Random Sampling vs Random Assignment Flashcards
Practice random sampling, random assignment, and scope of inference with short study scenarios, claim corrections, and explained answers.
Informazioni su questo mazzo
Practice random sampling vs random assignment with 48 original flashcards for introductory statistics and research-methods learners. Trace who entered a study, how conditions were assigned, and which conclusion the design supports.
The cards map procedures to design roles, short fictional studies to one bounded scope decision, and flawed claims to precise corrections. Six foundations lead into the four sampling-and-assignment combinations. Further scenarios are interleaved with claim repair, sampling-frame and response limits, assignment units, and evidence limits. Answers start with the conclusion and briefly explain why.
Use an introductory design-based inference convention: random sampling supports generalization to the population covered by the selection procedure, while random assignment supports causal comparison under sound conduct and appropriate analysis. A design alone does not show that an effect exists. Volunteer experiments do not automatically represent a wider population; randomized groups need not match exactly. Low response alone does not prove bias, and an average effect need not apply to every person.
This focused deck excludes calculations, p-values, confidence-interval construction, power, a full sampling-method taxonomy, advanced observational causal methods, and full-course or exam coverage. It avoids mechanical reverse definitions and asks one inference decision at a time, so practice stays focused on interpretation. Observational causal inference with additional assumptions is acknowledged, not taught here.
Questions, answers, scenarios, order, and metadata were independently authored with AI assistance from common statistical facts. Fact checks used Penn State’s collecting-data overview, sampling definitions, nonresponse discussion, and experimental-design principles. The broader causal-inference boundary is informed by Hernán and Robins. No source exercises, examination questions, source passages, or figures were copied. The original AI-generated cover is a decorative sampling-and-grouping still life, not a statistical diagram.
Original text, organization, metadata, and generated cover are offered under CC0 1.0 to the extent applicable rights exist. This dedication does not cover cited sources or third-party rights. This is an independent study resource, not affiliated with or endorsed by the cited institutions, authors, or any course or examination provider.
Carte in questo mazzo
Carta 1
Domanda
A study draws names from a complete membership list, then uses fresh coin flips to allocate the selected members to two tasks. Which step is random assignment?
Risposta
The coin flips that allocate members to tasks.
Sampling chooses who enters a study; assignment chooses which condition each experimental unit receives.
Carta 2
Domanda
A survey aims to describe all city residents but draws names from a library membership list. What is its sampling frame?
Risposta
The library membership list.
The target population is all city residents. Residents outside that list cannot be selected through this frame.
Carta 3
Domanda
Researchers give each participant one of two instruction sheets and compare task accuracy. What makes this an experiment?
Risposta
The researchers impose the instruction-sheet condition.
An observational study records conditions without assigning them. An experiment is randomized only if a chance process assigns its conditions.
Carta 4
Domanda
In an experiment comparing packing methods, what defines the experimental unit?
Risposta
The smallest unit that receives its own packing-method assignment.
It could be a parcel or a whole work team; the assignment procedure determines the unit.
Carta 5
Domanda
People who use a calendar app report fewer missed appointments. What does an association mean here?
Risposta
Calendar-app use and missed appointments vary together in the observed data.
That pattern alone does not show that the app causes fewer missed appointments. Other differences between users and nonusers may matter.
Carta 6
Domanda
A sound randomized experiment compares two typing tutorials. Before its results are available, what does random assignment establish?
Risposta
A basis for investigating a causal effect of the tutorials.
It does not establish that an effect exists or which tutorial performs better. Those conclusions require outcome data and appropriate analysis.
Carta 7
Domanda
A random sample of a college’s current students all take part. They are randomly assigned to two note layouts, under otherwise comparable conditions. For which population can this design investigate a causal effect on recall?
Risposta
The college’s current students.
Random sampling supports that population scope; random assignment supports the causal comparison. Sound conduct and appropriate analysis are still needed.
Carta 8
Domanda
A simple random sample of a town’s households all report their usual parcel-delivery service and satisfaction. No service is assigned. What kind of relationship can this design estimate for the town’s households?
Risposta
An association between usual delivery service and satisfaction.
Random sampling supports population inference. Self-selected services may differ in customer mix, so this design alone does not establish a service’s causal effect.
