Transcript Chapter 1

Chapter 1
Measurement, Statistics, and
Research
What is Measurement?



Measurement is the process of
comparing a value to a standard
Statistics is a mathematical tool used
for interpretation
Precision is essential: if the
measurement is not PRECISE, the
results cannot be TRUSTED
What is Measurement?

To be acceptable the data must be
– Valid, Reliable & objective



Valid (must be compared to known
value or method)
Reliable – is the measurement
consistent?
Objective – free from BIAS?
Steps in Measurement
Process
1.
2.
3.
4.
Object to be measured is identified and
defined
The standard to which the object is to be
compared is identified and defined
A comparison of the object to the
standard is made
A quantitative statement of the
relationship between the standard an
object is made (statistical evaluation)
Variables and Constants


A variable is a characteristic that can
assume more than one value
A constant can assume only one value
Types of Variables


Continuous variable – can assume any
value (ht, wt)
Discrete variable – limited to certain
values: integers or whole numbers
(2.5 children?)
Classification of Data
or Level of Measurement




Nominal Scale: mutually exclusive (male, female)
Ordinal Scale: gives quantitative order to the
variable, but it DOES NOT indicate how much better
one score is than another (RPE of 2 is not twice of
1)
Interval Scale: has equal units and zero is not an
absence of the variable (temperature)
Ratio Scale: based on order, has equal distance
between scale points, and zero is an absence of
value
Research Design & Statistical
Analysis


Research is a technique for solving
problems. Identifying the problem is critical
Types of Research:
– Historical
– Descriptive
– Experimental: involves manipulating and
controlling variables to solve a problem

Hypothesis:
– an educated guess
– based on prior research
– Can be tested
Hypothesis Testing





Research Hypothesis (H1): predicts
relationships or differences between groups
Null Hypothesis (H0): predicts NO
relationship or differences between groups
The statistical analysis reports the
PROBABILITY that the results would if H0
were true
If the probability (1 in 100) or (5 in 100)
that the null is true, we REJECT the null and
ACCEPT H1
NOTE: We never PROVED EITHER!
Independent & Dependent
Variables




Independent Variable: totally free to vary.
(balance is independent of VO2)
Dependent Variable: NOT free to vary (ht
and wt)
The INDEPENDENT VARIABLE is controlled
by the researcher (effects of exercise on
body fat)
The DEPENDENT VARIABLE is the variable
being studied (effects of exercise on body
fat)
Internal Validity

Internal Validity:
– refers to the design of the study
– All potential intervening variables must be
controlled (rat studies are easier to
control)
– Failure to use a control group harms
internal validity
– Instrument Error reduces internal validity
– Investigator Bias reduces internal validity
External Validity



External Validity refers to the ability to
generalize the results of a SAMPLE to the
POPULATION (rat studies don’t always
generalize to humans)
If a sample is not RANDOM it may not
represent the population
The process of generalizing from a SAMPLE
to a POPULATION is statistical inference
Statistical Inference







A Population is a group with a common
characteristic
A population is usually large and it is difficult to
measure all members
To make inference about a population we take a
representative sample (RANDOM)
In a random sample each member of the
population is equally likely to be selected
A stratified sample is a sample that is selected
according to existing subcategories (rep, dem, ind)
A sample cannot accurately represent the
population unless it is drawn without BIAS
In a bias free sample selection of one member does
not affect to selection of future subjects
Parameters and Statistics



A parameter represents the population
A statistic represents the sample
The difference between a statistic and
a parameter is the result of sampling
error
Probability and Hypothesis Testing



Statistics is the science of making
educated guesses
Statistics allow us to make a
statement and then cite the odds that
it is correct
A random sample of 200 females have
a mean ht of 5’ 2” ± 2”. The odds are
95 to 5 that this mean is correct.
Probability and Hypothesis Testing



A random sample of 200 females have a
mean ht of 5’ 2” ± 2”.
This means that the odds are 95 to 5 that
the true mean is between 5’ and 5’ 4”
If a sample results in a mean of 5’ 3” we
accept a hypothesis that the ht is 5’ 3”
because it lies within the limits (5’ and 5’ 4”)
Theories and Hypotheses




A theory is a belief regarding a
concept or a series of related concepts
Many hypotheses can be TESTED
If a sufficient number of results
confirm the theory it is accepted as
true
Mental practice improves performance
Misuse of Statistics






Abdominal Exercise devices?
Toothpaste?
Examples of statistics that may or may not
be true
Lack of random sample, small sample size,
research is PAID
Outliers: extreme scores (more than 3 SD)
Mean income (Income is a skewed
distribution)