THE FINGER-READING EFFECT WITH

CHILDREN: TWO UNSUCCESSFUL

REPLICATIONS
BY YUNG-JONG SHIAH
ABSTRACT: This article reports two attempted replications of nger-reading
under controlled conditions. In Experiment 1, 18 children ages 7 to 12 years were
tested individually to determine if by touching they could identify a 2-digit number
in 1 of 4 different colors imprinted on a sheet of paper in the absence of visual
cues. No signicant nger-reading effect was found, and there were no signicant
relationships between nger-reading scores and questionnaire responses concerning
imagery ability and paranormal beliefs. In Experiment 2, an attempt was made to
screen talented participants in 3 steps: a selection study (SS), a conrmation study
(CS), and a training study (TS); 2,200 children ages 6 to 13 years were invited to
participate in the SS, of whom 1,771 did so; 1,655 completed 10 trials, and 820
obtained 1 or more hits in recognizing a 2-digit number with color; 728 of these
completed all 20 trials in the CS with the same target stimuli. None scored any
hits, so there was no TS. Controls against sensory leakage were less stringent in
the SS than in the CS, which could account for the vast scoring difference. The 2
experiments failed to select children with the potential to develop nger-reading
ability. Suggestions for future research are provided.
The present study was designed to replicate the nger-reading effect
(Lee, 1998, 1999, 2002, 2003; Lee & Chang, 2001; Lee, Chen, & Tang, 2000;
Lee, Tang, & Kuo, 2004; Tang, Lee, & Hsu, 2000). The authors previous
nger-reading studies have been reported in detail elsewhere (Shiah &
Tam, 2005). There, the average success rate of target recognition (p <.05)
after 4 consecutive days of 2 hr training by means of touching a sheet of
paper containing a two-digit number or a complex character in one of four
different colors was 24% (41 out of the 173 participants who completed
training). The dropout rate was about 20% (43 participants).
The successful children reported that distinctive visual images
accompanied their successful guesses. These visual images appeared as if
they were seen in the real world. The children subjectively reported seeing
answers on a transparent and sometimes an opaque screen (Lee, 1998,
1999; Lee et al., 2000; Shiah & Tam, 2005; Tang et al., 2000). They appeared
to recognize complex characters or other complex symbols easily after
seeing a distinctive or opaque screen masking their normal visual imagery.
However, the results of that nger-reading study are subject to
criticism (Du, 2005). All the trials for all participants were not clearly
described, as well as the method of analysis. Moreover, the nger-reading
effect may have resulted from fraud due to a lack of stringent safeguards
(Shiah & Tam, 2005). Thus, it is not safe to assume that nger-reading is a
110 The Journal of Parapsychology
real ability. As the nger-reading effect has not yet been tested and replicated
by other researchers, the author attempted to replicate the nger-reading
effect himself. In a recent paper, modied nger-reading procedures with
robust controls were described (Shiah, 2005). The present paper draws on
these newly developed procedures.
EXPERIMENT 1
The main hypotheses for this experiment concern childrens
performance when directly touching the targets. As Lee found that 24% of
his children had a signicant result in identifying a two-digit number in one
of four different colors, Hypothesis 1 is that at least 25% of participants will
achieve signicant, above-chance scoring on such targets after 8 hr training.
Hypothesis 2 concerns learning, that is, whether improvement in scoring
will be observed from the beginning to the end of training.
Hypotheses 3 and 4 concern the existence of a relevant visual
experience in the form of a screen. Lee (Shiah & Tam, 2005; Tang et
al., 2000) found that some participants subjectively reported having seen
a screen with an image of the right answer overlapping their normal
vision while touching the target, no matter whether their eyes were open
or closed. To the best of our knowledge, this visual experience is quite
different from real vision. Hypothesis 3 predicted improved scoring of
recognizing two-digit numbers with colors after subjectively experiencing
a visual screen.
Lee argued that children who experience a vivid screen bearing
the right target will score better than those who do not experience a screen.
Thus, Hypothesis 4 predicted that participants scoring will signicantly
improve after they acquire the ability to see a screen while touching a target,
and better than children not reporting having seen a screen.
Exploratory questions included the following. Hypothesis 1
concerned only the proportion of successful individuals, but what will be
the overall mean of the scores? Second, how many hits will there be on
numbers and colors separately, and will these scores improve from the
beginning to the end of training?
What is the nature of the visual imagery in nger reading? Visual
imagery has been widely explored as a vehicle for expressing ESP ability
(Blackmore & Rose, 1997; George, 1982; Honorton, 1975; Honorton,
Tierney, & Torrey, 1974; Price, 1973; Schechter, Solfvin, & McCollum,
1975). The ganzfeld technique, which involves participants experiencing
target-related visual imagery under conditions of sensory deprivation
(Irwin & Watt, 2007), often elicits ESP performance at levels signicantly
above chance (Bem & Honorton, 1994; Bem, Palmer, & Broughton, 2001;
Palmer, 2003; Storm & Ertel, 2001; Utts, 1995). However, some researchers
do not agree with this conclusion (Hyman, 1985, 1994; Milton & Wiseman,
1999, 2001). However, no such studies with children appear to have been
111 The Finger-Reading Effect With Children
reported, and the literature does not yet suggest how many children have
reported this experience. So, what percentage of children will report mental
imagery in the present study? Will children who visualize a screen describe
it in similar ways? What is the nature of visual experience in participants
who successfully identify the targets? What is the relationship between the
screen experience and hits on numbers and colors separately?
Another exploratory question concerns ESP belief and nger-
reading success. Use of so-called sheep-goat questionnaires is one of the
early methods developed to understand the relationship between ESP
belief and psi-hitting (Schmeidler, 1952; Thalbourne, 1981). It was found
that sheep (who believe in the possibility of ESP) have higher ESP scores
than goats (who do not believe in the possibility of ESP). What are the
relationships between hit rates, mental imagery, and paranormal beliefs?
Finally, some researchers seem consistently to achieve positive results
in ESP experiments; others appear to obtain negative or chance ndings
(Smith, 2003; Watt & Ramakers, 2003). For example, Watt and Ramakers
(2003) found that participants tested by psi-believing experimenters had
higher scores on the psi task than those tested by disbelieving experimenters.
This nding is called the experimenter effect. Will there be an experimenter
effect on participants nger-reading performance?

