Asking Questions Changes Behavior: Mere Measurement Effects on Frequency of Blood Donation

Article (PDF Available)inHealth Psychology 27(2):179-84 · April 2008with661 Reads
DOI: 10.1037/0278-6133.27.2.179 · Source: PubMed
Abstract
This research examined the impact of completing a questionnaire about blood donation on subsequent donation behavior among a large sample of experienced blood donors. Participants (N=4672) were randomly assigned to an experimental condition that received a postal questionnaire measuring cognitions about donation or a control condition that did not receive a questionnaire. Number of registrations at blood drives and number of successful blood donations were assessed using objective records both 6 months and 12 months later. Findings indicated that, compared to control participants, the mean frequency of number of registrations at blood drives among participants in the experimental group was 8.6% greater at 6 months (p<.0.007), and was 6.4% greater at 12 months (p<.035). Significant effects were also observed for successful blood donations at 6 months (p<.001) and 12 months (p<.004). These findings provide the first evidence that the mere measurement is relevant to promoting consequential health behaviors. Implications of the research for intervention evaluation are discussed.
Asking Questions Changes Behavior: Mere Measurement Effects on
Frequency of Blood Donation
Gaston Godin
Laval University
Paschal Sheeran
University of Sheffield
Mark Conner
University of Leeds
Marc Germain
He´ma-Que´bec
Objective: This research examined the impact of completing a questionnaire about blood donation on
subsequent donation behavior among a large sample of experienced blood donors. Design: Participants
(N 4672) were randomly assigned to an experimental condition that received a postal questionnaire
measuring cognitions about donation or a control condition that did not receive a questionnaire. Main
Outcome Measures: Number of registrations at blood drives and number of successful blood donations
were assessed using objective records both 6 months and 12 months later. Results: Findings indicated
that, compared to control participants, the mean frequency of number of registrations at blood drives
among participants in the experimental group was 8.6% greater at 6 months ( p .0.007), and was 6.4%
greater at 12 months ( p .035). Significant effects were also observed for successful blood donations
at 6 months ( p .001) and 12 months ( p .004). Conclusion: These findings provide the first evidence
that the mere measurement is relevant to promoting consequential health behaviors. Implications of the
research for intervention evaluation are discussed.
Keywords: cognitions, questionaire, mere measurement, behavior, blood donation
Predicting and changing behaviors that affect health outcomes is
a key goal for health psychologists. The usual paradigm in research
on health behavior change involves (a) identifying cognitions that
predict the focal behavior among one’s target population using a
prospective survey, (b) designing an intervention to change the key
cognitive predictors based on the survey findings, and (c) evalu-
ating the impact of the intervention on subsequent behavior (e.g.,
Schaalma et al., 1996). To identify cognitions that should be
measured in the survey, researchers typically draw on established
health behavior models such as the theory of planned behavior
(TPB; Ajzen, 1991), the model of interpersonal behavior (MIP;
Triandis, 1980) or social cognitive theory (SCT; Bandura, 1997).
These models construe people’s intention (e.g., “I intend to donate
blood during the next six months”) as the key predictor of action.
Thus, surveys designed to find out what cognitions predict the
focal behavior—to inform ensuing interventions designed to
change that behavior—usually ask participants about their behav-
ioral intentions. But what if the simple act of asking intention
questions changes behavior? This would suggest that inviting
people to answer intention questions in relation to a particular
health behavior could in itself constitute a useful strategy for
promoting performance of behavior. The aim of the present re-
search was to test this idea in relation to frequency of blood
donation.
In fact, studies in social and consumer psychology have ob-
served that asking participants to report their intentions causes
changes in subsequent behavior (e.g., Feldman & Lynch, 1988;
Greenwald, Carnot, Beach, & Young, 1987; Morwitz, Johnson, &
Schmittlein, 1993; Sherman, 1980). This phenomenon is known as
the mere measurement effect (Morwitz et al., 1993), though other
terms such as measurement reactivity, self-erasing errors of pre-
diction, and self-generated validity also have been used. In Sher-
man’s (1980) original demonstration of the effect, one group of
participants was asked to predict how likely they would be to
perform a socially desirable or socially undesirable behavior (vol-
unteering for the American Cancer Society or singing the Star
Spangled Banner down the phone, respectively); a second group
made no prediction about their behavior. Findings showed that
participants who were asked to predict their behavior subsequently
were more likely to perform the socially desirable behavior (31%
vs. 4%) and were less likely to perform the undesirable behavior
(40% vs. 68%) compared to control participants who made no
prediction. The mere measurement of participants’ intentions af-
fected their future behavior.
