Week 3 · Chapter

Chapter 3 — Studying Children: Research, Observation, and Ethics

Child Development Free sample edition

The module's primary reading · ~40–50 minutes · Objective 2 (describe research and observation methods with children — descriptive, correlational, and experimental designs; longitudinal and cross-sectional designs; running records, anecdotal records, checklists, and time and event sampling — and the ethics of consent, assent, and confidentiality), building on Objective 1 (explain child development's domains, periods, and core questions (nature and nurture, continuity and discontinuity), the ecological systems model, and the major theories of development, describing what each theory explains and what it leaves out) · Read it first this week; the lecture, the slide deck, the tutorial, the quiz, and the Child Observation Workshop all build on it.

In plain language

This week is about how we learn anything true about children, and how we do it without harming them. You meet the research methods, meaning the planned ways researchers gather evidence, from watching children at play to running careful experiments. You learn five observation records, meaning written accounts of what a child did, that teachers and researchers use every day. You practice keeping what a child did apart from what you think it means. You also learn the ethics of studying children, set out by the Society for Research in Child Development in principles approved in 2021, and the course's own observation rules.

Why you already care about this

Think about the last time you watched a child for a few minutes. Maybe it was a 2-year-old at a family dinner, a 4-year-old in line at the grocery store, or 3 kids on a playground. You probably noticed something and wondered why. Maybe you also caught yourself deciding what the child was feeling before you knew for sure.

That quick leap is human. It is also where most mistakes about children begin. A parent hears "your child was difficult today" and has no idea what actually happened. A teacher reads a note that says "she was bored" and cannot tell whether the child yawned, wandered off, or asked for a new book. Careful description protects children from being judged on one moment.

This week gives you the tools that careful people use. You will see how researchers decide whether two things are related or whether one causes the other. You will also see how educators write down what a child did so that a second observer would agree.

By the end of this chapter you will be able to:

name the steps of the scientific method;
tell descriptive, correlational, and experimental research apart, and say why a correlation does not show a cause;
tell a longitudinal design from a cross-sectional design;
choose among the running record, anecdotal record, checklist, time sampling, and event sampling;
write an observation in two columns that keep description apart from interpretation;
follow the ethics of consent, assent, confidentiality, and the course's observation rules.

Carry this line with you:

Write what the child did, not what you think it meant.

Where we are

Already covered: Week 1, Introduction: What Child Development Is, gave you the three domains, the periods, the core questions, and the running case. Week 2, Theories of Development, gave you the lenses: Erikson's psychosocial stages, behaviorism and social learning, Piaget and Vygotsky named, and ethology.
This week: how children are studied: research methods and designs, five observation records, the line between description and interpretation, and research and observation ethics.
How they connect: a theory needs evidence, and this week shows you where the evidence comes from and how to record it fairly.
Where we're going next: Week 4, Prenatal Development and Birth, begins the course of development in time order, and every record you read from then on uses this week's methods.

Week 1 introduced the running case: an illustrative, unnamed child followed through eight observation records from birth to age 15. Week 2 read record R4, the child at 2½ years saying "Me do it" at a childcare center, through three theories. This week you look at R4 again, this time as a record. You ask how it was written, what it kept, and what it left out.

1. Ask a question the scientific way

Suppose you notice that the toddlers at a library story time seem to sit longer when the reader uses puppets. That is a hunch. A hunch can be right or wrong. The scientific method is how you find out.

The scientific method is a careful process for answering a question: ask a question, make a prediction, gather observations, look at what they show, and share the result so others can check it.

The last step matters more than it looks. A result that no one else can check is a private opinion. Sharing your method lets another person repeat it and see whether they get the same result.

Worked example — the story-time hunch. Walk the steps with the puppet idea.

