Review and Discussion of Children's Conceptions of Computers

Author(s): Michael T. Rücker and Niels Pinkwart

Source: Journal of Science Education and Technology , APRIL 2016, Vol. 25, No. 2 (APRIL
2016), pp. 274-283
Published by: Springer

Stable URL: https://www.jstor.org/stable/43867796

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

Springer is collaborating with JSTOR to digitize, preserve and extend access to Journal of
Science Education and Technology

This content downloaded from

5.179.1.3 on Fri, 05 Apr 2024 06:53:40 +00:00
All use subject to https://about.jstor.org/terms
 J Sci Educ Technol (2016) 25:274-283 /&'
DOI 10.1 007/s 1 0956-0 1 5-9592-2 CrossMark

Review and Discussion of Children's Co

Michael T. Rücker1® • Niels Pinkwart1

Published online: 8 December 2015

© Springer Science+Business Media New York 2015

Abstract Today's children grow up surrounded by these devices, which naturally influences their interactions
computers. They observe them, interact with them and, as a and learning in contemporary computerised classrooms.
consequence, start forming conceptions of how they work In the context of computer science education, Pea
and what they can do. Any constructivist approach to (1986) was among the first to identify various program-
learning requires that we gain an understanding of such ming misconceptions, and he proposed the idea of a "su-
preconceived ideas and beliefs in order to use computers as perbug," a more fundamental misconception regarding the
learning tools in an effective and informed manner. In this general capabilities of the machine, which he identified as
paper, we present five such conceptions that children the root cause for many of the problems plaguing his
reportedly form about computers, based on an interdisci- novice programmers. More recently, Sorva (2013) has
plinary literature review. We then evaluate how persistent conducted an extensive review of student misconceptions
these conceptions appear to be over time and in light of in programming and also addresses the issue of conceptions
new technological developments. Finally, we discuss the related to the underlying machine that actually executes the
relevance and implications of our findings for education in code. Computers have long since moved beyond the CS
the contexts of conceptual pluralism and conceptual classroom and have become essential tools for teaching and
categorisation. learning in numerous other subjects. In the context of a
constructivist approach to learning, it is clear that using
Keywords Literature review • Conceptions • these tools in an effective and informed manner requires
Constructivism • Computational devices • Computer that we gain an understanding of students' preconceived
science education • Technology education ideas and beliefs about them.
However, it is a very challenging task to investigate
such fundamental conceptions of computers, given the
Introduction rapid change and development of new technologies and
modes of interaction. The ways we interact with computers
Today's children grow up surrounded by computers. They today have almost nothing in common with those of 60, 40
observe them, interact with them and, as a consequence,or even 20 years ago. We are not dealing with conceptions
regarding relatively stable theoretical constructs, but con-
start forming conceptions of how they work and what their
basic capabilities are. Thus, children will arrive in our
stantly evolving human-made artefacts. Thus, it is not even
classrooms with more or less sophisticated ideas about clear whether any such conception can indeed persist for a
longer period of time.
In this paper, we present a literature review to assess the
current state of research on children's conceptions of
E3 Michael T. Rücker
computers. In "Terminology and Method" we briefly dis-
rueckerm@informatik.hu-berlin.de
cuss the terminology and methodology of our review pro-
1 Computer Science Department, Humboldt-Universität zu cess. In "Children's Conceptions of Computers", we
Berlin, Unter den Linden 6, 10099 Berlin, Germany present five distinct conceptions that we were able to

<0 Springer

This content downloaded from

5.179.1.3 on Fri, 05 Apr 2024 06:53:40 +00:00
All use subject to https://about.jstor.org/terms
 J Sci Educ Technol (2016) 25:274-283 275