Subsequent studies have demonstrated mere measurement
effects in both laboratory and field settings. For instance,
Greenwald et al. (1987) reported that students who were asked
about their intentions to vote in the following day’s elections
were more likely to do so than were students who were not
queried about their voting intentions. Similarly, Morwitz et al.
Gaston Godin, Canada Research Chair on Behaviour and Health, Fac-
ulty of Nursing, Laval University; Paschal Sheeran, Department of Psy-
chology, University of Sheffield; Mark Conner, Institute of Psychological
Studies, University of Leeds; Marc Germain, Microbiology and Epidemi-
ology, He´ma-Que´bec.
Correspondence concerning this article should be addressed to Gaston
Godin, Canada Research Chair on Behavior and Health, Laval University,
Quebec, Canada G1K 7P4. E-mail: gaston.godin@fsi.ulaval.ca
Health Psychology Copyright 2008 by the American Psychological Association
2008, Vol. 27, No. 2, 179 –184 0278-6133/08/$12.00 DOI: 10.1037/0278-6133.27.2.179
179
(1993) found that a simple question about purchase intent
increased rates of purchase of both automobiles and personal
computers (see Chandon, Morwitz, & Reinartz, 2004; Dholakia
& Morwitz, 2002; Spangenberg & Greenwald, 1999, for equiv-
alent findings). Finally, intention questions influenced students’
subsequent brand choices in laboratory experiments of con-
sumer behavior (Fitzsimons & Williams, 2000; Morwitz &
Fitzsimons, 2004).
The dominant explanation of the mere measurement effect is
that asking a behavioral intention question heightens the acces-
sibility of the person’s attitude towards the behavior, which in
turns increases the likelihood that the behavior will be per-
formed (Morwitz & Fitzsimons, 2004). Morwitz and Fitzsimons
showed that responding to a query about one’s purchase intention
increases the activation level of one’s preexisting brand attitude.
When the respective brand attitude was both highly accessible and
positively valenced, participants were likely to choose that brand.
When the activated attitude was negatively valenced, on the other
hand, this led to a decrease in the choice of this brand. Subsequent
experiments by Fitzsimons and Williams (2000) indicated that the
operation of mere measurement effects is nonconscious and auto-
matic. Findings from a study that used a process dissociation
procedure to delineate the relative impact of controlled versus
effortless processing (Jaccoby, 1991) showed that the mere mea-
surement effect does not arise from careful consideration of exist-
ing knowledge and attitudes, but rather is due to the automatic
activation of these mental structures.
Rationale for the Present Research
Further research on mere measurement effects is warranted, for
two reasons. First, previous studies have examined consumer,
political, and laboratory task intentions (see references above).
However, we were unable to locate any studies that have examined
this phenomenon in relation to consequential health behaviors.
Because health behavior change is an important goal for health
psychologists, and mere measurement is potentially a simple and
inexpensive intervention strategy, it is therefore worthwhile re-
searching whether the simple act of asking an intention question
changes behavior.
The present research concerned blood donation. Blood transfu-
sion is a universal medical intervention that saves 4.5 million lives
in the USA each year (America’s Blood Centers, 2006). However,
the number of active blood donors is small (approximately 3% of
the population; e.g., Davey, 2004), and these donors give blood
much less frequently than is (a) desirable in terms of blood supply,
or (b) feasible in terms of their medical eligibility (e.g., Ferguson,
1996). Moreover, the number of first-time donors is decreasing
(Wu et al., 2001). Thus, effective interventions to increase the
frequency of blood donations are needed urgently (Davey, 2004).
The second reason why further tests of the mere measurement
effect are warranted relates to issues surrounding the methods and
findings of previous studies. Most previous studies of mere mea-
surement effects use student samples. Students have superior cog-
nitive, test-taking abilities and less crystallized attitudes compared
to more representative samples (Sears, 1986) and thus may be
more liable to respond to this phenomenon. It is also the case that
several studies have relied on self-report measures of behavior that
leave open the possibility that the observed effects are due to
self-presentational, consistency, or memory biases. Finally, re-
search to date has not clearly established the durability or robust-
ness of mere measurement effects. Chandon et al. (2004) pointed
out that most studies in this area examined a single instance of
behavioral performance, and involved a relatively short time in-
terval between measurement of intention and behavior. Their study
examined repeat purchases at a Web-based grocer over a 9-month
period. Findings showed the expected difference in rates of pur-
chase between mere measurement and control conditions at one
month. However, this difference had disappeared by three months
and remained absent right up to the 9-month follow-up. Thus, there
is some doubt about whether mere measurement effects will be
observed among a non-student sample using objective measures of
health behavior, and whether these effects persist over repeated
measurements and long time periods.
