DVs in attention research

1
DVs in attention research

• Two most important are RT and accuracy
• Speed-Accuracy trade-off usually, the faster
  the response, the less accurate it is
• Slower responses tend to be more accurate
• Often, look at both to see whats going on.

2
Theios (1975)

• Subject named a visually presented digit
• DVs were RT and accuracy
• IV was stimulus probability (.2 to .8)
• Eg, 6 occurs only 20 of the time, but 7
  occurs 40 and 9 occurs 80
• Theios found no significant differences in RT
• But failed to take into account error rates on
  the simple task of naming digits.

3
FIG Kanto8e 8-7 Theios (1975)
4
Interactions Lorig et al (1991)

• DV was event related potentials (ERPs)
• High tone on 75 of trials (frequent)
• Low tone on 25 of trials (rare)
• Observers counted low ignored high tones
• Also varied musk 0, 20, (unaware) and 88
  (aware).

5
FIG Kanto8e 8-12 Lorig et al
6
FIG 2X3 example 1
7
MEs and INTs in above Figures

• 2x3 e1 ME of dos. context, no INT (On the
  average, percent correct recall (PCR) increased
  with increasing drug dosage. On the average, PCR
  was higher for good than poor context. However,
  the effect of varying context on performance was
  identical at each level of dosage.)

8
FIG 2X3 example 2
9
MEs and INTs in above Figures

• 2x3 e2 ME of dos. context INT (On the
  average, PCR increased with increasing dosage. On
  the average, PCR was higher for good than poor
  context. Varying context had no effect on PCR
  when 0 mg of the drug was taken, but had an
  increasingly differential effect on PCR as the
  dosage increased.)

10
FIG 2X3 example 3
11
MEs and INTs in above Figures

• 2x3 e3 ME of dos. context INT (On the
  average, PCR was different when taking the drug
  than when not taking the drug Compared to the
  no-drug condition, PCR was higher when taking 100
  mg and lower when taking 200 mg. On the average,
  PCR was slightly higher for poor than good
  context. There was a slight advantage in PCR when
  subjects were given poor context in all but the
  200 mg dosage condition.)

12
FIG 3X3 example 1
13
MEs and INTs in above Figures

• 3x3 e1 ME of dos. context INT (On the
  average, PCR increased with increasing dosage. On
  the average, PCR increased with improved context
  for remembering. The advantage of medium context
  over poor context was identical at all three
  dosages, however the advantage of good context
  over both of the other levels of context differed
  across the levels of drug dosage. Specifically,
  the advantage of having good context was greater
  when taking the drug than when not taking the
  drug.)

14
FIG 3X3 example 2
15
MEs and INTs in above Figures

• 3x3 e2 ME of dos. context, no INT (On the
  average, PCR increased with increasing dosage. On
  the average, PCR was identical with poor or
  medium context, but was higher with good context.
  However, the advantage of good context over the
  other two levels of context was identical at each
  level of drug dosage.)

16
FIG 3X3 example 3
17
MEs and INTs in above Figures

• 3x3 e3 ME of dos. context INT (On the
  average, PCR increased with increasing dosage. On
  the average, PCR was identical with poor or
  medium context, but was higher with good context.
  The advantage of good context over the other two
  levels of context was identical when the drug was
  taken, but was greater when no drug was taken.)

18
FIG 3X3 example 4
19
MEs and INTs in above Figures

• 3x3 e4 ME of dos. context INT (On the
  average, PCR increased with increasing dosage. On
  the average, PCR was lowest for medium context,
  slightly higher for poor context, and highest for
  good context. For the placebo group, medium
  context led to slightly higher PCR than poor
  context, and good context led to much better PCR
  than the other two levels of context. This
  advantage of good context over the other two
  levels of context was reduced for the 100 mg
  group, and there was no difference in PCR between
  poor and medium context. For the 200 mg group,
  there was an advantage in PCR for good context
  over poor context, and an even greater advantage
  over medium context.)

20
FIG 3X3 example 5
21
MEs and INTs in above Figures

• 3x3 e5 no ME of dos or context, but INT (On the
  average, PCR was the same across all three drug
  groups, and was the same for all three levels of
  context. However, the lack of two significant
  main effects was obscured by the presence of a
  complex cross-over interaction, in which
  increasing the drug dosage improved PCR for those
  with poor context, harmed PCR for those with
  medium context, and had no effect on PCR for
  those with good context.)