identify based on different

our levels of abstraction but
review. In are nevertheless
"Discussion"often
discuss how interdependent.
persistent these Hence, research on children's mental
conceptions appear t
is, how models of computers, specific
much they appear to be influenced by ne computer processes or larger
nological computer We
developments. systems may also provide
also valuable insights into
address two po
we regard as centralrelated overarching
for any or underlying conceptions.
fruitful application
findings to educational A preliminary
practice.search on the subject revealed that the
"Summary an
sion" concludes the relevant literature
paper is dispersed across
with multiple fields of
a brief summ
outlook. research, without a common terminology or publication
forum. Consequently, we did not a priori limit our search to
certain databases, publication formats or even research
Terminology and Method fields, which is usually common practice in systematic
literature reviews. We judged such a strict approach to be
This paper is concerned with children's ideas and beliefs unsuitable to capture the majority of the relevant literature
about how computers operate, what they are made of, and in this case.
what their intrinsic capabilities are. Apart from established Instead, we employed an iterative procedure. We started
dictionaries, a broadly accepted definition for computer is out with various search terms, like conception , conceptu-
surprisingly hard to find. Brunjes (1977) asked the question alisation or mental model , which we combined with terms
"What Is a Computer?" and concluded that like computer or technology , and entered them into various
search engines, like available library catalogues, Worldcat
no current definition will ever truly define what a
and Google Scholar. This led to an initial set of relevant
computer is because computers continually grow and
publications. From there, we started to traverse those
change. [...] Experience will provide the best defini-
publications' cite graphs. That is, for each publication, we
tion of all, or, if not a definition, at least an under-
looked at the literature it cited and used systems like
standing, which is really more important! (p. 85)
Google Scholar and Web of Science to find publications
In light of ongoing research on biological or quantum that cited it. This led to more relevant findings and, occa-
computation, Brunje's statement from nearly four decades sionally, additional search terms, like intelligent artefact ,
ago seems almost prophetic. Nevertheless, we feel that that we had not considered previously. We reiterated this
conducting a transparent literature review requires that we process until, eventually, it stopped yielding any new rel-
at least provide a working definition, if only to give the evant findings.
reader a chance to object. So for the time being, we will Naturally, most publications discovered in that way
aim to use the term computer in a rather inclusive sense to were ultimately not considered relevant to our interest.
denote any device - electronic or otherwise - that can be Determining a publication's relevance consisted of two
instructed to automatically execute computations. Such a steps. The first was a rather superficial survey including
definition has to be used cautiously, nevertheless, and we only the respective title and abstract. In most cases, this
will return to it in "Conceptual Categorisation". was enough to determine that a publication did not cover a
To denote people's mental representations, ideas and topic relevant to our interest. Such publications were not
beliefs about a particular phenomenon or artefact, two included in further iterations of the search. In particular, the
prominent terms are found in the literature: concept and majority of the literature on misconceptions in CS educa-
conception. However, the former is also sometimes used to tion, which primarily focusses on specific algorithmic
denote allegedly objective representations of what some- constructs or data structures (e.g. Kolikant 2001; Eckerdal
thing really is (cf. Weiskopf 2009, p. 149 for a discussion). and Thune 2005; Sanders et al. 2006), was discarded in this
Obviously, the mental representations of interest here are step.
not of such an objective kind. Thus, we will refer to them The second step was performed after the search had
as conceptions throughout. Related terms are preconcep- been completed and involved reading the full text. In doing
tion and misconception, which are commonly used to so, we looked for explicit descriptions of conceptions by
denote people's erroneous or unscientific conceptions. the respective authors. Subsequently, we read the articles
Another relevant term in this context is mental model. A again and looked for further, implicit evidence for these
general definition of it, however, is almost impossible to conceptions in the other accounts. This revealed that still
obtain, as it is used differently in different contexts and many of the remaining publications did not contain find-
fields. See Jones et al. (2011) for an interdisciplinary sur- ings relevant to our interest. This included, in particular,
vey and synthesis attempt. It is often unclear where exactly normative assessments of students' prior knowledge (e.g.
the line between mental model and conception should be Simon et al. 2006; Hammond and Rogers 2007) or purely
drawn. Franco et al. (1999) suggest that they exist on statistical word analyses (e.g. Oleson et al. 2010).

Springer

This content downloaded from

5.179.1.3 on Fri, 05 Apr 2024 06:53:40 +00:00
All use subject to https://about.jstor.org/terms
 276 J Sci Educ Technol (2016) 25:274-283

We considered reliable
technological findings,
novelty. Many of the quotesthose
are indicative of
con
which evidence could be found in at least three different a conception that computers can think and the author
concludes that some children seemed to believe that
publications. Conceptions such as "all computers are giant
machines" (Wolfe 1968, p. 38) or that there is a "chemical
"[c]omputers are 'smarter' than men" (p. 37) and that the
that can make the computer work" (Hyson and Morris "computer is a replica of a man's brain" (p. 38).
1986, p. 22) were thus considered artefacts. The final resultOver a decade later Mawby et al. (1984) conducted
individual interviews with 20 children aged 8 to 12. When
was a set of five conceptions, which we will present in the
asked how a computer works, several made references to a
following section. The corresponding evidence is contained
in 16 publications: 10 journal articles, three conference
brain or mind. The authors conclude that many children
papers, one technical report, one book chapter and onewere unsure whether computers can think and to what
monograph, which cut across various research disciplines,
extent this thinking is similar to human thinking.
including education, developmental psychology and soci- Turkle (2005), in her seminal 1984 publication, makes
ology. Table 1 provides an overview and indicates the
very similar observations. She studied over 200 children,
respective evidence they contain. aged 4 to 14 and reports that they often disagreed and
discussed about the mental capabilities of computers and
whether they are truly alive. However, regardless of what
Children's Conceptions of Computers answer they finally arrived at, their discourse about com-
puters was predominantly psychological rather than phys-
Computers are Intelligent ical, including aspects like intelligence, intentions,
motivations and consciousness (pp. 48^9). Further, she
The conception that computers are intelligent, thinking observed
or that the more the children interacted with com-

even conscious entities is arguably the most widelyputational devices, the more elaborate and nuanced their
researched and reported in the literature. Generally
psychological discourse about them became (pp. 51-52).
speaking, it includes attributing to the computer some formHughes et al. (1987) conducted semi-structured inter-
of mind or brain (human or otherwise) as well as various views with over 100 children, aged 6-12, on two different
mental states like motivations, intentions or even emotions.
occasions 16 months apart. Among other things, they asked
The computer is often anthropomorphised and seenthe
as subjects whether they thought computers could think,
remember, want something or do things by themselves. In
some kind of living entity that is better understood in terms
of psychology rather than technology. accordance with Turkle, they found that the number of
Wolfe (1968) presents one of the earliest accountspositive
of answers to these questions significantly increased
from the first to the second interview occasion. Interest-
this conception. He presents a sample of quotes by seventh-
ingly though, on both occasions there were no significant
grade children with virtually no prior computer experience
whatsoever, illustrating their ideas and beliefs about this
differences between the age groups. This suggests that the