The present study examined mere measurement effects on fre-
quency of blood donation. Participants were randomly assigned
either to a mere measurement condition that received a postal
questionnaire measuring cognitions about blood donation, or to a
control condition that did not receive a questionnaire. Donation
behavior was assessed both 6 months and 12 months later. Thus,
the present research involved repeated behavioral assessments
over a longer period than has been examined in studies of the mere
measurement effect to date. The prediction tested here was that the
mere measurement of cognitions about giving blood would in-
crease the frequency of blood donation over the subsequent year.
Method
Study Population and Data Collection
Participants were blood donors aged 18 to 70 years who met the
medical criteria specified by He´ma-Que´bec, the blood supply
agency of the Province of Que´bec, Canada. Participants all had
given an allogenic blood donation at a blood drive in the week
before the onset of the study (April 21-26, 2003), and were
randomly assigned to the experimental (questionnaire) or control
conditions. The present study formed part of a larger research
project that aimed to understand the psychosocial determinants of
repeated blood donation. Thus, to garner the requisite question-
naire responses relevant to achieving this aim, a larger proportion
of the sample were assigned to the experimental condition (n
2900) compared to the control condition (n 1772). Assignment
to conditions was random, but a specified quota of the sample
received the questionnaire.
Participants in the experimental group were mailed a question-
naire that measured several social cognition constructs (e.g., be-
liefs, attitudes, intention) regarding blood donation during the next
six months. The survey procedure was similar to that specified by
Dillman (2000). The research was approved by the Ethics Com-
mittees of both Hema-Quebec and Laval University. A letter
signed by the CEO of He´ma-Que´bec was mailed to participants in
the experimental group informing them that they would shortly
receive a questionnaire on blood donation. One week later, the
questionnaire and a stamped addressed envelope were distributed.
One week and three weeks after the questionnaire was mailed,
reminders and thank you notes were sent systematically to all 2900
participants. The last step of Dillman’s (2000) procedure involves
distributing a second questionnaire; however, this step was not
180
GODIN, SHEERAN, CONNER, AND GERMAIN
carried out to reduce expense. Of the 2900 questionnaires mailed,
2389 were completed, representing a response rate of 82.4%.
Questionnaire
The questionnaire measured variables specified by an extended
version of the theory of planned behavior. Three items were used
to index intention: “I intend to give blood during the next six
months”, “I will give blood during the next six months”, and “I
will try to give blood during the next six months”, all measured on
5-point scales (very unlikely-very likely). The three key predictors
of intention according to the TPB—attitude, subjective norm and
perceived behavioral control—also were measured. The additional
predictors of intention included in the present theoretical frame-
work were self-efficacy, moral norm, role beliefs, self-identity,
anticipated regret, and satisfaction with last blood donation. These
variables were all assessed using three or more items. The ques-
tionnaire comprised 77 items in total.
Dependent Variables
Two objective measures of behavior were extracted from blood
agency data. The first measure was the number of registrations at
a blood drive. This measure comprises both individuals who gave
blood and those who wanted to give blood but were not allowed to
do so because they failed specified medical criteria. This is the key
dependent variable because it takes account of the fact that people
had acted towards the goal of giving blood. The second measure
was the number of successful blood donations, which excludes
individuals whose blood were refused at respective blood drives.
For both of these measures of blood donation behavior, the short-
term effect was assessed at six-months (this was the time period
specified in the questionnaire measures), and the long-term effect
was assessed at one-year. Donation information was obtained in a
manner that ensured anonymity, and only medically trivial infor-
mation was provided about the control participants. Participants’
identities were not known to the researchers, and the blood supply
agency did not have access the questionnaire data. An anonymous
research code was used to link individuals with their behavioral
data.
Results
Overview of the Analyses
An intention-to-treat approach was adopted in the analyses. That
is, data from all participants in the experimental condition were
included whether or not relevant participants completed the ques-
tionnaire. The first set of analyses was a randomization check that
compared experimental versus control group on relevant demo-
graphic and background variables. The main analyses examined
the effects of mere measurement on the number of registrations
and successful donations at the 6-month and 12-month assess-
ments. The final set of analyses was explanatory and examined the
impact of completing versus not completing the questionnaire on
donation behavior.
Randomization Check
Comparisons between the experimental and control group on
background and demographic characteristics, i.e., age, sex, lan-
guage (French versus English), region of recruitment (Montreal,
Quebec, otherwise), and previous donation behavior, indicated that
randomization was, for the most part, successful (see Table 1).
