Research Study
What is EVIDENCE BASED PRACTICE?
A systematic review of critical appraisal and synthesis of the most relevant research.
Clinical Expertise
Patient Preferences and Values
Initiatives to Advance EBP
To Err is Human: Institute of Medicine (IOM): Building a Safer Health System
Initiatives driving the movement
IOM’s goal: By 2020, 90% health care will be evidence-based
US Preventive Services Task Force (sponsored by AHRQ): analyzing evidence and publishing guidelines (Guide to Preventive Clinical Services)
Magnet Recognition Program (ANA): mandate nursing research and use of EBP
Goal of EBP
OPTIMUM PATIENT OUTCOMES
Once you begin to look for
evidence-based projects,
you’ll start to see them everywhere!
SOURCES OF EVIDENCE
Research Findings
Agency quality monitoring data
Data from national databases
Expert opinions
Scientific principles
Research prOCESS
Conducting Research: Steps of the Research Process
1. Identify issue or question
2. Formulate research question
3. Review the literature
If further research is needed:
4. Determine theoretical framework
5. Design the study
6. Select the sample
7. Collect data
8. Analyze data
9. Interpret results
If indicated, change practice!
Step ONE: Identify issue
Step Two: formulate Research Question
Conducting Research: Formulating the research question
Research Questions (quantitative):
Identify the target population
State an intervention or treatment (independent variable)
List the variables to be measured/outcomes (dependent variables)
Sample Research Question
Do pediatric patients who are given a
reward when they cooperate during
nursing procedures tend to be more
cooperative during those procedures
than unrewarded peers?
12
Sample Research Question
What are the relationships among spiritual well-being, sleep quality, and health status in HIV-infected men and women?
13
Conducting Research: Formulating the research question
Research Question:
What are the effects of weekly quizzes on the grades of nursing students?
What is the target audience?
What is the independent variable?
What is the dependent variable?
Conducting Research: Formulating the research question
Research Question:
Do nursing students who participate in study groups earn higher grades on final exams?
What is the target audience?
What is the independent variable?
What is the dependent variable?
Conducting Research: Formulating the research question
Research Question:
Is there a difference in patient satisfaction scores between patients who have had nursing students care for them and patients who have not had nursing students assigned to them?
What is the target audience?
What is the independent variable?
What is the dependent variable?
Avoid “Yes” or “No” Question Formats
Practice
Identify the Variables
Do nursing home residents who are praised
by nursing personnel for self-feeding
require less assistance in feeding than
those who are not praised?
19
Identify the Variables
Do pediatric patients who are given a
reward when they cooperate during
nursing procedures tend to be more
cooperative during those procedures
than unrewarded peers?
20
Identify the Variables
For nurses working in psychiatric wards, does a 5-day training course in aggression management positively influence nurses’ tolerance towards patient aggression and their feeling toward aggressive patients?
21
Identify the Variables
Do students participating in the Crawford Intervention sessions have higher physical activity and lower dietary fat intake compared with students not in the intervention group?
22
Identify the Variables
What are the effects of Behavioral Change
Intervention (BCI) in self-efficacy and
outcome expectations for exercise and
functional activity, physical activity
participation, and physical performance
of older adults in Thailand?
23
Identify the Variables
Does a telephone therapy intervention for patients with prostate cancer cause improvements in their psychological distress and coping skills?
24
Identify the Variables
What are the effects of animal-assisted therapy on the nutritional intake of individuals with Alzheimer disease?
25
Conducting Research: Formulating the research hypothesis
Research Hypothesis H1
The results a researcher expects to get about the relationships among variables
Null Hypothesis H0
The prediction that there will be no relationships among the variables
Conducting Research: Formulating the research hypothesis
Research Question
What are the effects of weekly quizzes on the grades of nursing students?
Research Hypothesis H1
Weekly quizzes will improve the grades of nursing students.
Null Hypothesis H0
Weekly quizzes will not influence the grades of nursing students.
Research Question
What are the effects of weekly quizzes on the grades of nursing students?
Research Hypothesis
Weekly quizzes will improve the grades of nursing students.
28
Research Hypothesis
The results a researcher expects to get about the relationships among variables
Null Hypothesis
The prediction that there will be no relationships among the variables
29
Research Question
What are the effects of weekly quizzes on the grades of nursing students?