METHOD
Experimenters
There was a main experimenter (the author) and ve
coexperimenters, all students at Edinburgh University. The main
experimenter (the author), one coexperimenter, and one participant were
in the room during the experimental session.
Participants
Eighteen participants, 8 boys and 10 girls, joined in the training
process, 5 of whom went through the entire training program. The
participants were tested individually. The age range was from 7 to 12 years
(M = 9.67, SD = 1.94). Fourteen participants were right-handed and 4 were
left-handed, according to the Edinburgh Handedness Inventory (Oldeld,
1971). Participants sex, age, and handedness are presented in Table 1.
None had a history of nerve injury, nger trauma, or

learning disability,
and their nger pads were free of calluses.

The participants were recruited from elementary schools in
Edinburgh, Scotland. An announcement included an introduction to
the nger-reading task, the purpose of the training, and a consent form
for parents to sign. The participants were informed that the study was
approved by the Psychology Department Ethics Committee of Edinburgh
112 The Journal of Parapsychology
University as suitable for children, and the experimenters had the
necessary Disclosure Certicate for Scotland, showing they had no
criminal record.

TABLE 1
PARTICIPANTS SEX, AGE, AND HANDEDNESS

Participant Sex Age Handedness
S1 Female 12 Right
S2 Female 7 Right
S3 Female 7 Right
S4 Male 10 Left
S5 Male 11 Left
S6 Male 7 Right
S7 Female 11 Right
S8 Male 10 Right
S9 Female 10 Left
S10 Male 11 Left
S11 Female 7 Right
S12 Female 11 Right
S13 Male 7 Right
S14 Male 12 Right
S15 Female 12 Right
S16 Female 11 Right
S17 Male 10 Right
S18 Female 8 Right

Based on Lees original research, it was assumed that participants
would obtain the best performance when they participate over an intensive
program over 4 successive days of 2-hr training. No one could join the
experiment without agreeing to complete the total 8-hr regimen. The
duration of all sessions was usually 1 week. All participants were paid 2.50
for each 2-hr session.
Questionnaire
The questionnaire included items about personal characteristics,
imagery in various modalities, and belief in the paranormal. (The
questionnaire is presented in Appendix 1.) All items were taken from
a questionnaire developed by the Koestler Parapsychology Unit, except
for the four questions on tactile imagery, which were developed by the
author.
113 The Finger-Reading Effect With Children
Target Preparation
The nger-reading task and stimulus material was the same as that
described in the authors previous paper (Shiah, 2005). All target materials
were prepared in advance by a research assistant who otherwise was not
involved in the experiment. The coexperimenter who handled the target
envelopes had no relationship or contact with the assistant who prepared
the targets.
The target materials were produced in a strictly standard way.
The targets were 75 two-digit numbers ranging from 11 to 97 in one of
four colors: red, green, blue, or black. The numbers were the same ones
used by Lee. The targets were generated by a pseudorandom algorithm
developed by Paul Stevens, who at the time was a research fellow at the
Koestler Parapsychology Unit of Edinburgh University. The algorithm
uses a sort routine based on the Microsoft Visual Basic RND function and
seeded by the computer timer at the start of the program. The targets
were selected with replacement; thus, the same target could be repeated,
and each trial was independent of the others. Mean chance expectation
(MCE) was 1 in 300. As previous experience indicated that a participant
can usually complete 20 trials in 2 hr, the computer program generated
the targets in sets of 20. All the targets for the experiment were generated
at one time.
Each target was printed exactly in the middle of a 5 cm 8 cm

sheet of paper (A4, white, 75 g/m
2
) manufactured by H. E. Copier. Each
digit was about 0.6 cm 0.5 cm (Times New Roman, 24 points). The
digits were printed using an HP Ofcejet G85 color printer, which was
conrmed to produce zero elevation (Shiah, 2005). Each target sheet was
inserted exactly in the middle of a 15.2 cm 8.9 cm Niceday opaque,
gummed manila envelope with no markings (70g), manufactured by
Guilbert (product code: 182543). The short side of the paper contacted
the bottom of the envelope. The target digits faced the front of the
envelope, right side up.
To aid double-checking of results, the 20 envelopes in each set
were numbered faintly 120 by pencil on the outside and not detectable by
touching. After the envelopes were sealed, each set of envelopes was inserted
in an opaque plastic bag, which was put in a 22.9 cm 16.2 cm Niceday
envelope (90g) otherwise identical to the smaller envelopes (product
code: 183422). This large envelope was signed by the research assistant
and sealed by cellotape at its two ends, such that any tampering would be
detected by the coexperimenter who cut open the envelopes during the
experiment. The large envelope was not opened until the session began.
All the envelopes and plastic bags were tested under sunlight to prove that
the targets could not be seen.
Each target sheet, small envelope, plastic bag, and large envelope
was used only once. Thus, no participant could receive any cues from any
114 The Journal of Parapsychology
markings made on these materials by previous participants. The targets
were stored in a separate room, to which neither the participants nor the
coexperimenter had access. Participants did not see any targets or their
containers until receiving feedback. The research assistant who prepared
the samples saved detailed information about the target list in a secure
computer and on a disk stored in a sealed opaque envelope. Only the
research assistant had access to these records. All the target materials were
double-checked after the experiment to check for possible recording errors
or cheating by replacement of target sheets.
Test Barriers and Lab Setup
During the actual test, the target sheet was placed inside a black
plastic bag (see Figure 1), which is normally used for handling photographic
negatives (double-lined, 43 cm 40.5 cm, Tecnodia Co., Ltd. Japan). It has
two layers, each with its own zipper. Virtually no light can enter the bag, as
was shown by a light detector. The bag is large enough that the participant
can move the hand freely within the bag.
The participant was separated from the bag by an 80 x 80 cm black
screen placed on the table in front of which the participant was seated. The
screen has two holes with cuffs, through which the participants forearms
are inserted. The holes are 8 cm in diameter and 1.5 cm from the bottom
of the screen. The distance between them is 15 cm. The screen is located on
the table between the participant and the bag. Peeking can only take place
if any gaps in the barrier-cuff and the bag-cuff are lined up exactly, as the
participant could then conceivably peer through any small gap. However,
this is impossible without the observers immediately noticing participants
contorting their body.
The experimenter and coexperimenter were positioned on each
side of the barrier to observe the participant during the nger reading (see
Figure 2).
Procedure
The nger-reading training was conducted between June 2005 and
February 2006. Before it began, the questionnaire was administered to the
children individually in a semistructured interview format to make sure
they understood the questions. The terms telepathy, clairvoyance, precognition,
and psychokinesis were explained.
The participants were tested individually. The training sessions
did not exceed 2 hr, due to childrens limited attention span. There was a
15-min break in each hour, during which participants were rewarded with
drinks or snacks.
115 The Finger-Reading Effect With Children
F
i
g
u
r
e

