Experimental Psychology PSY 433

Download Report

Transcript Experimental Psychology PSY 433

Experimental Psychology
PSY 433
Chapter 5
Research Reports
Goals of the Final Report
 Communicate to the scientific community.
 Clearly describe your project in sufficient detail
to permit replication.
 Convince readers that your findings support
your conclusions.


How strong is the evidence?
Does it justify your statements about theory?
 Summarize your contribution to the ongoing
debate on an important question.

Pay special attention to your abstract!
You Are Telling a Story
 Introduction -- state your research question,
review the literature, make your predictions
(hypotheses).
 Methods – describe how you explored the
question in sufficient detail to permit
replication.
 Results – describe your findings and test your
hypotheses using statistics.
 Discussion – analyze your results and put
them back into the context of your question.
Parts of the Article
 Title page
 Abstract
 Text:




Introduction
Methods
Results
Discussion
 References
 Appendix, Tables and Figures
Abstract
 This may be the only part of your paper that
most people read, so make it count!
 Write this last.
 Tell the story of your study, with one sentence
per report section.
 Do not exceed 120 words.
Use of Tenses
 For the final report, revise the sections
that were written for your proposal
because they will be graded again.
 Your proposal was written in the future
tense (e.g., “subjects will…”), but for the
final report…


Put the Methods section in the past
tense.
Report your results in the past tense.
Reporting Results
 Only include the results that are relevant to
your research question, not all data collected.
 Go from the general to the specific.
 Provide tables for:



Multiple analyses.
Complex experiments (factorial designs).
ANOVA
 Organize your results section around your
hypotheses, testing one at a time.
Describing Data
 Don’t forget to include descriptive
statistics (means, SDs).
 “The mean number of words recalled
was calculated for each group. The
means and the standard deviations for
each group are shown in Figure 1.”
 “Recall was higher for the drug group
(M = 15, SD = 5.43) than for the
placebo group (M = 10, SD = 4.98).”
Reporting Inferential Statistics
 “The data were analyzed using an
independent t-test. The t-test showed no
significant difference between the mean of
the placebo group and the mean of the
drug group, t(34) = 1.35, p = .782.
 “Using one-way ANOVA, gender
differences were found to be significant,
with females scoring higher on the
average than males, F(1, 23) = 23.89, p
=.025.”
 Show more complex analyses in a table.
Report Exact p-Values
 The old approach simply tested results
against a standard of p<.05 by looking up the
critical value in a table.


Significance was an all-or-nothing judgment.
Only the critical value (cutoff) was known, not
the exact p-value for your statistic.
 Today, SPSS gives exact p-values for every
result. Report those exact values (p=.031).
 NEVER report p > .05 for a non-significant
result. It implies use of p > .05 as a standard.
Ethics of Reporting Statistics
 Don’t change your hypotheses (prediction) to
fit what you actually discovered. Instead say
you were surprised by your results.
 Decide how many subjects to test in advance.



Don’t stop collecting data because you already
have significant results.
Don’t add more subjects because your results
are almost significant and would become so
with a few more subjects.
State your reason for ending data collection.
Avoid “p-Hacking”
 p-hacking is the practice of trying different
approaches to data analysis until you find one
that gives significant results. It is unethical.
 Collect at least 20 observations per condition.
 Report all experimental conditions, even
failed manipulations (studies that didn’t work).
 List all variables collected in a study, even if
they are not analyzed in your paper.
 If there is any doubt, report results with and
without excluded subjects, covariates.
Changes in Reporting
 The internet is making possible different
approaches to report writing.
 Because journals are no longer limited in
space, authors can supply complete data
sets, stimuli (materials) and alternative
analyses.

This represents a movement toward greater
transparency.
 Exact, not conceptual replications are needed
results are marginal.
References
 Format varies depending on the type of material
being referenced (e.g., book, article, web site).
 Only list the sources actually mentioned in the text of
your report.


Everything listed in the references must be cited in text
Everything cited in text must be listed in the
references.
 When you mention a source referenced in another
paper say: “as cited by…” and cite the source you
actually read, not the original quoted by someone
else.
Tables and Figures
 Tables go first – always use APA format.
 Tables contain numbers or words.
 Figures are pictures and typically present
graphs of data, sample stimuli, equipment
setup, diagrams of experiment flow,
flowcharts of cognitive processes or diagrams
of theoretical models.
 Tables have titles that go at the top. Figures
have captions that go at the bottom.
 Include at least 1 of each in your final report.
Discussion
 First, state what you discovered during your
experiment.

Do not repeat results but interpret them and
state whether your hypotheses were
confirmed.
 Tell whether your findings are consistent with
what others have found.
 Describe any threats to validity and problems
with your experiment (confounds, bias,
limitations of generalizability, problems).
 Conclusion – what are the consequences?
Contents of Final Report
 Must contain all sections listed in the APA
Publication Manual, including:








Title page
Abstract
Introduction
Methods
Results
Discussion
References
Tables and Figures