false

Textbook Notes
(368,473)

Canada
(161,904)

York University
(12,849)

KINE 2049
(57)

Merv Mosher
(57)

Chapter 5

Unlock Document

Kinesiology & Health Science

KINE 2049

Merv Mosher

Fall

Description

Chapter 5:Experimental Research
Methods Overview
Purpose of methods section: explain the details of how the investigation was
conducted, it is after the introduction
Introduction: explanation of what the researcher is investigating and
description of the scope
Literature review is then summarized to provide background information
that leads to the problem
Methods section must be in enough detail to let another person replicate the
study, must have enough detailed info
Important info about the facilities and the equipment is used and presented
The methods section contains info about the procedure: length, duration,
types of treatments
Methods determine the “type” of research conducted
Editors Viewpoint: Who’s on First, What’s second?
Compares methods section to sorting out a line up in baseball
It is the authors responsibility to make sure that there is very little confusion
for the reader and all relevant info is presented
The purpose of this editors viewpoint is to show what info is needed to score
points with fans (readers)
You must describe your subjects/sample (race, age, ethnicity)
Provide clear operational definitions
Clarify where subjects were included/excluded
Identify independent and dependent variables
Talk about validity and reliability of measures taken in the experiment
Provide a step-by-step outline of how the study was carried out
In survey studies: describe the series of steps designed to maximize the
return rate
Qualitative studies: document procedures such as in-depth interviews should
be written in detail
In the methods section you describe how the subjects were selected
Subset of population = sample (greek u symbol)
Measure all member of pop = census (greek X symbol), can be an estimate of
u) =parameter
Statistics: facts about a sample
The sample:
The researcher must clearly identify the population of interest, must be clear
and not have any ambiguity regarding who is included
Next you must identify all members of the population selected
The researcher then has 2 questions:
1. How large the sample should be
2. How the people should be picked in the sample
Number of people in sample depends on: precision, reduction of error
probability and reliability As the size of the sample size increases, the reliability increases BUT no
always:
Some researches may cost a lot more if the sample size is increased, and the
cost may not be worth the reliability
And sometimes, it does not matter how large the sample size is, if it is biased
it does not represent the population
Example: The Literary Digest Poll
- Polled 10 million voters on who would win Roosevelt or Landon
- They predicted that Landon would win but Roosevelt won
- They sent out ballots via mail and people who were listed in the
directories were car owners, thus the sample was mostly upper class
people, who favored Landon
- People who were most likely to return mailed ballots were upper class
and more educated
- The poorer population was not shown in the poll, who favored Roosevelt
- The literary digest failed to select a proper cross section
Sample selection:
Must use a simple random process to select subjects from the population
Every member of the population must have an equal chance of being
selected
One persons selection should not affect the selection of another ex. Adding
another person to the sample because they are a sibling of a person in the
sample = violation of the independent condition
5 Types of Sampling Techniques:
Systematic Sampling:
When an entire list of people that could be in the sample is listed, you can use
a systematic way into selecting the sample
Ex. In a list of 1000 names, if you want 50 subjects, you’d choose every 20 th
person on the list
This leads to random sampling as long as the list contains all members of the
population
Stratified Random Sampling
Population is first divided into relevant subgroups and then random
sampling techniques are applied to each subgroup
Ex. If you want 100 students, you would pick students from each group
(senor, junior, sophomore) to come up with a stratified sample. This way the
sample represents the student population paying attention to student level
as well
Stratification may be done on more than one variable but normally is done
only on variables that are relevant to the dependent variable
Stratified sampling techniques ensure that small samples are still
representative of the population in what the y are investigating
Quota Sampling Is like stratified sampling, but does NOT involve random selection from the
various levels stratified
Ex. You survey the FIRST # of people that come to the survey
This may not be representative because it is not random sampling
This method would exclude students who didn’t come to do the survey
This method is convenient and requires you to have a list of all students, it is
not likely to be representative
Cluster Sampling
Designed for surveying large populations where is it hard to list all members
of the population
Ex. Wanted to survey elementary kids in your province, instead of a random
sample, you’d chose to do a random cluster sample:
Randomly select 5 counties out of 100 randomly select 5 school districts
randomly select 5 schools
-All of the students in the 5 schools would be chosen
- This method would violate the “independence” rule because the kids at each
school are connected to each other and are being chosen because they go to
the same school
Accidental Sampling
Involves the use of a self-selected sample, mostly used in televion surveys,
magazine surveys
The people that participate in these surveys are those who are interested and
willing to take the actions to become involved
This method is completely unacceptable in research reports
Ex. A radio station tell to call a # to talk about their views on abortion
- those with the strongest opinion will call
- does not represent people with moderate views and also excludes people
who did not know about it (do not listen to that radio station)
Summary on Sampling
