Data Management Basics

Overview of Data Management Process

• Identify study question

• Select study approach

• Design study and collect data

• Analyze data

• Report findings

• Current Stage: We are now entering Stage 4 (Analyze data).

Objectives of Data Management

• State basic principles of data management.

  • Importance of managing data efficiently and effectively throughout a research project.

• Define data management and its importance.

  • Data management involves the practices and policies necessary to handle data safely and securely.

• List and describe basic elements of a data management plan (DMP).

  • Critical aspects needed to ensure data integrity and accessibility.

• Define codebook and its significance.

  • Codebook serves as an essential tool within the DMP.

• Discuss the importance of data cleaning and variable recoding.

  • Ensures the quality and usability of the collected data.

Data Management and Codebooks

• Definition of Data Management:

  • Refers to the entire process of record-keeping before, during, and after a research study.

• Elements of a Data Management Plan (DMP):

  • Data Type:

  • Description of types and estimated amount of scientific data generated.

  • Contextual Details (Metadata):

  • Specifies how the document describes the data.

  • Storage, Backup, and Security:

  • Outline of how and where the data will be stored and secured.

  • Provisions for Protection/Privacy:

  • Addressing privacy and confidentiality issues.

    • Includes data security, data safety, and clinical data.

  • Policies for Re-use:

  • Guidelines on how other researchers may use the data.

  • Access and Sharing:

  • Methods for allowing access to the data by other researchers and how to facilitate discovery of the data.

  • Achieving and Providing Access:

  • Plans for preserving data and providing long-term access.

  • Roles and Plan Oversight:

  • Assignment of responsibilities for data management and required resources.

Codebooks

• Definition of Codebook:

  • A guide specific to a study, describing each variable and how collected data will be entered into a computer file.

  • Addresses anticipated data problems like missing responses.

Example of a Codebook Entry

• ICDC LLCP 2021 Codebook Report:

  • Overall Version Data:

  • Weighted with _LLCPWT

  • Label:

  • State FIPS Code

  • Section Name:

    • Record Identification

  • Section Number:

    • 0

  • Question Number:

    • 1

  • Column:

    • 1-2

  • Type of Variable:

    • Num

  • SAS Variable Name:

    • _STATE

  • Question Prologue:

  • Question: "State FIPS Code"

  • Value Labels and Frequencies:

  • 1 (Alabama): 4,586 Frequency (1.05% Weighted %)

  • 2 (Alaska): 5,493 Frequency (1.25% Weighted %)

  • … (Additional States and Frequencies Included)

Data Entry

• Software Programs for Data Entry:

  • Spreadsheet:

  • A file storing data in cells of a row-by-column table (single table).

  • Database System:

  • Stores data in tables where each row represents one record, allowing linked relationships across tables.

• Double-Entry Method:

  • Two individuals enter the same data into separate files; discrepancies are compared and resolved for accuracy.

Relational Databases

• Definition:

  • Storing data in tables at specific ‘levels’ enhances the organization between stable traits and time-varying data for longitudinal studies.

  • Also aids in data cleaning and reduces redundancy, thereby minimizing storage space needs.

• Examples:

  • ODK Server, Microsoft Access.

Data Structure

• Table Organization:

  • Individual and Household Data

  • Individual ID, Date of Birth (DOB), Name, Race, Sex, Community ID, Community Name, House Number, Community ID Latitude, Longitude, Household ID, Education, Survey ID, Collection Date, Entry Date, Interviewer, Census Data.

Data Cleaning

• Definition:

  • The process of correcting typographical or other errors in data files.

• Example of Data Cleaning:

  • Before Cleaning and After Cleaning for variables such as SEX and responses related to survey data.

Recoding Variables

• Definition:

  • The process of generating values for a new variable based on one or more existing data columns within a file.

• Example of Recoding:

  • Total SPPB Score:

  • SPPB ≤ 9 indicates Frail (coded as 1)

  • SPPB > 9 indicates Nonfrail (coded as 0)

Derived Variables

• Definition:

  • A new variable created during data analysis from existing variables.

• Example Calculation for Body Mass Index (BMI):

  • Formula: $BMI = rac{WT ext{<em>LB}}{HT ext{</em>IN}^2} imes 703$

  • Original Variable Examples:

    • HTIN: 66, WTLB: 155 → BMI: 25.0

    • HTIN: 73, WTLB: 253 → BMI: 33.4

    • HTIN: 59, WTLB: 112 → BMI: 22.6

Overall Objective Check

• Defining data management and exploring its importance.

• Listing and describing the basic elements of a data management plan.

• Defining codebook and discussing its significance in the data management plan.

• Discussing the importance of data cleaning and variable recoding processes.