Stats midterm

Research

Research

The systematic investigation into and study of materials and sources in order to establish facts and reach new conclusions

Clinical research

is a structured process of investigating facts and theories and exploring connections

Research trinity

Design, measurement and analysis

Design

Establishes the “integrity” of the research.

Measurement

Determining “constructs”, operational definitions

Analysis

Establishes conclusion validity

Statistics

Formulas, calculations, equations, etc., to analyze the results of research studies, or other information of interest

Empirical

Information is obtained via observation and measurement

Likert scale

1-5 (Strongly agree - Strongly disagree)

Sources of knowledge

1. Tradition

2. Authority

3. Trial & error

4. Deductive & inductive reasoning

   1. Scientific research

Anecdote

A short story about an interesting or funny event or occurrence

Tradition

Skills and knowledge handed down over time \n – Advantage: Each subsequent generation starts with \n a more advanced state of knowledge \n - Greater component of training than ongoing \n education

Authority

Specialist in the area \n Licensed psychologist, accountant. \n Will have more detailed knowledge within \n their realm. \n BUT: \n Only advantageous if specialist is well-informed on \n current best evidence

Trial and Error

Altering practice based on trying something \n new.... \n BUT: \n - Not systematic, haphazard approach. \n - Difficult to determine why or why not a solution \n has worked

Deductive & Inductive Reasoning

Application of logical reasoning to come up with a solution.

Deductive reasoning

General observation → specific conclusion

Inductive reasoning

Specific observation → general conclusion

The Scientific Method

Incorporating inductive and deductive reasoning \n Systematic and controlled process \n Involves a sequence of events

- Make an observation \n - Formulate a question & hypothesis \n - Design an experiment..... \n - Execute the experiment

- Analyze the results...... \n - Draw conclusion

Key elements of scientific method

Skepticism \n Open-mindedness \n Objectivity

Quantitative research

Measurement of outcomes using numerical data under \n standardized conditions

Qualitative research

- Deriving conclusions from open-ended questionnaires, surveys, interviews and observations. \n - Not applying numerical descriptions to observations.

Basic research

- Pure or fundamental research \n - Direct practical applications? \n - Investigates underlying psychological, anatomical, or \n physiological mechanisms

Applied research

- Research that has direct practical application \n – Based on a general premise (assumption) that is not directly tested in the study

Translational research

Application of basic scientific findings to clinically relevant issues, and simultaneously, the generation of scientific questions based on clinical dilemmas.

Types of research

1. Descriptive \n 2) Exploratory \n 3) Experimental

Descriptive research

Describing a group of individuals on a set of variables to document their characteristics.

Example: Describing personality traits among in-residence first-year university students.

- Case study

- Normative study (Establishing baseline statistics for specific groups of people)

Exploratory research

Examining the interaction of various factors on people’s behaviour/characteristics.

- Epidemiology

- Cohort study (Following a group of people over time)

- Case-control study (Selecting groups of individuals based on the presence of the condition investigated)

- Correlational-predictive research (Examining relationships between variables)

- Methodological (Evaluating/demonstrating the reliability, validity, and/or validity of measurement instruments and/or procedures)

Correlational study

Identify the dependent variable (this is the variable that we are trying to predict)...... \n Identify the independent variable (i.e., the x-axis variable)...... \n Make a scatterplot of the data

Experimental research

Designed to compare groups to establish a cause-and-effect \n relationship

- RCT – Randomized Controlled Trial (Controlling factors that can affects of the results of a research study)

- Quasi-experimental designs (Same as experimental designs, but with less control)

- Pre-experimental designs

Developing a research question

-Introspection

-Find the exception to the rule

-Matter of degree

Hypothesis

Educated “guess”, sometimes truly a guess

Research hypothesis

Based on knowledge

Statistical hypothesis

Probability of occurrence

Null vs alternative hypothesis (type 1 and 2 error probability)

Scientific law

A statement based on repeated experimental observation

“Newton law of motion”

Scientific theory

Psychology relies much more on theories than laws

Well substantiated explanation of an event

Key concepts of hypothesis

-Variables: Elements in a study that can vary of assume different values

-Qualitative or quantitative (numerical)

Constant: Factor that does not change. Can be used for experimental control: Eliminating confounding items

Conceptual definition

General definition of a variable

Operational definition

Definition of the same variable (conceptual) in context

Personality

Conceptual definition: Characteristic thoughts, emotional responses, and behaviours that are relatively stable in an individual over time across circumstances.