Carta 9
Domanda
Museum volunteers are randomly assigned to two exhibit-label formats. A well-run comparison finds convincing evidence of a difference in quiz scores. For whom is a causal interpretation directly supported?
Risposta
The participating volunteers’ comparison of the two formats.
Random assignment supports a causal interpretation for these participants. It does not by itself justify extending the result to all museum visitors.
Carta 10
Domanda
The first shoppers leaving one store are asked which shopping-list app they chose and how many items they forgot. What conclusion does this design support by itself?
Risposta
A description of the app-use association among the surveyed shoppers.
Neither chance-based selection from a stated population nor random assignment occurred. Population generalization and a causal app effect are not secured by this design.
Carta 11
Domanda
Repair the claim: “Our simple random sample must match the population’s proportions exactly.”
Risposta
A simple random sample can differ from the population by chance.
Random selection provides a basis for inference and assessing sampling uncertainty; it does not guarantee an exactly representative realized sample.
Carta 12
Domanda
A school randomly surveys families from its online-portal list, and all selected families reply. Families without portal accounts are absent. Why is a claim about all school families unsupported by this sampling procedure alone?
Risposta
Families without portal accounts had no chance of selection.
The frame undercovers the target population. Complete response among selected families does not repair that coverage gap.
Carta 13
Domanda
Twelve tutorial groups are randomly assigned to use one of two discussion formats; every student follows their group’s format. What was randomly assigned?
Risposta
The twelve tutorial groups.
Each group is an experimental unit. Students within a group were not assigned independently.
Carta 14
Domanda
A company randomly assigns recruited testers to two search interfaces, but reports no task results. Can it yet say that the new interface reduces search time?
Risposta
No. The effect’s existence and direction remain unknown.
Random assignment supports a causal comparison once outcomes are analyzed; it is not evidence that one interface works better.
Carta 15
Domanda
Repair the claim: “We gave every available visitor a random ID number, so we obtained a random sample of all visitors.”
Risposta
Random labels do not make recruitment random.
Selection must use a chance-based procedure that actually determines who enters. Labeling people after convenience recruitment does not change how they were selected.
Carta 16
Domanda
All workshop attendees use a new checklist and make fewer mistakes on a second attempt than on their first. Why can this comparison alone not isolate the checklist’s effect?
Risposta
Practice is an alternative explanation for the improvement.
Without a suitable comparison condition, the checklist change is entangled with repeating the task. A before-and-after difference alone does not isolate its cause.
Carta 17
Domanda
A random sample receives a transport survey, but only 15% respond. Does that response rate prove the estimate is biased?
Risposta
No. It raises a nonresponse concern, not proof of bias.
Bias depends on how response is related to the survey outcome. The respondents may differ from nonrespondents in ways that matter.
Carta 18
Domanda
An archive randomly samples documents from its complete collection, then randomly assigns each to one of two indexing workflows. Both are applied as planned. Can the design investigate a causal difference in indexing time for that collection?
Risposta
Yes, with analysis appropriate to the sampling and assignment.
Selection supports inference to the collection; assignment supports comparison of the workflows’ effects. No difference is established until results are evaluated.
Carta 19
Domanda
Repair the claim: “Observational evidence can never support any causal inference.”
Risposta
Observational causal inference can be possible with additional assumptions and methods.
This deck uses an introductory design-based rule: an observed association alone does not establish causation. Advanced causal identification is outside its scope.
Carta 20
Domanda
A convenience group of puzzle-club members is randomly assigned to receive hints or no hints. The trial runs as intended with complete outcome data. Can this design investigate whether hints affect these participants’ completion times?
Risposta
Yes. Random assignment supports a causal comparison for these participants.
Convenience recruitment limits population generalization; it does not erase the experiment’s random assignment.
Carta 21
Domanda
A researcher assigns arriving participants alternately to audio or written instructions, starting with audio. Is this random assignment?
Risposta
No. Alternation is a predictable allocation rule.
A chance process must determine assignment. Arrival order may be related to participant characteristics, even if the two groups end up equally large.
Carta 22
Domanda
A sound randomized experiment supports an effect of a reminder on submitting one form on time. Why does that not establish improved long-term organization?
Risposta
Long-term organization was not the measured outcome.
