An overview of Research & Common Terms Flashcards
Inductive Reasoning
Makes broad generalizations from specific observations
Deductive Reasoning
Starts out with a general statement, or hypothesis, and examines the possibilities to reach a specific, logical conclusion
Hypothesis
a written, declarative statement in the present tense of a prediction of the relationship between two or more variables. Used in quantitative research (not in qualitative)
Null Hypothesis
Predicts there is no relationship between the variables stated in the hypothesis. Not acknowledging a null hypothesis threatens overall study validity. Careful sample selection using power analysis can help prevent these errors.
Type I Errors
When the null hypothesis is rejected when it should not be. This is of more concern to researchers than type II errors
Type II Errors
When the null hypothesis is not rejected when it should be since a relationship between the studied variable does exist.
Quantitative Research
Grounded upon the scientific method of theory testing and focuses on more
objective evidence.
* focused on determining the cause and effect of a phenomenon, as researchers believe that events
are not random but can be linked to a specific cause
* research is tightly controlled throughout
* uses numbers and statistics to produce evidence
* the people being studied are referred to as “subjects”
* relies on unbiased observation and reliable/valid measurement of variables
Independent and Dependent variables
cause and effect, the hypothesis
Independent variables
treatment, intervention, or the experimental variable
Dependent variables
outcome, the ruling out of extraneous effects
Extraneous variable
Unplanned effect (cause) on an outcome in a research study
Hawthorne effect
an example of an extraneous variable: refers to the
the psychological effect when study subjects change their behavior from how
they would normally behave because they know they are being
watched/observed
Methods for survey research
interviews and questionnaires. May address knowledge, beliefs,
perceptions, attitudes, feelings, experiences, behaviors, etc. Questions should be clear, simple,
ordered from general to specific.
Questionnaire
can be written (paper) or electronic (online or computer generated)
o Pros: inexpensive, fast, easy to analyze, allow for anonymity,
o Cons: low response rates, does not allow for clarification
Attitude scales
Likert scale and Semantic Differential scale are structured, self report measures to learn more about the attitudes of the subjects
Likert Scale
attitude statements of usually 5 to 7 points, ranging from
Strongly Disagree to Strongly Agree. Scores of all questions are summed to
obtain one total score.
Semantic Differential Scale
Not as commonly used as the Likert scale. Pairs
of 2 opposite adjectives are placed on lines with a 7-point scale between
them. Subjects mark the point that best represents their attitude.
Quantitative Research Designs
Descriptive and Correlational (do not involve intervention on the part of the researcher)
Descriptive studies
Describe characteristics of a topic Ie: individuals, groups, situations and/or frequency of occurrence of certain phenomena
Correlational studies
Examine relationships between variables or between subjects and the strength and direction of the relationship
Do Quasi-experimental involve a researcher intervention?
Yes
Quasi-experimental studies
Examine causality but there is NO random assignment and/or control group
Experimental studies
Examine causality and DO have random assignment of subjects, with both an experimental and a control group
Clinical Trials
Research studies, experimental in design, which assess the effects of specific new or revised clinical interventions.
Examples include clinical treatments, medical equipment use on clients, medications. In these studies, safety and efficacy of the intervention(s) are studied.
RCT (randomized controlled trials)
Viewed as the gold standard for evidence-based practice decisions
Cross sectional studies
Look at one phenomenon at one point in time and across multiple separate populations that differ in a clear characteristic such as age, developmental status, etc. Allows for a picture of a phenomenon as it exists in the present.
no follow up intented
Advantage: They are relatively economical, easy to design and implement, quick to conduct and obtain results
Disadvantage: Do not capture changes that occur over time
Longitudinal Prospective Cohort Study
Follow subjects over a period of time in the future. Data is collected at two or more different points in time.
Prospective Study
The study of interest is identified in the present and subjects are followed for a given time frame in the future
Cohort Study
A type of longitudinal study in which subjects come from a similar background or were born in the same time period
Retrospective Study
uses past data which has already been collected about events that have already occurred. Record review is a possible source of data. This is a good type of study for exploratory research but not for experimental research.
Pilot study
smaller scale version of a planned study which identifies and prevents a problem that could occur in the larger study. Testing of methods and procedures of the planned study are done for feasibility and accuracy of the planned larger study. Sample population for the pilot study is similar to that of the larger one.
Double-blind experiment
Neither the subjects not the researchers know who the control group and who the experimental group is. They are the goal standard for experimental research.
Control Group
Group of subjects in an experimental study that is comparable to the experimental group but does not receive the experimental intervention/treatment or is given an alternate treatment, such as the traditional one. Using a control group provides a baseline to measure the effects of the treatment.
The purpose of a control group
To decrease threats to external validity and to increase confidence that the findings of the study can be generalized to other populations, etc. Also, using a control group helps to decrease both possible researcher bias and participant Hawthrone effect.