Ask: do toddlers at story time stay seated longer when the reader uses puppets?
Predict: toddlers will stay seated longer on puppet days.
Observe: on several story-time days, with and without puppets, record how long each toddler stays seated.
Analyze: compare the puppet days with the other days.
Report the result: describe exactly what you did, so another observer at another library could try it.

In words: the scientific method turns "I noticed something" into "here is what I did, what I found, and how you can check it."

Misconception check — a prediction read as a finding. Students often treat the prediction as if it were the result: "puppets keep toddlers seated" before anyone has watched. It is tempting because the hunch feels true. The cure: a prediction is step two, and it is only a guess until the observations in steps three and four support it or do not.

2. Describe what children do

Before you can explain a behavior, you need a good description of it. Descriptive methods are research methods that describe what children do without testing a cause. There are five you will meet often.

Naturalistic observation is watching children's behavior in the everyday settings where it normally happens, without setting anything up. Think of a playground's morning: you watch from a bench and record what happens.
Structured observation is setting up the same situation for every child in a study, so that a chosen behavior can be observed and compared across children. Every child might be handed the same puzzle and watched for the same few minutes.
A case study is a close look at one child or one family over time.
An interview is a set of questions asked of a child or an adult, face to face.
A survey asks many people the same written questions.

Each method sees something and misses something. Naturalistic observation shows real life, but the behavior you care about might not happen while you are watching. Structured observation makes sure it happens, but the setting is less natural. Interviews and surveys reach thoughts you cannot see, but people do not always remember or report accurately.

Worked example — matching the method to the question. Read each question, then pick the method that can answer it.

Question: how do children take turns on the slide when no adult steps in? Naturalistic observation fits, because the point is everyday behavior.
Question: how do 3-year-olds and 5-year-olds handle the same puzzle? Structured observation fits, because every child needs the same task.
Question: how do families plan bedtime on school nights? A survey or an interview fits, because a bedtime plan is not something you can watch from a park bench.

Misconception check — one observation read as the whole child. Students often treat one observation as a description of who the child is: "I saw her alone at recess, so she is a loner." It is tempting because the moment was vivid. The cure: one observation describes one moment in one setting. A fair description needs many observations, in more than one setting, and it never becomes a label.

3. Test whether two things go together

Picture a researcher who visits many families at dinner. She records how often each family reads together at dinner and how many words each child uses. Suppose she finds that the two go up together. That is a correlation.

Correlational research measures whether two things vary together; it can show that they are related but cannot show that one causes the other.

Why can't it show a cause? Because a third thing might explain both. Families with more time in the evening might read more and also talk more at dinner. Families with more books at home might do both. The reading might cause some of the word growth, or it might not. A correlation alone cannot tell you.

Worked example — naming the third things. Take the dinner study and ask what else might explain the link.

The link: more reading at dinner goes with more words.
A third thing: how much time the family has in the evening.
Another third thing: how much the adults talk with the child at other times.
A reversed direction: children who already use more words might ask for more reading.

In words: the dinner study shows that reading and words go together, and any of those other things could be part of why.

Misconception check — correlation read as causation. Students often read "goes together" as "causes": "reading at dinner builds vocabulary." It is tempting because the story makes sense. The cure: correlation is not causation. Name at least one third thing, and one reversed direction, before you trust a causal claim.

4. Test whether one thing causes another

To test a cause, a researcher has to control what changes. Experimental research does that.

Experimental research is research in which the investigator changes one thing and compares conditions that children were randomly assigned to, so that it can test whether that one thing causes a difference.
Random assignment is placing each participant in a study's conditions by chance, so that the groups start out alike and a later difference can be traced to the condition rather than to who was in each group.

Random assignment is the key. If children choose their own group, the groups may differ from the start. If a coin flip decides, those differences tend to even out.

You met a classic study last week. Bandura, Ross, and Ross (1961) compared children who watched an adult act aggressively toward an inflatable Bobo doll with children who did not. The researchers set up the conditions themselves, which is the shape of an experiment, and the children who watched the aggressive adult were more likely to imitate that aggression.

