Response Rate Data Research Articles

50% Response rates: half-empty, or half-full?

When the Black Women's Health Study, a prospective cohort of over 59,000 women who have been followed since 1995, invited all of its participants to provide a DNA sample for future research, only 51% of those participants agreed to do so. Responders were significantly older and more health conscious than non-responders. The Black Women's Health Study is a unique resource, but this low level of response and its resulting self-selection bias are now the norm in contemporary epidemiologic, and especially cohort, studies. Epidemiology desperately needs new approaches that work better and cost less. The literature on predictors of response focuses too narrowly on participant characteristics and does not identify any clear steps studies can take to increase participation. To improve research quality, cost-efficiency, and long-term sustainability of studies, epidemiology can and should approach, analyze, and leverage response-rate data more creatively and extensively than most studies have done to date.

Bouts, Changeovers, and Units of Operant Behavior

Responding of various types occurs as bouts of activity separated by pauses. Deprivation and reinforcement affect the tendency to initiate and persist engaging in bouts but have little effect on performance once engaged. Such a two-state conception encourages time-based measurement even when the response is recorded as a brief, discrete event like a lever press. The two-state conception also emphasizes that response rate, as commonly calculated, is a mixture of different units controlled by different sets of variables. Focusing on the components of the two states--especially bout-initiations (or changeovers)--might help solve some puzzles presented by total response-rate data and might help clarify how schedules of reinforcement affect response rate.

Falsification of matching theory: changes in the asymptote of Herrnstein's hyperbola as a function of water deprivation.

Five rats pressed levers on variable-interval schedules of water reinforcement at various levels of water deprivation. In one phase of the experiment, three deprivation conditions that replicated conditions in Heyman and Monaghan (1987) were arranged, along with three less extreme deprivation conditions. In a second phase, water deprivation was arranged so that subjects were exposed to a greater range of access to water per day. Herrnstein's hyperbola described the rats' response-rate data well. The y asymptote, k, of the hyperbola appeared roughly constant over the conditions that replicated those of Heyman and Monaghan, but decreased markedly when less extreme deprivation conditions were included. In addition, k varied systematically when the second method of arranging deprivation was used. These results falsify a strong form of matching theory and confirm predictions made by linear system theory.

CONFIRMATION OF LINEAR SYSTEM THEORY PREDICTION: CHANGES IN HERRNSTEIN'S k AS A FUNCTION OF CHANGES IN REINFORCER MAGNITUDE

Eight human subjects pressed a lever on a range of variable-interval schedules for 0.25 cent to 35.0 cent per reinforcement. Herrnstein's hyperbola described seven of the eight subjects' response-rate data well. For all subjects, the y-asymptote of the hyperbola increased with increasing reinforcer magnitude and its reciprocal was a linear function of the reciprocal of reinforcer magnitude. These results confirm predictions made by linear system theory; they contradict formal properties of Herrnstein's account and of six other mathematical accounts of single-alternative responding.

MULTIPLE SCHEDULE COMPONENT DURATION: A RE‐ANALYSIS OF SHIMP AND WHEATLEY (1971) AND TODOROV (1972)

The tendency for relative response rate to approach matching as multiple schedule component duration decreases has been interpreted as confirming a prediction of Herrnstein's multiple schedule equation. However, the equation predicts that absolute response rate will decrease in both multiple schedule components as component duration decreases. The absolute response-rate data of two studies of component duration do not support this prediction; absolute rate either increased or remained relatively constant.

PERFORMANCE ON VARIABLE‐INTERVAL SCHEDULES ARRANGED SINGLY AND CONCURRENTLY1

Extensive parametric data were obtained from pigeons responding on variable-interval schedules arranged on three, two, and one response keys. Number of responses on the keys, the time spent responding on the keys, and the number of reinforcements obtained on the keys were measured. Response rates on each key were an increasing function of the reinforcement rate on that key, and an inverse function of the reinforcement rate on the other keys. In terms of preference, both response and time-allocation ratios undermatched ratios of obtained reinforcements, and the degree of undermatching was consistent both within, and between, two- and three-schedule data. When absolute response-rate data were analyzed according to Herrnstein's (1970) quantitative account, obtained values of assumed constants were not consistent either within or between conditions. However, a power-function modification of Herrnstein's account fitted the data well and provided similar exponent values to those obtained for the undermatching of preference ratios.