1
.

T
h
e

b
a
g

a
n
d

s
c
r
e
e
n
116 The Journal of Parapsychology
Figure 2. The laboratory setup
Warm-up training. The experiment began with one practice session.
After the experimenter turned off the light, participants were required to
sit, close their eyes, and breathe deeply with a calm and peaceful mindset
for at least 5 min. After the light was put back on, the participants were
asked to practice image making. The experimenter showed a red apple
or other simple object to the participants, who were asked to look at it very
carefully and remember every detail. Then they closed their eyes to visualize
the object exactly as if they were looking at it. It was found in the authors
previous experiment that all the children seemed able to perform this
task. Once the children could do this successfully, they tried, for example,
to visualize the apple changing color three or four times, say from green
to blue to black. Participants also saw a demonstration describing touch
reading as a way to identify the target.
Formal training. After the participants completed three to ve practice
trials, the experimenter gave the coexperimenter one large sealed envelope
containing a plastic bag with 20 small envelopes. The coexperimenter opened
the large envelope, took one small envelope from the plastic bag, put it in
the black bag, and then closed the zipper. The rest of the small envelopes
were kept in the large envelope until required. The participants were clearly
informed at this time about the meaning of randomization.
Next, the coexperimenter opened the black bag so the participants
could get their hands into it. He then closed the zippers. The participants
117 The Finger-Reading Effect With Children
were then told to open the sealed envelope and remove the target sheet.
They had been taught to tear off the very end of the right or left side of the
envelope to allow removal of the target sheet.
Next, the participants scanned the target with their ngers. They
were asked to focus on touching the target and to imagine that they could
see the target while doing so. They were also told that there was a fold in the
top left corner of the target sheet as a cue they could use to help them locate
the target number. They were asked to tell the coexperimenter whatever
they saw and felt. There were no time restrictions, and participants were
free to use whatever scanning pressure and speed they chose.
Participants were told that taking their hands out of the black bag
or pulling at the bag or cuffs was not allowed, and they were to avoid any
unnecessary movement of their arms. They could only take their hands out
of the black bag after they told the coexperimenter that they had made their
nal response. In the meantime, the coexperimenter and the experimenter
recorded the participants responses and response times.
After the participant nished the trial, the coexperimenter removed
the target sheet from the black bag and showed it to the participant. If
the participant wanted to have a break during the experiment, the
coexperimenter sealed the big envelope and put it in another room.
The experimenter then locked this room so that no one could access the
targets.
RESULTS
Participants responses, including the experimenters record, the
coexperimenters record, and the original data, were double-checked by an
independent researcher, as were the data keyed into the computer.
The length of each session varied from 30 min to 2 hr. Usually, it
took the participants 1 hr to attempt 10 items. Thus, we chose 80 trials as
the best estimate of meeting the criterion of 8 hr of training suggested by
Lee. However, only ve participants completed 80 trials. The major reason
for failing to meet this requirement is that the children felt the program
was somewhat tedious and time-consuming.
Tests of Finger-Reading Success

