Jeromy Anglim's Blog: Psychology and Statistics


Wednesday, December 9, 2009

Writing a Method Section in Psychology: Lab Reports, Theses, Articles

This post discusses how to write a method section in psychology. It has relevance to writing journal articles, theses, and lab reports. It includes discussion of: (a) key elements to include in each section; (b) specific issues to consider; and (c) links to online and offline resources.


GENERAL THOUGHTS
The method is perhaps the easiest section of a report to write. If you are able to conduct the study, enough should be known in order to write about what was done. The main challenge involves designing a good study in the first place. However, even with a clear idea of what was done, several issues remain: what information should be included? Where should it placed? How should it be organised? What language should be used?

This post reflects my general thoughts. In a separate post I set out a process for extracting principles from specific journal articles to guide your writing. This post should be seen as a supplement to this.


WHAT TO INCLUDE
A method section in an APA Formatted Psychology Journal Article typically includes three sections:
  • Participants
  • Measures / Apparatus
  • Procedure
Participants: 
Include: Overall Sample size and if appropriate group sample size; Sample source; method of recruitment; relevant sample characteristics (age, gender, and others); exclusionary criteria; whether the sample was representative; acceptance rates for participants solicited; number of participants or observations excluded and the reasons for exclusion. If participants are assigned to conditions, explain how participants were assigned. It is often good to include information on the statistical power implications of the planned or obtained sample size. If inducements to participate are offered, these should be explained.

Measures / Apparatus:
The amount of information required varies based on the type of report.

The following also sets out details on what to report for certain types of measures and apparatus that I commonly encounter:
Self-Report Psychological Tests: The response options; the numeric values assigned to the response options; how scale scores are calculated (i.e., mean, total, or something else); sample items; a reference to where the full scale can be obtained;

Ability-Based Tests: Things to include: what the test claims to measure; evidence for its reliability and validity (e.g., references to correlations with other tests and various indices of reliability); sample items; reference to the source; possibly a justification of appropriateness for the sample.

Tasks: 
If computerised; programming language, operating system, specifications of the computers relevant to the design (e.g., accuracy of reaction time measures, response latency of the display). For further discussion, see my article deconstruction of a few method sections.


Procedure: 
Design: The APA Style recommends that this section include information about the design. this includes: If there were factors in the study, say whether the factors were between-subjects or within subjects, the number and nature of the levels of each, and how participants were assigned to any between-subject condition. If assignment is non-random, explain this in more detail.


Sequence: It should outline the sequence of components in the experiment. Bem has suggested that it can be good to present the procedure in terms of the chronological experience of an individual participant. This should generally include some indication of the total duration of the study and of the individual components.


Instructions: All instructions should be accessible. However, given the nature of the publication, there may be a need to present just the aspects of the instructions which are critical to the design. The mode of delivery (source: experimenter, computer, other device; modality: audio, visual, text) and whether it was standardised.

ASSORTED ISSUES
Style
A clear and concise style is best.


Grammar
The Method is written in the past tense. When describing something that the experimenter made occur, the passive voice is typically used.





Naming variables, values, and operations
Bem advises one to "[n]ame all groups, variables, and operations with easily recognized and remembered labels" (p.7). This is excellent advice for both the method and the rest of the report. Readability is the aim. Consistent use of readily understandable labels will aid readability.


How much detail should be included?
If a detailed description has been presented in a previous chapter, journal article, or other published work, it is often acceptable to refer the reader to this previous work.
The amount of material to include also depends on the nature of the work. A thesis is generally expected to provide more details. If material seems marginally important, it can be placed in an appendix. Many journals provide options for online Appendices. If the information will help someone wanting to replicate the study, it is best that this information be made accessible.
When deciding what information to include it is useful to consider what information is most important. What information would be required to comfort a critical reader?

Power Analysis?
There has been a lot of discussion about the importance of power analysis in designing a study (e.g., Cohen's A Power Primer; The free software G-Power 3). In particular, power analysis has major implications for decisions about sample size. Although many studies make no mention of the statistical power of their study as an implicit assumption, there is a good argument to include a few lines on statistical power in the Participants Section. Remember: (a) statistical power is a property of a particular hypothesis and not a property of a study. (b) statistical power is contingent on population effect size, sample size, and alpha, and (c) population effect size is rarely known exactly.

Information that results from the study
The question often arises of whether certain results should be placed in the results or the discussion.
Reliability of a test should generally be reported for each test used in the results section.
Some descriptive information of the sample is included in the participants section.

Ethics
Discuss any ethical issues related to the research and what was done to ensure the ethical conduct of the study.

What about computer programs?
My attitude is that intellectual property considerations permitting, computer programs and software used in experiments should be made available in an accessible location. This location should be noted in the method. Researchers should try to make it as easy as possible for others to build on what they have done.

How should information be distributed between apparatus and procedure?
Where should the following information go: instructions; details about a task; sequencing of trials, randomisation, block and trial structure, and so on. Clearly some of it concerns the task itself but other aspects pertain more to the way that the task is used in the experiment.

ADDITIONAL RESOURCES
Online Resources
Offline Resources
Related Posts