Polit ln ch01

Copyright ©2010 by Pearson Education, Inc.Upper Saddle River, New Jersey 07458

All rights reserved.

Statistics and Data Analysis for Nursing Research, Second EditionDenise F. Polit

Statistics and Data Analysisfor Nursing Research

Second Edition

CHAPTER

Introduction to Data Analysis in an Evidence-Based Practice Environment

1




Research and Evidence-Based Nursing

• Evidence-based practice (EBP): Use of best clinical evidence in making patient care decisions

• Best source of evidence: Systematic research




What is Research?

• Research is systematic inquiry that uses orderly, disciplined methods (including statistics) to answer questions or solve problems

• The aim of nursing research is to generate evidence that can contribute to EBP for nurses




EBP Questions

• How reliable is the evidence? • What is the magnitude of effects? • How precise is the estimate of effects?

• Answering these questions requires an understanding of statistics




What is Statistics?

• Statistics is the term for a collection of mathematical methods of organizing, summarizing, analyzing, and interpreting information gathered in a study




Data and Data Analysis

• In the context of a study, the information gathered to address research questions is data

• In quantitative research, data are usually quantitative (numbers)

• Quantitative data are subjected to statistical analysis




Where Do Data Come From?

• Example 1: Interviews/questionnaires – Question: On a scale from 0 to 10, please rate

your level of fatigue– Answer (Data):

Person 1: 7 Person 2: 3 Person 3: 10 Etc.




Where Do Data Come From? (cont’d)

• Example 2: Observation– Observer: What is the patient’s pressure ulcer

stage (1-4)?– Rating (Data):

Person 1: 1 (Stage 1) Person 2: 4 (Stage 4) Person 3: 2 (Stage 2) Etc.




Where Do Data Come From? (cont’d)

• Example 3: Biophysiologic measurements– Recorder: Patient’s oxygen saturation– Measurement via pulse oximeter (Data):

Person 1: 98% Person 2: 95% Person 3: 89% Etc.




From Problem to Data Analysis

• Research problem identified (e.g., high stress in hospitalized children)

• Research question asked: Will a music intervention reduce stress?

• Concepts defined/measured • Data collected • Data analyzed through statistics

– Of course, a lot of things happen in between each step!




Concepts and Variables

• Concept: An abstraction inferred from characteristics or behaviors

– Examples: height, sex, respiratory rate

• In quantitative studies, concepts are called variables

• A variable is something that takes on different values

– Height, sex, and respiratory rate vary from one person to the next




Types of Variables

• Important distinction in most studies (with implications for analysis):

– Independent variable: The hypothesized cause of, or influence on, an outcome

– Dependent variable: The outcome of interest, hypothesized to depend on, or be caused by, the independent variable




Examples of Independent and Dependent Variables

• Independent variable (IV): Smoking• Dependent variable (DV): Lung cancer

IV DV ?




Independent Variables

• Some IVs are existing characteristics (like people’s smoking behavior)

• Others are created by researchers when they introduce an intervention to some people (e.g., a nursing intervention for smoking cessation)

IV DV?




Research Questions

• Research questions are the queries researchers seek to answer through the collection and analysis of data

• Research questions communicate the research variables and the population (the entire group of interest)

– Example: In hospitalized children (population) does music (IV) reduce stress (DV)?




Variable Definition

• In studies, variables need to be defined• Conceptual definition: The theoretical

meaning of the underlying concept• Operational definition: The precise set of

operations and procedures used to measure the variable




Discrete Versus Continuous Variables

• Variables have different qualities with regard to measurement potential

– Discrete variables– Continuous variables




Discrete Variables

• Discrete variable: Categories are indivisible, with a finite number of values between two points

• Examples:– Number of siblings: 0, 1, 2, etc. (1.2 is not possible)

– Number of hospital beds (129.034 beds is not possible)




Continuous Variables

• Continuous variable: Can (in theory) assume an infinite number of values between two points

• Examples:– Time elapsed since birth

(e.g., 55.0359 years)




Measurement

• Measurement: Involves assigning numbers to qualities of people or objects to designate the quantity of the attribute, according to a set of rules




Levels of Measurement

• Level of Measurement: A system of classification with four types of measurement rules that affect the kind of statistical analysis that is appropriate:– Nominal– Ordinal– Interval– Ratio




Nominal Measurement

• Nominal Measurement: – Lowest form of measurement– Numbers are used simply as labels to name categories