Worked example — turning the story-time hunch into an experiment. Set it up step by step.

The one thing that changes: puppets or no puppets.
Random assignment: a coin flip decides which story-time sessions use puppets.
Everything else held the same: the same reader, the same room, the same length of story.
The comparison: how long toddlers stay seated in each condition.

In words: because chance, not the reader's mood, picked the puppet days, a difference in sitting time points toward the puppets.

Honesty note: many questions about children cannot be tested by experiment. No one may randomly assign children to poverty or to a harsh home. That is why so much child development research is correlational, and why you will keep asking "go together, or cause?"

Misconception check — any comparison read as an experiment. Students often call every study that compares two groups an experiment. It is tempting because both compare groups. The cure: an experiment changes one thing and uses random assignment. Comparing children whose families already chose different routines is correlational, even if it looks like a test.

5. Add time: longitudinal, cross-sectional, and sequential designs

Development is change over time, so researchers need a way to study time. There are three main designs.

A longitudinal design is a research design that follows the same children over time and measures them again and again, so it can show change within the same children.
A cross-sectional design is a research design that compares children of different ages at one time; its differences are between groups of children, not change within the same children.
A sequential design is a research design that combines the other two by starting several age groups at once and following each group over time.

A longitudinal design shows real change, but it takes years and some families drop out. A cross-sectional design is fast, but children of different ages also grew up at different times.

Worked example — the running case as a design. The running case follows the same child through 8 records.

Same child, many times: that makes it longitudinal.
Two of its looks: record R5 is at 4 years and record R6 at 7 years.
The gap: 7 − 4 = 3, so 3 years separate those two looks at the same child.
The cross-sectional version: compare a group of 4-year-olds with a different group of 7-year-olds at one time.

In words: the running case can show how one child changed between 4 and 7; the cross-sectional version can show only how two groups differ.

Limitation: the running case is one illustrative child, so it shows how longitudinal records work, not what all children do.

Misconception check — group differences read as change. Students often read a cross-sectional difference as growth: "7-year-olds share more than 4-year-olds, so children learn to share between 4 and 7." It is tempting because the ages line up. The cure: those were different children. Only following the same children over time, in a longitudinal or sequential design, shows change within the same children.

6. Choose an observation tool

Teachers and caregivers observe children every day, and they use five common records. Each one fits a different question.

A running record is a timed, continuous, word-for-word account of what a child does and says, written in the present tense and in time order, with no interpretation inside the record.
An anecdotal record is a brief account of one significant event in a child's behavior, written soon after the event happens.
A checklist marks whether each of a list of behaviors was seen during an observation, without describing how or when.
Time sampling records what a child is doing at set intervals, such as at the end of every few seconds or minutes.
Event sampling records each time a chosen behavior occurs, often with what happened before and after it.

Worked example — record R4 two ways. Here is Dataset A's record R4 (illustrative), written first as a running record. The child is 2½ years old, at a childcare center's block area, on Day 1.

Running record of record R4 (illustrative): the child, 2½ years, childcare center block area, Day 1, 9:05 to 9:07
Time	What the child did and said
9:05	Stacks a block on a tower. Another child stacks blocks on a separate tower beside the child.
9:06	Talks quietly while stacking. The other child also talks quietly. Neither child speaks to the other. No shared plan.
9:07	A caregiver reaches toward the tower. Pushes the caregiver's hand away. Says, "Me do it."

The running record has 3 timed lines, from 9:05 to 9:07, and 9:07 − 9:05 = 2 minutes of continuous recording.

Now the same moment as an anecdotal record, written soon after:

Day 1, block area. The child, 2½ years, was building a tower when a caregiver reached toward it. The child pushed the caregiver's hand away and said, "Me do it."

What each kept:

The running record keeps all 3 timed lines, both children's quiet talk, and the order of events.
The anecdotal record keeps 1 event. It loses the side-by-side building and the minute-by-minute order.

