PH 292: Health & Social Behavior First-Year Seminar: Starting Your Literature Search

What causes disease?
For any "disease" or condition, you could start by considering interactions among environmental and social factors:

Poor diet, resulting from food choices, "causes" nutritional deficiency or obesity in a population.

But consider:

• Is it "caused" by historical distribution of land use (redlining and racism in the US; colonialism in some other countries)?
• Or by the regulatory environment, including crop subsidies, food inspections, etc.?
• What about the role of NGOs, IGOs, aid networks?
• What about infrastructure, such as food distribution networks, transportation, etc.?
• Is the status of women/girls a factor?
• What about mental health issues?
• What is the role of commercial activity?
• What about the healthcare and health insurance system?

. . . Keep moving upstream!

Also, consider that how issues are framed is influenced by our assumptions and biases.

Below are some examples of questions or lines of inquiry.
Considering what question a research article addresses may help you determine if it is relevant to your needs:

• Are their racial or ethnic disparities in type 1 diabetes mellitus prevalence?
• Compare and contrast personal versus "upstream" factors relevant to these disparities.
• Describe examples of things that could reduce these disparities, differentiating between personal and upstream factors. 
• If a policy or program increases disparities, what are possible reasons for this? Differentiate between personal and systemic factors.
• Describe a plan/program/policy to reduce these disparities.
• Justify why systemic or upstream factors contribute more to these disparities than personal factors.

What is a good Research Question? It is a question that:

• identifies a relevant issue in your field;
• pursues relatively uncharted research territories to address the problem;
• piques the interest of others.

This blog post has tips on how to write a good research question, including examples of bad, good, and great questions.

To reduce bias, it may be best to pose your question in a neutral manner. Examples:

• Neutral question: What is the impact of school-based physical activity interventions on time spent engaged in moderate to vigorous physical activity among children, in comparison to children not exposed to school-based interventions?
• Non-neutral question: Which school-based interventions are effective in increasing the amount of time children engage in moderate to vigorous physical activity?

However, a research question often reflects power and agency. Compare the following:

• Why are African American children expelled from school more than white children?
• Why do schools expel disproportionate numbers of African American children?

What is the question being addressed in the study you are reading? Compare:

• "Our intervention worked toward fixing Problem X."
• "The most effective interventions for fixing Problem X are: ..."
• "The effects of our intervention on Problem X are: ..."

When you read an article, answer (briefly) the following (source):

• What question are the authors addressing?
• What did they do to answer the question?
• What was their rationale for doing what they did to address their question?
• What were the findings?
• How were the authors think their findings mean?

Finding a systematic review that addresses the question you are interested in can be very helpful: take a look at the search strategy and databases used in the systematic review for tips on your search.

It may be useful to have a structure to help guide you when searching. 

When you formulate a research question, consider these elements:

• Phenomenon: What happened?
• Subject or Population: Who did it?
• Time: When did it happen?
• Location: Where did it happen?
• Cause or Motivation: Why did it happen?
• Process: How did it happen?

Note: It is possible that not all of the above elements will be appropriate for your search topic.
This worksheet (docx) will take you through the process of formulating a search.

PICO is another popular way to structure a search.
PICO stands for:

• Patient or Population;
• Intervention;
• Comparison or Control;
• Outcome.

Worksheet: Developing an Efficient Search Strategy Using PICO (docx): A  fillable form that provides a PICO example and prompts you to document synonyms and MeSH headings.

Embase has a PICO search form to guide you.

SPIDER is a search structure used in qualitative research. SPIDER stands for:

• Sample;
• Phenomenon of Interest;
• Design;
• Evaluation;
• Research Type.

Once you have a search strategy, you can try the Polyglot Search tool, which helps to translate a PubMed search into the correct syntax for several other databases, including Embase, Scopus, Web of Science, and more.

Words matter!

• Do you want articles on labor or articles on labor? Or is it labour?
• Do you want articles on HIV (a virus) or articles on HIV diseases (such as AIDS)?
• What's the difference between diet, food, eating, etc.?
• Is lead a noun or a verb?

Indexing facilitates more precise search statements, especially for topics that are vague or ambiguous.

• Using index terms helps you avoid the need to think of every possible synonym or alternate spelling of your search terms.
• Indexing means the citations in the database are assigned terms from a "controlled vocabulary." Most topical article databases use a controlled vocabulary.
• But be aware of the fact that indexing schemes are shaped by the cultural milieu from which they originate.
• Index terms are sometimes called descriptors or thesaurus terms; in PubMed they are called Medical Subject Headings, or MeSH.

• Search Tips/Search Operators (pdf)
• Search Workflow
• Quick Article Evaluation Worksheet

Off-campus access is limited to current UC Berkeley faculty, staff and students. Choose one of the following methods:

Library Proxy (aka EZproxy):
When you click on a link to an article, database, etc., from a library web page. you will be prompted to authenticate via CalNet.
If you click on an article (etc.) link found via a search engine or a non-UCB Library webpage, you should use this bookmarklet to access the licensed resource.

Virtual Private Network (VPN):
Download and install the VPN client to allow access the UC Berkeley licensed resources.
Make sure you select Library Access - Full Tunnel VPN when you log on.
VPN FAQ

Students: Problems setting up Library Proxy or VPN? Contact your librarian, or Student Technology Services: (510) 642-HELP, or sts-help@berkeley.edu.