Chapter Nine: Descriptive Research

Descriptive Research Designs Observation Studies Correlational Research Developmental Designs Survey Research

Observation Studies Characteristics: - might involve humans, animals, plants, nonliving objects - tends to have a particular prespecified focus - behavior being studied is quantified in some way - involves considerable advance planning, meticulous attention to detail, and a great deal of time - provides a quantitative alternative to qualitative approaches, such as ethnographies and grounded theory studies

Maintaining Objectivity in Observation Studies Define the behavior studied precisely and concretely so that it is easily recognized when it occurs. Divide the observation period into small segments and record whether the behavior does or does not occur in each segment. Use a rating scale to evaluate the behavior in terms of specific dimensions. Have two or three people rate the same behavior independently, without knowledge of one another’s ratings. Train the raters to use specific criteria when counting or evaluating the behavior; continue training until consistent ratings are obtained for any single occurrence of the behavior.

Correlational Research A correlational study examines the extent to which differences in one variable or characteristic are related to differences in one or more other variables or characteristics. In correlational studies, researchers gather data about two or more characteristics for a particular group of people or other appropriate units of study in order to determine whether and in what way these characteristics might be interrelated. Correlational data is plotted on a scatter plot. Correlation does not, in and of itself, indicate causation.

Developmental Designs Two Developmental Designs: 1. Cross-sectional study: people from several different age groups are sampled and compared. 2. Longitudinal study: a single group of people is followed over time, and data related to the characteristics under investigation are collected at various times.

Cross-Sectional vs Developmental Studies Cross-sectional studies are easier to conduct because all of the data can be collected at one time. In a longitudinal study, data is collected over a period of months or years and participants may drop out. In a longitudinal design, when people respond repeatedly to the same measurement instrument, the characteristic being measured may change because of their practice with the instrument. In a cross-sectional design, different age groups may represent different life experiences. This poses a threat to internal validity. In a cross-sectional design, correlations between characteristics at different age levels can’t be computed.

The Cohort-Sequential Developmental Design Addresses some of the weaknesses of longitudinal and cross-sectional designs. The researcher begins with two or more age groups (the cross-sectional piece) and follows each age group over a period of time (the longitudinal piece). Like a longitudinal study, the cohort-sequential study allows calculation of correlations between measures taken at two different time periods; therefore, predictions can be made across time.

Survey Research Involves acquiring information about one or more groups of people — about their characteristics, opinions, attitudes, etc. – by asking them questions and tabulating the answers. Goal is to learn about a large population by surveying a sample of that population. Also called a descriptive survey or normative survey. Simple design – the researcher poses a series of questions, quantifies the responses, and draws inferences about a particular population from the responses of the sample. Captures a fleeting moment of time; by drawing conclusions from the transitory collection of data, extrapolation can be made about state of affairs over a longer period of time. Relies on self-report data.

Types of Survey Research face-to-face interview telephone interview written questionnaire the Internet

The Face-to-Face Interview Structured Enables the researcher to establish rapport with participants Yields the highest response rates in survey research Time and expense involved may be prohibitive

Telephone Interviews Structured Less expensive and time-consuming than face-to-face interviews Accessible participants Response rate lower than for face-to-face interviews but higher than for mailed questionnaires

Paper-and-Pencil Questionnaires Can be sent out to large groups of people over a large geographical area Participants can respond to questions with assurance of remaining anonymous and thus may be more truthful than in face-to-face or telephone interviews Have a low return rate Often make use of checklists and rating scales

Using Checklists and Rating Scales Checklist: a list of behaviors, characteristics, or other entities under investigation. Rating Scale: used when a behavior, attitude, or other phenomenon of interest needs to be evaluated on a continuum (“never” to “always”) - Likert Scale

Conducting Interviews in a Quantitative Study Guidelines: Conducting Interviews in a Quantitative Study Identify questions in advance. Consider how participants’ cultural backgrounds may influence responses. Make sure interviewees are representative of the group. Find a suitable location. Get written permission. Establish and maintain rapport. Focus on the actual rather than on the abstract/hypothetical. 8. Don’t put words in people’s mouths. 9. Record responses verbatim.