Null Hypothesis
Weekly quizzes will not influence the grades of nursing students.
30
Step three: review the literature
Step Four: Determine Theoretical Framework
Conducting Research: Determine theoretical framework
A theory is an idea or explanation
related to a particular subject.
Conducting Research: Determine theoretical framework
Every research study is based on theory
Predetermined nursing theory
Researcher’s own theory
Theories from other disciplines
Often times it is not mentioned in report/article
Conducting Research: Determine theoretical framework
Research is based on theory. Research finds out if the parts of the theory hold up.
Theory is initial inspiration for research study
Research helps answer the question “Were the predictions based on my theory correct?”
Step five: Design the Study
Conducting Research: Design the study
Differences Between Two Research Methods
Quantitative Research
Qualitative Research
Conducting Research: Design the study
Quantitative Research
“A formal, objective, rigorous, systematic process for generating information
about the world.”
Conducting Research: Design the study: Quantitative Research
Truth is absolute.
A single reality can be found by careful research.
Researcher must be objective.
Conducted to test theory, examine relationships among variables, and determine cause-and-effect relationships.
Conducting Research: Design the study: Qualitative research
Qualitative Research
“A systematic, subjective approach
used to describe life experiences
and give them meaning.”
Conducting Research: Design the study: Qualitative research
Truth is relative; there are multiple truths. What is true for one person may not be true for another.
“Truth” can only be discovered by working to understand a person’s perspective.
Conducted to promote understanding of human experiences and situations
Conducting Research: Design the study: Qualitative research
Qualitative research questions
The population (What are the characteristics of the patients or clients?)
The situation (What conditions, experiences, or circumstances are we interested in understanding?)
Conducting Research: Design the study: Qualitative research
Examples of qualitative research questions
What is the lived experience of being a nursing student?
What is it like to receive a terminal diagnosis?
What is it like to be a graduate nursing student?
Quantitative Designs
Quantitative Design
Important terms:
Rigor: degree of precision of measuring tools, representativeness of sample, maximum control of variables
Control: researcher control over variables
More important terms:
Bias
Something that distorts the findings of a research study
For example:
Members of this class would bias a study
about the public’s understanding of the
value of nursing care.
More important terms:
Manipulation
Controlling a treatment or intervention
For example:
The researcher controlled the length of time
between administration of pain medication.
More important terms:
Generalizable
The degree to which a study’s findings
are true for the entire target population
For example:
Because the sample was large enough and members of the study groups were randomly selected, the findings of the study are assumed to be true for everyone.
Four quantitative research designs:
Descriptive (non-experimental)
Correlational (non-experimental)
Quasi-experimental
Experimental
Types of Studies Described by Time Factors
Cross Sectional Design
Longitudinal (prospective, cohort)
Retrospective (case-control)
Descriptive Designs
Exploration and description
May describe characteristics of people
May determine the frequency of something
May categorize data
Used to examine variables and
characteristics of subjects.
Example of Descriptive Study
A researcher wanted to learn about women’s experiences during menopause. She measured the frequency of the symptoms, the average age at menopause, and the percentage of women using medications to alleviate symptoms.
Measurements used in Descriptive Studies
Frequencies
Averages
Percentages
Proportions
Incidence
Prevalence
Rate
Risk
Types of Descriptive Studies
Case Study Design
Descriptive
Comparative Descriptive
Descriptive Correlational
Descriptive Design
Example
Homeless African-American women
compared to homeless white women for
differences in health promoting behaviors.
Case Study Descriptive Design
The study of one patient, family, or organization.
Example
The family of a blind child was studied
to learn how this family coped with
their child’s blindness through
his growth and development stages.
Cross Sectional Descriptive Design
A study that follows the same subjects at only one point in time instead of across time.
Example
Surgeons and OR nurses in one hospital were surveyed to determine degree of compliance with universal precautions.
Descriptive Longitudinal Studies (Prospective, Cohort)
Conducted on a cohort of people over a period of time going forward
Usually very long-term projects
Very expensive
Sample must be free of the disease/illness at the start of the study
Descriptive Longitudinal Design
A study that follows the same subjects over a long period of time.
