Research methods and research design

Research methods

Systematic research can be conducted in a number of ways. Researchers select which approach is most appropriate to investigate the question or problem they have identified. A research method is the systematic way of collecting the data to measure or assess the variable(s) being studied. For child development research, researchers can watch or observe children in a laboratory or in a more natural setting such as home, school, or a neighbourhood playground. They can ask children questions, give children standardized tests, conduct case studies, examine human behaviour across different cultures, or carry out physiological research.

Observation

Casually watching a group of children at play is not a reliable method to collect data for a research study. Effective observation skills include knowing what to look for; where, when, and how to make the observations; and, how to record them in a systematic manner.

Suppose a researcher wants to study parent-child interactions. Will she want to study all interactions or just those interactions that happen when the parent is leaving the child to go to work? Does the researcher want to study younger or older children? Does she want to observe the interactions in a laboratory setting, at home, in a childcare centre or in a playground? Will the observations be written down in a narrative form or on coded observation sheets? Should video cameras or tape recorders be used?

Researchers often choose laboratory settings when they want to control for specific factors that may affect children’s behaviour but are not the focus of the research study. A laboratory setting is one where many of the complex and confounding factors of the child’s real world are eliminated or reduced. For instance, a researcher who wants to investigate how different types of construction materials may influence children’s activity in building structures may set up a structure and environment in a laboratory setting to eliminate factors such as other children or familiar toys. He can introduce materials that the children have not seen before and study the effect of different ways to set up the toys or different guidance techniquesfrom adults.

Laboratory research does have limitations. For the most part, children observed in a laboratory situation are aware that they are being observed. The artificial or unnatural setting implicit in a laboratory may be a factor influencing their behaviour. Also the child behaviour researchers may wish to investigate may be difficult, if not impossible, to examine in a laboratory setting. For example, it would be difficult and unethical, to study children’s reactions to family conflict ortragedy in a laboratory setting.

In naturalistic observations, researchers observe children’s behaviour in real life settings and do not attempt to manipulate or control the setting. A researcher investigating children’s reactions to construction materials might simply observe children in an early learning setting or kindergarten class.

Interviews and Questionnaires

Researchers sometimes ask children (or parents and other adults who know the children) questions. Interviews and questionnaires can be used to find out about children’s attitudes or experiences.

Interviews can be unstructured, or they can be very structured. Unstructured interviews typically use open-ended questions. The interviewer may ask a child, “What makes you angry?” or a parent, “What does your child usually do when she is mad?” The answers may be the basis for follow-up questions. In contrast, structured interviews use specific, close-ended questions such as “How many hours did your child sleep last night?” or “What did you eat for breakfast this morning?” Further structure may be imposed by asking people responding to the questions to select one option from a number of different choices, which may be on a scale or continuum. In a question about how often a child gets into fights with other children, the person responding might be asked to select the most accurate option from the scale: “several times a day”, “at least once a day”, “at least once a week”, “once a month”, “never”.

A questionnaire or survey is similar to an interview, except that the individual responding to the questions reads the questions and marks or writes down the answers. Most surveys are highly structured. Respondents are asked to answer questions from a set of specific options. Questionnaires and surveys can be administered to large numbers of people. A good survey must be carefully designed to ask specific questions, which are unambiguous. Validity (i.e., does the question measure what we intend it to measure) of survey questions should also be tested in several independent research studies.

Case Studies

A case study is an in-depth study of an individual or group of individuals. Another name for this is the clinical method and it has its origins in psychoanalytic theory. Freud’s work with individuals stresses the importance of understanding their life histories. Case studies may bring together information gathered by a number of methods, including observations, interviews, and test scores, which together can help to build a complete picture of a person’s development or the functioning of a small group of people, such as a family.

Case studies can provide in-depth and sometimes dramatic portrayals of people’s lives but it is difficult to generalize from this information.

Ethnography is another type of case study research method, which is often used in anthropological studies. Investigators interested in human development may choose this method, particularly if they are interested in studying the impact of culture. It is a descriptive, qualitative approach, similar to that used to study individuals, but its focus is on understanding a culture or distinct social group.

The ethnographic method uses participant observation. The researcher usually lives with the cultural community for a period of time (months, sometimes years), and participates in all aspects of daily life. Field notes, which include observations, self-reports from members of the social group, and the investigator’s interpretations are collected over this period of time. The notes become a description of the community, which tries to articulate its unique social values and processes.

Standardized Tests

Standardized tests require people to answer a series of oral or written questions. The individual’s total score provides information about that individual. A researcher can compare an individual’s scores to those of a large group of people to determine the individual’s place relative to others. There are standardized tests to measure intelligence, academic achievement, personality, and aggression.

Research Design

Researchers need to draw up an overall plan or research design to investigate the questions or test the hypothesis they have identified. Once the question is clear, researchers must make decisions about who the participants will be, what measures will be used to measure or assess performance or development, and how the data collected will be analyzed.

