NURS341 Study Guide - Winter 2018, Comprehensive Midterm Notes - Empirical Research, Triangulation, Schizophrenia
12 Oct 2018
School
Department
Course
Professor
NURS341
MIDTERM EXAM
STUDY GUIDE
Fall 2018
- Measurement bias – Influences affecting how the data were collected (weight on a
scale that is always 2 kg too high, inaccurate blood pressure cuff, non-valid
questionnaire)
- Performance bias – influences affecting the study (ex. practitioners knowing they were
part of the stud, partiipats’ aareess of treatets/plaeo, additioal are to oe
or both groups that was not intended)
- Reporting bias – influences affecting what results are selected for publication by the
researchers or journal editors (discouragement with negative results, difficult
interpreting mixed results, increased enthusiasm to publish positive results or those that
only support one hypothesis)
Sample
- Describes the participants in the study
Sample size – small samples can influence interpretation of results
o Calculations to determine whether the sample size is sufficient to detect
clinically meaningful difference
o Sample size usually calculated on basis of one/two primary outcomes
o Rare outcomes require larger samples
Power – ability of a study to detect meaningful differences
o Typically set at 0.80
o Determines sufficient sample size
Data collection
- Describes the method for gathering the study information
- To judge whether a measurement tool/survey has been validated, look for terms validity
and reliability (psychometric properties)
- At a minimum, content validity of a tool is normally reported
- Tool considered valid if it accurately measures what it is intended to measure
3 common types of reliability:
- Inter-rater reliability – consistency of two people gathering the same data that would
get same results
- Test-retest reliability – determine whether similar results will be obtained if the same
tool is used at different times
- Internal consistency – deterie hether ites o a sure hag together or are
actually different concepts muddled together
Sensitivity – how good a test is at detecting who has a condition/disease
Specificity – how good a test is at detecting who does not have a condition/disease
Data analysis and results
find more resources at oneclass.com
find more resources at oneclass.com
- Presents the statistical procedures or quantitative methods used to analyze the data
and findings
Descriptive statistics
- Express characteristics or summarize data
- Mean – average score
- Median – score in the middle of a range of scores
- Mode – most frequency score
- Range – describes the variability and is the difference between the highest and lowest
scores
- Standard deviation – another indication of variability and is calculated from the average
squared differences (deviations) about the mean
Inferential statistics
- Depict inference, relationships, and probabilities
- Common statistical concepts and terms:
- Statistical significance – the statistical calculation shows a relationship between the
variables that is unlikely to be due to chance only (due to experimental intervention)
o Look for p < 0.05, probability of 5% (0.05) or less that the results are due to
chance alone
- Clinical relevance – Are the results practical or meaningful
- T-test – used to analyze the difference between 2 mean scores
- Analysis of variance (ANOVA) – used to analyze mean score differences between 3 or
more groups
- Multivariate analysis of variance (MANOVA) – compare mean scores of 2 or more
groups on 2 or more variables at the same time
o Ex. do confidence scores for managing insulin doses for strenuous exercise and
for travel differ between people in whom diabetes has been newly diagnosed
and those in whom diabetes is diagnosed more than 2 years previously
- Chi-square test (x2) – compares the actual number with an expected number
- Confidence interval (CI) – range of values within which the true value is expected to be
found
- Correlation – relationship between 2 variables
- If 2 variables are correlated, does not mean that one caused the other, means that one
variable is related to the other variable
- Correlations can be positive or negative with a range of scores from negative association
(-1) to positive association (+1)
- R of 0.3 is low, 0.5 is medium, 0.8 is high
- Correlation between 2 raters (inter rater reliability) should be at least 0.80
- Effect size – strength of a relationship between 2 variables or an estimate of the impact
of an intervention
find more resources at oneclass.com
find more resources at oneclass.com
Document Summary
Measurement bias influences affecting how the data were collected (weight on a scale that is always 2 kg too high, inaccurate blood pressure cuff, non-valid questionnaire) Performance bias influences affecting the study (ex. practitioners knowing they were part of the stud(cid:455), parti(cid:272)ipa(cid:374)ts" a(cid:449)are(cid:374)ess of treat(cid:373)e(cid:374)ts/pla(cid:272)e(cid:271)o, additio(cid:374)al (cid:272)are to o(cid:374)e or both groups that was not intended) Reporting bias influences affecting what results are selected for publication by the researchers or journal editors (discouragement with negative results, difficult interpreting mixed results, increased enthusiasm to publish positive results or those that only support one hypothesis) Power ability of a study to detect meaningful differences: typically set at 0. 80, determines sufficient sample size. To judge whether a measurement tool/survey has been validated, look for terms validity and reliability (psychometric properties) At a minimum, content validity of a tool is normally reported. Tool considered valid if it accurately measures what it is intended to measure.