Causal evidence about one form’s submission does not automatically extend to other behaviors or longer time periods.
Carta 23
Domanda
A random sample of all local sports-club members fully answers a survey about chosen practice times and enjoyment. What prevents interpreting an enjoyment difference as the causal effect of practice time?
Risposta
Members chose their practice times.
Schedule flexibility or other differences may affect both that choice and enjoyment. Random sampling supports inference about the association in club members, but does not randomize practice time.
Carta 24
Domanda
Repair the claim: “We randomized 200 online volunteers to two tasks, so they represent all adults.”
Risposta
Random assignment does not make the volunteers representative of all adults.
It governs allocation within the recruited group. Extending the causal comparison to all adults needs a separate justification for that population scope.
Carta 25
Domanda
A city posts an open survey link, then randomly chooses 300 completed responses to analyze. Does the second step make this a random sample of all city residents?
Risposta
No. It is a random subsample of the submitted responses.
Residents first selected themselves into the response pool. Randomly reducing that pool does not give all residents a known selection chance.
Carta 26
Domanda
Recruited learners are grouped by prior experience, then randomly assigned to two tutorials within each group. What role do the experience groups play?
Risposta
They are blocks in the experiment.
Blocking organizes treatment assignment to account for a relevant preexisting difference. It does not turn the recruited learners into a random population sample.
Carta 27
Domanda
A designer asks friends about their chosen keyboard layout and typing comfort. Can the design alone establish that changing layouts would change comfort?
Risposta
No. It records an association in the interviewed friends.
Layout choice may reflect prior experience or existing comfort. Neither convenience recruitment nor observation of chosen layouts supplies random assignment.
Carta 28
Domanda
A company randomly surveys its complete current employee roster with full response. Can the sampling procedure alone support a claim about people it will hire next year?
Risposta
No. Future hires are outside the sampled population.
A complete frame for current employees does not supply a probability sample of future employees.
Carta 29
Domanda
A sound randomized experiment supports a positive average effect of a task guide. Does that show the guide helps every participant?
Risposta
No. A positive average effect does not establish a benefit for each person.
Individual effects can differ, and the group comparison does not reveal every person’s individual effect.
Carta 30
Domanda
A factory randomly samples packages from today’s complete output and randomly assigns each to one of two cushioning methods. A sound analysis finds lower average damage with one method. How far does the sampling support that causal finding?
Risposta
To today’s output represented by the sampling procedure.
The randomized comparison supports a cushioning effect under the tested conditions. Other factories, future output, or different shipping conditions need additional justification.
Carta 31
Domanda
Repair the claim: “In our random survey, people who bought premium pens wrote faster, so the pens caused faster writing.”
Risposta
Premium-pen ownership was associated with writing speed.
Owners may differ in practice or other factors. Random selection of people does not randomly assign pen ownership or remove that confounding.
Carta 32
Domanda
Visitors who volunteer at a science fair are randomly assigned to two assembly diagrams. Complete results and a sound analysis support a difference in assembly errors. Can the finding automatically describe every science-fair visitor?
Risposta
No. Volunteer recruitment does not justify that population extension.
The randomized comparison supports a causal diagram effect for the participants under the study conditions; generalizing further needs additional evidence or assumptions.
Carta 33
Domanda
Repair the claim: “A randomly sampled survey is automatically a randomized experiment.”
Risposta
A random sample determines who is observed; a randomized experiment also randomly assigns conditions.
Surveying existing behavior without imposing conditions remains observational, even when participant selection is random.
Carta 34
Domanda
Twenty work teams are randomly assigned to two training formats. Outcomes are recorded for 200 workers. Why should the analysis account for teams?
Risposta
Workers in the same team share an assignment and may have related outcomes.
The 200 measurements are not 200 independent treatment assignments. Analysis must reflect the group-level randomization.
Carta 35
Domanda
A random sample of all active accounts on one platform fully reports use of its optional folder feature and ease of finding files. What population relationship can the study investigate?
Risposta
The association between folder use and reported ease among active accounts on that platform.
The random sample supports this population scope. Users selected their own feature use, so a causal feature effect is not established by this design alone.
Carta 36
Domanda
Repair the claim: “Our convenience survey has 50,000 responses, so selection bias is impossible.”