• Example: Sex, coded with 2 arbitrary numbers

0 1 0 1 2 4

It does not matter what the codes are, the numbers have no quantitative meaning (although codes like 0 and 1 are more sensible)

• Numbers cannot be treated mathematically




Ordinal Measurement• Ordinal Measurement:

– Uses numbers to designate ordering on an attribute– Conveys some information about amount– But does not indicate distance between values

• Example: Degree of pain 1 = None 2 = Some 3 = A lot _|__________|_______________________________|_ 1 2 3

Distances are not equal, and are not known Averages do not make sense




Interval Measurement

• Interval Measurement: – Also uses numbers to designate ordering on an attribute and

conveys information about amount– Distance between values are assumed to be equal– Averages can be computed

• Example: Ambient temperature (Fahrenheit) |___|___|___|___|___|___|___|___|___|___| 70 71 72 73 74 75 76 77 78 79 80

The difference between 70 and 75 degrees is the same as the difference between 75 and 80 degrees




Ratio Measurement

• Ratio Measurement: – Uses numbers to designate ordering, conveys

information about amount, distances are equal– AND there is a real, rational zero – Averages can be computed

• Example: Medication dose (e.g., number of milligrams, number of pills)




Measurement Level Comparison

• At each successive measurement level, there is more information, and greater analytic flexibility

• If you start with ratio measures, you can collapse information to a lower-level measure, but the reverse is not true

• Higher-level scales are usually (though not always) preferred




Comparison of Levels

Nominal Ordinal Interval Ratio

Classification ✓ ✓ ✓ ✓Magnitude ✓ ✓ ✓Equal Interval ✓ ✓True Zero ✓Math Permissible Count Count,

RankCount, Rank, Add, Subtract

Count, Rank, Add, Subtract, Multiply, Divide




Types of Statistical Analysis

• Calculation– Manual versus computerized

• Purpose– Descriptive versus inferential

• Complexity– Univariate, bivariate, multivariate




Descriptive Statistics

• Researchers collect their data from a sample of study participants—a subset of the population of interest

• Descriptive statistics describe and summarize data about the sample

– Examples: Percent female in the sample, average weight of participants




Inferential Statistics

• Researchers obtain data from a sample but often want to draw conclusions about a population

• Parameter: A descriptive index for a population– Example: Average daily caloric intake of all 10-year-old

children in New York

• Statistic: A descriptive index for a sample– Example: Average daily caloric intake of 300 10-year-

old children from three particular NY schools




Inferential Statistics (cont’d)

• Researchers use statistics to make inferences about parameters

• Inferential statistics, based on laws of probability, help researchers draw objective conclusions about a population, using data from a sample

• Inferential statistics are often used to test hypotheses (predictions) about relationships between variables




Data Analysis Plan

• Many activities must occur between the collection of data and analyses to address research questions

• Researchers need to develop a careful data analysis plan that lays out the analytic steps to be taken




Preanalytic Steps

• Researchers typically begin by:– Making decisions about how to code their data

—including codes to use to designate missing values

– Entering data into a computer file (e.g. an SPSS file)

– Then giving variables abbreviated variable names, and inputting additional information about the variables (e.g., what the codes mean)




Preanalytic Steps (cont’d)

• Many additional steps remain before substantive analyses begin, including efforts to:– Clean the data (correct errors)– Examine the extent of missing data– Correct problems relating to missing data– Assess whether assumptions for advanced

analyses are met– Assess whether there are biases– Understand the data and characterize the

study sample




SPSS and Statistical Analysis

• SPSS (Statistical Package for the Social Sciences) is among the most popular statistical software packages for analyzing research data

• It is user friendly and menu driven• The datasets offered with this textbook are

set up as SPSS files




The Data Editor in SPSS

• The data editor in SPSS offers a convenient spreadsheet-like method of creating, editing, and viewing data

• There are two “views” within the data editor: – Data View: Shows the actual data values– Variable View: Shows variable information for

all variables




Data View in the Data Editor

• The columns represent one variable each; unique variable names (no more than eight characters long) are shown at the top of each column

• Each row is a case, representing an individual participant

• The data view tab is at the bottom




Variable View in the Data Editor

• Variable View shows a wealth of information about how variables are coded, how they will be labeled in output, level of measurement, and so on

• The Variable View tab is at the bottom




Versions of SPSS

• New versions of SPSS are created regularly, to offer improved options for analysis and presentation

• Examples in this book were created in SPSS Version 16.0

• The student version of SPSS is available for analyzing relatively small datasets (no more than 50 variables and no more than 1,500 cases)

Documents

Polit ln ch01