But before these decisions can be made, researchers must decide on the overall research strategy and approach. When designing a research study on human development, investigators must decide whether to use a correlational or experimental strategy. They must also decide what approach or time span of development that the inquiry will cover.

Correlational Strategy

In the correlational strategy, researchers are looking at the relation between two or more events or characteristics. This is useful because the more strongly two events are related or associated to each other, the more one can be predicted from the other. The relationship or association is called a correlation.

A correlational strategy is useful for investigating a number of questions:

• How can parents help their children succeed in school?
• Do early childhood education experiences have an impact on later school achievement?
• Is a child’s health related to a family’s social and economic standing?

A correlational design strategy provides an opportunity to look at the relationships between the participants’ circumstances, experiences and characteristics and their behaviour and development. However, there is one major limitation of the correlational approach: it cannot be used to determine cause and effect.

A study that considers different parental communication styles and children’s academic achievement, may find that children are more likely to do better in school on math and reading comprehension standard tests if their parents typically communicate and interact with their children in a responsive manner. The study will show a relationship or correlation between responsive parental interaction and positive school achievement. However, it does not indicate that the parents’ communication styles cause the school achievement. It is possible that parents are more responsive to children who are more competent; or there may be other factors or variables, which are not part of the study, which affect both parental interaction and children’s achievements. Genetic background, family income, or cultural values could all be powerful influences.

In correlational studies, investigators often use a numerical measure to describe the relationship they have found between two measures or variables. The measure, called a correlational coefficient, is based on a statistical analysis that is used to describe the degree or strength of the association between two variables. A correlation coefficient can range in value from -1 to +1. The size of the number shows the strength of the relationship. A correlation of zero means that there is no relationship or association between the variables. The closer the value is to +1 or -1, the stronger the relationship. The negative (-) or positive (+) sign of the number indicates the direction of the relationship. A negative number indicates an inverse relationship. For instance if researchers found a correlation of -.5 between a parent’s harsh communication style and children’s academic achievement, it would indicate harsh communication is associated with poorer academic results, or negatively associated with positive academic results. On the other hand, a correlation of +.6 between responsive parental communication and higher academic achievement, shows a positive relationship between the two events. Note that the correlation of +.6 between responsive parental communication and their children’s academic achievement is stronger relationship than the correlation of -.5 between harsh communication and academic achievement. The higher the correlation coefficient (whether positive or negative), the stronger the relationship between the two variables is.

Experimental Strategy

Experimental design strategies allow researchers to determine the causes of behaviour or development. Experiments can establish cause and effect while correlational studies can only show relationships. Cause is the factor being manipulated and effect is the behaviour or developmental outcome, which changes because of the manipulation. In a correlational study, nothing is manipulated, while in an experiment the researcher deliberately changes one factor to see the effect on development.

• The independent variable is the manipulated factor in an experiment. The term independent means that the variable can be changed independently of other factors.
• The dependent variable is the factor that is measured in an experiment. It may change as a result of the manipulation of the independent variables.

In order to ensure that the experiment is really assessing the impact of the independent variable on the dependent variable, researchers must ensure that all other possible variables are held constant or the same for the participants in an experiment. In a traditional scientific experiment involving the cause and effect of sunlight (independent) on the growth (dependent growth) of a bean seed, there are a limited number of variables to control. The investigator would ensure the same seeds, soil, containers, and amount of water are used. Then by varying the amount of sunlight the bean seeds received and systematically recording the growth of the bean plants, the experiment would yield information about the cause and effect of sunlight on bean plant growth.

Human development is much more complicated, with numerous factors involved, making experimental research design much more complex.

Investigators may choose not to use experimental designs for a number of reasons. The focus of the investigation might be new and there is little knowledge about which variables to manipulate. It may be unethical to manipulate the variables (such as exposing children to potentially harmful substances to study the cause of disease), or it may be impossible to manipulate the factors

Randomized Controlled Trials

In experimental studies, because it is virtually impossible to control all factors that might influence human behaviour and development, researchers must take specific precautions to control for unknown characteristics of the participants that could have an impact on the findings. The random assignment of research study participants to control and experimental (or treatment) groups reduces the likelihood that the experiments results will reflect other factors. It reduces the possibility of the two groups differing on factors such as age, intelligence, health problems, or other unknown factors.

Cross-sectional vs longitudinal studies

Cross-sectional studies happen at one point in time. In order to compare children of different ages, the researchers collect results from different groups of children. For example, a researcher might measure the dependent variable on children who are two, four and six years of age to compare whether the independent variable affects children differently based on age. However, in this design, researchers are not able to consider how an individual changes or develops over time. For that consideration, researchers conduct longitudinal studies. Longitudinal studies span a longer period and track the same individuals to assess them at different ages. In the example, above, the same children would be measured at age two, four and six to note changes over time. Cross-sectional studies are more time-efficient since the researcher typically can collect the data in a shorter period. Longitudinal studies, while more time-consuming, allow researchers to analyse how the passage of time affects the individual and how early events impact later ones. One potential challenge researchers may face is how to reduce attrition – the issue of losing track of participants over time and therefore having a shrinking sample.