Full text of DOI:10.1111/j.1467-842X.2000.tb00508.x

Brief Report
Perception of social value predicts participation in
school-based research
Bernadette Pound
Abstract
Victorian Public Health Training Scheme, Royal Melbourne Hospital
Objective: To investigate factors affecting
the participation of schools in a
Michaela Riddell, Graham Byrnes and Heath Kelly
serosurvey.
Victorian Infectious Diseases Reference Laboratory
Methods: A telephone interview was
conducted with a representative of 80
schools (response rate 92%). The schools
had been randomly selected to participate
in a seroprevalence survey evaluating a
measles vaccination campaign of Victorian
school-aged children in 1998.
s schools are ideal places for the Methods
recruitment of samples represent-
Study population
A
ative of school-aged children, they
The population for this study was theVic-
torian schools that had been approached to
participate in the measles seroprevalence
survey. Potential participant schools for the
seroprevalence survey had been identified
as part of a multi-stage random cluster sam-
ple. Firstly, 22 local government authorities
Results: Univariate analysis suggested
that responses to the interview were not
influenced by school level (primary/
secondary), geographic location, funding
source or participation in the
may receive many requests to participate in
research. This study identifies some of the
issues which were important to school rep-
resentatives when approached to participate
in a large seroprevalence study.
seroprevalence survey. There was,
however, a strong association of
A national ‘catch-up’ mass vaccination
campaign of all primary school-aged chil-
dren, using the measles, mumps, rubella
(LGAs) were randomly selected with prob-
participation in the seroprevalence survey
with the perception of value to students
and the value to the community. Factor
analysis identified two issues: the societal
value and practical issues, which
explained most of the variance in
ability of selection proportional to size of
LGA. In the second stage, one primary and
one secondary school were randomly
selected from the LGA, with probability pro-
portional to size. If a school refused, another
randomly selected school from the same
LGA was invited to participate. When an
insufficient number of students participated
from a school (12 per primary school and
(
MMR) vaccine, was conducted in 1998 aim-
ing to increase the population immunity
against measles infection in children aged
1
5
-12 years. Twelve months after its com-
2=
participation (pseudo R 0.84).
pletion, the Victorian Infectious Diseases
Reference Laboratory undertook a large
seroprevalence survey of the Victorian
school population in order to evaluate the
success of the campaign. Participating stu-
dents were invited to answer a questionnaire
and to provide a blood and saliva sample.
Recruitment of schools for the measles
seroprevalence study was more difficult than
expected. Eighty-seven school representa-
tives were approached to obtain 51 schools
Conclusion: The perception of the
benefits of the study strongly influenced
the decision by school representatives to
participate in the seroprevalence survey.
Implications: Recruitment of schools for
health-related research may be improved if
the number of research projects in schools
is monitored and possibly restricted, and
the social value of the research is
emphasised.
3
3 per secondary school), as was the case in
five LGAs, an additional school from the
same LGA was randomly selected.
School representatives from all 87 schools
invited to participate in the seroprevalence
study were approached to respond to a tele-
phone questionnaire investigating their atti-
tude to this research. At each school a con-
tact person was invited to respond to the tele-
phone questionnaire on behalf of the school.
In almost 90% of schools this was the prin-
cipal or vice-principal.
(Aust N Z J Public Health 2000; 24: 543-5)
(
59%) willing to participate. This rate of re-
fusal had not been anticipated when design-
ing the study. There is little published litera-
ture, and certainly none particular to Aus-
tralia, that investigates issues of importance
for the successful recruitment of schools. We
therefore devised a short questionnaire to
explore issues affecting participation in the
seroprevalence study.
Questionnaire
The themes covered by the questionnaire
emerged during recruitment of schools for
Correspondence to:
Submitted: April 2000
Dr Heath Kelly, Head, Epidemiology, Victorian Infectious Diseases Reference Laboratory,
0 Wreckyn Street, North Melbourne, Victoria 3051.
Fax: (03) 9342 2665; e-mail: heath.kelly@mh.org.au
Revision requested: June 2000
Accepted: September 2000
1
2000 VOL. 24 NO. 5
AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH
543

Pound et al.
Brief Report
the seroprevalence survey. The questionnaire consisted of several
types of questions. General questions investigated school repre-
sentatives’ past experience of research, the types of research that
each school had participated in, as well as general attitudes to
health-related research in schools. Specific questions seeking
knowledge of measles and measles vaccination and the decision-
making process involved in participation or non-participation in
the measles seroprevalence survey were asked.
