Statistical interpretation and advanced statistical methods' Let the data speak

1
Statistical interpretation and advanced
statistical methods.Let the data speak!

• State of Play
• What test and where?
• Interpreting statistics
• Advanced methods

2
State of Play

• Literature review
• Results report
• to be mailed 14/10 week 10
• Interpretative report
• due 21/10
• Quizzes
• extended to exam date
• Exam
• second week(details out soon)

3
State of Play Interpretive Report

• ? provide a short introduction
• ? make a clear statement of the hypotheses tested
  (these may be different from those in your small
  group results report if you wish)
• ? give a brief description of the way in which
  the hypotheses were tested and the outcomes
• discuss the significance of these findings in the
  light of the published literature, citing sources
  used
• ? give an account of any shortcomings in the
  methodology of the research project
• ? provide suggestions for future research or
  refinements to the research methodology
• ? list references in APA style.

4
State of Play Interpretive Report Rubric Pass

• Satisfactory introduction including major
  research questions
• Satisfactory statement of three hypotheses from
  the results report and the way in which they
  were tested
• Satisfactory description of interpretations and
  conclusions about hypotheses
• Satisfactory discussion of the significance of
  these findings in the light of the published
  literature
• Satisfactory critique of methodology
• Satisfactory suggestions for future research and
  refinements
• Satisfactory use of statistical language
• Satisfactory use of referencing and citing

5
State of Play Interpretive Report Rubric Credit

• ? Good introduction including major research
  questions
• Good choice of relevant hypotheses and their
  description
• Good description of interpretations and
  conclusions about relevant hypotheses
• Balanced and informed discussion of the
  significance of these findings in the light of
  the published literature
• Good critique of methodology
• Good suggestions for future research and
  refinements
• Good use of statistical language
• Good use of referencing and citing

6
State of Play Interpretive Report Rubric
Distinction

• ? Excellent introduction including major
  research questions
• Excellent choice of relevant hypotheses and their
  description
• Excellent description of interpretations and
  conclusions about relevant hypotheses
• Balanced and highly informed discussion of the
  significance of these findings in the light of
  the published literature
• Excellent critique of methodology
• Incisive suggestions for future research and
  refinements
• Articulate use of statistical language
• Excellent use of referencing and citing

7
State of Play Quizzes

• extended to exam date
• some 150 questions
• final quiz contains week 8,9, 10
• available late week 10
• check all your marks late week 10

8
State of Play Exam

• 2 hours at Ridley Centre, Wayville
• 50 multiple choice questions(50)
• 10 short answer style questions(50)

9
State of Play Exam possible questions A

• Provide 4 reasons why educational research is
  conducted?
• Describe briefly one national or international
  test on student ability
• What is informed consent in research in education
  in primary aged children and why is it important?
• Describe the experimental setup and research
  goals of the Meyer et al personalisation study?
• What are 3 ethical issues that might be
  associated with the conduct of educational
  research in primary aged children
• Outline two problems that are encountered in
  organising research in primary schools
• Outline two problems encountered when collecting
  information from junior primary children
• Describe two qualitative methods.
• Who are the consumers of educational research and
  what role might they have in education?
• A parent wants to know why their child is below
  average on the national numeracy test. What
  factors should you consider before giving your
  reply?

10
State of Play Exam possible questions B

• Describe one categorical variable and one
  continuous variable from a study with which you
  are familiar.
• In representing research results describe two
  types of charts that might be used and explain
  why.
• Why in inferential statistics is the p value(or
  alpha) always quoted?
• Research questions are not the same as
  hypotheses. Discuss briefly.
• Give two examples of the need to re-categorise
  variables from a study with which you are
  familiar.
• Outline briefly one classic myth in educational
  research.
• What is generalisability? Explain it in relation
  to a study with which you are familiar.
• Why do educational research when it is so hard?
• What does it mean when it is said that statement
  made with inferential statistics are always
  equivocal?
• Why is it important to examine the literature in
  a field before conducting research?

11
What tests and where?Comparing Means for
Different Groups

• Comparing two means Independent samples t-tests
• males and females on activity at lunch time
• males and females on getting sports scores off
  the web(or horoscopes)
• More than two means Analysis of Variance-ANOVA
• local street class and means of lunch time
  activity
• extent of internet access and web site recognition

12
What tests and where? Correlations among
variables.

• Is web site recognition related to author
  recognition?
• Are lunch time and play time activities times
  associated?
• Use curve estimation to match a line

13
What tests and where? Chi-square

• Street class and before school activities
• WWW attitude and level of access at home

14
Interpretation On the Shoulders of Giants

• adding a small finding to a field of research
• restrained conclusions
• limited generalisation
• blending with arguments from the literature
• relevance to important research questions

15
Interpretation reasoning and argument

• Use of the hypothetical(English)
• If X then Y
• Given X then
• Accepting for the moment then
• One possible conclusion
• Further research

16
http//www.esa.int/SPECIALS/Mars_Express/SEMGKA808
BE_0.html
17
Interpretation the breakthrough-stomach ulcers
cured!

• Robin Warren Barry Marshall
• He shared the 10-million-Swedish kronor (A1.7
  million) prize with Professor Marshall, 54, for
  their 1982 discovery of the Helicobacter pylori
  bacterium, which showed that a bacterium rather
  than stress causes stomach inflammation and
  ulcers.
• Traditional medical thinking had held that stress
  and diet, such as spicy foods, were among the
  most common causes of ulcers.
• Refhttp//news.ninemsn.com.au/article.aspx?id656
  62

18
Advanced methods

• Strength of findings
• effect size, power, strength of correlation
• Linear regression
• Line of best fit..predicting lunch time activity
  from play activity
• Reliability of a scale
• web efficacy-Chronbach alpha
• Factor Analysis
• grouping of self correlating variables into
  factors

19
References

• References
• chapter 8, 9, 10 of Levine D. and Stephan D.
  "Even you can learn Statistics" bundled with the
  Activates package for SPSS, Prentice Hall 2005