Professional Documents
Culture Documents
STATISTICAL I.A.
Part 1 (Intro)
Done by STATS + IA + me
ORDER OF PRESENTATION
Introduction to statistics/Statistical Descriptive Statistics
Studies Standardized Scores
Designing a good Statistical Studies Measures of Association
Planning
SBA Session
Development
Sampling
Instrumentation
Pilot
Final Assembly
What is Statistics?
Consider the Telescope…… it helps us to see things that are too far away
Outlined below are the five most common errors in statistics that lead to
misleading/poor statistical results.
START EARLY- Things will not always go as planned, so plan for the unexpected by
giving yourself and the students adequate time to work through any unforeseen
circumstances that MAY arise.
This helps to plan lessons and sessions that may aid in the completion of
the I.A. in a timely manner.
Allot I.A. Sessions – Though it is true that sessions are already limited when
considering what must be taught, having I.A. sessions gives you a chance to
Make one general explanation/announcement rather than many
Students and teacher will come with the frame of mind to address (ATTACK) the I.A.
You can observe patterns/trends/problems/concerns etc. and address them early.
You will be able to assess the general needs of the class better.
Planning a Good Statistical
I.A.
Set and Maintain DEADLINES for
Drafts with specified contents
(e.g. draft 1, up to Methodology, draft, 2 Calculations & graphs and draft 3,
Analysis, Interpretation, conclusion & recommendation. )
NB! If you have the flexibility, set any of the following as a graded assignment
Final Projects
Read widely – Ideas can come from anywhere.
A GUIDE TO A GOOD
STATISTICAL I.A.
DEVELOPEMENT
Selecting a Title
Any real world event that can be addressed or transformed into a single question
that can be used to capture data on the specific problem/topic, where the data
collected is varied, i.e. no one answer to a question can be given. All data points
in the pool are not the same.
Selecting a Problem/Topic
(An Iterative Process)
a) On average, how much time do you spend each night on homework?
b) Many kids are concerned about not having enough free time. Do you
agree?
Experimental
Quasi-Experimental
Non-Experimental
Type of Study (cont’d)
Experimental Designs
Non-Experimental Designs
by design?
a) Comparing the incidence of bullying (because of weight) at a school with the incidence at a
similar school that has no antibullying program.
b) An Integrated Mathematics student who finds out whether each of several middle-school students
has been bullied (because of their weight) and then measures each student’s self-esteem.
c) A researcher selects 15 rats and place them in groups Junk Food + Soft drink, and “whole food” +
water and group 3 “regular rat food”. Groups 1 and 2 sleep times, meal time and exercise
schedule was enforced by the researcher. Group 3 was “free-for-all”.
Methodological Approach
Qualitative
Quantitative
There are two main groups we collect data on, these are population and sample. For
each, varied analytical tools are used.
A population describes a group where all the members belong to a well defined set on
which we are collecting data. We use parameters to analyse a population.
However populations are usually large, hence it is harder to collect information from
all its members. So we often times use a sample to make inferences about the population.
unbiased sample, and must also represent the population. This will lessen the
errors.
Rule of Likeliness
Example
There are 400 balls made on a daily basis at factory A. 300 of the balls are red
and 100 are blue. If we chose 100 of the balls at random, how many of the balls
would need to be red for the sample to be unbiased of the population?
Solution
Given that the percentage of red balls in the population is 75, then percentage of
red balls in the sample must also be 75. Therefore the number of red balls that
will make the sample unbiased is 75.
SAMPLING
Methods
Example: Choosing any three National Hardo Bread from the shelf of Hardo
Bread
SAMPLING
Methods (cont’d)
Cluster – this is where we simply assign members from the sample frame to
one and only one group. These groups are called clusters. We then draw an
analysis (sample) from one and only one cluster, so only individuals from
that particular cluster are used in analyzing the population.
SAMPLING
Methods (cont’d)
How to sample
• Nature of population
• Resources available (time and money, personal, supplies)
Sampling Frame
This is the list of members from which a sample was taken.
A GUIDE TO A GOOD
STATISTICAL I.A.
INSTRUMENTATION
Instrument Design
Questionnaire vs Survey
Questionnaire
Written
Used only to collect data (often times to be listed)
Survey
Process from instrumentation to analysis
May be presented written or orally
Instrument Design
Deducing Variables
designing an instrument. This helps to identify the focus of the data gathering
Does the incidents of students being bullied increase the more they are
Introduction
Why this project
Problem
Nature
Type/approach
Aim
Research Question
Final Assembly (cont’d)
Methodology Type
Design
Justification
Type
Data Management
Limitations How
Sampling
When
Type
Codes
Method
Validity
Frame
Extraction
Instrumentation