In words: use a running record when the order and the context matter, and an anecdotal record when one event is what you need to remember.

The other three tools answer other questions. A checklist tells you quickly whether "stacks three or more blocks" was seen, but not how. Time sampling tells you what the child was doing at each set moment, but it misses what happens between moments. Event sampling catches every push or every shared toy, but it ignores everything else.

Misconception check — a checklist read as a full record. Students often think a checked box tells the whole story. It is tempting because a checklist looks complete. The cure: a checklist records only whether something was seen. It cannot tell you what came before, how long it lasted, or what the child said.

7. Keep description apart from interpretation

Here is the habit this whole course rests on. Write what the child did in one column. Write what it might mean in another.

Objective description is the part of an observation record that states only what the child did and said, in words any observer present would agree on.
Subjective interpretation is an observer's reading of what a behavior might mean, written in a separate column as a possibility rather than a fact.

The test is easy to say and hard to do. Could a second observer, standing beside you, agree with every word in the left column? "Throws the block and cries" passes. "Was angry" does not, because anger is a guess about feelings you cannot see.

To make agreement possible, you define the behavior before you watch. "Pushes away" is defined as "uses a hand to move an adult's hand off or away from the materials", so two observers would mark the 9:07 line the same way.

Worked example — sorting R4. Put each statement in its column.

Record R4 (illustrative) sorted into description and interpretation
What the child did and said	Possible interpretations
Pushes the caregiver's hand away. Says, "Me do it."	May show a wish to act on one's own, which Erikson's autonomy vs. shame and doubt describes.
Builds beside another child; neither speaks to the other.	May show a child practicing a skill, which a behaviorist would read through what happens next.
(left blank: no behavior supports it yet)	The child was annoyed with the caregiver.

"Pushes the caregiver's hand away. Says, 'Me do it.'" is objective description, so it stays in the left column.
"The child was annoyed with the caregiver" is subjective interpretation. It moves to the right column as a possibility, not a fact.
The theories from Week 2 fill the right column, each written as "may show".

In words: the left column is what anyone could see; the right column is what theory suggests, held loosely.

Misconception check — an interpretation written inside the observation. Students often write "the child was angry" where they mean "the child threw the block and cried". It is tempting because the feeling seems plain in the moment. The cure: write the behavior and the words. Move the feeling to the interpretation column, and write it as "may show".

Pause and choose one of these two.

(a) Describe a moment when someone summed you up with a feeling word, such as "you seemed upset", and say what they actually saw.

(b) Describe a moment when you summed up someone else with a feeling word, and rewrite it as what you actually saw.

8. Protect the child: consent, assent, and the observation rules

Children cannot protect themselves from a careless researcher, so research with children has firm rules. The Society for Research in Child Development sets them out in its Ethical Principles and Standards for Developmental Scientists, approved in 2021.

Informed consent is permission to take part in research given by a parent or guardian after the purpose, the procedures, and the right to stop have been explained.
Assent is a child's own agreement to take part in research, asked for in words the child can understand; it is separate from a parent's or guardian's consent.
Confidentiality is keeping research or observation records private so that no child or family can be identified from them.
Every child keeps the right to stop at any time.

This course adds its own rules, Dataset F (illustrative), which every Child Observation Workshop follows. The default material is the running case's pinned records. An optional live observation happens only in a public place where watching children is ordinary, such as a park, a library story time, or a children's museum, and only from a distance. It can also happen in a licensed program with the program's written permission. You never interact with, photograph, or record a child. You write no name or identifying detail. You never assess, screen, or label a real child. A family member observing their own child follows the same rules and shares nothing that identifies the child.

Before an optional observation in a licensed program, you read the permission script to the program director, word for word:

"I'm a student in a child development course, and I'm learning to observe children's play. May I observe the children in one of your play areas for about twenty minutes? I won't interact with the children, I won't take photos or recordings, and I won't write down any names or anything that identifies a child or family. I won't assess or evaluate any child. Is that all right?"