However, participants in the experimental group were slightly
older than control participants (Ms 44.7 and 43.8, respectively,
p .001). Moreover, age was significantly correlated with fre-
quency of blood donation at 6-months (r .22, p .001).
Consequently, age was covaried in subsequent analyses. Homoge-
neity analyses conducted on the slope for the covariate indicated
that analysis of covariance (ANCOVA) was appropriate.
Effects on Blood Donation Behavior
The mean intention score among participants in the experi-
mental condition was high (M 4.50; SD 0.69 on a 5-point
scale) and attitude scores were also positively valenced (M
4.06; SD 0.59 on a 5-point scale). These values suggest that
participants in the experimental condition held cognitions to-
wards blood donation that could engender mere measurement
effects (Morwitz & Fitzsimons, 2004). Consistent with this
idea, the proportion of participants who registered at least once
for blood donation at 6 months was 49.2% in the control group
and 53.7% in the experimental group, respectively. At 12
months, these proportions were 65.2% (control) and 69.9%
(experimental), respectively. Chi-square analyses indicated that
these differences were significant at both time-points,
2
(1,
N 4672) 8.91 and 11.15, respectively, p .005. The
effects of condition also remained significant in logistic regres-
sions of behavior on condition that controlled for age (Betas for
condition .15 and .19, respectively, p .01).
Although incidence of blood donation is important, more im-
portant from the point of view of blood supply is the number of
registrations and successful donations. ANCOVA showed that,
compared to control participants, the mean number of registrations
at blood drives was significantly higher among participants in the
experimental group at 6 months, F(1, 4669) 7.18, p .007, and
at 12 months, F(1, 4669) 4.46, p .035 (see Table 2). These
Table 1
Demographic and Background Characteristics for Experimental
and Control Conditions
Variables
Control group
(N 1,772)
Experimental group
(N 2,900)
pnMor %nMor %
Age (years) 1772 43.8 12.1 2900 44.7 11.8 .01
Sex (%) Ns
Female 686 38.7 1110 38.3
Male 1086 61.3 1790 61.7
Language (%) Ns
English 171 9.6 268 9.2
French 1601 90.4 2632 90.8
Region (%) Ns
Montreal 792 44.8 1255 43.3
Quebec 295 16.7 517 17.8
Other 683 38.6 1128 38.9
Blood donation
during past
24 months 1772 2.26 2900 2.40 Ns
181
MERE MEASUREMENT EFFECTS
findings indicated that asking questions increased the frequency of
registration in the experimental versus control conditions by 8.6%
and 6.4% in the short-term and long-term, respectively.
Comparison of the experimental and control conditions on the
number of successful blood donations affords the same conclusion
about the impact of asking questions on behavior. Donations were
significantly more frequent in the experimental group at both 6
months, F(1, 4669) 13.24, p .0003, and 12 months, F(1,
4669) 8.42, p .004. Thus, the present findings show clear
evidence of mere measurement effects. Although the magnitude of
the impact of mere measurement on donation behavior was small
in terms of standard estimates of effect size (ds ranged from .08 to
.12), these differences were significant in both statistical and
practical terms. Finally, it is worth noting that the results were not
substantively altered in analyses that controlled for either the
number of times participants’ blood was refused by the agency or
past donation behavior.
Explanatory Analyses
The aim of the explanatory analyses was to examine whether
completing the questionnaire, as opposed to receiving but not
completing the questionnaire, was necessary to obtain mere mea-
surement effects. ANCOVA therefore was used to compare the
number of registrations at both assessments for participants who
completed the questionnaire, participants who did not complete the
questionnaire, and participants who were not sent the questionnaire
(controls). Findings indicated clearly that completion of the ques-
tionnaire was a prerequisite for mere measurement effects. Partic-
ipants who did not complete the questionnaire registered less
frequently compared to participants who completed the question-
naire and participants who did not receive a questionnaire at both
the 6-month assessment (LSM 0.41, 0.84 and 0.69, SE 0.04,
0.02, and 0.02, respectively; p .001 for all comparisons), and the
12-month assessment (LSM 0.88, 1.63, and 1.41, SE 0.06,
0.03, and 0.03, respectively, p .001 for all comparisons). It is
also notable that the difference between experimental and control
conditions was even greater when participants who did not com-
plete the questionnaire were excluded from the analyses. Findings
were identical (in terms of the pattern of means and significant
differences) in equivalent analyses using number of successful
donations as the dependent variable.