Example
Surgeons and OR nurses in one hospital were surveyed every month for one year to determine changes in the degree of their compliance with universal precautions.
Descriptive Retrospective Studies (Case-Control)
Conducted on the past records of a group of people
Less expensive than longitudinal studies
Starts with a group of people with a disease and another group without the disease
Looks for differences between the two groups in the past
Correlational Designs
Investigate relationships among variables
Measure strength of relationship (-1 to +1)
Measure direction of relationship (positive or negative)
Example of Correlational Study (a descriptive correlational, retrospective study)
Researchers studied the records of infants who died of SIDS and a group who didn’t to compare them on birth weight, maternal characteristics, fetal heart variability, and sleep-wake cycles
Types of Correlational Studies
Descriptive Correlational Design
Predictive Correlational Design
Descriptive Correlation Design
Used to describe variables and examine relationships among these variables.
Example
Homeless adolescents were studied
to discover the reasons they give for being
homeless, their resilience, and risky behavior.
Predictive Correlation Design
Used to predict the value of one variable based on the values obtained for other variables
Example
The self-esteem, perceptions of childbearing, race, social status, and relationships with family members of adolescent girls were examined to identify those at risk for adolescent motherhood.
Experimental Research
Examines causal relationships between independent variable on dependent variable
Most rigorous research design
Highly controlled
Maximum control of variables
Example of Experimental Study
Researchers compared post-acute intermediate care in an inpatient nurse-led unit and conventional post-acute care on general medical wards. They looked at variables such as patients’ length of stay, functional status, and mortality. All patients were randomly assigned to the unit or ward.
Experimental Research
3 main characteristics:
Controlled manipulation of independent variable
Treatment or intervention
Random assignment of subjects to control and experimental groups
Types of Experimental Studies
Pretest-posttest design (could also be quasi-experimental)
Clinical trial
Quasi-Experimental Research
Examines causal relationships OR effects of one variable on another
Involves a treatment or intervention and measuring the effects of the treatment or intervention
Experimental design without random assignment of subjects
Example of Quasi-experimental Study
A researcher evaluated the effectiveness of a violence-prevention intervention for nursing assistants working in long-term care. The intervention was implemented in three nursing homes, and three other nursing homes served as the comparison.
More important terms:
Causality
when an independent variable causes
an effect in a dependent variable
For example:
Smoking causes cancer.
More important terms:
Multicausality
when there is more than one independent variable that causes an effect in a dependent variable
For example:
Smoking and oral contraceptives cause
increased risk of cancer.
Degree of Rigor and Control
DESCRIPTIVE high rigor, low control
CORRELATIONAL high rigor, low control
QUASI-EXPERIMENTAL high rigor, partially controlled
EXPERIMENTAL high rigor, totally controlled
Step six: select the sample
Sample Selection
Population
(Target Population)
The entire set of people
who meet the sampling criteria
Sample Selection
Subjects
People who are being studied
Sample Selection
Sample
the selected group of
people taking part in the study
Sample Selection
Accessible population
The portion of the target population to which the researcher has access
Sample Selection
Sampling criteria
Eligibility criteria
Inclusion criteria
Exclusion criteria
Sample Selection
Representativeness
The sample, the accessible population,
and the target population must
be as similar as possible
Sample Selection
Sampling Bias
The systematic over-representation or
under-representation of some segment
of the population
Sample Selection
Sampling Bias
For example:
We are conducting a study on nursing students and collected data from the first 10 students who entered the library. What bias are we introducing into the sample?
Choosing a Sample from the Sampling Frame
Random sampling methods (for quantitative studies)
Simple random sampling
Stratified random sampling
Cluster sampling
Systematic sampling
84
Table of Random Numbers
| 061501 | 896939 | 173249 | 851495 | 080678 |
| 088722 | 507157 | 949463 | 056582 | 050831 |
| 240067 | 095888 | 119144 | 975852 | 075242 |
| 007007 | 884184 | 103728 | 791761 | 009531 |
| 541119 | 594649 | 519266 | 374786 | 227599 |
| 706416 | 424993 | 000243 | 315604 | 741784 |
| 210478 | 436022 | 815218 | 528465 | 008957 |
| 643146 | 983722 | 610063 | 602459 | 763468 |
| 456307 | 417707 | 680497 | 420151 | 542222 |
| 940208 | 478597 | 053814 | 634305 | 351413 |
| 365415 | 270855 | 142284 | 335173 | 922755 |
Practice
Which type of sampling is this?