Worked example — checking R4 against Dataset F. Hold the record up to each rule.

Names: the record names no child, adult, or program.
Roles: it identifies people by role and age: "the child, 2½ years", "a caregiver", "another child".
Permission: it was written with permission.
The script: it asks for about twenty minutes and makes 4 promises: no interacting, no photos or recordings, no names or identifying details, and no assessing or evaluating any child.

In words: R4 follows the rules because it describes behavior, by role and age, with permission, and nothing in it could point to a real child.

Misconception check — assent read as consent. Students often think a parent's yes settles everything. It is tempting because the parent is the legal decision-maker. The cure: consent comes from the parent or guardian, and assent comes from the child. A child who says no, or who turns away, keeps the right to stop.

9. Say the reading in one plain sentence

Every reading of a record in this course ends the same way: one plain sentence that names the method, what it kept, and what it left out. The sentence keeps you honest. It stops you from claiming more than the record shows.

Worked example — the plain sentence for R4. Gather the pieces.

Method one: the running record kept the order and the side-by-side building.
Method two: the anecdotal record kept the one event that matters most, "Me do it".
The limit: neither record says what the child felt.

The plain sentence: the running record of R4 keeps the order and the side-by-side building that the anecdotal record drops. The anecdotal record keeps the one event that matters most, "Me do it". Neither record says what the child felt.

Caveat: a plain sentence about a record is still a reading of one moment. It is never a verdict on the child.

Misconception check — the plain sentence read as a label. Students sometimes end a reading with a judgment of the child: "this is a stubborn child." It is tempting because a label feels like a conclusion. The cure: name the method and the evidence, not the child's character. A record describes a moment; it never sorts a real child into a type.

Key terms

Scientific method — A careful process for answering a question about children: ask a question, make a prediction, gather observations, look at what they show, and share the result so others can check it. (first used: 1. Ask a question the scientific way)
Descriptive methods — Research methods that describe what children do without testing a cause; they include naturalistic observation, structured observation, case studies, interviews, and surveys. (first used: 2. Describe what children do)
Naturalistic observation — Watching children's behavior in the everyday settings where it normally happens, such as a playground or a family home, without setting anything up. (first used: 2. Describe what children do)
Structured observation — Setting up the same situation for every child in a study so that a chosen behavior can be observed and compared across children. (first used: 2. Describe what children do)
Correlational research — Research that measures whether two things vary together; it can show that they are related but cannot show that one causes the other. (first used: 3. Test whether two things go together)
Experimental research — Research in which the investigator changes one thing and compares conditions that children were randomly assigned to, so that it can test whether that one thing causes a difference. (first used: 4. Test whether one thing causes another)
Random assignment — Placing each participant in a study's conditions by chance, so that the groups start out alike and a later difference can be traced to the condition rather than to who was in each group. (first used: 4. Test whether one thing causes another)
Longitudinal design — A research design that follows the same children over time and measures them again and again, so it can show change within the same children. (first used: 5. Add time: longitudinal, cross-sectional, and sequential designs)
Cross-sectional design — A research design that compares children of different ages at one time; its differences are between groups of children, not change within the same children. (first used: 5. Add time: longitudinal, cross-sectional, and sequential designs)
Sequential design — A research design that combines the other two by starting several age groups at once and following each group over time. (first used: 5. Add time: longitudinal, cross-sectional, and sequential designs)
Running record — A timed, continuous, word-for-word account of what a child does and says, written in the present tense and in time order, with no interpretation inside the record. (first used: 6. Choose an observation tool)
Anecdotal record — A brief account of one significant event in a child's behavior, written soon after the event happens. (first used: 6. Choose an observation tool)
Checklist — An observation tool that marks whether each of a list of behaviors was seen during an observation, without describing how or when. (first used: 6. Choose an observation tool)
Time sampling — An observation method that records what a child is doing at set intervals, such as at the end of every few seconds or minutes. (first used: 6. Choose an observation tool)
Event sampling — An observation method that records each time a chosen behavior occurs, often with what happened before and after it. (first used: 6. Choose an observation tool)
Objective description — The part of an observation record that states only what the child did and said, in words any observer present would agree on. (first used: 7. Keep description apart from interpretation)
Subjective interpretation — An observer's reading of what a behavior might mean, written in a separate column as a possibility rather than a fact. (first used: 7. Keep description apart from interpretation)
Informed consent — Permission to take part in research given by a parent or guardian after the purpose, the procedures, and the right to stop have been explained. (first used: 8. Protect the child: consent, assent, and the observation rules)
Assent — A child's own agreement to take part in research, asked for in words the child can understand; it is separate from a parent's or guardian's consent. (first used: 8. Protect the child: consent, assent, and the observation rules)
Confidentiality — Keeping research or observation records private so that no child or family can be identified from them. (first used: 8. Protect the child: consent, assent, and the observation rules)
Permission script — The fixed request this course's students read to a program director before an optional observation in a licensed program, asking for about twenty minutes and promising no interaction, no photos or recordings, no identifying details, and no assessment of any child. (first used: 8. Protect the child: consent, assent, and the observation rules)