Discussion
The present study is the first, to our knowledge, that assessed the
impact of mere measurement on the frequency of performance of
a consequential health behavior, namely, blood donation. There
were grounds for skepticism about whether mere measurement
effects would be observed—because the present study used the
longest term assessment of behavioral impact in research to date,
and eschewed the problems associated with using student samples
and self-reported measures of behavior. Even though the odds
therefore seemed to be stacked against obtaining mere measure-
ment effects, there was clear evidence of the operation of these
effects. Findings in relation to both the number of registrations at
blood drives and the number of successful donations indicated that
receiving a postal questionnaire about blood donation engendered
more frequent performance compared to not receiving that ques-
tionnaire. Moreover, the behavioral effects did not diminish over
time; differences between the experimental and control conditions
were significant at 6-months and remained significant at 12-
months. Thus, the present findings indicate that mere measurement
effects have both reliable and durable effects on frequency of
blood donation.
Most previous research has been conducted in laboratory set-
tings where it is possible to ensure that the participants complete
the questionnaires as instructed. These studies leave open the
question whether it is the mere receipt of a questionnaire measur-
ing intentions (and other constructs) that generates a mere mea-
surement effect, or whether completion of the questionnaire is
necessary. This is a potentially important issue in health behavior
contexts, where one needs to know whether distributing a ques-
tionnaire or ensuring its completion would be necessary to pro-
mote the target behavior. Findings from the present study provided
an answer to this question: Questionnaire completion is a prereq-
uisite for mere measurement effects. Because the mechanism of
mere measurement effects is increased accessibility of attitude
towards the behavior (Fitzsimons & Williams, 2000; Morwitz &
Fitzsimons, 2004), this suggests that the mere receipt of a ques-
tionnaire is not enough to activate respective attitudes; rather, it is
necessary to engage with the questionnaire (complete it) to engen-
der the attitude accessibility levels needed to generate effects on
subsequent behavior (see also Simmons & Prentice, 2006). The
implication is that future studies that use the mere measurement
effect to promote health behaviors will need to ensure good re-
sponse rates to the questionnaire to maximize impacts on behav-
ioral performance.
The present findings have important implications for research
on predicting health behaviors and for the evaluation of behavior
change interventions. Our results are relevant to a recent debate in
Health Psychology initiated by Ogden’s (2003) pragmatic and
conceptual analysis of social cognition models. Ogden argued that
Table 2
Frequency of Registrations at Blood Drives and Successful Donations at 6-Month and 12-Month Assessments for Experimental and
Control Conditions
Variables
6-Month assessment 12-Month assessment
Control
LSM SE
Experimental
LSM SE p
Control
LSM SE
Experimental
LSM SE p
Number of registrations at blood drives 0.70 0.02 0.76 0.02 0.007 1.41 0.03 1.50 0.03 0.035
Number of successful blood donations 0.59 0.02 0.67 0.01 0.001 1.21 0.03 1.33 0.03 0.004
Note. LSM Least squared mean (i.e., mean adjusted for group difference in age); SE Standard error.
182
GODIN, SHEERAN, CONNER, AND GERMAIN
the process of completing a questionnaire measuring behavior-
relevant cognitions determines, rather than merely predicts, future
behavioral performance. Ajzen and Fishbein (2004) refuted this
criticism and cited one of their own studies (Ajzen, Brown, &
Carvajal, 2004) where undergraduates filled out a questionnaire
based on the theory of planned behavior either before or after the
behavior measure. Ajzen and Fishbein (2004) concluded that
“[t]here was no evidence whatsoever that responding to the ques-
tionnaire affected later behavior. . .” (p. 433). Clearly the present
findings are more consistent with Ogden’s position than that of
Ajzen and Fishbein. However, Ajzen et al.’s (2004) data also
clearly demonstrate that mere measurement effects are not ubiq-
uitous. Further research is therefore needed to delineate what
factors (e.g., sample or behavior characteristics) determine the
magnitude of the impact of questionnaire completion on subse-
quent behavior.
Mere measurement effects also have ramifications for behavior
change interventions and randomized controlled trials that include
questionnaire measures. Our findings suggest that quasi-
experimental designs or designs with pre-tests in both intervention
and control groups should be used with caution. Whenever possi-
ble, it would seem desirable that intervention evaluations include
a control condition equivalent to that used in the present study in
order to quantify mere measurement effects. When inclusion of
such a control group is not feasible (e.g., due to financial or
practical constraints), then it will be important to ensure that the
effect size that is anticipated for the intervention is greater than the
average effect size associated with the impact of mere measure-
ment (d .10 in the present study)— otherwise, effects on behav-
ior might be obscured. Indeed, mere measurement effects could
have undermined the impact of numerous interventions in the
literature that have reported small or non-significant impacts on
behavior change.