In order to select a sample of patients with asthma, a researcher first randomly selected a sample of pulmonary clinics and then a sample of asthma patients from each clinic.
Which type of sampling is this?
In order to select a sample of nursing students, a researcher randomly selects 10 nursing schools on the east coast and then 5 nursing schools in one zip code are randomly selected.
Which type of sampling is this?
In order to select a sample of nursing students, a researcher first obtained a list of all nursing schools in the US by BSN or RN-BSN, then randomly selected 10 schools from each type, and then a sample of students divided by BSN students and RN-BSN students and randomly selected 100 students from each school.
Which type of sampling is this?
In order to select a sample of ResU nursing students, a researcher put each student’s name in a hat and drew names.
Which type of sampling is this?
In order to select a sample of nursing students at ResU, a researcher obtained a list of all students, then randomly selected every 10th student on the list.
Sample Selection
Non-random sampling (for quantitative studies)
Convenience sampling
Quota sampling
Snowball sampling
Consecutive sampling
Purposive sampling
Practice
Which type of sampling is this?
In order to select a sample of nursing students at WSCN, a researcher visited a class and asked for volunteers.
Which type of sampling is this?
In one nursing school 20% of the students were Asian, 33% were Hispanic, 19% were African American, and 28% were non-Hispanic white. In order to select a representative sample of nursing students at WSCN, a researcher accepted the same proportions of students (by ethnicity) for the study.
Sample Selection
Power analysis
The number of subjects you need for the study.
Sample size is affected by the research design, number of variables, sensitivity of the measuring tools, effect size and data analysis techniques.
Step seven: collect the data
Continuous Variables
A variable with an infinite number of possible values falling along a continuum
Example: age
Discrete Variables
A variable with a finite and small number of possible values falling along a continuum
Example: grade in elementary school
Collect the Data
The measurement of concrete variables is considered to be “direct measures.”
Examples
BP, height, weight, EKG, bladder scans, gtts per minute of IV fluid
Collect the Data
The measurement of abstract variables is called “indirect measures.”
Generally, multiple indirect measures are needed to measure an abstract variable for a study.
Collect the Data
Examples of Indirect Measures
IQ tests
Stress scales
Scores on a tool that measures Coping
Patient satisfaction
Collect the Data
Measurement Error
the difference between true measure and what is actually measured
Collect the Data
A researcher used a blood pressure cuff that was calibrated 1.8mmHg too high. The blood pressures of all subjects of the study were 1.8mmHg higher than the true measurement.
This is an example of a systematic error.
Collect the Data
Another researcher used a blood pressure cuff that was calibrated correctly, but didn’t place the stethoscope correctly for all subjects. The blood pressures of some subjects of the study were too high or too low.
This is an example of a random error.
Collect the Data
Levels of Measurement
(Types of Variables)
Nominal
Ordinal
Interval
Ratio
Steps Eight and Nine will be discussed
Next week
Review of reliability and validity
Reliability
Reliability is the ability of an instrument to measure a concept or construct consistently.
Three attributes of reliability:
Stability of the instrument (produces the same results with repeated testing)
Homogeneity (or internal consistency) of the instrument (all items measure the same construct)
Equivalence (produces the same results when equivalent instruments are used)
Reliability
Types of reliability:
Interrater reliability: two or more instructors will give the same scores
Test-retest reliability: no matter how many times a test is given, the student will make the same grade
Equivalent forms reliability: different versions of a test produce the same scores
Internal consistency: all items/questions of a test produce the same scores
Reliability
Measured with correlation coefficients
(called reliability coefficients)
Validity
Validity is the degree to which an instrument accurately measures what it is supposed to measure.
Validity
How valid is the use of a yardstick to measure length?
How valid is the use of a thermometer to measure temperature?
How valid is the use of an IQ test to measure intelligence?
How valid is the use of the Kozlarek personality test to measure personality traits?
Validity
How valid is the use of the Patel test to rate faculty effectiveness?