Hypothesis 1 predicted that at least 25% of participants would achieve
a signicant result in recognizing number plus color within 8 hr. The null
probability for a correct response on number plus color is .0033. Because the
number of trials is large, exact binomial probabilities were estimated using
the Poisson approximation (Robinson, 1985). The criterion for signicance
was p >.05, one-tailed. None of the ve participants who completed the
training showed signicant nger-reading ability in recognizing number plus
color by a binomial test; neither did any of the other participants. Only two
118 The Journal of Parapsychology
participants got any hits at all on number plus color; one obtained one hit
and the other two hits. Thus, Hypothesis 1 is not supported. In addition, the
overall mean was not signicant. Likewise, no signicant results were found
in recognizing colors and numbers separately (see Table 2).
TABLE 2
HITS ON COLORS AND NUMBERS SEPARATELY
Participant Trials Total hits (%) on color
(MCE = 25%)
Total hits (%) on
number (MCE = 1.33%)
P1 108 28 (26%) 1 (0.9%)
P2 46 8 (17%) 0 (0%)
P3 54 15 (28%) 0(0%)
P4 35 13 (37%) 0 (0%)
P5 143 39 (27%) 2 (1.4%)
P6 40 11 (28%) 0 (0%)
P7 73 14 (19%) 0 (0%)
P8 39 10 (26%) 0 (0%)
P9 91 26 (29%) 1 (1%)
P10 20 4 (20%) 1 (5%)
P11 35 9 (26%) 1 (3%)
P12 9 2 (22%) 0 (0%)
P13 45 9 (20%) 2 (4.4%)
P14 370 102 (28%) 1 (0.3%)
P15 27 9 (33%) 0 (0%)
P16 63 20 (32%) 0 (0%)
P17
182
37 (20%) 3 (1.7%)
P18 15 5 (33%) 0 (0%)
Total trials 1,395 361 (26%) 12 (0.9%)
Hypothesis 2 predicted that participants would signicantly
improve their nger-reading ability from the beginning to the end training.
Across sessions, the hit rate on number plus color did not show any trends
or patterns of improvement with practice. Thus, the data do not support
Hypothesis 2. Likewise, an across-session analysis of hit rates on colors and
numbers separately revealed no trends or patterns of improvement. Thus,
there was no support for Hypotheses 3 and 4.
It is possible that the large number of dropouts biased the results.
This concern was examined by an additional analysis. First, as shown in Table
3, we broke down the sample into approximately equal groups based on
two-session units (40 trials), and we recorded the mean hit rates for number
plus color, color separately, and number separately for each group.
119 The Finger-Reading Effect With Children
TABLE 3
MEAN HIT RATES (STANDARD DEVIATIONS) AS A FUNCTION OF SESSIONS COMPLETED
Group Mean hit rate
for number plus
color
(MCE = .0033)
Mean hit rate for
color
(MCE = .25)
Mean hit rate for
number
(MCE = .0133)
Less than 2
sessions (n = 8)
0
(SD = 0)
.28
(SD = 0.058)
.010
(SD = 0.019)
2 or 3 sessions
(n = 5)
0
(SD = 0)
.23
(SD = 0.065)
.009
(SD = .020)
All 4 sessions
(n = 5)
.0036
(SD = .005)
.28
(SD = 0.013)
.011
(SD = 0.005)
As shown in Table 4, a one-way ANOVA comparing all three groups
revealed no signicant differences for either color or number separately.
There was too little variability in scores for a valid analysis of number plus
color.
TABLE 4
ONE-WAY ANOVA COMPARING THE THREE GROUPS

Source/Hits SS df MS F p
Color Between
groups
.0083525 2 .0041763 1.519 .251
Within groups .0412475 15 .0027498

Total
.0496000 17

Number Between
groups
.0000085 2 .0000043 .015 .985
Within groups .0042620 15 .0002841

Total .0042705 17

Finger-Reading, Imagery, and Belief
Individuals were classied as having good imagery, and believing
in psi and having had psi experiences, if one or more of the relevant
ratings were 5 or above on the 7-point scales. Sixteen of the 18 participants
(89%) reported that they were good at creating a visual image of a
familiar scene, and 17 participants (94%) reported that they could see
such a scene clearly. Eleven participants (61%) reported that they were
very good at creating auditory imagery, 8 (44%) reported that they were
120 The Journal of Parapsychology
very good at creating olfactory imagery, 7 (39%) reported that they were
very good at receiving gustatory imagery, and 11 (61%) reported that
they were very good at creating tactual imagery and had very good tactile
sensitivity. Nine participants (50%) reported getting a visual image when
touching something they could not see. As for paranormal beliefs and
experiences, 13 participants (72%) reported that they believed in ESP
and 7 participants (39%) reported that they believed in psychokinesis.
Six participants (28%) reported having had an experience of telepathy;
3 (17%) an experience of clairvoyance, and 12 (67%) an experience of
precognition.
There were no signicant relationships between nger-reading
hit rate and any of the imagery or paranormal beliefs items, as shown in
Table 5.
TABLE 5
CORRELATIONS OF RESPONSES TO IMAGERY AND BELIEF QUESTIONS
WITH FINGER-READING HITS (N = 18)
Question Hit rate on
colors
Hit rate on
numbers
Hit rate on
numbers with
colors
Imagery:
Easy to create visual imagery .144 -.130 .140
Visual imagery is clear .147 -.206 -.306
Good auditory imagery .179 -.277 -.393
Good olfactory imagery .416 -.029 -.031
Good gustatory imagery .373 -.335 -.264
Good tactual imagery .071 -.108 .055
Good tactile sensitivity -.122 -.011 .044
Good visual imagery when
touching something you
cannot see
-.192 -.186 .123
Number of hours per day
using a keyboard
.013 -.307 .226
Paranormal beliefs and
experiences:
Belief in ESP .040 -.387 .026
Belief in psychokinesis .249 -.419 .236
Experience of telepathy .049 -.126 .231
Experience of clairvoyance .099 -.198 .220
Experience of precognition -.432 .300 .284
121 The Finger-Reading Effect With Children
Experimenter Beliefs
Questionnaire results for experimenters belief in psi and psi
experiences are shown in Table 6. All the experimenters are considered
believers in ESP, as they all gave ratings of 5 or higher on this question. As
there was no evidence of psi among the participants, it is logical to infer
that there is no experimenter effect on their performance.
TABLE 6
EXPERIMENTERS BELIEF IN PSI AND PSI EXPERIENCES (N = 6)
Question M (SD)
Belief in ESP 5.86 (0.90)
Belief in psychokinesis 5.58 (1.14)
Experience of telepathy 5.57 (1.27)
Experience of clairvoyance 3.86 (1.77)
Experience of precognition 5.29 (2.14)