Risposta
A large convenience sample can still have selection bias.
More responses do not automatically correct systematic differences between those included and the target population.
Carta 37
Domanda
A researcher surveys willing members of one art class about their chosen sketching tools and confidence. What population conclusion is justified by recruitment alone?
Risposta
No automatic generalization beyond the participating students.
Willing students in one class were not randomly selected from all art students. Any wider population claim needs a separate justification.
Carta 38
Domanda
Participants are randomly assigned to two editing tools, but many frustrated users of one tool leave before rating it. Why is comparing only the remaining ratings a concern?
Risposta
Outcome-related dropout can make the retained groups incomparable.
Initial random assignment does not guarantee that the people with observed ratings still form a fair comparison. Missing outcomes need attention.
Carta 39
Domanda
A sampler chooses one of two fixed halves of a 20-person roster by a fair coin flip. Everyone has a 50% inclusion chance. Is this a simple random sample of 10 people?
Risposta
No. Only two possible half-roster samples can be selected.
A fixed-size simple random sample gives every possible subset of that size equal selection probability. Equal individual inclusion chances alone are insufficient.
Carta 40
Domanda
Repair the claim: “We matched users and nonusers of a planning app on age, so any productivity difference must be caused by the app.”
Risposta
Matching on age alone does not isolate the app’s causal effect.
Other relevant differences may remain. Matching one measured characteristic is not random assignment and does not supply every assumption needed for observational causal inference.
Carta 41
Domanda
A correctly randomized experiment happens to place more experienced participants in one group. Does that imbalance alone prove assignment was not random?
Risposta
No. Chance imbalance is possible after valid random assignment.
Randomization balances characteristics in expectation over repeated assignments, not exactly in every realized experiment.
Carta 42
Domanda
A publisher randomly samples titles from its complete current catalog and randomly assigns their online listings to two thumbnail styles. Conduct is sound and analysis respects the design. Can it investigate a causal style effect on clicks for current catalog titles?
Risposta
Yes, for current catalog titles under the tested conditions.
Sampling and assignment support different parts of that scope. Extending to future titles or claiming a positive effect requires additional evidence.
Carta 43
Domanda
Repair the claim: “Random assignment guarantees an unbiased result even if one group is measured with a different, faulty timer.”
Risposta
Random assignment does not fix unequal measurement error.
A faulty timer used only for one group can distort the comparison. Sound outcome measurement remains necessary after randomization.
Carta 44
Domanda
A simple random sample of all district teachers fully reports their chosen lesson-planning method and preparation time. Does sampling alone justify saying one method saves time?
Risposta
No. It supports estimating the population association, not isolating a method’s causal effect.
Teachers chose their methods; experience, workload, or other differences may explain the time comparison.
Carta 45
Domanda
A town samples residents only from a property-owner register. Would surveying every listed owner remove the coverage problem for a claim about all adult residents?
Risposta
No. Adults absent from the register would still be excluded.
Complete coverage of the frame is not complete coverage of the target population; renters or other unlisted residents may matter.
Carta 46
Domanda
A volunteer community orchestra randomly assigns musicians to two rehearsal aids. Everyone follows the assigned aid, and outcomes are measured consistently. What causal question can the design investigate directly?
Risposta
Whether the aids affect the participating musicians’ measured performance under these conditions.
Random assignment supports this comparison. The volunteer orchestra was not a random sample of all musicians.
Carta 47
Domanda
Repair the claim: “Random assignment makes sharing the experimental hint sheet with the no-hint group irrelevant.”
Risposta
Sharing can change the treatment comparison.
The assigned groups may no longer differ in access as intended, and one participant’s assignment may affect another’s outcome. Randomization alone does not prevent this problem.
Carta 48
Domanda
A workshop assigns Monday volunteers to video instructions and Tuesday volunteers to written instructions. No random sampling occurs. Can a task-score difference isolate the instruction format’s causal effect?
Risposta
No. Instruction format is entangled with day and potentially different volunteers.
The study imposes conditions, but does not randomly assign them. The comparison alone cannot isolate the format’s effect.
48 carte
Random Sampling vs Random Assignment Flashcards
Nibomo si apre e puoi iniziare a studiare.