Sampling
A process that selects representative units of a population for a study, as it is rarely feasible or necessary to sample the entire populations of interest to obtain accurate and meaningful information
Eligibility criteria
Descriptors of the population which provide a basis for selection of the sample. Include age, gender, socioeconomic status, level of education, religion and ethnicity
Target Population
The entire set of cases whom the researcher would like to make generalizations about. Due to constraints involved with this, the researcher usually uses an accessible population (one that meets the population criteria and is also available)
Random sampling
ensures the sample is representative of the population from which it was chosen is representative of the population from which it was chosen, allows for generalizability of the findings.
Mathematical probability methods are used to guarantee that each person in a population has an equal chance of being selected for the sample.
The selection process is tightly controlled and is focused on eliminating bias.
Simple
random numbers are used to select subjects from the total population
Stratified
subgroups are selected from the population based on certain characteristics and a representative sample from each subgroup or stratum is then randomly chosen
Cluster
entire groups are randomly selected in stages and subjects are then randomly selected only from those groups or clusters
Systematic
predetermined sampling intervals are used to select subjects from the population
Measures of Central Tendency
Mean
Median
Mode
Mean
average value of a data set
Median
Midpoint of a data set
Mode
Most frequent value of a data set
Four Levels of Measurement
Nominal
Ordinal
Interval
Ratio
Nominal
Lowest or least rigorous measurement (ex: gender, blood types)
Ordinal
Rank ordering (ex: pain scale)
Interval
Consistent distance between ranks (ex: age, thermometer reading)
Ratio
True or natural zero point (ex: volume, speed)
Correlations
Focus on the relationship between variables.
0 indicates the absence of a relationship.
+1 is a perfect negative (or inverse relationship)
-1 is a perfect negative (or inverse relationships)
- Scatter plots: the closer the dots are to forming a straight line, the higher the correlation.
If the data are distributed from lower left to upper right, a positive correlation exists
If data are distributed from upper left to lower right, a negative correlation exists.
Dots all over the graph represent that there is no correlation present.
Inferential Statistics
uses data from samples to make inference about a population
T-tests
Examines differences between means of two or more groups or sets of values
ANOVA
Analysis of variance; compares the differences between two or more groups or sets of values
Chi-square test
compares sets of frequency or percentage of data
Level of significance
A statistical measure of the probability of rejecting a null hypothesis when it it true (saying there is a relationship among variables when there is not)
When statistical significance is obtained, this indicates that results were likely due to causation rather than chance.
Most common levels of significance are .05 and .01, with .05 being the usual minimum acceptable level
.01 and .001 are used when the decision to be made has important consequences for treatment.
The decision for setting the criteria for level of significance to be used in a study should be made before data collection begins to ensure research integrity
- Keep in mind: sample size has a great effect on the level of statistical significance (small sample sizes can result in low significance levels)
Interrater Reliability
When two or more independent raters use the same tool and are in agreement regarding their ratings
Quantitative Validity
(Internal validity)
The degree to which change in the effect (dependent variable) can be attributed only to the cause (independent variable), and not to extraneous variables
Quantitative Validity
(External validity)
The generalizability of the findings of an experimental study to other people and settings.
Qualitative Research
Focus on discovering and interpreting the subjective meaning of an experience to an individual or group. Participants are referred to as “participants” (not subjects, as in quantitative research). Subjects are not randomly selected and rather, are handpicked for their perceived representation of the population of interest (purposeful sampling)
Phenomenological
Describes experiences or phenomena from the point of view of the individuals involved; referred to as “lived experiences”
Ethnopgraphic
Focuses on the understanding of the culture of a group of people. The researcher becomes an active culture.
Bracketing
Method used to control for researcher bias to help ensure clear and accurate observations. Researchers reflects on their personal thoughts, feelings, etc. and set them aside [puts them in brackets]
Trustworthiness
Measure of truth and rigor
Four criteria of characteristics:
Credibility, dependability, confirmability, transferability
Data Saturation
A point in qualitative studies when there are no new ideas noted in the data analysis and it is noted that saturation of themes/categories has occurred
Field Notes
Reflective notes documented by the researcher’s reflection on strategies and methodologies used, analysis of observations, and personal feelings.
Focus group
Small group of 6-12 people for their similarities to discuss thoughts, feelings, etc.
Researchers observe participants for verbal and nonverbal behaviors and interactions.
Report
Written in an informal style. Direct quotes used, which adds to the credibility and trustworthiness. Themes are identified.
Interview
They are more personal than questionnaires. Can be structured, unstructured, or semistructured.
Pros: Higher response rate than questionnaires, increased depth of information.
Cons: Cost, lack of anonymity, possible interviewer bias.