DISCUSSION
The present experiment attempted to replicate Lees promising
nger-reading results under more stringent conditions. No signicant
results were found, nor were there any signicant relationships between
and imagery or paranormal beliefs and experiences. Thus, the nger-
reading effect was not replicated.
How can these results be interpreted? They might mean that
nger-reading ability does not really exist. Another possibility is that the
effect is real but could not be demonstrated due to the small sample size of
only 18 participants, all the more so because only 5 of the 18 participants
completed the truly intensive training recommended by Lee. Finally,
one should note Tarts observation that receiving immediate trial-by-trial
feedback improves ESP performance only with talented participants who
have demonstrated psi ability (Tart, 1966, 1975, 1977a, 1986; Tart, Palmer,
& Redington, 1979). Thus, another possible explanation might be that the
unselected participants in the present study might have lacked sufcient psi
ability to benet from the training.
Methodological Limitations
Three safeguards were not fully implemented in the study because
the experimenter did not become aware of them until the study had been
completed (rst case) or was already in progress (other cases). First, the
fold made in the paper for orientation purposes was made after the targets
had been generated. Thus, the independent researcher who prepared the
122 The Journal of Parapsychology
targets could have deliberately or unconsciously made, for example, bigger
folds for large numbers than for small ones. This possibility was ruled out
because no obvious patterns in the folds were found by an independent
researcher. Second, most participants were not asked to show that their
hands were empty before inserting them through the cuffs. Thus, these
participants might have concealed target sheets previously used in the
experiment. Another independent researcher has checked all items against
the original data and found no evidence of target replacement by any of
the participants. Specically, concealment in the hand would have required
multiple folds in the target sheets, and this was never found. Third, only 87
of 1,395 trials were videotaped, and no peeking was found in any of these
trials. Although all of the participants were continuously observed by two
researchers, there could have been moments when their attention lapsed.
Finally, it should be mentioned that the study had the problem of
optional stopping. That is, the number of trials was not prespecied and
varied markedly from one participant to another. Thus, the participants, who
received immediate feedback could have chosen to terminate the testing
early to achieve optimal scores. This aw introduces a statistical artifact that
could have biased the results. However, given that no participants achieved
signicant scoring, this concern can be disregarded.
EXPERIMENT 2
In order to maximize nger-reading results, the author drew on the
model of Tart applied in his learning studies by selecting possibly talented
participants through three testing phases: a selection study (SS), a conrmation
study (CS), and a training study (TS) (Honorton, Davidson, & Bindler, 1971;
Tart, 1966, 1977b; Tart et al., 1979). It was also hoped that more participants
would complete the full training regimen suggested by Lee.
The experiment was conducted between February 2006 and May
2006. The SS consisted of 10 trials and the CS consisted of 20 trials. Both
took the participants about 3 hr to complete. Based on Lees nding that
some participants seemingly showed nger-reading ability after only 20 min
practice with immediate trial-by-trial feedback (Shiah & Tam, 2005), the
longer period of 3 hr seemed to allow ample time for gifted participants to
demonstrate their ability.
Hypotheses
The experiment was designed to test most of the formal hypotheses
that were tested in the previous experiment. All the hypotheses apply to
identication of number plus color.
H1. The SS and CS will select gifted participants from a large
sample.
123 The Finger-Reading Effect With Children
H2. Participants will demonstrate signicant above-chance scoring
in the TS.
H3. Participants who achieve a signicant result in the TS will have
on hitting trials a visual image, particularly of a transparent screen, while
touching the target.
H4. Participants who see a screen in the TS will show signicantly
improved scoring compared to children not reporting a screen.
An important prediction of Tarts ESP learning theory is that
there should be a positive correlation between pretraining ESP talent and
improvement during training (Tart et al., 1979). Therefore:
H5. There will be a signicant positive relationship between scores
in the SS and scores in the CS plus TS.
H6. There will be a signicant positive relationship between scores
in the CS and scores in the TS.
SELECTION STUDY (SS)
The purpose of the SS was to screen a large number of participants
to nd those who had possible nger-reading ability.
Method
All of the experimental procedures for all of the session types were
approved by the Ethics Committee of Psychology at Edinburgh University

for the protection of human

participants.
Participants. A total of 2,200 children, 6 to 13 years old, from the
Hemei primary school in Taiwan were invited to participate. Of the 1,771
who began testing, 1,655 completed 10 trials. The mean age of this latter
group was 9.33 (SD = 1.63); there were 836 boys and 819 girls. None of
them had a history of nerve or brain injury, nger trauma,

or learning
disability (including dyslexia), and their nger pads were free of calluses.

