated binocularly; vertical and horizontal lines tended to be visible for longer than did 45-deg lines. Measures of fragmentation frequency and the latency to the first disappearance did not vary reliably with orientation under any conditions. Binocular afterimages lasted longer than did monocular ones, but generally showed the same pattern of fragmentations. These results are compared with those from experiments using optical stabilization and steady fixation, in which orientation differences have been reported.     

Averaging: An empirical validity criterion for magnitude estimation

Ss judged grayness of neutral value Munsell chips under two response conditions: magnitude estimation and graphic rating. In addition, they judged average grayness of pairs of chips under the same two response modes. The averaging data were evaluated in terms of a simple model for subjective averaging. The graphic rating data fitted the model, but the magnitude estimates showed consistent discrepancy. It was concluded that the Ss were averaging, but that magnitude estimation distorted their judgments.     

Transfer of responses to open and closed shapes in discrimination learning by cats

Two sets of 32 cats each were matched for performance in learning to discriminate an open and a closed shape and were then trained on a transfer task. The consistent groups were rewarded for choosing the same class of shape (open or closed) in learning and transfer. The inconsistent groups were rewarded for choosing the open figure in one task and the closed figure in the second, or vice versa. After learning the transfer task, all of the Ss relearned the original discrimination task under the same conditions that prevailed in initial training. The inconsistent groups made more errors on both the transfer and retention problems; both these differences were significant at the 0.1% level of confidence. These results and those obtained on preference tests support the hypothesis that cats classify shapes as open or closed in terms of perimeter and number of sides.