Research study: Operational definition (more specific: The big five)

The big five

Oppenness

Conscientiousness

Extroversion

Agreeableness

Neuroticism

Deductive hypothesis

Based on a theoretical foundation

Minimal direct observation

Inductive hypothesis

Based primarily on experimental findings

Examining patterns displayed in research papers, public health data, etc

Research hypothesis

States what the experimenter thinks will happen

Variables will be different or related to each other OR no differences or relationships… sometimes the same as the null hypothesis

How to evaluate the hypothesis

-Does it correspond with reality

-Is it parsimonious (phenomena behave in the most economic way) (making minimum number of assumptions)

-How specific is it

-Barnum effect (tendency of people to believe general descriptions of their personality are highly accurate)

-Is it falsifiable or refutable?

-Research study must be able to accept or refute hypothesis

Experimental design

Independent variable

Dependent variable

Independent variable (x)

Refers to variable that can be manipulated (ex: Treatment group) with or without

Dependent variable

The outcome variable

What will be measured

Ex: Number of psychotic episodes per week

Non-experimental designs

Explanatory (or predictor) variable

Criterion (or response) variable

Explanatory variable

It is what a researcher manipulates or observes changes in

Criterion (or response) variable

Outcome variable in non-experimental designs (like dependent)

How do you get participants?

Volunteer

Must know all relevant information

Variables

Anything that, when measured, will produce 2 or more different scores

Qualiative

Research participant information that is NOT measured in “amounts” but classifies the participant based on the measure

Muscle strength example: Strong, moderate, weak

Pain example: Intense, moderate, mild

Quantitative

A number is assigned to a measure obtained from the research participant (amount)

Muscle strength example: Amount of weight lifted

Pain example: Visual analog scale (0-10 scale)

Ethics

Application of moral principles concerning what an individual considers right and wrong to help guide one’s decisions and behaviour

Utilitarian

Greatest good for the greatest number

Altruistic perspective (ethics)

Involves helping others without personal benefit

Selfless actions

Egoism

Individuals should act in accordance with their own self-interests

Eric Poehlman did data fabrication and was sent to U.S. federal prison

Research ethics governed by

-Federal and institutional policy

-Research misconduct

Research misconduct

A serious breach of good research practise that includes but is not limited to falsification, fabrication, and plagiarism committed willfully or through gross negligence when planning, carrying out or reporting on research

Factors affecting ethical research

-Scientific competition

-Societal norms

-Political influences

-”Trendy” research

Scientific competition

Competing with others with similar interests

“Who will find the answer first”

Societal norms

-Undefined norms within a discipline or an organization

-Will foster scientific competition

-3 P’s: Policy, procedure, practise

Political influences

-Ability to obtain external funding for research

-Research activities are “tailored” around funding hopes

-Research activities that may “impress” influential people

-Status, monetary gain, decision making power

“Trendy” research

-Blindly “doing what other people are doing”

-Research questions often go unanswered

With ethics, the researcher must:

-Look over previous relevant work (Identify and acknowledge)

-Identify potential sources of error (limitations)

Internal responsibility

Researchers

External responsibility

Personnel not associated with study

ex: Human subjects committee, scientific journal personnel

Human subjects committee

-Serves to protect the rights of human participants from physical, psychological, or social harm

-Non-institutional members of the community

-Provides approval to being study

Journal editors and reviewers

-Determine the quality of submitted work

-Process prone to ethical violations (Editors found to sometimes violate to gain political favour) (Reviewers may reject without adequate justification due to competitive interests)

Final responsibility of ethical research

-Scientific community at large

-Readership of journals

-Identifying weaknesses in published work

-Letter to editor

Research ethics board - Review levels

-Exempt (Lowest): Research poses less than minimal risk (sometimes not reviewed at all)