It is important to acknowledge that the present study was de-
signed to examine the behavioral impact of asking people their
beliefs about blood donation, and does not speak to the processes
responsible for the observed effects on behavior. Thus, our re-
search assumes that heightened accessibility of favourable atti-
tudes explains mere measurement effect (Morwitz & Fitzsimons,
2004); however, we did not test this hypothesis. Similarly, we
assumed that the measurement of behavioral intentions, in partic-
ular, is responsible for the increases in donation behavior—
because it is well established that intention questions change
behavior (Morwitz & Fizsimons, 2004). However, measures of
satisfaction with previous donation and anticipated regret were
also taken in the present study. Dholakia and Morwitz (2002)
showed that asking questions about customer satisfaction affected
subsequent purchase behavior, and Abraham and Sheeran (2003)
found a marginally significant ( p .07) difference in subsequent
behavior for participants who received a questionnaire that contained
anticipated regret items compared to participants who received the
same questionnaire minus the regret items. Thus, it is possible that the
questions about satisfaction or regret caused the effects on behavior,
not the intention questions. The fact that we are unable to pinpoint
precisely which questionnaire items drive measurement effects on
blood donation is an important limitation of the present research.
Further studies are needed to answer questions such as, ”What are the
key constructs that should be measured in order to obtain mere
measurement effects?”, “What is the minimum number of items
needed to observe the effect?”, and “What instructions for question-
naire completion should be given to participants in order to maximize
mere measurement effects on behavior?”
It also should be acknowledged that the present study used a
sample of experienced blood donors, and leaves open the question
whether mere measurement effects of equivalent magnitude would
be obtained among less experienced blood donors or in relation to
other health behaviors. Because we assumed that the mechanism of
mere measurement effects is the activation of favorable attitudes
(Fitzsimons & Williams, 2000; Morwitz & Fitzsimons, 2004), it
follows that the strength of respective effects should depend upon
the distribution of attitudes towards the behavior among the target
sample, and the proportion of the sample that complete the ques-
tionnaire (i.e., the proportion whose attitudes are activated). Mere
measurement effects on subsequent behavior should be enhanced
as the proportion of the sample that holds positive attitudes and
fills in the questionnaire both increase. It is, however, an empirical
question whether the distribution of attitudes and rate of question-
naire completion indeed exert these effects in field settings, and so
this question remains to be answered in future research.
Notwithstanding these limitations and unanswered questions,
the mere measurement of behavior-relevant cognitions had a sig-
nificant and durable impact on blood donation in the present study.
Although the standardized estimate of effect size is small, the
practical importance of this intervention is substantial. Our find-
ings imply that distributing the questionnaire to blood donors
registered with the blood agency would produce 30,000 additional
blood donations. This is the equivalent of having one extra
month’s donations per year that could be translated into 90,000
more life-saving blood transfusions (America’s Blood Centers,
2006). Interventions to promote health behavior change have not
yet developed to the point where we can afford to ignore effects of
this magnitude (see Webb & Sheeran, 2006). Mere measurement
interventions clearly hold the potential to become an important
additional strategy for promoting public health. However, further
research on the efficacy and boundary conditions of mere mea-
surement effects is needed from health psychologists in order to
realize this potential.
References
Abraham, C., & Sheeran, P. (2003). Acting on intentions: The role of
anticipated regret. British Journal of Social Psychology, 42, 495–511.
America’s Blood Centers (2006). 56 Blood Facts. Retrieved April 7, 2006
from http://www.americasblood.org/index.cfm?fuseactiondisplay
.showPage&pageID28
Ajzen, I. (1991). The theory of planned behaviour. Organisational Behav-
iour and Human Decision Processes, 50, 179–211.
Ajzen, I., Brown, T. C., & Carvajal, F. (2004). Explaining the discrepancy
between intentions and actions: the case of hypothetical bias in contin-
gent valuation. Personality and Social Psychology Bulletin, 30, 1108
1121.
Ajzen, I., & Fishbein, M. (2004). Question raised by a reasoned action
approach: Comment on Ogden (2003). Health Psychology, 23, 431– 434.
Bandura, A. (1997). Self-efficacy: The exercise of control. New York:
W. H. Freeman.
Chandon, P., Morwitz, V. G., & Reinartz, W. J. (2004). The short- and
long-term effects of measuring intent to repurchase. Journal of Con-
sumer Research, 31, 566 –572.