Table 1 Overview of evidence

Source Background Int. Omn. Mech. Wire Prog.
found in reviewed publications
Wolfe (1968) Math. Educ. XX X
Mawby et al. (1984) CS Educ. XXX XX
Turkle (2005) PsychVSociol. XX X
Hyson and Morris (1986) Dev. Psych. X
Hughes et al. (1987) Dev. Psych. XX XX
Denham (1993) Dev. Psych. X
Scaife and van Duuren (1995) Dev. Psych. X
van Duuren and Scaife (1995) Dev. Psych. X X
van Duuren and Scaife (1996) Dev. Psych. X
van Duuren et al. (1998) Dev. Psych. X X
Jervis (2003) Tech. Educ. X
Jervis (2005) Tech. Educ. X X
Papastergiou (2005) CS Educ. X
Bernstein and Crowley (2008) Cogn. Psych. X
Levy and Mioduser (2008) Tech. Educ. X X
Diethelm et al. (2012) CS Educ. X

Springer

This content downloaded from

5.179.1.3 on Fri, 05 Apr 2024 06:53:40 +00:00
All use subject to https://about.jstor.org/terms
 J Sci Educ Technol (2016) 25:274-283 277

observed effects averse

are not
to physical due
explanations (see alsoto childr
"Computers are
development Mechanical" below).
but rather due However,
towhile theirother things like
increa
with computers, as bicycles
over or wind-up
the cars can16
eventually
months
be understood in bet
interviews, home terms
computer access
of gears and springs, computational amon
artefacts are
"rose from 7 to 40 %"
simply (p.
too opaque, 29).
too mentally challenging to be under-
In a series of studies, van Duuren and Scaife further stood in physical terms. "But children do not lack intel-
investigated whether and to what extent children attribute lectual curiosity or inventiveness. They turn to a way of
various mental states and cognitive features to computers understanding where there is more to say-that is, a psy-
and robots. An analysis of written stories by 230 children chological way of understanding" (p. 62). This argument is
aged 7-11 and 38 adults showed that subjects at all ages consistent with the observation that usage and experience
were prone to describing computers and robots in terms of rather than age correlate with children's conceptualisation
animate and intelligent behaviour (van Duuren and Scaife of computers as intelligent. Unless one is actually con-
1995). In two other studies, the authors investigated chil- fronted with computers, there is little incentive to analyse
dren's willingness to attribute a brain as well as various and explain their behaviour.
brain-related items like thinking, dreaming or remembering Levy and Mioduser (2008), in a small study with 6
to a person, a computer, a robot and a doll. It was found children, aged 5-6, specifically investigated what
that with increasing age (from 5 to 11), children became explanatory frames they used when analysing a robot's
more likely to make such attributions to the computer and behaviour. The authors report that, with increasing task
robot (Scaife and van Duuren 1995; van Duuren and Scaife difficulty, the children became more likely to use psycho-
1996). Unfortunately, none of the studies were controlled logical explanations instead of technological ones.
for prior computer experience or usage. The behaviour of computational artefacts can be very
In fact, more recent findings by Bernstein and Crowley complex and unpredictable - just like that of living beings.
(2008) support the hypothesis that computer experience is a And just as it is virtually impossible (not only for children)
relevant factor. The authors presented 60 children, aged to explain the behaviour of living beings in terms of neural
4-7, with various entities, including a calculator, a com- activity and muscle contractions, it is equally futile to try
puter, a humanoid robot and a rover, and asked subjects to and explain the behaviour of computational artefacts in
rate them with respect to their biological and intellectual such physical terms. A psychological framework provides
properties. The ratings were found to be independent of concepts like intentions, motivations and beliefs, which in
age, but positively correlated with prior experience. turn provide a means to analyse and discuss not only the
The lack of prior experience could thus explain the complex behaviour of humans, but also that of computers.
findings reported by Hyson and Morris (1986). The authors
conducted interviews with 15 4-year-olds who had little to
no prior experience with computers, and report that "[t]he Computers are Omniscient Databases
majority of children did not believe that a computer can
talk, get sick, go to sleep, or even think" (p. 24). This conception can essentially be summarised as: comput-
Taken together, these findings clearly suggest a devel- ers know everything, and they know everything by heart. The
opmental trend. Children who have had very little first- computer is seen as a giant database containing a seemingly
hand contact with computers are initially undecided as to
infinite amount of information, including the answers to
whether they can think for themselves or not. As they virtually all questions. Consequently, it does not really
become more engaged with the artefacts' capabilities, their compute anything at all. It just stores and retrieves data.
cognitive features become more and more salient and Again, some of the earliest pieces of evidence for this
children's discourse about them thus becomes more and conception can be found in the quotes reported by Wolfe
more psychological, up to the point where they may even
(1968). The author summarises that some children appar-
grant them a certain status of aliveness. At some point, ently believed that "[c]omputers contain in memory [...]
however, a line is drawn between the artefacts and livingmost of the facts known to man" (p. 37), and that, if you
human beings. Here, Turkle (2005) suggests that this linehad
is a problem, "you just 'ask' the computer for the
answer" (p. 38).
drawn on the basis of emotion: computers can think, but
only humans can feel. Another approach to draw this line isHughes et al. (1987) report on children who asserted
that computers would know about crimes being committed
by means of a computer's programmability: computers can
or have to be programmed, humans have free will (cf. or the current whereabouts of relatives. On the first inter-
"Computers are Programmable"). view occasion, "[t]he most commonly mentioned 'good
Regarding the epistemological origins of this concep-
things' about computers were that they 'help you' or 'tell
you things' (45 mentions)" (p. 18).
tion, Turkle (2005) argues that children are by no means

