# Observation and Surveys

Psychology I · Research Techniques · https://tryals.app/learn/psychology-i/observation-and-surveys

## Watching Behavior

**Observational research** studies behavior as it happens without manipulating any variable.

| Method | Setting | Key strength | Key risk |
| --- | --- | --- | --- |
| **Naturalistic** | Real world | High realism | Low control |
| **Structured** | Lab | Consistency | Lower realism |

**Naturalistic observation** watches behavior in real-world settings with zero intervention. **Structured observation** uses standardized settings like labs with fixed rules. A major threat is **observer bias**, where expectations alter interpretation. This is measured by **inter-rater reliability**, the exact degree of agreement between independent observers.

## Asking and Reporting

**Surveys and questionnaires** have people report on their thoughts using **Likert scales** (like 1-5 rating scales).

| Survey Bias | What it means |
| --- | --- |
| **Social desirability** | Answering favorably over honestly |
| **Wording effects** | Phrasing changes responses |

> **Common pitfall:** Assuming surveys capture absolute reality. Self-reports pass through memory filters and social pressure. Always cross-check surveys with behavioral observations whenever possible.

## Practice questions

7 of this lesson's 11 practice questions, with answers. The full set is in the app.

### 1. The face-to-face figure is the better estimate of a stigmatised behaviour, because people answer more carefully when asked in person.

**Answer:** False

**Why:** False. Social desirability suppresses admission of stigmatised behaviour, so the face-to-face figure is a *floor* and the higher anonymous figure is closer to the truth. Knowing a bias's direction is what lets you bound the answer even when you cannot remove it.

Page: https://tryals.app/practice/psychology-i/observation-and-surveys/the-face-to-face-figure-is-the-better-estimate-of-a-stigmatised

### 2. A survey achieves a 12% response rate from a perfectly drawn random sample. Why does the careful sampling not save it?

A. Because the non-responders can be assumed identical to responders, making the effort wasted
B. Because the 12% who replied are self-selected, and the sample was not
C. Because 12% is too small a number of people to analyse statistically
D. Because low response rates increase random error, widening the confidence interval

**Answer:** B. Because the 12% who replied are self-selected, and the sample was not

**Why:** Randomisation happened at invitation; selection happened at reply. If the people who bother to answer a survey on civic engagement are the engaged ones, the estimate is biased and no sample size fixes it, the same lesson as the self-selected online study, arriving through the back door.

Page: https://tryals.app/practice/psychology-i/observation-and-surveys/a-survey-achieves-a-12-response-rate-from-a-perfectly-drawn-random

### 3. Select every statement that is TRUE.

A. Structured observation trades some realism for easier comparison across participants.
B. Naturalistic observation maximizes realism but sacrifices control.
C. Inter-rater reliability checks how much independent observers agree.
D. Wording effects can shift responses to a question meant to stay the same.
E. Self-report survey data never needs cross-checking against behavior.

**Answer:** A. Structured observation trades some realism for easier comparison across participants.; B. Naturalistic observation maximizes realism but sacrifices control.; C. Inter-rater reliability checks how much independent observers agree.; D. Wording effects can shift responses to a question meant to stay the same.

**Why:** Realism-vs-control, structured comparison, wording effects, and inter-rater reliability are all true. Self-report DOES need cross-checking because it passes through memory and self-presentation - the one false option overstates its accuracy.

Page: https://tryals.app/practice/psychology-i/observation-and-surveys/select-every-statement-that-is-true

### 4. Naturalistic observation watches behaviour where it normally happens; structured observation stages the situation in a controlled setting. Why does gaining control cost you something that cannot simply be added back?

A. Because structured observation uses smaller samples, so its findings are statistically weaker.
B. Because participants in a controlled setting will deliberately mislead the researcher.
C. Because naturalistic observation is unscientific, so combining the two produces no real gain.
D. Because control is achieved by removing the ordinary context.

**Answer:** D. Because control is achieved by removing the ordinary context.

**Why:** Control is obtained by stripping out variation, and the stripped-out variation *is* the ordinary context, so if playground aggression depends on crowding and escape routes, the playroom has removed the things doing the work. The strongest evidence is finding the same result in both, since the two methods' weaknesses are different and no single alternative explanation covers both.

Page: https://tryals.app/practice/psychology-i/observation-and-surveys/naturalistic-observation-watches-behaviour-where-it-normally-happens

### 5. Order the steps of guarding an observational study against observer bias.

**Answer:**

1. Recognize that an observer's expectations could shape what they notice
2. Train several independent observers on the same coding scheme
3. Have them all code the same behavior separately
4. Compute their inter-rater reliability to check agreement

**Why:** Acknowledge the bias, train independent observers, have them code separately, then compute inter-rater reliability. Agreement can only be measured after multiple observers have independently rated the same behavior - the check is the last step.

Page: https://tryals.app/practice/psychology-i/observation-and-surveys/order-the-steps-of-guarding-an-observational-study-against-observer

### 6. Sort each described study by observation method.

**Answer:**

- Naturalistic observation: Watching children's play on a real public playground, no intervention, Filming shoppers in an actual supermarket without altering anything
- Structured (controlled) observation: A lab playroom with fixed toys and a fixed time limit for every child, A standardized stranger-approach task run identically for each infant

**Why:** Naturalistic observation watches behavior in its real setting with no intervention (maximizing realism); structured observation standardizes the setting for comparability (sacrificing some realism). The presence of a fixed, controlled setup is the discriminator.

Page: https://tryals.app/practice/psychology-i/observation-and-surveys/sort-each-described-study-by-observation-method

### 7. Even a well-designed survey does not simply capture what people think and do. What gets in the way?

A. The Likert scale, which cannot represent genuine opinion
B. Social desirability and question wording
C. Nothing, provided the sample is large enough
D. Sampling error, which no survey design can reduce

**Answer:** B. Social desirability and question wording

**Why:** The instrument participates in the answer. This is why surveys are cross-checked against behavioural data and why question wording is itself a research literature.

Page: https://tryals.app/practice/psychology-i/observation-and-surveys/even-a-well-designed-survey-does-not-simply-capture-what-people-think