Only 24% of all respondents to the telephone interview reported
a formalised process for consideration of applications to under-
take research in their schools. All respondents agreed that health-
related research was important and supported school-based re-
search for curriculum development and health education. How-
ever, respondents from schools that participated in the measles
seroprevalence survey were significantly more likely to see schools
as appropriate places to recruit subjects for health-related research
(98% vs. 84%, p=0.05). Significantly fewer respondents from non-
participant schools supported health-related research involving
collection of clinical specimens (86% vs. 100%, p=0.03).
Table 1 lists issues which had a significant influence on the
decision to participate in the seroprevalence study. Respondents
from participant schools were significantly more likely to be posi-
tively influenced by the social value of the study to both students
and community and to perceive some benefit from participation.
Although most respondents considered the number of times they
had been approached to participate in school-based research had
no influence on the decision to participate, respondents from non-
participant schools were more inclined to feel negatively about
this issue. More respondents from non-participant schools con-
sidered that the concurrent commitment to other research had a
negative influence on participation in the study.
The questionnaire was designed as a structured telephone in-
terview and was refined following consultation with school prin-
cipals not involved in the seroprevalence study. However, not all
respondents answered all questions and most respondents chose
to elaborate on their answers to some questions. Data were col-
lected over a two-week period in November–December 1999.
Data analysis
2
3
Data were analysed using EpiInfo 6.0 and Stata 6.0. Fisher’s
exact test was used for the yes/no questions and a chi-squared
likelihood test for trend for the ordered responses. A number of
related questions were posed to schools in the study and this made
factor analysis an appropriate tool for extracting summary infor-
mation. The relation between the postulated factors and observa-
tions were estimated using a maximum likelihood estimator.
Not surprisingly, respondents from non-participant schools were
more likely to see the collection of clinical samples as being un-
acceptable to parents. Respondents from participant schools were
more likely to have been positively influenced when considering
the effort required by the school for their participation (52%) com-
pared to respondents from non-participant schools who were more
likely to be negatively influenced (67%). However, about one-
third of respondents from both participant and non-participant
schools felt that the effort required of them by the seroprevalence
survey did not influence their decision to participate.
Results
Of the 87 school representatives approached to be involved in
the measles seroprevalence survey, 80 (92%) completed the in-
terview. Three did not respond and four refused. Participation in
the measles seroprevalence survey did not predict agreement to
be interviewed. There was no significant difference in the pro-
portion of government to non-government, primary to secondary
or Melbourne metropolitan to non-metropolitan school representa-
tives that agreed to be interviewed.
Table 1: Influences on the decision to participate in the school-based measles seroprevalence survey.
Participant Schools
Non-participant Schools
Influences on decision
to participate
Positive No influence Negative
Positive No influence Negative
p Value
(χ for trend)
2
(%)
(%)
(%)
(%)
(%)
(%)
Value to the community (n=71)
Value to students (n=70)
43 (88)
44 (90)
39 (82)
16 (34)
37 (79)
6 (12)
0
5 (23)
6 (29)
2 (10)
17 (77)
13 (62)
14 (74)
16 (84)
18 (95)
0
<0.001
<0.001
<0.001
0.003
3
(6)
2
2
0
1
(4)
(4)
2 (9)
Benefit from participation (n=67)
Interest in this project (n=66)
7 (15)
30 (66)
9 (19)
3 (16)
2 (10)
0
1
1
(5)
(5)
Confidence in the organisation
of the study (n=66)
(2)
(9)
<0.001
Commitment to other
research (n=68)
13 (28)
30 (64)
36 (77)
4
1
0
(5)
14 (85)
12 (60)
6 (29)
8 (40)
0.006
0.014
Number of times approached to
participate in studies (n=67)
4
(9)
7 (15)
Acceptability to parents (n=71)
Collection of blood samples (n=72)
Collection of saliva samples (n=70)
20 (42)
19 (40)
32 (68)
40 (85)
27 (60)
9 (19)
0
0
0
1
10 (39)
10 (40)
15 (65)
20 (91)
14 (61)
15 (60)
8 (35)
<0.001
0.006
0.004
0.019
2
3
(4)
(6)
13 (28)
4
2
(9)
(4)
Use of anaesthetic cream to
collect samples (n=67)
16 (36)
(5)
1
(5)
Effort required of school (n=67)
24 (52)
17 (37)
5 (11)
0
7 (33)
14 (67)
<0.001
5
44
AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH
2000 VOL. 24 NO. 5

Brief Report
Participation in school-based research
Factor analysis revealed two factors to be significant in under-
standing the responses (p<0.005), while there was no significant
advantage to including a third (p=0.136). After varimax rotation,
the nature of these factors is indicated by their loadings on the
observed responses given in Table 2. Responses relating to the
merit of the study (value to students, value to the community,
benefits to the participants, confidence in the organisation of the
study) loaded more heavily on factor 1 than factor 2. Responses
related to practical issues (effort required by the school, number
of times approached for research, commitment to other research)
loaded preferentially with factor 2. We have labelled the factors
Table 2: Loading of factors (unobserved variables) on
the decision to participate (observed variables) in factor
analysis.
