Differential response latencies

As with studies of depressed people, a great deal of between study variability existed in the studies using induced depression. Given the low number of studies investigating this topic, and the extremely different observed means and standard deviations, it is difficult to interpret aggregate effect sizes. For example, Challis and Krane (1988) find a mean difference of -116ms between negative and neutral studies while Niedenthal and Setterlund (1994) find an average difference of 27.42, which is relatively close to their obtained effect size for both the left eye (37.53ms) and right eye (16.34ms), given the much larger magnitude of Challis and Krane's results. Similarly the obtained variation of the difference in reaction times is an order of magnitude larger for Challis and Krane's study than Niedenthal and Setterlund's. Differences in the number of stimuli of each type used and type of mood induction procedure (Challis and Krane use a Velten procedure and Niedenthal and Setterlund use a musical mood induction), among other factors, may have contributed to these differences. As may be expected, for the contrast involving reaction times to negative vs. neutral words, the two obtained effect sizes are statistically significantly heterogeneous, Q=6.06, df=1, p<0.05. Yet, for positive vs. neutral words, the obtained effect sizes are more uniform, Q=.97, df=1, p>0.05, and similarly for positive vs. negative words, Q=.91, df=2, p>0.05.

Because of the oppositely signed effect sizes, aggregate differences in reaction times appeared small, and near zero. For the negative vs. neutral condition, D.=4.16ms, V.= 204.07, D._lb=-23.04, D._ub=32.16, d.=-.15ms, v.=.07, d._lb=-.68, d._ub=.39. For the positive vs. neutral condition, results were more uniform, and thus a statistically significant facilitation on positive words, like that observed with nondepressed people, was observed, D.=-36.66ms, V.= 344.10, D._lb=-73.10, D._ub=-0.30, d.=-.47, v.=.17, d._lb=-1.27, d._ub=.32. Differently from other conditions, positive words were facilitated statistically significantly more than negative words, primarily due to the pooled variance, D.=-38.74ms, V.= 286.64, D._lb=-71.92, D._ub=-5.55. The difference is not statistically significant when standardized effect sizes are examined, d.=-.31, v.=.05, d._lb=-.76, d._ub=.13.