Planned contrasts

Planned contrasts were performed to test the hypotheses that depressed people would be faster with respect to negative words than nonnegative words on the valence identification task, but slower on the lexical decision task, and that these differences would be greater than those for nondepressed people. Depression was operationalized as having a relatively high level of dysphoria within the past week, as well as at the mass testing session. That is, the depressed group was formed by people scoring over 15 at the initial mass testing session, and over 16 at the time of the experiment on the BDI. Similarly, the nondepressed group was formed by people scoring below (<6) at the mass testing session, and between 1 and 6 at the actual testing session on the BDI. The 150ms condition was used for all contrasts as this condition was hypothesized to be fast enough to enforce rumination on a word's affective valence but slow enough to be perceived without a great deal of extraneous noise. Alpha was controlled using a bonferroni adjustment for each family of contrasts.

The first set of contrasts examined hypothesized differences between the performance of depressed and nondepressed people. Alpha was controlled at 0.013 for these contrasts using a Bonferroni correction. Planned contrasts on the 150 ms condition reveal that depressed people were reliably slightly slower on negative words than on neutral words with respect to the performance of nondepressed people on the lexical decision task, D= 7.5 ms, F(1,74)=7.37, p=.008, R²=.09, observed power=.76. The performance of depressed people on positive v. neutral words was not statistically significantly different from that of nondepressed people, D= 5.3 ms, F(1,74)=2.00, p=.162, R²=.03, observed power=.286. On the valence identification task, the difference in reaction times to negative and neutral words was not statistically significant for depressed and nondepressed people, D=2 ms, F(1,73)=.14, p=.714, R²=.002, observed power=.051. Yet the difference between depressed and nondepressed people on positive vs. negative words was statistically significant for this condition, D=17 ms, F(1,73)=6.45, p=.013, R²=.08, observed power=.701.

The second set of contrasts involved expected differences within the depressed group. Alpha was again controlled at 0.013 using a Bonferroni adjustment. Depressed people were slightly but not statistically significantly slower on negative words than neutral words on the lexical decision task, D= 2.4ms, t(29)=-1.6, p=.12. There was also no statistically significant difference with respect to their performance on positive vs. neutral words D= 2.6ms, t(29)=-.91, p=.37. On the valence identification task, depressed people were statistically significantly faster with respect to negative than positive words, D=11.7 ms, t(29)=3.19, p=.003. Depressed people were statistically significantly faster with respect to negative than neutral words D=14.3 ms, t(29)=5.5, p<.001.

These data are depicted numerically in Table 2, p. 63 and graphically in Figure's 3, p. 64, and 4, p. 65, which show the mean reaction times for depressed and nondepressed people to each type of word for the valence identification and lexical decision tasks. The figure shows that the depressed group was delayed in responding to positive words, with respect to negative and neutral words, on the Valence-Id task, as predicted by the integrative theory. The depressed group was delayed in responding to negative words with respect to positive words on the lexical decision task by a small amount. These effects were not observed for the nondepressed group. The table and figure include reaction times to nonwords on the lexical decision task which generally appeared to be approximately one standard deviation larger than reaction times to words of any valence, for either group.

 

Figure  3.: Reaction times for lexical decision task, 150ms condition

 

Figure  4.: Reaction times for valence identification task, 150ms condition

 

Table: Reaction Time Data for the 150ms Condition

			NonDepressed		N	Depressed
Task	Duration	Valence	Mean	St Dev	N	Mean	St Dev	N
Lexical Decision	150	Positive	.48	.17	46	.51	.32	30
Lexical Decision	150	Negative	.46	.18	46	.54	.29	30
Lexical Decision	150	Neutral	.50	.15	46	.51	.25	30
Lexical Decision	150	Nonword	.64	.24	46	.68	.32	30
Valence Identification	150	Positive	.68	.25	45	.83	.31	30
Valence Identification	150	Negative	.66	.27	45	.66	.24	30
Valence Identification	150	Neutral	.84	.34	45	.86	.29	30