Influences on decision to participate Factor 1 Factor 2
Study
merits
Practical
concerns
Value to students
0.99
0.79
0.75
0.72
0.06
0.18
0.28
0.46
0.31
0.33
0.44
0.10
0.04
0.05
0.05
0.33
0.32
0.43
0.45
0.56
0.30
0.28
0.22
0.18
Benefit from participation
Value to community
Confidence in the organisation
Number of times invited to do research
Commitment to other research
Interest in the project
Effort of school
‘
Study Merits’ and ‘Practical Concerns’ respectively. A score on
these two factors was calculated for each school using Bartlett’s
4
estimator. When these scores were then used as explanatory vari-
ables for participation in a logistic regression, they were found to
Acceptability to parents
Blood collection
2
explain the participation decision very well (pseudo R =0.837).
Both were highly significant predictors of participation (study
Saliva collection
merits p=0.003, practical concerns p=0.008).
Anaesthetic cream
Discussion
There are several published reports describing recruitment strat-
egies for school-based research.5-9 While some of these reports
We recommend that education bureaucracies monitor and con-
sider restricting the number of research projects in schools repre-
sented by them. We further recommend that the positive social
and community values of a study should be emphasised when
attempting to recruit schools for health-related research. Schools
approached less frequently may then be more likely to participate
in research projects, even those involving interventions like
venepuncture.
suggest the value of emphasising the legitimacy or importance of
the project to the school contact,5,7,8 there is little published lit-
erature investigating issues relating to recruitment from the per-
spective of school representatives.
Reported rates of school recruitment vary widely. For exam-
ple, a school-based nutritional health promotion program reported
6
1
00% recruitment at the school level while a program aimed at
improving cardiovascular health achieved only 60% recruitment.7
In our study only 59% of school representatives contacted for the
seroprevalence survey agreed to participate. The requirement for
venepuncture from students contributed to unsuccessful recruit-
ment of schools.
Acknowledgements
The authors would like to thank SteveAlbon for assistance with
questionnaire development, the school principals who agreed to
pilot the questionnaire and all school representatives who re-
sponded to the questionnaire.
Prior to school recruitment we contacted the Department of
Education, the Catholic Education Office and the Association of
Independent Schools for permission to approach schools which
they represented. Within the Department of Education, the School
Community Support Branch is responsible for assessing all re-
search proposals involving schools. Following a favourable re-
view by the branch, researchers can attempt school recruitment.
There is no central control on the number of projects approved by
the Education Department and no procedure restricting the number
of times a school is approached to participate in research projects.
In most cases, the decision to participate in research was driven
by the attitude of the principal, vice-principal or year level
co-ordinator, fewer than a quarter of whom reported a formalised
process for the consideration of research projects. Tangible ben-
efits to participating schools, such as computer software or cur-
riculum development and support, are offered in some health pro-
motion research. In the absence of such incentives, we found that it
was the value of the study itself, as well as its benefit to the school
and participants, that had a positive impact on school recruitment.
References
1
. Immunise Australia program: Measles control campaign. Commun Dis Intell
1
998;22(8):156.
2. Dean A, Dean J, Coulombier D, Brendel K, Smith D, Burton A, et al. Epi Info:
word processing, database, and statistics program for epidemiology on
[
microcomputers].Version 6. Atlanta (Georgia): Centers for Disease Control
and Prevention; 1994.
3
4
5
.
.
.
Stata: [statistical software]. Release 6.0. College Station (TX): Stata Corpo-
ration; 1999.
Bartlett M. The statistical conception of mental factors. Br J Psychol
1
937;28:97-104.
Ellickson P. Getting and keeping schools and kids for evaluation studies. J
Comm Psychol 1994(CSAP special issue):102-16.
6. Harrington KF, Binkley D, Reynolds KD, Duvall RC, Copeland JR, Franklin
F, et al. Recruitment issues in school-based research: lessons learned from
the High 5 Alabama Project. J Sch Health 1997;67(10):415-21.
7. Lytle LA, Johnson CC, Bachman K, Wambsgans K, Perry CL, Stone EJ, et al.
Successful recruitment strategies for school-based health promotion: Experi-
ences from CATCH. J Sch Health 1994;64(10):405-9.
8. Petosa R, Goodman RM. Recruitment and retention of schools participating
in school health research. J Sch Health 1991;61(10):426-9.
9
.
Olds S, Symons C. Recommendations for obtaining cooperation to conduct
school-based research in schools. J Sch Health 1990;60(3):96-8.
2000 VOL. 24 NO. 5
AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH
545