<0 Springer

This content downloaded from

5.179.1.3 on Fri, 05 Apr 2024 06:53:40 +00:00
All use subject to https://about.jstor.org/terms
 278 J Sci Educ Technol (2016) 25:274-283

The following comment byforMawby

Reported evidence et but
this conception is scarce, al.
it (19
especially telling: exists. Again, one of the earliest accounts can be found in
Wolfe (1968), where one of the children described how a
Too often, [children] spoke as if computers k
computer works by way of "[m]echanical hands" that
specific facts, such as the product of 23 time
move information around and literally "write on the card"
rather than having general algorithms that gen
(p.36).
specific answers to specific questions. Many o
Mawby et al. (1984) also present several children's
younger children seemed almost to view the
quotes that are indicative of a conception of computers as
puter as a natural object, which "just knows" th
mechanical, likening them to a clockwork in which "each
and has the intrinsic ability to answer questions
part moves a different part" (p. 12), or using terms such as
30)
"engines" (p. 18), "machinery" (p. 20), or "little gears"
This question of whether children think computers (p. 32). Interestingly, some of these seemed to be blended
produce answers by computation or by just retrieving it with a conception of computers as intelligent such that it is
from memory was further investigated by van Duuren et al. the mechanics that enabled the computer to think. A similar
(1998). The authors conducted two studies, the first of blend can be found in Jervis (2005), where a child asserted
which we will discuss in "Computers are Programmable". in an interview: "[they think] with little gears" (insertion in
In their second study, they asked 20 adults and 60 children, the orig., p. 10).
aged 5-11, how they thought a computer had produced the Turkle (2005) proposes a plausible argument for the
result of a mathematical calculation. It was found that epistemological origin of this conception. She argues that
40 % of the 8-year-olds and 45 % of the 11 -year-olds
many other devices and toys that children interact with, e.g.
wind-up
thought the computer had had the answer already stored in cars or bicycles, can indeed be understood in
memory, whereas none of the 5-year-olds apparently mechanical terms, in terms of moving parts, gears and
thought so. springs. "Children try to use the same kind of reasoning
Work by Papastergiou (2005) illuminates the issue from with computer toys and computers. They try to understand
a different angle. She conducted a study with 340 students, how these work in physical terms. But this turns out not to
aged 12-16, regarding their mental models of the Internet. be so simple" (p. 61). In their frustration, her argument
The results show that a total of 214 students (~63%) continues, children eventually turn to a psychological
seemed to believe that a single computer - either the user's frame of explanation (cf. "Computers are Intelligent").
own or a remote one on the network - stores the entire However, many mechanical devices like wind-up watches,
Internet (pp. 347-349). The author concludes that many cash registers or typewriters have pretty much gone extinct
students apparently believed "that the capacity of a com- over the last decades. They have been replaced by com-
puter's storage media is unlimited" (p. 349). Diethelm puters. It may have been exactly these devices that children
et al. (2012) also report on students' conception ofknew thefrom their everyday lives and that provided a source
Internet as "one big central computer" (p. 72). of inspiration for their conception of computers as
In summary, there exists considerable evidence, albeit mechanical. So it is, in fact, questionable whether Turkle's
mostly indirect, that some children are prone to concep- argument is still valid today.
tualising computers as essentially giant databases. The In summary, reported evidence of children conceptual-
storage capacities of computers are seen as unlimited ising and computers as mechanical devices is rather scarce and
thus offer at least the potential for storing and retrieving is mainly found in older sources, the only exception being
every piece of information in existence. Consequently, theastudy reported by Jervis (2005). Given the above
single computer might very well store the entire Internet, argument, it is indeed questionable whether there is still a
while others may simply know the answers to virtually mentionable
all number of children today who conceptualise
questions, without the need to actually compute anything. computers in terms of mechanical parts.

