2/7/23, 7:05 PM Survey Research: Spr23 PSYC 005 #38039 RESEARCH METHODS IN PSYCHOLOGY-Online

https://canvas.pasadena.edu/courses/1131870/pages/survey-research?module_item_id=15132919 1/3

Survey Research

The following lecture on Survey Research was written by Heidi Iseman, a graduate student inClinical Psychology at Cal State University, Fullerton. She is currently working on her master'sthesis. She and her thesis advisor are using survey research to conduct the research.

Overview of Survey ResearchWhat is a survey and why would you use one when conducting psychological research? Surveyresearch can be either qualitative or quantitative and involves asking respondents to provideself-report responses to questions about their thoughts, feelings, and behavior. Researchers whouse survey research prefer to have random samples that are as large as possible so their findingsrepresent the population as much as possible. Surveys are a great research method to use if youneed a large number of participants.

Survey research started around the beginning of the 20 century when English and American“social surveys” were used to learn more about social issues, such as poverty. Surveys are widelyused by researchers within the government, as well as among academics in many fields, such aspsychology, sociology, political science, market research, and public health.

Why is survey research useful? In addition to its important uses in many different fields, survey research has provided prevalenceinformation about mental disorders in the United States. One such survey is The NationalComorbidity Survey: http://www.hcp.med.harvard.edu/ncs (https://www.google.com/url?q=http://www.hcp.med.harvard.edu/ncs&sa=D&usg=AFQjCNHGgRvKIdEBbW5iqCRTVDJMgBDeaA) . This large-scale mental health survey has gathered information about the prevalence of mentaldisorders and how they are associated with important outcome variables. This type of study helpsinform policymakers, as well as mental health providers.

Constructing SurveysWhen a respondent answers a survey question, there is a psychological process that he or shegoes through. One cognitive model of responding to a survey question involves first interpretingthe question, then retrieving information, forming a judgment, formatting a response, and lastly,editing the response (Sudman, Bradburn, & Schwarz, 1996). The chapter provides an example ofthis process using a survey question about alcohol use. First, respondents need to be able toaccurately interpret the question. This is why it is important for researchers to use language thatis free from jargon and easy to understand. Next, the respondent would need to retrieveinformation from his or her memory regarding how much he or she drinks. Different factors mightinfluence how well the participant is able to retrieve information, such as cognitive ability and thepresence of a substance use disorder. Then, the respondent would use the information retrievedto arrive at a judgment of how much he or she actually drinks. It is important to remember that thisis a judgment and therefore is subjective. The survey researcher has to trust that the respondentis able to accurately judge their drinking behavior. After the participant is makes a judgment, he or

th

2/7/23, 7:05 PM Survey Research: Spr23 PSYC 005 #38039 RESEARCH METHODS IN PSYCHOLOGY-Online

https://canvas.pasadena.edu/courses/1131870/pages/survey-research?module_item_id=15132919 2/3

she would then format their responses to fit the answer choices made available to them. Lastly,the respondent may edit their response based on the options available.

One important factor to consider when interpreting survey responses is context effects. Contexteffects refer to how the context in which the question is presented, rather than the actual content,influence responses. An example of a context effect is an item-order effect, where the order inwhich the question is presented influences how participants respond. In other words, responsesare affected by earlier questions. Respondents may be psychologically primed or have certainschema activated when one question is presented before another. Participants can also beinfluenced by the order in which response options are presented, which is why counterbalancingis an important technique. Counterbalancing involves randomizing the response options so thefindings are not unduly influenced by order effects.

There are different types of items that can be included in a survey. One type is an open-endeditem, where respondents are asked a question where they can provide their own unique answer.One example provided in the text is “What is the most important thing to teach children to preparethem for life?” This question is open-ended because the respondent has the freedom to providehis or her own response. These types of items are useful when researchers are conductingqualitative research or are at the early stages of research project development.