Participants who had diabetes were also excluded because of

associated
peripheral neuropathy. The parents of the children signed an informed
consent form before the experiment.
Materials and procedure. The participants were asked to complete 10
trials. Due to limited manpower and time, the trials were not videotaped.
The children were tested in groups of 30 in a classroom by one research
assistant. The screen and bags were not used in this study. Instead, each of
the 10 target sheets was folded twice, bound by a rubber band, and put in a
drawer in the participants desk. The participants were instructed to touch
each target sheet in the drawer without looking at it. At the end of each
trial, they wrote down their response on a record sheet before taking the
next item from the drawer. Otherwise, the procedure was the same as in the
formal sessions of Experiment 1.
124 The Journal of Parapsychology
Undoubtedly, sensory leakage was a problem in this study. One
research assistant could not carefully observe each participants responses
and behaviour to rule out peeking. Although the target was folded twice to
prevent seeing and peeking, a possibility was that the research assistant and
participants might see a mark on the outside without the barriers of the bag
and screen.
Results
Surprisingly, 820 of the 1,655 participants scored at least one hit on
number with color, as seen in Table 7. Thus, Hypothesis 1 was conrmed
for the SS.
TABLE 7
HITS FOR NUMBER WITH COLOR IN THE SS (N = 820)
Hits Number of participants
1 319
2 164
3 112
4 64
5 61
6
7
8
9
10
34
25
17
14
10
CONFIRMATION STUDY (CS)
The purpose of the CS was to identify and separate those participants
who scored signicantly in the SS.
Method
Participants. The 820 participants who got one hit or more on
number with color in the SS (p < .05, binomial analysis, two-tailed) were
invited to join the CS. Of these, 728 completed their session. They ranged
from 6 to 13 years in age (M = 9.12, SD = 1.52). There were 350 boys and
378 girls.
Materials and procedures. The procedures and materials were the
same as used in the more tightly controlled Experiment 1 except that the
target sheets were folded and bound as in the SS of Experiment 2 and
125 The Finger-Reading Effect With Children
placed in a large envelope (of the same type as used in Experiment 1); the
folded sheets were not placed in individual small envelopes. Also, there was
no videotaping.
Results
Not a single one of the children obtained a signicant result in the
CS for number plus color. Thus, Hypothesis 1 was not conrmed for the CS,
and the other hypotheses could not be tested. Also, there could be no TS.
DISCUSSION
The results of the SS and CS provide a striking contrast of
extremes, even considering that the probability of a hit on number
plus color is low. Over half of the children tested in the SS attained a
signicant result, but none of these achieved a signicant score in the
CS. One possible explanation for these remarkable results is that sensory
leakage led to peeking in the SS, which was precluded in the CS because
of the reintroduction of the screen. The SS results also show that children
from this population have a strong tendency to cheat on a nger-reading
task if given the opportunity.
Si-Chen Lee (personal communication, April 22, 2006) pointed
out that another possible explanation might be the effect of different
environmental settings, the most notable instance of which was the addition
of the screen in the CS. Based on his more than 10 years of experience
with nger-reading training, Lee believes that changes in the setting take
more time for children to get used to than is provided for in the training
protocol. The reasons why different settings might decrease participants
performance are still unclear; however, many children in the CS reported
that the screen increased the difculty of recognizing targets. Some also
reported that it was more difcult to have visual experiences while touching
the targets in this condition.
GENERAL DISCUSSION
No signicant overall scoring was observed in either of the two
nger-reading studies, which seems to contradict the ndings reported
by Lee. Thus, these studies seem to indicate that nger-reading ability
does not exist. However, given the fact that some researchers (Carpenter,
2004, 2005; Palmer, 2003; Walach & Schmidt, 2005) have argued that ESP
phenomena are real, there are some possible alternative explanations for
the present studys unsuccessful replication of Lees ndings. One of these is
that ESP, although real, is elusive and weak (Alcock, 2003; Kennedy, 2001).
For instance, only 1% of unselected adult participants have performed
successfully in remote viewing tasks (Utts, 1995).
126 The Journal of Parapsychology
According to Tarts ESP learning theory, only talented participants
can benet from feedback interventions. It might be that no participants
in these two studies had psi ability, so that none could benet from the
training or practice. Indeed, Lee (1988) has stated that if no talented
participants are found, it might take hundreds of hours to train nger-
reading ability, which means that it might take months or even years
to learn nger-reading. It is necessary to further explore the required
period of training. There is no way to objectively dene when to stop
training if no progress is being made, and researchers might just have
to make an arbitrary decision. There should be a cutoff that researchers
use consistently so the procedure can be dened for someone who might
want to replicate the study.
More generally, we still do not have a clear indication of whether
ESP can be trained by any method. Proposed ESP training processes need
to be specied to provide testable assumptions.
REFERENCES
Alcock, J. E. (2003). Give the null hypothesis a chance: Reasons to remain
doubtful about the existence of psi. Journal of Consciousness Studies,
10, 2950.
Bem, D. J., & Honorton, c. (1994). Does psi exist? Replicable evidence of
an anomalous process of information transfer. Psychological Bulletin,
115, 418.
Bem, D. J., Palmer, J., & Broughton, R. S. (2001). Updating the ganzfeld
database: A victim of its own success? Journal of Parapsychology, 65,
207218.
Blackmore, S. J., & Rose, N. (1997). Reality and imagination: A psi-
conducive confusion? Journal of Parapsychology, 61, 321335.
Carpenter, J. C. (2004). First sight: Part one, a model of psi and the mind.
Journal of Parapsychology, 68, 217254.
Carpenter, J. C. (2005). First sight: Part two, elaboration of a model of psi
and the mind. Journal of Parapsychology, 69, 63112.
Du, L. (2005). Taiwan: New university president has links to paranormal
research. Science, 308, 1852.
George, L. (1982). Enhancement of psi functioning through mental-
imagery training. Journal of Parapsychology, 46, 111125.
Honorton, C. (1975). Psi and mental imagery: Keeping score on Betts scare.
Journal of the American Society for Psychical Research, 69, 327332.
Honorton, C., Davidson, R., & Bindler, P. (1971). Feedback-augmented EEG
alpha, shifts in subjective state, and ESP card-guessing performance.
Journal of the American Society for Psychical Research, 65, 308323.
Honorton, C., Tierney, L., & Torrey, D. (1974). The role of mental
imagery in psi-mediation. Journal of the American Society for Psychical
Research, 68, 385394.
127 The Finger-Reading Effect With Children
Hyman, R. (1985). The ganzfeld psi experiment: A critical appraisal. Journal
of Parapsychology, 49, 349.
Hyman, R. (1994). Anomaly or artifact? Comments on Bem and Honorton.
Psychological Bulletin, 115, 1924.
Irwin, H. J., & Watt, C. A. (2007). An introduction to parapsychology (5th
ed.). Jefferson, NC: McFarland.
Kennedy, J. E. (2001). Why psi so elusive? A review and proposed model.
Journal of Parapsychology, 65, 219245.
Lee, S.-C. (1998). Mechanism of nger reading (the third eye) and
physiological measurements. Journal of Chinese Somatic Science, 8,
105113 (in Chinese).
Lee, S.-C. (1999). Unication of mind and material: The macroscopic
quantum phenomena. Journal of Chinese Somatic Science, 9, 124128
(in Chinese).
Lee, S.-C. (2002). The dialogue with the spirit. Buddhism and Science, 3, 68
79 (in Chinese).
Lee, S.-C. (2003). Traveling around the visible and invisible worlds. Buddhism
and Science, 4, 3244 (in Chinese).
Lee, S.-C., & Chang, L.-S. (2001). The connecting model between keywords
and information eld in nger reading experiments. Buddhism and
Science, 2, 6077 (in Chinese).
Lee, S.-C., Chen, C.-T., & Tang, D. (2000). The observation of extraordinary
phenomena when special keywords were tested in the experiment
on character recognition by ngers. Buddhism and Science, 1, 817
(in Chinese).
Lee, S.-C., Tang, D., & Kuo, H. (2004). The possible connection between
information eld and dark matter and dark energy. Buddhism and
Science, 5, 4758 (in Chinese).
Milton, J., & Wiseman, R. (1999). Does psi exist? Lack of replication of
an anomalous process of information transfer. Psychological Bulletin,
125, 387391.
Milton, J., & Wiseman, R. (2001). Does psi exist? Reply to Storm and Ertel
(2001). Psychological Bulletin, 127, 434438.
Oldfield, R. C. (1971). The assessment and analysis of handedness: The
Edinburgh inventory. Neuropsychology, 9, 97113.