Computers are Mechanical

Computers are Wire Networks
The conception of computers as mechanical devices per-
tains to their basic hardware configuration, i.e. how Under
they this conception the computer is seen as a network of
different components, e.g. chips, batteries, memory units or
are built: with gears, springs and levers. The inner work-
even various fantasy elements, which are linked together,
ings of a computer are thus seen more like an intricate
clockwork. Data and processes exist as physical entities
usually by wires or tubes. The links can either be systematic or
form
inside the computer. They move around and are stored ina completely untraceable tangle. The central aspect is
physical places. that components are connected somehow. The components

Springer

This content downloaded from

5.179.1.3 on Fri, 05 Apr 2024 06:53:40 +00:00
All use subject to https://about.jstor.org/terms
 J Sci Educ Technol (2016) 25:274-283 279

themselves are often through

of wires. A plausible reasoning
lesser import. thus might go as They lar
black
boxes. It is theirfollows:
connections that
since computers are very complex determine
electronic
puter's capabilities or devices, they have to have some very complex wiring
functions.
Many of the quotes
inside. reported byinitially
What exactly is wired to what, however, Mawby et
include references to remains
wires a complete mystery,
as athe tangle. As children
primary grow com
computer's internal older and acquire more knowledge
make-up (pp. about the 18,
components19, 31). A
of them are examples of
inside a computer, they conceptual
incorporate this knowledge into blends
computers "connect their existing conception.
wires to The wires start to disentangle
make it think" as (p.
programming means they now connect specific components.
"putting wires Eventually,together"
the
Hughes et (1987) components
al. also may even take on the dominant
report that, role as thebeing ask
computer wires are demoted
works, a "substantial number to mere passive information transmit- mentio
electrical ters. If, when
components, and how children
such as actuallywires
reach that tipping (27 men
tricity point,
(16), batteries (9)however,and
remains an open
plugs question. (6)" (p. 24).
mentioned tapes, memory units or microchips
Computers are Programmable
recently been introduced in class. How exact
might have conceptualised the connections bet
components, however, In its most essential form, this conception
cannot be can be described
reconstructed
authors' descriptions.as follows: the behaviour and capabilities of computers are
Findings reported determined
by Denham by humans and can be changed by humans.
(1993) are mo
What
mative in this respect. In these capabilities
a series are and how exactly
of one goes about
three studies
analysed children's changing
drawings them can, in turn, be constrained
of the by another inner w
computers. The firstconception
two of computers.
were However, this
a is pilot
not necessarily study in
children about the agethe case.of 12, and a replication

with 132 children, aged 9-12.

Turkle (2005) discusses programming Subjects in the context of wer
children's early attempts to
imagine themselves shrinking in understand
size the origins
and of aenter
puter through one of computer's
itscapabilities.
rear Oneports.
7-year-old explained: "The "People ques
was 'What would youtellexpect
it everything. Theyto put ideassee into the machine" (p. 56).
when you s
And an 8-year-old
looked around?"' (p. 349) The described produced a "special 'feelings cas-
drawing
very similar componentssette,"' which would to enable those
a computer to have mentioned
emotions
(p. 56). Again,
far the most significant were we find examples
communication/lof blends: computers
port, memory, and input/output
think or feel by virtue of their programming. (I/O) It is thefunctio
origin
Furthermore, older subjects
of their capabilities, whatever
were those may more be. Turkle (2005)
likely
structured connections,compares whereas
this to children's developing younger conceptions of ones
employ a disorganised animacy and"'muddle'
aliveness: of wire" (p
third study introduced constraints to the task
In all of this talk about the machine's origins, chil-
122 children aged 9-12, who were "asked t
dren are struggling to develop the idea of an "outside
themselves programming the computer to prin
push" for psychological activity much as they
on the screen" (p. 351). Again, wires, tubes
struggle to develop the distinction between inside and
mission links were the most frequent mean
outside pushes for physical motion, (p. 56)
children conceptualised the letters arriving on
sometimes with a According to Turkle
stopover at (2005), this distinction between
a memory unit.
Jervis (2003) also "inside" and "outside" children's
analysed pushes can ultimately formdrawin
the
internals of basis for in
computers, the distinction
a study between people and computers.
with 26 7-ye
26 1 1 -year-olds. Again,
Findings reported
wires by van Duuren
were et al. (1998)
found
support to
elements in many of the
this. In their firstproduced drawings
study, involving 20 adults and 60 children
stantial number aged 5-11, the to
"resorted authors'tangled
presented the subjectswires'
with a film as a
(p. 15). featuring four robots that were controlled by humans, and a
In general, the specific expression of this conception fifth identical robot with a seemingly autonomous beha-
seems to be heavily influenced by children's prior knowl- viour. The authors then investigated to what extent children
edge of computer components. Younger children simply were able to identify the robots' loci of control as external
may not have heard of "chips" or "memory units". But or internal. They found that while 5-year-olds were largely
they probably do know about electricity, about batteries unable to distinguish the different types of robot, the locus
and mains connections, and they know that electricity runs of control became a more salient feature for children with

<£) Springer

This content downloaded from

5.179.1.3 on Fri, 05 Apr 2024 06:53:40 +00:00
All use subject to https://about.jstor.org/terms
 280 J Sci Educ Technol (2016) 25:274-283

increasing age. humans a certain control over the computer.

