32
Joachim Funke
material, procedure, hypotheses, result, and a short discussion. Then, in a
next section, a general discussion selects interesting results and connects
them with results from other studies.
3.1 Experiment 1: Active Intervention vs. Pure Observation
Independent and dependent variables. In this first experiment (for more
details see Funke & Miiller, 1988) learning by active interventions was
compared to learning by pure observation of the system’s development (Fac-
tor 1: intervention vs. observation, I vs. O). In addition, the effect of a
diagnostic tool (subjects had to predict the next system’s state) was compared
to a nonprediction condition (Factor 2: prediction vs. no prediction, P vs.
NP). Dependent variables were QSC and QSL
Subjects, material and procedure. Subjects were 32 college students. In
each of the four conditions eight subjects were run individually. This allows
for detection of “large effects” (f = 0.40 in Cohen’s meaning of the word,
1977) with α = 0.10 and β = 0.30 for main effects. In the I- and О-condition
experimental twins were used. Each subject in the О-condition observed the
system data which another subject (the twin) under the !-condition had
produced (yoked-control design).
The system used was SINUS with parameters a=l,b = 0, c = 0.2, and
d = 0.9 in Eq. (2), (3) and (4). The system had to be manipulated during five
blocks of seven trials each. During the first four blocks subjects could freely
explore the system. During the fifth block all subjects (both the I- and the
О-group) had to reach and maintain a previously specified goal state. The
amount of system knowledge subjects had acquired (QSI, as measured by the
“causal diagram” at the end of exploration) and the control quality (QSC, as
measured via the distance of the actual to the specified goal states) served as
dependent variables.
Results. Funke and Miiller expected the І-group to be superior to the
“observers” with regard to amount of knowledge as well as to control quality.
Also, the “predictors” should accumulate more knowledge than the “non-pre-
dictors.” Path-analytical evaluation of the data supported the expectations
partially: The !-group was indeed better in controlling the system (significant
standardized path coefficient β = 0.42* from I to QSC), but seemed to know
less than the “observers” (β = -0.30* from I to QSI). “Predictors” had more
knowledge than “non-predictors” (mean QSI: 1.02 vs. 0.57, F(i,28) = 5.5O*).
Knowledge about the system was generally a good predictor of control
performance (β = 0.41* from QSI to QSC). Interestingly, there was a nega-
tive relationship between the time spent on the task and the quality of
performance.
Discussion. The results demonstrate the effects of task manipulations.
Active interventions allow better system control. However, this effect is not
accompanied by an increase in “extemalizable” knowledge. Similar dissocia-
tions have been reported by Broadbent, FitzGerald, and Broadbent (1986),