Grants and Research Office
Guidelines for Budgeting Statistical Consulting Team Services
Guidelines for Budgeting Statistical Consulting Team Services
in Grant Proposals
Within certain practical limits, you can buy and use as much of
these services as you want and can afford. It is a good idea to
meet with David Ronis or Deanna Isaman to develop a customized arrangement, selecting
services from the Stat Team that most effectively meet your project's
needs. Note also that the ideal level of Stat Team use might vary
from year to year. Thus a project may want to budget at a minimal
level in Year 1 and at a moderate level in Years 2 and 3. Various
levels of use of statistical services and their associated charges
are described below.
Scenario 0. No Use of Statistical Consulting Team Services
If you do not need any statistical expertise at all or if you
have all the statistical expertise you need without using the
Stat Team, then there is no reason to budget for the Stat Team.
On the other hand, if you want or need the help of the Stat Team
to prepare your grant proposal (for example, to do a power analysis),
then you presumably will want or need the help of Stat Team to
help carry out your project. Then scenario 0 is not appropriate
for you.
Scenario 1. Minimal Use of Statistical Consulting Team Services
If you used the Stat Team for a power analysis or other help with
preparing the grant proposal, then you will almost certainly want/need
some Stat Team help in carrying out the project, if only for revisiting
the power analysis. The range of time for minimal services varies
from about 10 hours per year for consultation on an extremely
simple project to about 50 hours per year for assistance with
an extremely complex project or cluster of projects. Ten hours
per year is a good minimum to budget in projects that plan very
little use of Stat Team services--such as an occasional consultation
(e.g., simple power analysis or assistance with interpretation
of a statistical test). Minimal use scenarios assume that the
project has its own staff that will handle all routine data management
and analysis tasks. If the amount of Stat Team time actually used
by a project exceeds the budgeted time by more than an hour or
two, there will be additional costs to the grant on a per hour
basis.
Scenario 1a. Minimal use for a small straightforward project.
10 hours/yr.
Scenario 1b. Minimal use for a moderately complex project. 30
hours/yr.
Scenario 1c. Minimal use for an extremely complex project. 50
hours/yr.
Scenario 2. Moderate Use of Statistical Team Services
The assumption in Scenario 2 is that the research project will
hire its own research assistant, perhaps 50-100% FTE and that
person (or group of persons) will do all or almost all of the
data management as well as the less complex analyses. Stat Team
will provide consultation on all statistics-relevant aspects of
the project, will do the most complex data management tasks, and
will conduct and write-up the more complex analyses. Stat Team
will also conduct power analyses.
This scenario involves having the investigator consider the
list of statistics-related activities on the project, noting which
activities would be more efficiently done by the project's research
assistant and which should be done by the Stat Team. The general
idea is to have the Stat Team provide considerable consultation
but only do the most complex analyses.
The budget for this obviously depends on how much work the Stat
Team will do. The maximum that could be anticipated would be no
more than 200 hours/year. Here are some rough estimates:
Scenario 2a. Moderate use for a small straightforward project.
30 hours/yr.
Scenario 2b. Moderate use for a moderately complex project. 90
hours/yr.
Scenario 2c. Moderate use for an extremely complex project. 160
hours/yr.
Services that the Stat Team Can Provide:
Conduct power analyses, collaborate with the PI in refining measurement
instruments, suggest data management methods/procedures, develop
data entry methods, suggest questionnaire formats, refer to available
data entry services, provide consultation on data entry procedures,
conduct confirmatory or exploratory factor analyses to guide data
reduction decisions, select, test and program data reduction by
computation of scales and indices with input from the PI, assess
the desirability of imputation of missing values with PI input,
impute values for missing data, assess the usefulness of transformations
to reduce deviations from assumptions of planned analyses, select
appropriate statistical techniques to achieve project aims with
PI input, reorganize data and transfer data between different
machines and different statistical programs in order to conduct
the analyses, conduct, summarize, and help interpret results of
statistical analyses, assist or collaborate with the PI on writing
up results for reports/publications, various relatively complex
analyses such as linear and logistic regression, structural equation
modeling, Analysis of Variance and Covariance (ANOVA and ANCOVA),
Repeated Measures ANOVA (RM-ANOVA), Multivariate ANOVA (MANOVA),
multilevel modeling, and analysis of survey data with complex
sampling designs.
Services that the Stat Team Cannot Provide:
Data entry, data coding, general data management tasks, and analysis
of qualitative data.
For additional information, please contact the:
Grants and Research Office
University of Michigan School of Nursing
400 N. Ingalls, Room 4236
Ann Arbor, MI 48109-0482
Phone: (734) 764-9555
Fax: (734) 615-1666
Page Last Modified: Thursday, 10-Aug-2006 12:53:24 EDT
© 2003
University of Michigan School of Nursing
|