Furthermore, those Conceptually, subject
tinguish between the
it has the powertwo types
to demote the machine of robot
to a mere machine.
justify this distinction. It was found "that
subjects in the oldest age groups who diffe
No. 5 from Discussion also had an unders
Nos. 1-4
programmable nature of robots" (pp. 877-8
This would Persistencea
indicate andcertain
Change developm
which can also be found in the findings re
Duuren and Scaife Depending on what
(1995). Theaspects ofauthors
a computer they are analy
con-
cerned with, the conceptions
stories about computers, robots presented andin the previousbicy
that, section can
with increasing age, be divided into two categories:
subjects intrinsic
became
include words like capabilities
"program" and internal hardware.
or Accordingly,
"program the con-
stories. ceptions of computers as intelligent entities or omniscient
It would seem that,databases describe their intrinsic capabilities,
as children i.e. what
grow o
gradually come to computers
understand can do. The conceptions that of computers a comas
mechanical devices
ities are not inherent in orthe wire networks describe their internal
machine itself
upon it by people hardware,
in an i.e. what
act computers are made of. Theprogramm
called conception
clear, however,
whether children
of computers as programmable devel
then sits at the intersection
ception entirely onof these
theirtwo categories,
own. as a computer's
Inprogramming
the above
determines (1995),
van Duuren and Scaife its capabilities (e.g. some
thinking or knowing),
of the
who were most likely to
which might involve use
changing terms
its hardware components like
"programming" in (e.g.
theirconnecting wires).
stories, had had ac
programming With respect to persistence,
experience. both the amount and
Observations b
(1987, p. 24),Jervis (2005,
respective publication pp.
dates of the reported evidence20-23)
sug-
Mioduser gest a difference
(2008, pp. 352-353) between these also
two categories (cf.
suggest
rarely come up Table 1).the
with Evidence for the internalof
idea hardware conceptions is
programm
selves. Hence, rather scarce and,
it may very well be with the exception of the two studies
that c
actually engage inpublished
programming
by Jervis (2003, 2005), is at least 20 yearsor
old. In at le
told about it in contrast, to
order evidencestart
for the intrinsicconceptualisin
capabilities conceptions
programmable is somewhat more common and is also found in more
devices.
Furthermore, recent
the wayaccounts. in
This would suggest that conceptions
which childre
programming related
itself is to the
a internal
somewhat hardware of computers are less
complex
own right. persistent than those related to their intrinsic
Where programming is seencapabilities. as
computer's Epistemologically,
capabilities, it is, this of
is extremely
course, plausible. In clo
to the conception"Computers
of a computer
are Mechanical" as
we have already argued howa w
already given the disappearance
examples of ofblends,
mechanical devices from our every-
where pr
day lives may have affected
conceived of as "putting wires children'stogether
conceptions. Fur-
thermore, computational devices
1984, p. 9), or as inserting keep shrinking in size,
a "feelings ca
which makes itother
2005, p. 56). However, increasingly difficult to imagine them
conception
unrelated to any containing numerous wires,
particular gears or levers. In contrast
conception of to a
also been reportedtheir physical appearance,
(Thune and however,Eckerdal
their theoretical
capabilities have
more, Mawby et al. (1984) observed stayed the same. Computers still do that
what
times tended to overextend the act of p
they have always done: they compute. If anything, their
practical capabilities into
include "anything entered have converged
the on the comput
above con-
board" (p. 10). ceptions. Modern computers indeed appear increasingly
We shall not discuss the issue of various conceptions of intelligent and omniscient, as they learn to understand and
programming in more detail here. Suffice it to say that speak natural language and are able to retrieve virtually
having a conception of computer programming - whatever any piece of information from vast online databases.
it may be - would seem to require a conception of a
computer as something that can be programmed. Humans Plurality and Context
can, and often must, tell the computer what to do. It is
humans who determine the computer's capabilities and Throughout "Children's Conceptions of Computers" we
functions. Therefore, the concept of programming gives have already given several examples of conceptual blends,

â Springer

This content downloaded from

5.179.1.3 on Fri, 05 Apr 2024 06:53:40 +00:00
All use subject to https://about.jstor.org/terms
J Sci Educ Technol (2016) 25:274-283 281

that is, of coherent desktop computers

combinations (M
of two di
tions. This Still,
implies that most
those of these stu
conceptions ar
exclusive, but can,predates
in fact,the
coexist withi
invention o
vidual. Other authors have argued for
e.g. smartphones, su
tablet
household
pluralism elsewhere, appliances.
and emphasised that tH
emergence of any specific
that areconception is
increasingly hi
do
on context or current communicative
Desktop computers, ing
1995; Weiskopf 2009; Dawson 2014).
"Persistence and Chan
For computers, such conceptual
related pluralism
to a computer's
plausible. Proudfoot (2011),
stable and for instance
insulated ag
researchers who refer to their
ments. own
However, artific
it does
creations as if theyconceptions
were indeed alive
also and t
apply t
ever, it would appearWhat
somewhat presumptuo
is needed here is
they would not alsoconceptual
be capable categorisatio
of thinking
about these artefacts in terms
they of
actually their prog
conceive o
ure, their storage eligible targets
capacities for
or their th
variou
nicating hardware Morris (1986)
components. reported
They simpl
in different distinguish
contexts computers
and for different pur
lowing the arguments of Mortimer
using (1995),
the keyboard e
as th
fact, posses very such
similar an approach
conceptions wou
as chi
difference may be Indeed,
that findings
they by
are more M
awa
their respective today's and
limitations, children apply
are thus mo
choosing the right one for
exactly a particular
those reasonin
274-275). remains unclear. Findi
Making students aware of that
suggest their conception
they might
Research
upon their respective in and
advantages the limit
cogn
key aspect of teaching
people within a a
can take plurali
variet
framework (Mortimer 1995). In
determining theorder to
catego
informed manner, current function,
however, origi
research first n
contexts in which relations
these between
conceptions are su
re
lematic. For instance, it is
2007). Inundeniable that
addition, artef
computers as intelligent
highlyand thinking enti
context-depende
powerful means tomultiple ways complex
analyse their to cate
"Computers are Vermaas 2013).
Intelligent"). Indeed
However, in t
programming, this very only
puters same in
conception
a second
learning obstacle (Pea
and 1986). Similarly,
uses are those of a r
puters as essentially that omniscient
is exactly databas
what t
adequate and useful situations.
in many everyday situ
To our kno
looking up thingsresearch
on Wikipedia or find
investigating i
restaurant. Yet, it may lead to
in general, think aboutmisconc
learning about computer
them not networks (Papa
only in terms
Diethelm et al. 2012).in Investigating
terms of theirthe com
rele
conceptions in other existing contexts certainly
conception of c
worthwhile goal. latter case.

