Psychology as a Science: Concepts, Research and Ethics

Review questions

From Observing Behavior to a Responsible Conclusion

How can we understand behavior without rushing to judge the person? Begin by separating what happened from what we think explains it. Silence in a learning circle describes behavior; lack of interest or difficulty understanding the question are explanations needing evidence. Psychology investigates behavior and mental processes and allows biological, cognitive and social questions about one situation. These levels are not permanent labels for people, and mentioning the brain, thought or society does not establish an explanation.

Historical schools show how questions and methods have changed. Structuralism focused on components of conscious experience, while functionalism asked what mental activity and behavior do. Behaviorism examined behavior in relation to environment and learning; psychoanalysis gave unconscious processes and early experiences a role. Humanism emphasized experience, choice and growth, while cognitive approaches investigated information processing through testable methods. Knowing this history helps distinguish questions but does not establish every proposition of an approach. Evidence is still needed for the particular claim.

Next, move from impressions to measurement. People may interpret “participates well” differently. An operational definition specifies the recorded action, its unit, the period, and inclusions and exclusions. In the discussion example, each relevant message counted once even when it contained two ideas; greetings, technical questions and unchanged repetition were excluded. Three of six messages met the rule: 50% of messages, not 50% of people. The denominator is part of the meaning, not a minor arithmetic detail.

Choose a method to fit the question. A message record answers questions about recorded actions; self-report shows how someone describes their experience. Silence alone does not reveal feelings, and satisfaction does not establish mastery. Counts, rates and proportions also differ. Dana answered three times in four opportunities and Rami three times in six: equal counts, but proportions of 75% and 50%. With no opportunities, record that the proportion is not calculable rather than zero. These results describe an activity, not personal worth or general ability.

Also ask whether measurement is consistent and whether it justifies the intended interpretation. Two counters agreeing on page openings supports count consistency but does not make opening a measure of understanding. This distinguishes reliability from validity. A task in which a learner demonstrates understanding may be needed. Adding appropriate evidence is better than renaming “Openings” as “Understanding score.”

Read the design together with the data. Descriptive work identifies a pattern in the supplied cases; correlational work examines relationships without establishing cause. One variable might affect the other, the direction might be reversed, or a third factor might affect both. An experiment changes an independent variable and measures a dependent outcome while controlling other differences where possible. Comparing two instruction formats does not isolate formatting if one also receives more working time. Random assignment helps comparisons, but it does not turn one club’s volunteers into a sample representing everyone: sample selection and assignment within it serve different purposes.

Numbers need bounded interpretation just as words do. Times of 2, 3, 3, 4 and 18 minutes have a mean of 6 minutes and a median of 3. The mean uses all values; the median identifies the middle position after sorting. The mean need not be most individuals’ time, and a distant value is not a reason to delete it without checking its record. For an even count, average the two middle values to obtain the median. Likewise, positive and negative associations describe direction rather than desirability, and no linear pattern does not exclude every possible relationship.

Finally, establish the scope of evidence and how it was collected. Distinguish original studies, reviews, textbooks and commentaries, and reading an abstract from reading a full article. A detail missing from the abstract is not necessarily absent from the study. Ask who participated, in which language, task and conditions, before transferring results. Replication with new data differs from recalculation using the same data: agreement can strengthen confidence, while disagreement invites examination of procedures and measurement, not automatic accusations or rejection of a field.

Ethics guides the chain from the beginning. Seek only information serving the question; explain purpose, access and withdrawal limits; and protect refusal without penalty. A key connecting codes to names makes data coded, not anonymous. Do not promise individual deletion when a response can no longer be identified, or treat participation as general permission to publish. The ethical framework studied here is aspirational, not a global law. Apply the learning to a fictional case or safe information available with permission: a specific description, a suitable measure, a justified comparison, a conclusion proportionate to evidence and respect for people.

Review questions

Test your understanding with these questions. Marking and saved results are available with a free account.