Davey, R. (2004). Recruiting blood donors: Challenges and opportunities.
Transfusion, 44, 597– 600.
183
MERE MEASUREMENT EFFECTS
Dholakia, U. M., Morwitz, V. G. (2002). The scope and persistence of
mere measurement effects: Evidence from a field study of customer
satisfaction measurement. Journal of Consumer Research, 29, 159 –167.
Dillman, D. A. (2000). Mail and internet surveys: The tailored design
model. New York: Wiley.
Feldman, J. M., & Lynch, J. G. (1988). Self-generated validity and other
effects of measurement on belief, attitude, intention, and behavior.
Journal of Applied Psychology, 73, 421–435.
Ferguson, E. (1996). Predictors of future behavior: A review of the psy-
chological literature on blood donation. British Journal of Health Psy-
chology, 1, 287–308.
Fitzsimons, G. J., & Williams, P. (2000). Asking questions can change
choice behavior: Does it do so automatically or effortfully? Journal of
Experimental Psychology: Applied, 6, 195–206.
Greenwald, A. G., Carnot, C. G., Beach, R., & Young, B. (1987). Increased
voting behaviour by asking people if they expect to vote. Journal of
Applied Psychology, 72, 315–318.
Jacoby, L. L. (1991). A process dissociation framework: Separating auto-
matic from intentional uses of memory. Journal of Memory and Lan-
guage, 30, 513–541.
Morwitz, V. G., & Fitzsimons, G. J. (2004). The mere measurement effect:
Why does measuring intentions change actual behavior? Journal of
Consumer Psychology, 14, 566 –572.
Morwitz, V. G., Johnson, E., & Schmittlein, D. (1993). Does measuring
intent change behavior? Journal of Consumer Research, 20, 46 61.
Ogden, J. (2003). Some problems with social cognition models: A prag-
matic and conceptual analysis. Health Psychology, 22, 424 428.
Schaalma, H. P., Kok, G., Bosker, R. J., Parcel, G. S., Peters, L., Poelman,
J., et al. (1996). Planned development and evaluation of AIDS/STD
education for secondary school students in the Netherlands: Short-term
effects. Health Education Quarterly, 23, 469 487.
Sears, D. O. (1986). College sophomores in the laboratory: Influences of a
narrow data on social psychology’s view of human nature. Journal of
Personality and Social Psychology, 51, 515–530.
Sherman, S. J. (1980). On the self-erasing nature of errors of prediction.
Journal of Personality and Social Psychology, 39, 211–221.
Simmons, J. P., & Prentice, D. A. (2006). Pay attention! Attention to
primes increases attitude assessment accuracy. Journal of Experimental
Social Psychology, 42, 784 –791.
Spangenberg, E. R., & Greenwald, A. G. (1999). Social influence by
requesting self-prophecy. Journal of Consumer Psychology, 8, 61–89.
Triandis, H. C. (1980). Values, attitudes and interpersonal behavior. In H.
Howe (ed.), Nebraska symposium on motivation, (pp. 195–259). Lin-
coln, University of Nebraska Press.
Webb, T. L., & Sheeran, P. (2006). Does changing behavioral intentions
engender behavior change? A meta-analysis of the experimental evi-
dence. Psychological Bulletin, 132, 249 –268.
Wu, Y., Glynn, S. A., Schreiber, G. G., Wright, D. J., Lo, A., Humphry,
E. L., Kleinman, S. H., & Garratty, G. (2001). First-time blood donors:
Demographic trends. Transfusion, 41, 360 –364.