On a parenthetical note, such research could also help

Conceptual Categorisation
inform classroom practices in a more direct way. An often
expressed, albeit rather broad, objective of CS and tech-
In "Terminology andnology Method" weto identify
education is to enable students have and defi
understand the various computational systems in their aut
as any device that can be instructed to
cute computations. Consequently,
environment the st
(e.g. Tucker et al. 2003, p. 3; ISTE 2007;
here include various devices like robots
Department for Education, UK 2013). This implies that (
Scaife 1995), electronic
students, among othertoys (Turkle
things, should be able to do precisely 200
calculators (Bernstein
what we have and Crowley
described above, that is, to look beyond an 2008)

â Springer

This content downloaded from

5.179.1.3 on Fri, 05 Apr 2024 06:53:40 +00:00
All use subject to https://about.jstor.org/terms
 282 J Sci Educ Technol (2016) 25:274-283

artefact' s primary function and,

children actually categorise ifand
as computers applicable
what cogni-
as a general-purpose computational
tive processes are involved when they do. device.
the related conceptual categories could thus re
Compliance with Ethical Standards
how well students are actually able to do s
insights into how this ability
Conflict of interest The authors might be
declare that they have impr
no conflict
education. of interest.

Summary and Conclusion References

In this paper, we have presented five conceptions that Bernstein D, Crowley K (2008) Searching for signs of intelligent life:
children reportedly form about computers, based on an an investigation of young children's beliefs about robot intelli-
gence. J Learn Sci 17(2): 225-247
interdisciplinary literature review. Accordingly, computers
Brunjes S (1977) What is a computer? J Med Syst 1(1):79- 85
can be seen as intelligent entities that are best understood Dawson C (2014) Towards a conceptual profile: rethinking concep-
in psychological rather than physical terms. They can also tual mediation in the light of recent cognitive and neuroscientific
be seen as unlimited and potentially omniscient databases, findings. Res Sci Educ 44(3):389-414
Denham P (1993) Nine- to fourteen-year-old children's conception of
which know the answers to virtually all questions, without
computers using drawings. Behav Inf Technol 12(6):346-358
having to actually compute anything. They can be seen as a Department for Education, UK (2013) National Curriculum in
complex network of wires and communication links in England: Computing programmes of study: key stages 1 and 2.
which the actual components might play a negligible role. URL: https://www.gov.uk/government/uploads/system/uploads/
attachment_data/file/239033/PRIMARY_national_curricu
They can be seen as containing a complex mechanical
lum_-_Computing.pdf. Visited on 12 Nov 2015
contraption. And they can be seen as programmable Diethelm I, Wilken H, Zumbrägel S (2012) An investigation of
machines that need to be told what to do by their human secondary school students' conceptions on how the Internet
masters. works. In: Proceedings of the 12th Koli Calling International
Conference on Computing Education Research. ACM, pp. 67-73
However, not all of these conceptions appear to be
Eckerdal A, Thune M (2005) Novice Java programmers' conceptions
equally persistent over time. While the conceptions related
of object and class, and variation theory. In: ACM SIGCSE
to a computer's intrinsic capabilities (i.e. thinking and
Bulletin. Vol. 37. 3. ACM, pp. 89-93
Franco
knowing) have persisted for nearly half a century now, theC et al (1999) From scientists' and inventors' minds to some
scientific and technological products: relationships between
conceptions related to a computer's internal hardware (i.e.
theories, models, mental models and conceptions. Int J Sci Educ
wires and mechanics) seem to be more easily influenced by
21 (3):277- 291
new technological developments and may even have dis- M, Rogers P (2007) An investigation of children's
Hammond
appeared completely by now. Epistemologically, thisconceptualisation
is of computers and how they work. Educ Inf
Technol 12(1):3-15
very plausible. The physical appearance of computers has
Houkes W, Vermaas PE (2013) Pluralism on artefact categories: a
changed drastically over the last few decades, while their
philosophical defence. Rev Philos Psychol 4(3):543-557
theoretical capabilities have remained the same. If any-M, Brackenridge A, Macleod H (1987) Children's ideas about
Hughes
thing, modern computers do indeed appear increasingly computers. In: Rutkowska J, Crook C (eds) Computers, cogni-
tion and development: issues for psychology and education. 1st
intelligent and omniscient.
ed. Wiley. Chap. 1, pp. 9-34
Finally, we have addressed two points that we regard
HysonasMC, Morris SK (1986) 'Computers? I love them!': young
central for any fruitful application of these findings to
children's concepts and attitudes about computers. Early Child
educational practice. First, it has to be assumed that Dev
an Care 23(1): 17-29
ISTE (2007) ISTE Standards: Students. Tech. rep. International
individual may hold several conceptions simultaneously,
Society for Technology in Education. URL: https://www.iste.
which may or may not be selected in a given context or
org/docs/pdfs/20- 14_ISTE_Standards-S_PDF.pdf. Visited on 12
situation. While a conception may be useful and appro-
Nov 2015
Jarvis T, Rennie LJ (1996) Understanding technology: the develop-
priate in some contexts, it may cause problems in others.
Education needs to take this into account and make stu- ment of a concept. Int J Sci Educ 18(8):977-992
Jervis A (2003) Children's thinking about computers. Paper presented
dents aware of such advantages and limitations, which at the British Educational Research Association Conference,
obviously requires that we first gain an understanding of Heriot-Watt University, Edinburgh
them ourselves. Second, it is unclear to what degree Jervis
the A (2005) 'It's good for plugging-in a budgie': children talking
about computers. Paper presented at the British Educational
findings presented in this paper can be generalised to more
Research Association Annual Conference, University of
recent computational devices like smartphones, tablets or Glamorgan
modern household appliances. In order to answer this
Jones N et al (2011) Mental models: an interdisciplinary synthesis of
question, research needs to investigate, which things theory and methods. Ecol Soc 16(1)