Palmer, J. (2003). ESP in the ganzfeld: Analysis of a debate. Journal of
Consciousness Studies, 10, 5168.
Price, T. W. (1973). Subjects control of imagery, agents, mood, and position
effects in a dual-target ESP experiment. Journal of Parapsychology,
37, 298322.
Robinson, E. A. (1985). Probability theory and applications. Boston:
International Human Resources Development Corporation.
Schechter, R., Solfvin, G., & McCollum, R. (1975). Psi and mental imagery.
Journal of the American Society for Psychical Research, 69, 321326.
128 The Journal of Parapsychology
Schmeidler, G. R. (1952). Personal values and ESP scores. Journal of
Abnormal and Social Psychology, 47, 757761.
Shiah, Y.-J. (2005). Toward a replication of the nger-reading effect.
Journal of Parapsychology, 69, 353375.
Shiah, Y.-J., & Tam, W.-C. C. (2005). Do human ngers see? Finger-
reading studies in the East and West. European Journal of
Parapsychology, 20, 117134.
Smith, M. D. (2003). The role of the experimenter in parapsychological
research. Journal of Consciousness Studies, 10, 6984.
Storm, L., & Ertel, S. (2001). Does psi exist? Comments on Milton and
Wisemans (1999) meta-analysis of ganzfeld research. Psychological
Bulletin, 127, 424433.
Tang, D., Lee, S.-C., & Hsu, L.-C. (2000). Comparison between normal
color sense and ESP color sense in color adaptation task: A case
study. Journal of Chinese Somatic Science, 10, 2128 (in Chinese).
Tart, C. T. (1966). Card guessing tests: Learning paradigm or extinction
paradigm? Journal of the American Society for Psychical Research, 60,
4655.
Tart, C. T. (1975). Learning to use extrasensory perception. Chicago: University
of Chicago Press.
Tart, C. T. (1977a). Consideration of internal process in using immediate
feedback to teach ESP ability. Journal of Parapsychology, 41, 255
256.
Tart, C. T. (1977b). Toward conscious control of psi through immediate
feedback training: Some considerations of internal processes.
Journal of the American Society for Psychical Research, 71, 375407.
Tart, C. T. (1986). Stopping on a hit: Preliminary studies of a method for
producing positive experiences in the parapsychology laboratory.
Journal of the American Society for Psychical Research, 80, 3148.
Tart, C. T., Palmer, J., & Redington, D. J. (1979). Effects of immediate
feedback on ESP performance: Second study. Journal of the American
Society for Psychical Research, 73, 151165.
Thalbourne, M. A. (1981). Extraversion and the sheep-goat variable: A
conceptual replication. Journal of the American Society for Psychical
Research, 75, 105119.
Utts, J. M. (1995). An assessment of the evidence for psychical functioning.
Journal of Parapsychology, 59, 289320.
Walach, H., & Schmidt, S. (2005). Repairing Platos life boat with
Ockhams razor: The important function of research in anomalies
for consciousness studies. Journal of Consciousness Studies, 12, 52
70.
Watt, C., & Ramakers, P. (2003). Experimenter effects with a remote
facilitation of attention focusing task: A study with multiple
believer and disbeliever experimenters. Journal of Parapsychology,
67, 99116.
129 The Finger-Reading Effect With Children
Psychology Department,
Kaohsiung Medical University, Taiwan
shiah@kmu.edu.tw
ACKNOWLEDGMENTS
The rst experiment was supported by a grant from the Koestler
Parapsychology Unit. The second experiment was supported by a grant from
the Society for Psychical Research, UK. The author wishes to thank Si-Chen
Lee for his support and his valuable comments on this paper. The author
is grateful to Caroline Watt for her helpful suggestions on a draft of the
manuscript, Andy McKinlay for his support, and Paul Stevens for designing
the computer generator. The author wishes to thank Wan-Yu Hung and
Chien-Ting Huang for preparing the targets. The author also gratefully
acknowledges Wendy Martin, Peter Ramakers and Mary McDonald for their
help as coexperimenters during Experiment 1. The author also gratefully
acknowledges John Palmer, editor of the Journal, and anonymous referees
for their helpful comments.
Abstracts in Other Languages
Spanish
EL EFECTO DE LECTURA POR LOS DEDOS CON NIOS:
DOS REPLICACIONES SIN XITO
RESUMEN: Este artculo informa sobre dos intentos para replicar observaciones
de lectura por los dedos (fnger-reading) bajo condiciones controladas. En el
primer experimento participaron 18 nios entre 7 y 12 aos de edad los cuales
fueron examinados individualmente para determinar si podan identifcar por
tacto un nmero de dos dgitos de uno de cuatro colores diferentes impresos en
una hoja de papel sin el uso de indicaciones visuales. No se obtuvo el efecto y
no hubo relaciones signifcativas entre la puntuaciones de lectura por los dedos
y respuestas a cuestionarios sobre habilidades relacionadas a imgenes mentales
y sobre creencias paranormales. En el segundo experimento se intento estudiar
participantes talentosos en tres etapas: un estudio de seleccin (ES), un estudio
confrmatorio (EC), y un estudio de entrenamiento (EE). 2,200 nios entre 6 y
13 aos fueron invitados a participar en el ES, de los cuales 1,771 participaron;
1,665 hicieron 10 pruebas, y 820 obtuvieron uno o ms aciertos reconociendo un
nmero de dos dgitos con color; 728 hicieron 20 pruebas en el EC con el mismo
objetivo. Ninguno tuvo aciertos, por los cual no hubo un EE. Los controles de
indicios sensoriales fueron menos estrictos en el ES que en el EC, lo cual podria
explicar la gran diferencia. Los dos experimentos para seleccionar nios con el
potencial para desarrollar habilidad de lectura por los dedos no fueron exitosos.
Se ofrecen sugerencias para futuras investigaciones.
130 The Journal of Parapsychology
German
DER EFFEKT DES LESENS MIT DEN FINGERN BEI KINDERN:
ZWEI NICHTERFOLGREICHE REPLIKATIONEN
ZUSAMMENFASSUNG: Der Artikel berichtet ber zwei versuchte Replikationen
des Lesens mit den Fingern unter kontrollierten Bedingungen. Im Ersten
Experiment wurden 18 Kinder im Alter zwischen 7 und 12 Jahren einzeln getestet,
um herauszufnden, ob sie in der Lage seien, eine zweistellige Zahl unter einer von
4 verschiedenen Farben aufgedruckt auf einem Blatt Papier ohne visuelle Reize zu
identifzieren. Kein signifkanter Effekt des Fingerlesens konnte festgestellt werden,
und es zeigten sich auch keine signifkanten Beziehungen zwischen den Scores im
Fingerlesen und den Fragebogenantworten in Bezug auf visuelle Vorstellungen
und paranormalen Einstellungen. Im Zweiten Experiment wurde der Versuch
unternommen, begabte Versuchsteilnehmer in 3 Schritten herauszufltern: einer
Selektionsstudie (SS), einer Besttigungsstudie (BS) und einer Trainingsstudie
(TS): 2.200 Kinder im Alter zwischen 6 und 13 Jahren wurden zur SS eingeladen,
1.771 nahmen daran teil; 1.655 absolvierten 10 Versuchsdurchgnge, von denen
820 einen oder mehr Treffer bei der Identifzierung einer farbigen zweistelligen
Zahl erzielten; 728 von diesen absolvierten alle 20 Versuche in der BS mit dem
gleichen Zielmaterial. Da keiner irgendwelche Treffer erzielte, unterblieb die TS.
Die Kontrollen zur Ausschaltung sensorischer Lecks waren in der SS weniger
streng als in der BS, was die grosse Trefferdifferenz erklren knnte. In beiden
Experimenten gelang es nicht, Kinder zu fnden, die das Potential zur Entwicklung
der Fhigkeit zum Fingerlesen zeigten. Vorschlge fr zuknftige Forschungen
werden gemacht.
French
LEFFECT DE LECTURE DIGITALE AVEC LES ENFANTS :
DEUX REPRODUCTIONS SANS SUCCES
SOMMAIRE : Cet article raconte deux essais de reproduire la capacit pour les
enfants de lire avec leurs doigts sous les conditions contrles. Dans lexprience
l, 18 enfants entre 7 et 12 ans ont t testes individuellement pour dterminer
sils pouvaient, uniquement par la touche et sans les yeux, identifer un numro
de deux chiffres imprimes en 1 des 4 diffrentes couleurs sur une feuille de
papier. Aucun effet important de lecture digitale a t constate. Aucune relation
importante entre capacit de lire avec les doigts et les rponses dun questionnaire
sur la reconnaissance dimage et les croyances du paranormal. Dans lexprience
2, un essai a t fait pour dpister des participants talentueux en trois tapes : une
tude de slection (SS), une tude de confrmation (CS), et une tude dinstruction
(TS); 2,200 enfants ages de 6 a 13 ans ont t invites a participer dans le SS, donc
1,771 lont fait ; 1,655 ont complte 10 preuves, et 820 ont dmontre au moins
1 fois ou plus leurs capacits de reconnatre un numro de deux chiffres avec la
couleur ; 728 de ces enfants la ont complte les 20 preuves dans le CS avec le
131 The Finger-Reading Effect With Children
mme stimulus. Aucun enfant a t capable de toucher la cible, donc il ny avait
pas de TS. Les contrles des fuites sensorielles taient moins appliques dans le SS
par rapport le CS, ce qui pourrait expliquer les importantes diffrences dans les
rsultats. Les deux exprience nont pas russi a slectionner des enfants avec le
potentiel a dvelopper une capacit de lire avec les doigts. Les suggestions pour
une recherche futures sont fournies.
APPENDIX 1
PARTICIPANT QUESTIONNAIRE
1. Name:
2. Address:
3. Phone: (work) (home)
4. Sex: Male Female
5. Age:
6. Educational background and/or vocational training:
7. On average how many hours do you currently use a keyboard (computer or
typewriter) per day?
8. How easy is it for you to create a mental image of a familiar scene?