Summary

The scientific method runs ask, predict, observe, analyze, share.
Descriptive methods show what children do; they do not test a cause.
Correlation shows that two things go together, not that one causes the other.
An experiment changes one thing and uses random assignment, so it can test a cause.
Longitudinal designs follow the same children; cross-sectional designs compare different children at one time.
Choose the record that fits the question: running record, anecdotal record, checklist, time sampling, or event sampling.
Keep two columns: what the child did and said, and possible interpretations.
Consent comes from a parent or guardian; assent comes from the child; both keep the right to stop.
Every observation follows Dataset F and identifies no real child.
End every reading in one plain sentence about the method and the evidence, never about the child's character.

Check your understanding

Work these before looking at the answers — they are ungraded, and the answers (with reasoning) are at the bottom of the page.

An illustrative youth soccer league finds that players whose families attend more games also show up to more practices. A coach says, "Family attendance makes kids practice more." What kind of research is this, and what is one third thing that could explain the link?
A researcher wants to know how often children in an illustrative after-school homework room ask an adult for help. She does not care what else they do. Which observation tool fits best, and why?
A family member writes this line about a toddler at a children's museum: "He got bored with the water table." Rewrite it as objective description, and say where the original sentence belongs.
In one month, an illustrative study tests the memory of 25 five-year-olds and 25 nine-year-olds and finds the older group remembers more. How many children were tested, what design is this, and what can it not show?

Answers (worked reasoning — read after attempting)

This is correlational research: it measures two things, family attendance and practice attendance, and finds they go together. One third thing is how much free time the family has, since more free time could mean both more games watched and more practices reached. Because nothing was randomly assigned, the coach's causal sentence goes further than the evidence.
Event sampling fits best, because the question is about one chosen behavior, asking an adult for help, and event sampling records each time it occurs. A running record would capture far more than she needs, and a checklist would show only that asking happened, not how often.
A description could read: "Splashes at the water table for a short time, then walks to the block area." The original sentence, "He got bored", is a guess about a feeling. It belongs in the interpretation column, written as "may show boredom", and only as one possibility.
25 + 25 = 50 children were tested. Different children of two ages were compared at one time, so this is a cross-sectional design. It can show that the two groups differ, but it cannot show that any one child's memory grew between 5 and 9, because no child was tested twice.

Next up in this module: the Readings & Resources page, the lecture tutorial, the practice set, and the quiz; next week the Apgar score is the first scoring method you read.

Built with The Course Maker · samples.thecoursemaker.com