Sample question
Rate your instructor’s availability:
a. always available
b. sometimes available
c. rarely available
Types of Validity
Content validity:
measures a representative sample
of the construct and only that construct
(it must measure all parts of the construct and not include parts of other constructs)
If a research exam contains questions about APA rules, is it a valid test for knowledge about research?
Types of Validity
Face validity:
the extent to which, on the surface, the tool looks like it is measuring the construct
Types of Validity
Criterion-related validity
Concurrent validity (correlation of one known test with another)
Predictive validity (correlation between the measure of the construct and a future measure of the construct-does it accurately predict an outcome?)
Types of Validity
Construct validity
measures a theoretical construct, attribute, or trait (something that cannot be directly observed)
by observing behavior or comparing a characteristic to a known relationship
Examples: racial prejudice, love, creativity
Validity
Measured with correlation coefficients.
Ethical considerations in research
Protection of Human Subjects
3 ethical principles relevant to research
Respect for Persons
Beneficence
Justice
Protection of Human Subjects
Respect for Persons
Each individual:
Is unique and free
Has the right and capacity to decide
Has value and dignity
Has the right to informed consent
Protection of Human Subjects
Respect for Persons
People are autonomous agents who have the right to self-determination
They must be informed about a study.
They can choose whether or not to participate in a study.
They may withdraw from a study at any time without negative consequences.
Protection of Human Subjects
Respect for Persons
Right to Privacy:
People have the right to determine the time, extent, and general circumstances under which their private information will be shared with or withheld from others.
Protection of Human Subjects
Anonymity
Confidentiality
Protection of Human Subjects
Respect for Persons
You must be able to show that potential subjects comprehended the risks and potential benefits.
Avoid the use of coercion or undue influence.
Informed consent
Protection of Human Subjects
Respect for Persons
When compensation is provided:
Fully explained and not coercive
Fair payment for time, travel, and inconvenience
Degree of health care to be made available
Protection of Human Subjects
Respect for Persons
People with diminished autonomy are vulnerable and must have additional protection.
Legal or mental incompetence
Terminal illness
Confinement to an institution
Pregnant women, fetuses and neonates
Children
Economically or educationally disadvantaged
Protection of Human Subjects
Respect for Persons
Children provide assent,
their parents provide permission.
Protection of Human Subjects
Beneficence
Do No Harm!
Maximize possible benefits,
minimize possible harm.
Protection of Human Subjects
Justice
Treat groups fairly and equitably
Use fair procedures in sample selection
Make sure people aren’t considered for inclusion because of their availability, compromised position, or vulnerability.
Protection of Human Subjects
Summary
Respect for Persons
Informed consent
Right to privacy
Additional protection for vulnerable people
Beneficence
Maximize benefits, minimize harm
Justice
Watch for unfair sample selection or treatment
Protection of Human Subjects
IRB Process
IRB Committee
concerned only with protection of human subjects
must approve research proposal before data can be collected
must monitor the study
USE OF CLINICAL GUIDELINES
Use of Clinical Guidelines
Use existing clinical guidelines to find gaps in empirical knowledge.
Unit 2: Analyzing Data
Statistical Tests of Significance
Parametric tests meet the assumptions that the dependent variable (s):
Have an interval or ratio level of measurement
Have a normal distribution in the population
Data can be treated as random samples
Whereas, Nonparametric tests don’t meet any of these assumptions and are done on nominal or ordinal data only!
137
Testing the Reliability of Measures
If the coefficient is 1.00, each item in the instrument is consistently measuring the same thing.
A slightly lower coefficient (0.80 to 0.90) indicates that the instrument is measuring quite well.
A coefficient below 0.70 indicates that there is a problem with some of the items and the researcher should not analyze the data collected with the instrument (scale).
A reliability of 0.70 means that 70% of the time you can trust the score to accurately reflect what is being measured
138
Level of Significance and Probability
Level of Significance
How does a researcher decide how much of a difference is a significant difference?
Level of Significance
An instructor is teaching students about the research process. She will give you a final exam to determine how much you have learned.
What score will indicate that students have learned enough to pass? 50%? 75%? 90%?
Level of Significance
Researchers have to decide the “passing scores” for their studies.
They use probabilities (p) to do so.
Level of Significance
A probability states the likelihood that scores or groups are the same.