Conducting Interviews in a Quantitative Study (con’t) Guidelines: Conducting Interviews in a Quantitative Study (con’t) 10. Keep your reactions to yourself. 11. Remember you’re not necessarily getting the facts. 12. As you write questions, think about how to quantify responses. 13. Consider asking questions that will elicit qualitative information. 14. Pilot-test the questions. 15. Restrict each question to a single idea. 16. Save controversial questions for the latter part of the interview. 17. Seek clarifying information when necessary.

Guidelines: Constructing a Questionnaire Keep it short. Keep the respondent’s task simple. Provide clear instructions. Use simple, clear, unambiguous language. Give a rationale for any item for which the purpose is unclear. Check for unwarranted assumptions implicit in the question.

Guidelines: Constructing a Questionnaire (con’t) 7. Word questions in ways that don’t give clues about preferred or more desirable responses. Determine in advance how you will code the responses. Check for consistency. Conduct one or more pilot tests to determine the validity of your questionnaire. Scrutinize the almost-final product one more time to make sure it addresses your needs. 12. Make the questionnaire attractive and professional looking.

Guidelines: Maximizing the Return Rate for a Mailed Questionnaire Consider the timing. Make a good first impression. Motivate potential respondents. Include a self-addressed, stamped envelope. Offer the results of your study. Be gently persistent.

Practical Application: Computerizing Data Collection in Descriptive Research Directly enter data as an observation is made. Use the computer as a tape recorder. Look for peripheral devices that can aid data collection. Administer a questionnaire on a computer. Use the computer to monitor the quality of the data being collected.

Sampling Designs in a Descriptive Study Probability Sampling: the researcher specifies in advance that each segment of the population is represented in the sample. Nonprobability Sampling: the researcher has no way of forecasting or guaranteeing that each element of the population will be represented in the sample. Some members of the population have little or no chance of being sampled.

Probability Sampling way that each member of the population has an Random selection: choosing a sample in such a way that each member of the population has an equal chance of being selected; assumes that the characteristics of the sample approximate the characteristics of the total population.

Probability Sampling Techniques ▪ simple random sampling: least sophisticated of all sampling designs; sample is chosen by simple random selection. ▪ stratified random sampling: the researcher samples equally from each one of the layers in an overall population. ▪ proportional stratified sampling: the researcher samples proportionally from each one of the layers in an overall population. ▪ cluster sampling: occurs when the population of interest is spread out over a large area; the large area is subdivided into smaller units; a subset of identified clusters is randomly selected. ▪ systematic sampling: involves selecting individuals according to a predetermined sequence, which must originate by chance.

Nonprobability Sampling Techniques Convenience sampling: also known as accidental sampling; takes samples that are readily available; appropriate for less demanding research problems. Quota sampling: a variation of convenience sampling; selects participants in the same proportion that they are found in the general population, but not in a random fashion. Purposive sampling: participants are chosen for a particular purpose; the researcher must always provide a rationale explaining the selection of a particular sample.

Choosing an Appropriate Sample Size The larger the sample, the better. For smaller populations (N=100 or fewer), survey the entire population. If population is around 500, sample 50%. If population is around 1,500, sample 20%. If population is over 5,000, a sample size of 400 is fine. The larger the population, the smaller the percentage needed for a representative sample.

Sampling Bias Bias: any influence, condition, or set of conditions that singly or in combination distort the data. Sampling Bias: any influence that may disturb the randomness by which the choice of a sample population has been selected. Strategies for identifying sampling bias: - Scrutinize the questionnaire for items that may be influenced by factors that distinguish respondents from nonrespondents. - Compare responses that were returned quickly with those that were returned later (may reflect the kinds of responses that nonrespondents would have given. - Randomly select a small number of nonrespondents and match their answers against those of respondents.

Some Final Suggestions Questions related to your research project: ▪ Why is a description of this population and/or phenomenon valuable? ▪ What specific data will I need to solve the research problem and subproblems? ▪ What procedures do I need to get the information? How should I implement the procedures? ▪ How do I get a sample that is truly reflective of the entire population about which I am concerned? ▪ How can I collect my data in a way that ensures no misrepresentations or misunderstandings? ▪ How do I control for possible bias in the collection and description of data? ▪ What do I do with the data once I have collected them? How do I organize them and prepare them for analysis?