Infants and Children in Context. Tara L. Kuther

Measures

      Some developmental scientists collect data by watching and recording children’s behavior. Developmental scientists employ two types of observational measures: naturalistic observation and structured observation.

      Scientists who use naturalistic observation observe and record behavior in natural, real-world settings. For example, Salo, Rowe, and Reeb-Sutherland (2018) observed 12-month-old infants playing with their parents. They recorded infants’ gestures and how often they participated with parents in paying attention to or interacting with an object (such as a toy). One year later, infants who used more gestures and engaged in more joint attention, especially responses to parents’ efforts to direct their attention, showed more advanced language development; they understood and produced more words.

      This researcher is using a video camera to observe and record the facial expressions a newborn baby makes while it sleeps.

      Thierry Berrod, Mona Lisa Production/Science Source

      Sometimes the presence of an observer causes those being observed to behave in unnatural ways or ways that are not typical for them. This is known as participant reactivity, and it poses a challenge to gathering data by naturalistic observation. One way of reducing the effect of observation is to conduct multiple observations so that the children get used to the observer and return to their normal behavior. Another promising method of minimizing participant reactivity is to use an electronically activated voice recorder (EAR) (Mehl, 2017). Participants carry the EAR as they go about their daily lives. The EAR captures segments of information over time: hours, days, or even weeks. It yields a log of people’s activities as they naturally unfold. The EAR minimizes participant reactivity because the participant is unaware of exactly when the EAR is recording. For example, researchers who study child trauma use the EAR to sample conversations between parents and children to understand how parent–child interactions influence children’s adjustment and how the family environment can aid children’s recovery from trauma (Alisic, Krishna, Robbins, & Mehl, 2016).

      Naturalistic observation permits researchers to observe patterns of behavior in everyday settings, such as whether an event or behavior typically precedes another. Such observations can help researchers determine which behaviors are important to study in the first place. For example, a scientist who studies bullying by observing children’s play may notice that some victims act aggressively before a bullying encounter (Kamper-DeMarco & Ostrov, 2017). The scientist may then decide to examine aggression in victims not only after a bullying incident but also beforehand. Naturalistic observation is a useful way of studying events and behaviors that are common. However, some behaviors and events occur infrequently, requiring a researcher to observe for very long periods of time to obtain data on the behavior of interest. For this reason, many researchers make structured observations.

      Structured observation entails observing and recording behaviors displayed in a controlled environment, a situation constructed by the experimenter. For example, children might be observed in a laboratory setting as they play with another child or complete a puzzle-solving task. The challenges of identifying and categorizing which behaviors to record are similar to those involved in naturalistic observation. However, the laboratory environment permits researchers to exert more control over the situation than is possible in natural settings. In addition to cataloging observable behaviors, some researchers use technology to measure biological functions such as heart rate, brain waves, and blood pressure. One challenge to conducting structured observations is that children do not always behave in laboratory settings as they do in real life.

      Self-Report Measures

      Interviews and questionnaires are known as self-report measures because the child under study answers questions about his or her experiences, attitudes, opinions, beliefs, and behavior. Interviews can take place in person, over the phone, or over the Internet.

      One type of interview is the open-ended interview, in which a trained interviewer uses a conversational style that encourages the child under study to expand his or her responses. Interviewers may vary the order of questions, probe, and ask additional questions based on responses. The scientist begins with a question and then follows up with prompts to obtain a better view of the person’s reasoning (Ginsburg, 1997). An example of this is the Piagetian Clinical Interview, which requires specialized training to administer. Consider this dialogue between Piaget and a 6-year-old child:

       You know what a dream is?

       When you are asleep and you see something.

       Where does it come from?

       The sky.

       Can you see it?

       No! Yes, when you’re asleep.

       Could I see it if I was there?

       No.

       Why not?

       Because it is in front of us. … When you are asleep you dream and you see them, but when you aren’t asleep you don’t see them.

       (Piaget, 1929, p. 93)

      Open-ended interviews allow children to explain their thoughts thoroughly and in their own words. They also enable scientists to gather a large amount of information quickly. Open-ended interviews are very flexible as well. However, their flexibility poses a challenge: When questions are phrased differently for each child, responses may not capture real differences in how children think about a given topic and instead may reflect differences in how the questions were posed and followed up by the interviewer.

      A structured interview poses the same set of questions to each child in the same way. On one hand, structured interviews are less flexible than open-ended interviews. On the other hand, because all children receive the same set of questions, differences in responses are more likely to reflect true differences among children and not merely differences in the manner of interviewing. For example, Evans, Milanak, Medeiros, and Ross (2002) used a structured interview to examine American children’s beliefs about magic. Children between the ages of 3 and 8 were asked the following set of questions:

       What is magic? Who can do magic?

       Is it possible to have special powers? Who has special powers?

       Does someone have to learn to do magic? Where have you seen magic? (p. 49)

      After compiling and analyzing the children’s responses as well as administering several cognitive tasks, the researchers concluded that even older children, who have the ability to think logically and perform concrete operations, may display magical beliefs.

      To collect data from large samples of people, scientists may compile and use questionnaires, also called surveys, made up of sets of questions, typically multiple choice. Questionnaires can be administered in person, online, or by telephone, email, or postal mail. Questionnaires are popular data collection methods with adolescents because they are easy to use and enable scientists to collect information from many people quickly and inexpensively. Scientists who conduct research on sensitive topics, such as sexual interest and experience, often use questionnaires because they can easily be administered anonymously, protecting participants’ privacy. For example, the Monitoring the Future Study is an annual survey of 50,000 eighth-, tenth-, and twelfth-grade students that collects information about their behaviors, attitudes, and values concerning drug and alcohol use (Miech et al., 2017). The survey permits scientists to gather an enormous amount of data, yet its anonymity protects the adolescents from the consequences of sharing personal information that they might not otherwise reveal. Questionnaires, however, rely on a child’s ability to read and understand questions and provide responses. It is not until late childhood and, more often, adolescence that questionnaires become feasible sources of data.

      Despite