PSYC 3210 :: Laboratory Report - Behavioral Data Analysis
To present the methods and results of your behavioral training experiments in a
simplified version of a scientific paper.
Due Date: Wednesday, 9/28, at the beginning of class
Although scientific results are communicated in many forms, including posters,
presentations, articles, and reviews, the primary research paper is carefully
scrutinized before publication by scientists in the field and serves as the
definitive record that will be referenced by future wokers. Results only become
findings when they are published in a peer-reviewed journal. Research papers
all have the following components:
- Title: A short description of the major finding
- Authors: A list of the people who contributed significantly to the research
- Abstract: A 100-200 word summary of the paper
- Introduction: Describes the background, motivation, and specific hypotheses of
the research. Cites previous papers to place the current work in the context
of the larger scientific literature.
- Methods: Describes the experimental and analytical procedures used to collect
and analyze the data. Must provide sufficient detail to allow someone with the
appropriate expertise to replicate the experiments.
- Results: Presents the data from the experiments in text and in accompanying
graphical figures. Results are presented in a chronological order, or in a
logical sequence that lays out the experimental evidence in support of the
claims, but generally without extensive interpretation.
- Discussion: Summarizes the results and restates how they support the
hypotheses and claims of the paper. Describes potential caveats and
limitations of the results. Interprets the finding in the context of the
scientific literature, challenging existing theories if needed and suggesting
further work to clarify or extend the finding to new situations.
- References: A bibliography of works cited in the paper.
Your task is to present your behavioral experiments using a simplified version
of a research paper that focuses on the Methods and Results sections. Our goal
is to work on clear, concise writing and clear, aesthetically pleasing figure
design. Each student must write his or her own report, but group members may
assist each other with data analysis and making figures.
Your paper must describe the following experiments:
- Hopper training
- Discrimination training
- Category generalization and feature tests
Your report will comprise the following sections:
In this section, explain the nuts and bolts of of the experiment. Use the
Herrnstein et al (1976) paper we discussed in class as a model. Give the subject
species, and describe the training/test apparatus and the location(s) where the
experiments were performed. Give each experiment its own subsection, and in each
subsection give a detailed description of
- The organization of the trials
- The stimuli or experimental manipulations
- What you (or the computer) measured in each trial
- Any statistical tests used to make comparisons
Your goal is to be clear, precise, and complete, while not including extraneous
information. For example, you should note that some of the discrimination
training took place in an open classroom with other birds also undergoing
training, but it’s not necessary to say what room and building we were in.
In this section, tell the reader what the results of the experiments were. Each
experiment should have its own subsection. In each subsection, clearly state the
hypothesis being tested, the specific predictions, and describe the results.
Your results section will have the following figures:
- Hopper training: line plot of response speed (distance divided by latency)
versus trial number. Follow the convention with time series data, which is to
plot time on the abscissa (horizontal axis).
- Discrimination training: line plot of average response accuracy or average
response rate versus trial block, with separate lines for S+ and S- stimuli.
Lines must be different colors or styles and either labelled in the figure or
in the caption. Make sure to report block size in the figure caption.
- Category generalization/features: spectrograms of all probe stimuli. This
should be a single figure with the spectrogram of each stimulus shown in a
separate panel. You can use screenshots from Audacity. The time, frequency,
and intensity scales need to be the same for all panels.
- Category generalization: bar plot of average accuracy for each stimulus
presented during the probe test, including the original training stimuli.
Bars should be clearly labeled and ordered in a logical manner, not
alphabetically. Indicate which stimuli are significantly different from
chance with stars above the histogram bar and note what p value you used
as a cutoff in the caption.
For experiment 1, use the data your group collected and recorded in class. For
experiments 2-4, use the data recorded by the computer. Instructions for retrieving
the data are in the Data visualization and Behavioral probe experiments
Important: you are NOT required to use R, ggplot, or any specific program to run
analyses or prepare figures. However, most of what you need to do, you’ve
already learned in R, so it ought to be fairly easy to work in that environment.
If there’s anything you’re having a hard time figuring out how to do, don’t
hesistate to contact the TAs or the instructor. Additional figure guidelines:
- All graphs must have clearly labeled axes.
- Graphs should be clear and uncluttered. Avoid overly large and overly small
symbols. Readers should be able to easily see the trend or pattern you are
trying to convey.
- Figures go on separate pages at the end of the text.
- Figures must have captions, which should be short and describe what’s shown in
the figure. You may bring attention to a notable trend or data point, but
avoid interpreting or explaining in the caption.
For this exercise, do not interpret results in the Results section. Your goal
is to clearly communicate what you were testing and what happened.
In the discussion, very briefly (one paragraph of 100 words max) summarize the
results of your experiments, and describe whether or not they supported your
hypothesis. If for any reason you were unable to obtain sufficient data to test
the hypothesis, indicate this and say why.
Additional notes and resources
Your report should be no longer than 11 pages and no shorter than 8, including
figures. Use 1.5 line spacing and 12 point, Times New Roman font. Submit on
paper at the beginning of class 9/28.
Balance precision and conciseness, and avoid writing that sounds scientific but
is not in a natural voice. For example, “The training of the starling’s
discrimination ability was accomplished through operant conditioning” is a
deeply obfuscated way of saying, “Starlings learned to discriminate between
songs through operant conditioning”. For more information, check out these
useful guidelines on sentence structure in scientific writing.
- Clear and concise writing (30%)
- (best) Text is concise, easy to read, free of spelling and grammatical
mistakes, and contains necessary details with little extraneous material
- (good) Text is clear and contains necessary information, but also
extraneous details or verbiage
- (heading the right direction) Text has deficiencies in organization,
content, and is sometimes difficult to follow
- (needs work) Text is often difficult to follow or is missing important information
- (poor) Numerous grammatical and spelling errors that severely impede communication
- Figure design (30%)
- (best) Figures are aesthetically pleasing and convey data effectively.
Colors, line and point styles, and text all work together.
- (good) Figures convey data effectively, but with aesthetic choices that
reduce legibility or that obscure the main point of the graph
- (heading in the right direction) Figures are generally correct but have some
missing elements or poor design choices (colors or symbols that are
difficult to tell apart, overly heavy or light lines, crowding, illegibly
- (needs work) Figures contain the correct data, but presented incorrectly or
without sufficient information to determine what is being shown
- (poor) Figure does not contain the correct data
- Experimental design and content (25%)
- Hypotheses are clearly stated
- Hypotheses make sense and satisfy instructions
- Predictions follow clearly from hypotheses
- Predictions are testable
- Tests match predictions
- Results presented address hypotheses and predictions
- Unanticipated results or problems with methodology are described
- Format and organization (20%)
- text within sections is logically organized and flows from paragraph to paragraph
- sections are clearly titled and in the correct order
- adheres to length, font, and other guidelines
- figures have captions and are appropriately referenced in the text