Springer

This content downloaded from

5.179.1.3 on Fri, 05 Apr 2024 06:53:40 +00:00
All use subject to https://about.jstor.org/terms
 J Sci Educ Technol (2016) 25:274-283 283

Kolikant YB-D (2001) Simon Gardeners and

B et al (2006) Commonsense computing: cinema
what students know tic
students' preconceptions before we teachof (episode 1:concurrency.
sorting). In: Proceedings of the Com
11(3):22 1-245 second international workshop on Computing education
Levy ST, Mioduser D (2008) Does it 'want' or 'was it programmed research. ACM, pp. 29-40
to..'? Kindergarten children's explanations of an autonomous Sorva J (2013) Notional machines and introductory programming
robot's adaptive functioning. Int J Technol Des Educ education. ACM Transactions on Computing Education 13(2)
18(4):337-359 Thuné M, Eckerdal A (2010) Students' conceptions of computer
Malt BC, Sloman SA (2007) Artifact categorization: The good, the programming. Tech. rep. 2010-021. Uppsala University, Com-
bad, and the ugly. In: Creations of the mind: Theories of artifacts putational Science
and their representation. Oxford University Press, Oxford, UK, Tucker A et al (2003) A Model curriculum for K-12 computer
pp 85-123 science: final report of the ACM K-12 Task Force Curriculum
Mawby R et al (1984) Structured interviews on children's conceptions Committee. ACM. URL: http://www.csta.acm.org/Curriculum/
of computers. Tech. rep. 9. Bank Street College of Education sub/CurrFiles/K- 12ModelCurr2ndEd.pdf. Visited on 12 Nov
Mortimer EF (1995) Conceptual change or conceptual profile change? 2015
Sci Educ 4(3):267-285 Turkle S (2005) The second self: computers and the human spirit.
Mumtaz S (2002) Children's conceptions of information and com- 20th anniversary edn. MIT Press, Cambridge, MA
munications technology. Educ Inf Technol 7(2): 155-168 van Duuren M, Scaife M (1995) How do children represent intelligent
Oleson KE et al (2010) Developmental human factors: children's technology? Eur J Psychol Educ 1 0(3):289- 301
mental models of computers. In: Proceedings of the Human van Duuren M, Scaife M (1996) 'Because a robot's brain hasn't got a
Factors and Ergonomics Society Annual Meeting. Vol. 54. 19. brain, it just controls itself - Children's attributions of brain
SAGE Publications, pp. 1450-1453 related behaviour to intelligent artifacts. Eur J Psychol Educ
Papastergiou M (2005) Students' mental models of the internet and 1 1(4): 365-376
their didactical exploitation in informatics education. Educ Inf van Duuren M, Dossett B, Robinson D (1998) Gauging children's
Technol 10(4):341-360 understanding of artificially intelligent objects: a presentation of
Pea RD (1986) Language-independent conceptual 'bugs' in novice 'counterfactuals'. Int J Behav Dev 22(4):87 1-889
programming. J Educ Comput Res 2(l):25-36 Weiskopf DA (2009) The plurality of concepts. Synthese
Proudfoot D (2011) Anthropomorphism and AI: Turing's much 169( 1 ): 145-173
misunderstood imitation game. Artif Intell 175(5-6): 950-957 Wolfe LR (1968) Computer concepts possessed by seventh-grade
Sanders I, Galpin V, Götschi T (2006) Mental models of recursion children. Arith Teach 1 5(1 ):35- 39
revisited. SIGCSE Bull. 38(3): 138-142
Scaife M, van Duuren M (1995) Do computers have brains? What
children believe about intelligent artifacts. Br J Devi Psychol
13(4):367-377

Ô Springer

This content downloaded from

5.179.1.3 on Fri, 05 Apr 2024 06:53:40 +00:00
All use subject to https://about.jstor.org/terms