Measurement and Eval Exam 1

Scientific method steps

1. Observation and question

2. Lit review / research

3. Experiment (Hypothesis)

4. Evaluate

5. Repeat

Domains of human performance

1. Cognitive

2. Psychomotor

3. Affective

4. Health

5. Growth

Psychomotor

Sport performance

Affective

Social emotional

Test

Tool used to measure and attribute

Measurement

Value you get from a test

Evaluation

Interpretation of a measurement

Qualitative

Descriptive measurement

Quantitative

Numeric measurement

Norm referenced

Evaluation of a test measurement is based on comparison to others

Criteria referenced

Evaluation of a test measurement is based on a set criteria

Formative test

An initial test to be used to compare to future tests

Summative test

A test used to evaluate at a given time compared to initial/ formative test

Continuous values

Decimal/ fraction

Discrete values

Whole numbers

Scales of measurement

• Categorical

• Numerical

Categorical data

• Nominal

• Ordinal

Numeric data

• Interval

• Ratio

Nominal data

Group into categories without rank

• Does not say anything about ability

Nominal data examples

• Sex

• Drivers license #

• Jersey #

• Race

Ordinal data

Ranked data, but is not in intervals

• No common order between values

Ordinal data example

Marathon race

Interval data

Ranked, set intervals, but no absolute zero

• Can have negative numbers

Interval data example

Temperature

Ratio data

Ranked values, set intervals, absolute zero

Peer reviewed research

1. Empirical research

2. Literature reviews

3. Meta analysis

Emperical research

Singular study

Literature review

Summary of multiple studies

• No no statistical analysis

• Qualitative

Meta analysis

Combines data from multiple studies to performa a large group analysis

• Talks about subjects

• Quantitative

Order to read articles

1. Abstract

2. Discussion

3. Methods

4. Intro

Research definition

Way to gather information

• make a second decision or judgement

• develop new knowledge

Applied research

Addresses immediate problems

• real-world settings

• direct application for practitioners

Basic research

Addresses theoretical problems

• limited direct application

Deduction/ deductive reasoning

• Start with general knowledge of population

• Assume all is same for every individual

Induction/ inferential statistics

• Start with specific examples

• Draw conclusion to larger population

Population

Parameter

Sample

Statistics

Parametric

Defines a curve

Parametric procedures

Sample represents a population

Non-parametric procedures

Assumptions

• Sample does not represent a population

Lepokurtic

• More in center, less on outside

• Skinny

Playtikurtic

• Less people in middle

• Wider

Negative skew

More people on the right side

Positive skew

More people on left side

Independent variable

What is being controlled/ manipulated

• How grouping individuals

• How many times are groups being tested

Dependent variable

What is being measured

• Test chosen must be measurable

3 Primary charges to research ethics

1. Respect

2. Benefice

3. Justice

Respect

Ability to participate in study and quit whenever

• wellfare of subject

• knowledge of results

Benefice

Potential benefits of study outweigh the potential disadvantages

Justice

Burdens of study are equally distributed among all populations

Protected populations

• minors

• pregnant women

• prisoners

• mentally handicapped

How to get minors involved in studies

Assent: minor consent

Consent: parent/ legal guardian approval

Internal Review Board

Determines if subjects are protected during a study

• research cannot start without approval

Research hypothesis

Proposed answer to the research question and tries to answer why

Statistical hypotheses

1. Null

2. Alternative

Null hypothesis

States that there is no difference, all is equal, no relationship

Alternative hypothesis

States there is at least one difference, there is a relationship

Directional alternative hypothesis

• one-tailed

• more powerful

• claims one group to be better than another

Non-directional alternative hypothesis

• two-tailed

• states there is a difference, but not sure exactly what

P value

p< 0.05

a priori

Refers to prior knowledge about a population

Reliability

Give same results over and over

Objectivity

Same results regardless of who performs the test

Validity

Tests what it is intended to test

• does test give the “correct” answer?

Can a test be reliable, but not valid

Yes

Can a test be unreliable and not valid

Yes

Can a test be unreliable, but valid

No

Internal validity

Used to ensure “soundness” of the research

• limits effects of other possible independent variables

Selection bias

Choosing sample that will create desired results

How to solve selection bias

• Random selection

• Random assignment

Random selection

Subjects are randomly chosen from population

Random assignment

Subjects are randomly assigned to groups

History in research

Issues prior to the study that could influence results

Maturation in research

Issues outside the study or during study that could influence results

Mortality in research

Subjects dropping out of study

Placebo effect

Subjects will have effect, even if not given actual treatment

Blind studies

Subjects don’t know what group they are in

• controlled or uncontrolled

Double-blind studies

Subjects and testers don’t know which group is which

Angel effect

tester thinks certain subjects are supposed to do well, so they grade them more leniently

Ceiling/ floor issues

Test is not robust enough to discriminate

External validity

Results are relatable to “real world”

Tye 1 error nickname

Alpha error

Type 2 error nickname

Beta error

Type 1 error

• Rejects a true null

• False positive

• Says there is an effect, when there isn’t one

• Reject null when shouldn’t

Type 2 error

• Fail to reject a false null

• False negative

• Says there is no effect, when there is one

• Accept null when shouldn’t

N

Number of subjects in a study

x̄

Mean

Normative

Compared to people around

Criterian

Based on a criteria