Impossible Effortless
9. If you can create a mental image of a familiar scene, how clearly can you see
the scene?

Not clear at all As clear as using
normal vision
10. How well can you hear a sound in a mentally imaged scene?

Not at all Very well
11. How well can you imagine smelling something?

Not at all Very well
12. How easily can you experience taste in a mentally imagined scene?

Not at all Very easily
13. How well can you imagine feeling something through touching?

Not at all Very well
14. If you touch something that you cannot see, such as in a darkened room, do
you get a visual image of it?

Not at all Very often
If yes, how clear is it?

132 The Journal of Parapsychology
15. How good do you think your touch sensitivity is?

None at all Very good
16. Do you have a job or hobby that involves touch sensitivity? Please specify:




17. Does ESP exist (extrasensory perception: reception of information without
the use of known senses or logical inference)?

Certain Uncertain Impossible
18. Have you ever had an experience which is best explained by telepathy (ESP
of the thoughts, feelings or behaviour of another person or organism)?

Yes Uncertain No
19. Have you ever had an experience which is best explained by clairvoyance
(ESP of distant physical events or concealed objects)?

Yes Uncertain No
20. Have you ever had an experience which is best explained by precognition
(ESP of future)?

Yes Uncertain No
21. Does PK exist (psychokinesis: mental infuence on the physical world)?

Certain Uncertain Impossible