Survey Research PDF

Preview

Summary

Advantages and disadvantages
Designs
Survey types and their advantages/disadvantages
What can go wrong
...

Description

Survey Research Survey research is useful to study psychological constructs that are difficult to observe directly: • Cognitions (e.g., values, stereotypes) • Emotions (e.g., life satisfaction, attachment) • Behaviours (e.g., routine daily life, sexual behaviour, drug use) What is survey research?  Survey research is a type of field study that involves the collection of data from a sample of people drawn from a well-defined population through the use of a questionnaire."  It is used to quantitatively describe a specific aspect of a given population  Asks people questions (self-report)  It collects data from a subset of the population

Research advantages  Enables us to study a large variety of topics in real-life natural settings.  Increases our confidence in generalising the findings to the general population.  Provides ideal conditions to examine the moderating role of personal, social, or cultural variables. Research disadvantages  Correlational nature. We usually do not manipulate the independent variables and therefore cannot infer causal relationships between the variables.

Cross-sectional design  Involves the collection of data at a single point in time from a sample drawn from a specified population. o A cross-sectional design is often used to assess o The frequency with which people enact certain behaviours (e.g., adultery). o The number of people who have particular attitudes or beliefs (e.g., attitudes toward monogamy).  But… a cross-sectional design can also be used to assess differences between subgroups in a population.  But… a cross-sectional design also provides the opportunity to assess relations between variables. In a Repeated Cross-Sectional Design (or Successive Independent Samples Design) data are collected from independent samples drawn from the same population at two or more points in time. o A Repeated Cross-Sectional Design is useful if we want to describe changes in attitudes or behaviours over time. A Repeated Cross-Sectional Design 



May also provide evidence to support a causal relation: o Changes over time in the independent variable should be mirrored by changes in the dependent variable. o Changes over time in the independent variable should be mirrored by changes in the dependent variable.

Panel Surveys (Longitudinal Designs)  Data are collected from the same individuals at two or more points in time. o Longitudinal designs enable us to test causal hypotheses in at least two ways:  Examine whether changes over time in an independent variable correspond to changes in a dependent variable.  Examine whether initial levels of an independent variable predict changes in a dependent variable. Experiments within Surveys  Respondents are randomly assigned to different versions of a questionnaire.  Differences in responses between the groups can be attributed to the specific elements that were manipulated.  Hypothetical scenarios  Manipulate a certain characteristic of the actor or the circumstances  Examine participants’ reactions and attitudes

Survey Sampling  Sampling is a procedure in which a specified number of individuals are drawn from a sampling frame that represents an actual list of the population.  Two general classes of sampling methods: o Probability Sampling o Nonprobability Sampling  Probability Sampling refers to selection procedures in which individuals are randomly selected from the sampling frame.  Nonprobability Sampling refers to selection procedures in which individuals are not randomly selected.  Advantages of probability sampling: o Increased representativeness o Generalisability  Advantages of nonprobability sampling: o Convenience o Cost

3.1.1 Simple Random Sampling 3.1.2 Stratified Sampling 3.1.3 Cluster Sampling Simple Random Sampling:



Individuals are drawn from the population at random, and all have the same chance of being selected.

Stratified Sampling:  The sampling frame is divided into subgroups (strata), and the sampling process is conducted separately on each stratum. Cluster Sampling:  A sample of individuals is drawn from groups (called "clusters") rather than one-byone. Then all individuals within a given cluster are sampled.  Advantage: reduces time and cost  Disadvantage: reduces accuracy/representativeness

What can go wrong Sampling Error:  A discrepancy between the sample data and the true population data, which is caused by random differences between the sample and the sampling frame.  The chances are dependent on the sample size and variance, and the population size.  We can calculate estimates of the sampling error and know the magnitude of uncertainty regarding our data. Nonresponse Error:  Often some individuals chosen for the sample are unable or unwilling to participate.  Error occurs when the sampled individuals who did not respond differ systematically from those who did.  To reduce error, we need to minimise the apparent costs of responding, and maximise the rewards for doing so. Coverage Error:  A discrepancy between the sampling frame and the population.  Increasing the sample size tends to reduce the sampling error, but it cannot correct methodological problems such as nonresponse and coverage errors.

Nonprobability sampling Convenience Sampling:  Participants are selected solely on the basis of convenience.  The problem: People who happen to be in a certain place or respond to advertisements do not represent the population. Purposive Sampling:  Convenience sampling among members of a particular subgroup within a population

Snowball Sampling:  A few members of a subgroup are located, and each is asked to suggest other members of the subgroup for the researcher to contact. Quota Sampling:  Members of various subgroups of the population are selected to create a sample that reflects known characteristics of the population.  Think about the demographic characteristics that might be related to the topic.  Plan in advance how many individuals will be recruited from each demographic category.  Base the quotas on accurate information about the composition of the population  The sample may be relatively representative (compared to a convenience sample but not to probability sampling)....