-Expedited (Common): Poses minimal risk typically encountered in daily life

-Full review: Research poses greater than minimal risk (reviewed by at least five members)

Nuremberg code

Developed by military tribunals

Established basic ethical principles

-Participants must be able to make voluntary decision

-Must not make any positive or negative inducement for participants to enroll in study

-Participants must provide informed consent

-Investigator must be compotent

Declatation of Helsinki

Emphasized review of research protocols by an independent committee not involved with the project

Developed by institutional review board

Belmont report

Established basic ethical principles in research in the United States

-Respect for persons (Provide sufficient information, treat people as people)

-People capable of making their own choices

-Informed consent (Risks and benefits) (Understand study and procedure) (Free from coercion) (Freedom to withdraw)

-Benefits must outweigh risks (Cost-benefit analysis)

-Researchers should do no harm

-Requires burdens to be distributed fairly, no favour or distain

Ethical treatment of data

-Falsification/Fabrication of data (eric poehlman) (Autism and vaccines)

-Massaging the data (data manipulated to get statistical significance)

-Dropping participants if outliers

-Reported only significant results

-Including some variables but not others

Role of quantitative methods (stats)

Applying numbers, equations, formulas, math techniques, graphical analyses

Research process

Theory → hypotheses → “Collect data” Observations → Empirical generalizations (Interpret observations)

Loop

Research question → research design → Findings/answers → Share/publish findings

Select design based on research goal

Non-experimental

Quasi-experimental

Experimental

Determine measures

Keys to effective measurement

Types of measurement error

Reliability and validity of measures

Recruit research participants

Types of sampling

Sampling issues

Population

Refers to entire group of individuals to which a law of nature applies

Ex: All uni students in canada, all canadians over 65, all people with cardiovascular disease

Parameters

Numbers that characterize the population (greek symbols) (u used for mean) (o used for standard deviation) (p used for correlation coefficient)

Sample

Relatively small subset of a population that is intended to represent that population (ex: All uni students in canada = sample: 50 undergrads from uni of guelph)

Statistics

Numbers that characterize the sample (english symbols = Sx, r, X (with a line on top))

Random sample

Individuals that make up the sample are randomly selected from population

May not be representative sample

Difficult to randomly select

Data

Scores on variables, or info expressed as numbers

Variables

Traits that can change values from case to case

Cases

Entities from which data are gathered (people, businesses, cities, countries)

The 2 approaches to statistical applications

Descriptive

Inferential

Descriptive statistical approach

-Describing basic characteristics of a person or group

-Summarize one variable (univariate)

-Summarize relationship between 2 variables (bivariate)

-Summarize the relationship between three or more variables (multivariate)

Inferential statistical approach

-“Inferring” results of an analysis to a larger population

-Involves statistical “hypothesis testing”

-Also involves descriptive statistics

-Establishes… cause and effect relationship

Discrete variables

Measured in units that cannot be subdivided (gaps)

Variable classified in whole units

Dichotomous variables (only 2 possible outcomes)

or categorical variables

Continuous variables

Measured in unit that can be subdivided infinitely

Variable taking on a value along a continuum (no gaps)

Infinite number of fractional components (in theory)

Limited by the precision of the testing instrument (ex: age)

Dependent variable

Measuring an outcome (ex: results of a federal election)

Independent variable

A factor influencing an outcome i.e., the dependent variable (Residential location regarding a SARS CoV2 infections (DV))

Precision measurement

Refers to exactness of measure

Sensitivity of testing instrument

Measurement scales

-Nominal

-Ordinal

-Interval

-Ratio

Nominal scales

Identification scores

Does not measure amount or value

Classifies groups or responses

Ex: Sex at birth, gender, do professional athletes make too much money? (Yes or no)

Criteria: Mutually exclusive, exhaustive (category for every score), homogeneous (include cases that are comparable)

Ordinal scale

-Scores that indicate rank order

-Scores only indicate a position between items

-Ex: Opinion surveys, clinical measurement of muscle strength