The most common level of significance for most nursing studies is 5%, or .05.
(p) = .05
Level of Significance
For example:
A researcher conducts a study to see if support hose prevents muscle pain in people with PVD.
75% of the experimental group (who wore the support hose) reported less muscle pain, while
38% of the control group experienced less muscle pain.
Was there a difference between the 2 groups?
Level of Significance
The researcher finds that the probability of subjects in Group A being different than the Group B subjects is 80%. (8 out of 10 Group A subjects experienced a different level of muscle pain than the Group B subjects.)
p = .20
Probability Theory
Probability is the chance that
a certain outcome
will occur from some random process
146
Normal (Bell) Curve
A normal curve is a theoretical frequency distribution of all possible values in a population…no real distribution of values from a study will exactly fit the normal curve.
Levels of significance and probability are based on the logic of the normal curve.
It is a way to look at the distribution of the values obtained from a sample.
147
Normal Bell Curve
148
“Tailedness” and the Normal Curve
One-tailed test of significance… the hypothesis is directional with extreme scores in only one tail of the curve
Two-tailed test of significance is used when the researcher does not know which way the scores will fall…either tail
149
One-Tailed Test
150
Two-Tailed Test of Significance
151
Inference and Generalization
An inference is a conclusion or judgment based on evidence
An inference is made from the statistical findings obtained from a specific sample and applied to the general population.
Generalization is extending the findings from a sample to a larger population.
Ex: A researcher may conclude that a significant difference was found between two samples, one in which the subjects had been catheterized during hospitalization and another in which the subjects had not. The findings are generalized from the study sample to all previously hospitalized patients.
152
Analyzing Quantitative Data
Selecting a statistical test is based on:
Research design
Levels of measurement
Number of groups
Four quantitative research designs:
Descriptive
Correlational
Quasi-experimental
Experimental
Descriptive Analysis
Used to describe samples
Used for data analysis in Descriptive Studies
Types of Descriptive Analyses
Descriptive statistics
Frequency distribution
Percentage distribution
Measures of Central Tendency (mean, median, and mode)
Measures of Dispersion (variability)
Range
Variance
Standard Deviation
Measures of Central Tendency
A measure of central tendency is frequently referred to as an “average”
Defined: Statistical procedures for determining the center of a distribution of scores.
Three are commonly used in statistical analyses:
Mode
Median
Mean
159
Mode
The mode is the numerical value or score that occurs with the greatest frequency
It tells us the most frequently occurring value on a scale item
Data can have more than one mode, and is then called “bimodal”
Best used with nominal data; but can also be used with ordinal, interval and ratio data
What is the mode in these two examples?
Example: 2, 2, 2, 4, 4, 4, 6, 8,10
Example: 12, 13, 14, 14, 14, 16, 18
http://www.mathgoodies.com/lessons/vol8/mode.html
160
Median
The median is the score in the exact center of the ungrouped frequency distribution—the 50th percentile
If the number of scores is uneven then exactly 50% of the scores are above the median and 50% of the scores are below the median
If the total number of scores is an even number, i.e. 10, the median is the average of the two middle scores
For example, what it the median of these scores?
1, 3, 2, 5, 4, 6, 4, 2, 1, 5
Can be used with ordinal data; also interval and ratio data
http://www.mathsisfun.com/median.html
161
Mean
The Mean is the most commonly used measure of central tendency. It is used for many statistical analyses. (comparing the group means)
The mean is the sum of the scores divided by the number of scores being summed.
The mean is the appropriate measure of central tendency for interval and ratio data. Cannot be used with nominal and ordinal data.
What is the mean for the following scores?
22, 28, 34, 16, 41, 9
http://www.algebralab.org/lessons/lesson.aspx?file=Algebra_StatMeanMedianMode.xml
162
Types of Descriptive Statistics
Frequency distributions
Ungrouped frequency distributions
Grouped frequency distributions
Percentage distributions
Measures of central tendency
Measures of dispersion
163
163
Example of an Ungrouped Frequency Distribution
164
Data are presented in raw, counted form
1: /
2: /////
3: ///
4: /
5: //
164
Example of a Grouped Frequency Distribution
Data are pre-grouped into categories
Ages 20 to 39: 14
Ages 40 to 59: 43
Ages 60 to 79: 26
Ages 80 to 100: 4
165
165
Example of Percentage Distribution
E.g., Costs of running a business:
Salaries: 41.7%
Maintenance: 8.3%
Equipment: 16.7%
Fixed costs: 8.3%
Supplies: 25%
166
166
167
Commonly Used Graphic Displays of Frequency Distribution
167
Measures of Dispersion
Measures of dispersion, or variability, are measures of individual differences of the members of the sample.