184
GODIN, SHEERAN, CONNER, AND GERMAIN
    • "The classification of health behavior might also explain why studies and comparisons differ between these two reviews (Rodrigues et al., 2015; Wood et al., 2015). The Wood review classified behavior as health, prosocial, consumer, undesirable/risky behavior, and " other, " resulting in some studies included by the Rodrigues review being classified in other behavioral domains (i.e., prosocial, k = 3; Cioffi and Garner, 1998; Godin et al., 2008 Godin et al., , 2010) and risky behaviors (k = 1; Levav and Fitzsimons, 2006). "
    Full-text · Article · Jun 2016
    • "Although this contribution rate was unequally distributed among our participants (with some participants contributing many order sets and others not contributing any), this contribution rate is higher than contribution rates (3-22%) reported in other studies [36,48] and represents an increase compared to participants' self-reported baseline contribution rate prior to starting this study (3.5%). The mere measurement effect must also be considered as a potential explanation for the increase in intention, use, and contribution rate [49]. "
    [Show abstract] [Hide abstract] ABSTRACT: Background: Wikis have the potential to promote best practices in health systems by sharing order sets with a broad community of stakeholders. However, little is known about the impact of using a wiki on clinicians' intention to use wiki-based order sets. Objective: The aims of this study were: (1) to describe the use of a wiki to create structured order sets for a single emergency department; (2) to evaluate whether the use of this wiki changed emergency physicians' future intention to use wiki-based order sets; and (3) to understand the impact of using the wiki on the behavioral determinants for using wiki-based order sets. Methods: This was a pre/post-intervention mixed-methods study conducted in one hospital in Lévis, Quebec. The intervention was comprised of receiving access to and being motivated by the department head to use a wiki for 6 months to create electronic order sets designed to be used in a computer physician order entry system. Before and after our intervention, we asked participants to complete a previously validated questionnaire based on the Theory of Planned Behavior. Our primary outcome was the intention to use wiki-based order sets in clinical practice. We also assessed participants' attitude, perceived behavioral control, and subjective norm to use wiki-based order sets. Paired pre- and post-Likert scores were compared using Wilcoxon signed-rank tests. The post-questionnaire also included open-ended questions concerning participants' comments about the wiki, which were then classified into themes using an existing taxonomy. Results: Twenty-eight emergency physicians were enrolled in the study (response rate: 100%). Physicians' mean intention to use a wiki-based reminder was 5.42 (SD 1.04) before the intervention, and increased to 5.81 (SD 1.25) on a 7-point Likert scale (P =.03) after the intervention. Participants' attitude towards using a wiki-based order set also increased from 5.07 (SD 0.90) to 5.57 (SD 0.88) (P =.003). Perceived behavioral control and subjective norm did not change. Easier information sharing was the most frequently positive impact raised. In order of frequency, the three most important facilitators reported were: ease of use, support from colleagues, and promotion by the departmental head. Although participants did not mention any perceived negative impacts, they raised the following barriers in order of frequency: poor organization of information, slow computers, and difficult wiki access. Conclusions: Emergency physicians' intention and attitude to use wiki-based order sets increased after having access to and being motivated to use a wiki for 6 months. Future studies need to explore if this increased intention will translate into sustained actual use and improve patient care. Certain barriers need to be addressed before implementing a wiki for use on a larger scale.
    Full-text · Article · May 2016
    • "We report on these findings in the companion process evaluation paper, but note that this may have unintentionally worked as a co-intervention with the feedback report, as it may have acted to focus participant attention on pain assessment. While this was an unexpected effect, not one we intentionally designed into the intervention, similar results have been observed previously , and there is a literature on what is called " mere measurement " or the " question-behavior effect " [29,30]. This is an intriguing finding that deserves follow-up work in future implementation studies. "
    [Show abstract] [Hide abstract] ABSTRACT: Background There is considerable evidence about the effectiveness of audit coupled with feedback for provider behavior change, although few feedback interventions have been conducted in long-term care settings. The primary purpose of the Data for Improvement and Clinical Excellence-Long-Term Care (DICE-LTC) project was to assess the effects of a feedback intervention delivered to all direct care providers on resident outcomes. Our objective in this report is to assess the effect of feedback reporting on rates of pain assessment, depression screening, and falls over time.Methods The intervention consisted of monthly feedback reports delivered to all direct care providers, facility and unit administrators, and support staff, delivered over 13 months in nine LTC units across four facilities. Data for feedback reports came from the Resident Assessment Instrument Minimum Data Set (RAI) version 2.0, a standardized instrument mandated in LTC facilities throughout Alberta. The primary evaluation used an interrupted time series design with a comparison group (units not included in the feedback intervention) and a comparison condition (pressure ulcers). We used segmented regression analysis to assess the effect of the feedback intervention.ResultsThe primary outcome of the study, falls, showed little change over the period of the intervention, except for a small increase in the rate of falls during the intervention period. The only outcome that improved during the intervention period was the proportion of residents with high pain scores, which decreased at the beginning of the intervention. The proportion of residents with high depression scores appeared to worsen during the intervention.Conclusions Maintaining all nine units in the study for its 13-month duration was a positive outcome. The feedback reports, without any other intervention included, did not achieve the desired reduction in proportion of falls and elevated depression scores. The survey on intention to change pain assessment practice which was conducted shortly after most of the feedback distribution cycles may have acted as a co-intervention supporting a reduction in pain scores. The processing and delivery of feedback reports could be accomplished at relatively low cost because the data are mandated and could be added to other intervention approaches to support implementation of evidence-based practices.
    Full-text · Article · Nov 2014
Show more