The only sampling that is accepted in a research report is simple random
sampling
But sometime it is not possible
Does this mean that any research that does not use simple random sampling
is invalid? Yes.
BUT. If the researcher can show that the sample chosen is not different than
the actual population, it is okay, but results must be seen with caution
Variables
Variable: a characteristic that will show different values under different
conditions (exercise intensity, emotional state, temperature)
A main goal of research is to see how variables will behave or react to
different conditions in the environment
Independent
Independent variable: those that are manipulated by the investigator
Cause-and-effect shows the relation in the affected variable when the
independent variable is manipulated Some projects may have more than 1 independent variable because more
than 1 variable may have an effect on the dependent variable/ behavior
being studied
The simplest experiment = independent variable with 2 levels and 1
dependent variable
Dependent
Dependent variable: measured in an experiment that is affected by the
independent variable, it is the outcome
(look at studies on page 77-78)
Operational Definitions
Defines the characteristic of a variable (independent of dependent)
Ex. Endurance athlete can be put in categories such as 5k, 10k
Ex. People being weighed or in an experiment can be put in categories:
100lbs, 200lbs,
In an experiment, menstrual cycle was operationally defined as
premenstrual: 3-5 days before menses, menstrual: day 1-3 during menses,
and postmenstrual;: 10-12 days after onset of menses
Operational definitions are important so that researchers can completely
understand the details of the research
Control Variables
Factors other than the ones chosen/studied that could affect the outcome
Control variable: variable that is held constant during an experiment
The number of control variables are greater than the # of independent and
dependent variables
Crucial aspect is to find and fix as many control variables
2 categories:
1. Experiemental:
-variables that could be independent variables if the researcher
wanted to make them
2. Statistical:
- control variables that might impact the results of an experiment
- can use the Analysis of Covariance (ANOVA) to “control” the
variables, but this procedure costs in terms of statistical concerns and
can only be used in certain situations
Control Variables
Adding an additional level to the independent variable and this group is
exposed to everything as the experimental group except the treatments
Confounding Variables
The effect of an uncontrolled variable in an investigation. The presence of a
confounding variable may threaten internal validity of the investigation
Ex. In the racket experiment, the angle of the racket when hitting was 90
degrees the first time and then 93 the next, therefore, the angle of the racket
is a confounding variable
“The effect of string tension on, rebound, was confounded by the difference
in racket angle” confounding variables can undo the results of an experiment by providing
alternate explanations for the findings
Variance
a unit of measure of score variability
Primary Variance (systematic variance):
Expected or desired variance in the dependent variable by manipulating the
independent variable. It is wanted variation.
Secondary Variance (extraneous variance):
Consistent and unwanted variation in the measurement of the dependent
variable.
Ex. A scale for measuring weight that is not at 0, it measures 2 pounds over
the actual weight
Results will consistently over or under estimate the true value of the
dependent variable
Error Variance:
Inconsistent and unwanted variance
Random variability in the data that is not caused by the independent variable
Variation that occurs that cannot be explained ex. People differ on
characteristics that we don’t exactly know why (gender, heredity, diet, age)
“left over” variance- variance that is left over after controlling as many
variables as you can
Could also happen by the researcher taking wrong measurements
The effect of error variance is unpredictable and reduces the accuracy
Procedures and research instruments are designed to reduce error variance
Primary Variance Consistent and wanted variation on
measures of the dependent variable
Secondary Variance Consistent and unwanted variation on
measures of the dependent variable
Error Variance Inconsistent and unwanted variation on
measures of the dependent variable
Maximizing Primary Variance
1. goal of research is to see the effect of different levels of an independent
variable on a dependent variable, thus the researcher must fix the
independent variable levels far enough so that change in the dependent
variable can be measures
Ex. In tennis racket study, the string tension was set at 50, 60, and 70 lbs. If
the researcher had set it at 51, 52, and 53 lbs, much change wouldbt have
been recorded
2. goal is that the levels of the independent variables should be set at realistic
levels, in which it could be applied to real life
Ex. The 3 levels used in the racket test are the normal range used by tennis
player, thus they are realistic and the research would be useful in real life
Thus, by following these 2 goals, the researcher maximized the ways the
research would be used in real life Controlling Secondary Variance
Secondary variance can be reduced by careful selection of subjects to
experimental groups
Randomization and matching control unwanted consistent differences
Blind experimental design
Properly calibrated equipment
Minimizing Error Variance
Minimize by care and precision when gathering data
Differences between subjects can be controlled by using samples that are
representative of the population
Appropriate and careful measurement procedures
C

More
Less
Related notes for KINE 2049

Join OneClass

Access over 10 million pages of study

documents for 1.3 million courses.

Sign up

Join to view

Continue

Continue
OR

By registering, I agree to the
Terms
and
Privacy Policies

Already have an account?
Log in

Just a few more details

So we can recommend you notes for your school.

Reset Password

Please enter below the email address you registered with and we will send you a link to reset your password.