ONR Services - Data Analysis

Primary Contact: Yelena Perkhounkova (elena-perkhounkova@uiowa.edu)

What is the role of the statistician for the College of Nursing?

I have an idea for a research grant. When should I talk to a statistician?

I’m developing a grant proposal. How should I prepare for my first meeting with a statistician?

How many subjects will I need? Or, do I have enough subjects?

What should I know about data management? How do I prepare data for data entry?

What data entry systems should I use?

I need statistical support for my ongoing research. What should I provide to the statistician?

What resources are available for qualitative data management and analyses?

What is the role of the statistician for the College of Nursing?

The statistician provides support to the CON faculty and postdoctoral fellows as they prepare grant proposals and conduct research. The statistician works directly with the faculty to formulate research questions and hypotheses, choose an appropriate study design, and develop plans for analysis, data entry, and data management. The statistician will conduct statistical analyses and help with report preparation and dissemination of results.

I have an idea for a research grant. When should I talk to a statistician?

The answer is “as soon as possible.” When you decide to write a grant proposal, you will meet with Dr. Linda Liu Hand, Director of the ONR. She will advise you to meet with the statistician in early stages of your grant proposal development.

I’m developing a grant proposal. How should I prepare for my first meeting with a statistician?

To make your first meeting with the statistician most effective, you need to have, prior to the meeting, a good understanding of what you want to accomplish with your research, what questions you want to answer, what relationships you want to investigate, whether similar research has been done and what the findings were, what populations of research subjects will be available to you, and what measurement instruments you want to use. Please bring to the meeting a brief outline of the following:

• Purpose of your research
• Research questions
• Description of subjects
• Measurement instruments
• Proposed design
• A list of outcome and independent variables
• Your earlier research in the topic (if any)
• Any research articles on the topic that would be helpful to the statistician

Back to question list

How many subjects will I need? Or, do I have enough subjects?

The first question is concerned with estimating a sample size for your future study, and the second question is concerned with justification of your existing sample. The latter is also called power analysis.

The answer depends on many things, including your research questions, variables, study design, statistical tests required to test your hypotheses, desired significance level and power, and availability of subjects. The statistician will need all the information about your outcome (dependent) variables that is available. If measurement instruments you plan to employ have already been used, you need to find descriptive statistics for your variables of interest, such as means and standard deviations, preferably for the population you plan to study. If earlier research reported effect sizes, the statistician will need that information, too. You will need to decide what sizes of experimental effects can be expected in your study, and what magnitudes of effects are important to you.

What should I know about data management? How do I prepare data for data entry?

The way data are collected and entered will determine how much time will be needed to prepare them for analyses. The early choices you make are very important and will depend on your research questions. Therefore, it is very important to discuss plans for data management with your statistician. The ONR’s data manager will discuss data entry options with you, help you to plan your databases and data entry forms, and train your RAs to do data entry.

If you need services beyond that, the following options outside CON are available for a fee:

• The Clinical Trials Statistical and Data Management Center(CTSDMC)
• Biostatistics Consulting Center

What data entry systems should I use?

The data entry systems that are currently used by CON researchers are: Access, Excel, and WebSurveyor. From any of these systems, data can be exported into SAS or SPSS for statistical analysis. Among these systems, Access has the most capabilities; however, it is also the most challenging to master. WebSurveyor has been gaining popularity recently. Just like Access, it can be used to set forms for data entry. The forms can be accessed from any computer connected to the Internet, which facilitates the data entry process. It is important to set up the forms properly so that the data can be exported into a statistical package ready for analysis. ONR’s data manager will assist you with this.

For large scale studies with multiple relational datasets, Access is still the best option. For an introduction to Access, try these resources provided by the University of Iowa free of charge:

• Books 24x7: The Unofficial Guide to Microsoft Office Access 2007
• Learning online with SkillSoft course: Getting Started with Access 2007

For questionnaire design and data collection, Snap Survey software is available in the Research Suite. Snap forms are designed to be optically scanned, and data can be analyzed within Snap or exported into other data management systems.

Back to question list

I need statistical support for my ongoing research. What should I provide to the statistician?

In communication with the statistician, it is important to provide as much of the following as possible:

• Relevant sections of your original grant proposal (Specific Aims though Research Design and Methods sections).
• A one-page summary of your study design and participants, measurement instruments, and variables, with all of the changes to the original proposal indicated.
• A brief summary of your accomplishments so far. The current stage of your research: design, grant application preparation, data collection, analysis, preparation of report or manuscript for publication, etc.
• A brief description of what you want the statistician to do for this project.
• Your data, including the description of your variables (e.g., codebooks).
• Research articles on the topic.

What resources are available for qualitative data management and analyses?

In some respect, qualitative data are managed and analyzed similarly to quantitative data. For needs specific to qualitative data, consider NVivo software. Check with Linda Curran to see when the next NVivo training workshop will be available.

Back to question list