Question 1Which statement only describes behavior in a specific situation? Husam stopped responding because he forgot the idea. Husam avoided responding to protect his image. Husam sent two questions in ten minutes. Husam thought the question was below his ability.
Question 2Which is the most accurate correction to this reasoning? A historical emphasis on choice does not establish every effect of choice. The idea’s historical influence removes the need to test its effect. Attention to personal experience rules out studying performance. Distinguishing humanism from behaviorism makes the claim true.
Question 3Which level best fits the main focus of this question? Biological: differences in nervous-system activity. Cognitive: how information is stored in memory. Structuralist: reported components of conscious experience. Social: rules governing turn-taking among participants.
Question 4Which alternative asks a testable question about initiative? Those who initiate care; those who do not are hiding their interest. Does the number of initiatives differ when response time is clear? Genuine initiative is something no one can observe. Rename absence of initiative as weakness in initiative.
Question 5To measure messages requesting clarification of instructions within ten minutes, which counting rule fits best? Count each message from a participant who seems keen to understand. Count each message asking what is required, excluding answers and technical questions. Count each message containing a question mark, including submission questions. Count learners who wrote at least once during the period.
Question 6Which information directly answers the team’s chosen question? The number of steps the participant completed correctly. An observer’s estimate of confidence from writing speed. The participant’s answer about how clear the wording seemed to them. The number of times the instructions page was opened.
Question 7What are the correct response proportions out of the available opportunities? Mira 40%; Tariq 30%. Mira 20%; Tariq 30%. Mira 40%; Tariq 60%. Mira 60%; Tariq 70%.
Question 8Which evaluation of the measure is supported by this information? Agreement between counters establishes word count as an explanation-quality measure. The count is unreliable because explanation quality was not tested. Explanation quality can be judged from the count without reading the content. Recording is consistent, but its validity for explanation quality is not established.
Question 9Which statement distinguishes sampling from random assignment within a study? Sampling determines who participates; random assignment determines the condition a participant receives. Sampling determines conditions; random assignment selects participants from the population. Sampling and random assignment are two names for the same procedure. Random assignment within any sample is enough to represent the population.
Question 10What are the mean and median of these times? Mean 6 minutes; median 7 minutes. Mean 7 minutes; median 8 minutes. Mean 7 minutes; median 4 minutes. Mean 7 minutes; median 6 minutes.
Question 11Which conclusion is supported by this information? Requesting hints caused the longer solution times. More requested hints were associated with longer times; the cause is unresolved. The relationship is negative because slower solving is undesirable. Longer solution times caused participants to request hints.
Question 12Which change directly addresses the stated confound? Increase sample size while retaining the different working times. Change the graph of the means without changing procedures. Provide the same working time and task in both formatting groups. Remove the slowest participants in each group after seeing results.
Question 13You read a research abstract that does not describe sampling. Which statement accurately describes the limits of your access? The study used a random sample because the text calls it a study. The researchers did not describe sampling anywhere in the article. I reviewed every recruitment procedure and found them appropriate. The abstract I read does not state how participants were selected.
Question 14How should the proposed generalization be handled? Accept it because random assignment in the workshop represents all adults. Limit it to the tested context and seek evidence for other groups and languages. Reject the finding within the workshop because the sample excludes some people. Accept generalization as long as the task has the same name in each language.
Question 15Which step is most directly relevant to understanding the stated difference? Treat the second result as sufficient evidence that the first team cheated. Replace the new data with another calculation on the original data. Examine equivalence of the translated instructions and outcome measurement. Prefer the earlier finding solely because it came first.
Question 16Which action corrects how the data are described and used in this case? Describe them as coded, restrict access, and explain withdrawal limits before participation. Describe them as anonymous while the facilitator keeps the linkage key as a precaution. Publish the entire coded table because agreeing to answer permits any use. Delete the key immediately but promise withdrawal of any individual answer at any time.