In contrast, closed-ended items involve presenting a question and a set of predeterminedresponses for the participant to choose from. These types of items are useful for quantitativeresearch and are used when researchers are aware of what types of responses they should begetting. Closed-ended items lend themselves well to quantitative research because they do notneed to be re-coded, as might an open-ended item. These types of items can be easily convertedonto a spreadsheet and statistically analyzed. When researchers use closed-ended items thatare quantitative in nature, a rating scale is often used. A rating scale provides an ordered set ofresponses that the participant can choose from. One of the most popular types of rating scales isa Likert Scale. The original Likert Scale was developed in the 1930s by a researcher namedRensis Likert. A Likert Scale involves presenting the respondent with a question where he or shecan either agree or disagree and then measuring their response along a 5-point scale rangingfrom Strongly Agree, Agree, Neither Agree nor Disagree, Disagree, andStrongly Disagree.

Effective survey items follow the BRUSO acronym. They are brief, relevant, unambiguous, andobjective. Questionnaire items should not be overly long, irrelevant to the research study, difficultto interpret, or subjective. It is important to carefully edit and screen the items included in thesurvey to ensure that the responses obtained are valid and reliable. The chapter provides someexamples of poor and effective items.

After the researcher has chosen the items to be included in the survey, he or she must format thesurvey. Typically, a survey begins with an introduction where the researcher describes thepurpose of the study, acknowledges academic affiliations or sponsorships, and recognizes andthanks the participant for taking part in the study. The researcher must obtain informed consentfrom the participant and should provide information, such as how long the survey will take tocomplete, the types of questions that will be asked, as well informing the participant that he or shecan withdraw from the study at any time.

Conducting SurveysOnce you have created your survey, the next step is to determine how you are going to conductit. First, you will have to find a sample of individuals who will take your survey. There are twomain forms of sampling. Probability sampling refers to when the researcher knows the

2/7/23, 7:05 PM Survey Research: Spr23 PSYC 005 #38039 RESEARCH METHODS IN PSYCHOLOGY-Online

https://canvas.pasadena.edu/courses/1131870/pages/survey-research?module_item_id=15132919 3/3

probability that each member of the population is likely to be part of the sample. Some examplesof probability sampling are simple random sampling, stratified random sampling, and clustersampling. These forms of sampling are further discussed in the chapter.

The following video describes the types of probability sampling and how they are conducted:

4.2 Probability Sampling Techniques (https://www.youtube.com/watch?v=jauhoR7w1YM)

(https://www.youtube.com/watch?v=jauhoR7w1YM)

Non-probability sampling refers to when the researcher does not know the probability that eachmember will be selected. One example of non-probability sampling is convenience sampling. Convenience sampling is when the researcher gathers whichever participants are the easiest torecruit.

This video from Pew Research discusses non-probability (online opt in) surveys – Methods 101:What are non-probability surveys? (https://www.youtube.com/watch?v=T9dmq6Lwh24)

One important concept to consider is sampling bias. Sampling bias occurs when a sample is notrepresentative of the population. Sampling bias can create inaccurate results that cannot begeneralized. Probability sampling was created as a means of reducing sampling bias.

The most common form of sampling bias is non-response bias. This refers to when individualswho do not take part in the survey differ in significant ways from the members of the populationwho did take part in the survey. Some methods of reducing non-response bias are to providepotential participants with reminders, notifying them ahead of time, incentivizing participation, andcreating surveys that are short and easy to complete.

How can you distribute your survey?Lastly, the chapter summarizes the four main ways that surveys are distributed: through mail, overthe phone, online, and in-person. In-person surveys yield the highest response rates but are themost expensive. Internet surveys and mail surveys are the least expensive, yet they have thelowest response rates. Internet surveys are rapidly gaining popularity and are projected to be themost commonly used way of obtaining survey data.

Below are some of the most commonly used websites to create and distribute online surveys:

http://www.qualtrics.com (http://www.qualtrics.com/)http://www.surveymonkey.com (http://www.surveymonkey.com/)https://requester.mturk.com/create/projects/new(https://requester.mturk.com/create/projects/new)