It gives some indication of how scores in a sample are dispersed around the mean.
It indicates how different the scores are, or the extent to which individual scores deviate from one another.
If individual scores are similar, the sample is homogenous, and vice versa….
168
Measures of Dispersion
Examples of Measures of Dispersion:
Range
Value obtained by subtracting the lowest score from the highest score. Ex: 9 – 1 = 8
Variance – larger the variance the greater the dispersion of scores
Standard deviation
The square root of the variance
The average difference (deviation) of values
Standardized scores
A common standardized score is the Z-score and it expresses the deviation from the mean
Scatter-plots
169
Scatterplots
Have two scales: horizontal axis (X) and vertical axis (Y)
Illustrate a relationship between two variables
170
170
Structure of a Plot
171
Example of a Scatterplot
172
172
Nonparametric Statistical Tests
Chi-squared test
Used with nominal data
Tests for differences among frequencies of groups
Mann-Whitney U-test
Median test
Kruskal-Wallis test
Spearman’s rank order correlation
173
Parametric Tests
t-Test
The most common analyses for significant differences between two samples is the t-Test
Among the most frequently used t tests are:
A test of whether the mean of a normally distributed population has a value specified in a null hypothesis.
A test of the null hypothesis that the means of two normally distributed populations are equal
Used to test for causality – by testing for significant differences between groups (to determine the effects of interventions)
174
Parametric Tests
Analysis of Variance (ANOVA)
Tests for differences between means of groups
Also, can be used for two or more groups
Used to examine multiple outcome variables
Results are given as a F score
175
Analysis of Relationships
Correlational Designs
Investigate relationships among variables
Measure strength of relationship (-1 to +1)
Measure direction of relationship (positive or negative)
Relationship Analysis
Measured by correlation coefficients such as
Pearson Product-Moment Correlation (r)
Used when variables are interval or ratio
Relationship Analysis
r = 0.60
r = -0.8
r = 0.0
Differences between
Variables
(nominal and ordinal data)
Chi Squared (x2)
Used for nominal or ordinal data
Variables are described as percentages instead of means
Chi Squared (x2)
To determine how closely observed
frequencies or probabilities match
expected frequencies or probabilities
Expected frequencies = the number of cases
that would be found in each cell of a
contingency table if the H0 were true
Chi Squared (x2) (actual data)
| Complication Incidence | 72 hours | 96 hours | Total |
| Had complications | 9 (18.0%) | 11 (22.0%) | 20 (20.0%) |
| No Complications | 41 (82.0%) | 39 (78.0%) | 80 (80.0%) |
| Total | 50 | 50 | 100 |
Chi Squared (x2) (expected frequencies)
| Complication Incidence | 72 hours | 96 hours | Total |
| Had complications | 10 (20.0%) | 10 (20.0%) | 20 (20.0%) |
| No Complications | 40 (80.0%) | 40 (80.0%) | 80 (80.0%) |
| Total | 50 (100%) | 50 (100%) | 100 (100%) |
Example of Chi Square study
Researchers examined the relationship between the background characteristics of homeless men and whether or not they were positive for hepatitis C virus (HCV) infection. They found that, for example, a significantly higher proportion of men who were veterans (p<.01), who were white or Latino (p<.001), and who had been in prison (p<.001) were HCV positive than men not in these categories.
Differences between
Groups
(nominal and ordinal data)
Mann Whitney U
to test for differences between
groups when the variables are
ordinal data
Mann Whitney U
Example:
A researcher wants to measure differences in happiness of nursing students at the beginning of their first semester and at the end of their last semester.
Needs help with similar assignment?
We are available 24x7 to deliver the best services and assignment ready within 3-4 hours? Order a custom-written, plagiarism-free paper
Get Answer Over WhatsApp
Order Paper Now
