Investigative question worksheet with answer key
8 problems on investigative question, ready to print for class or homework. The answer key prints on its own page, so hand out everything before it.
Name: ______________________________ Date: ______________ Period: ______
Problem 1
A counselor at a high school with 1,480 students says, "I want to know whether our students get enough sleep." The district wellness policy names 8 hours on a school night as its target. The counselor can pull 120 students out of class for one period, and she plans to select those 120 at random from the school's enrollment list.
(a) Explain why her sentence is not yet an answerable investigative question, and say which part of it is already fine. (b) Write the collection component, the part phrased in terms of the variable of interest. (c) Write the analysis component as a confidence interval question, and name the parameter. (d) Write the analysis component as a hypothesis test question, and name the parameter and the direction. (e) Write the conclusion component, and find what fraction of the school is in the sample.
Problem 2
For each situation, name the way of collecting data it calls for, choosing from census, survey, observational study, and experiment, and give the one feature of the situation that forces your choice. Use the most specific label that fits, since a survey is itself one kind of observational study.
(a) A district needs to know exactly how many of its 62 school buses fail this month's brake inspection. (b) A cafeteria manager wants to know whether opening a second, clearly labeled serving line shortens the mean time students wait for lunch. She can open the second line on some days and leave it closed on others, and she decides which days. (c) A student council has one homeroom period, enough to reach about 200 of the school's 900 students, and wants to estimate the proportion who would use a late bus. (d) A researcher wants to compare rates of a sleep disorder between adults who worked night shifts for at least five years and adults who never did. Nobody can be put on night shifts for five years for a study. (e) A safety team registers 900 newly licensed teen drivers, records which of them chose to take an optional advanced driving course, and follows all 900 forward for three years, counting crashes.
(f) Situations (d) and (e) are both observational studies, one prospective and one retrospective. Say which is which and how you can tell.
Problem 3
A city news site with about 210,000 daily readers publishes an article about a proposed transit levy and puts a poll box at the bottom: "Do you support the levy? Yes / No." Over one day, 4,812 readers answer and 3,051 of them click Yes. The next morning the site's headline reads "City backs transit levy, 63% say yes."
(a) Find the proportion who clicked Yes and the share of the daily readership that answered at all. (b) Which of these three investigative questions can this data set answer? (i) What proportion of the readers who answered the poll support the levy? (ii) What proportion of the city's residents support the levy? (iii) Did reading the article change readers' support for the levy? (c) Name the kind of sample this poll produced, and say what it does to the estimate. (d) The editor says the poll would be trustworthy if 50,000 readers had answered instead of 4,812. Respond. (e) Write an investigative question about the city that the site could actually answer, and say what it would have to collect.
Problem 4
Each investigative question below states a goal. For each, say whether the analysis component points to a confidence interval or a hypothesis test, name the parameter the question is about, and, where a test is called for, name the direction of the alternative.
(a) A quality manager wants a range of plausible values for the mean weight, in grams, of the almonds her machine puts in a bag. (b) A shipping contract promises that at most 5% of packages arrive late. The operations lead wants to know whether his late rate is above that. (c) A researcher wants to know whether the proportion of adults who compost differs between two counties. (d) A principal wants to know whether favorite sport, with three categories, is associated with class year, with four categories, among the students at her school.
(e) One of the four cannot be answered with an interval estimate in this course. Say which, and why.
Problem 5
A coach says, "I want to know whether our new warm-up works." His squad is 48 players. He has last season's warm-up and this season's new one, and he can decide which players do which.
(a) Name two different measurable responses the word works could mean here, and say why the study cannot leave both open. (b) He settles on 20-meter sprint time in seconds, recorded right after the warm-up. Name the explanatory variable and its levels. (c) Which way of collecting data does his question call for, and what one feature of the situation makes it available? (d) Name the experimental units and give the group sizes for an even split. (e) The athletic director wants to roll the new warm-up out to all 14 schools in the district. Explain why this study does not answer his question, and name the change that would.
Problem 6
A researcher is handed a data file on the 1,200 students at one high school. It records each student's GPA and whether they belong to each of 30 school clubs. He had no question in mind when the file arrived. He compares mean GPA between members and non-members for each club in turn, 30 comparisons, using a 5% significance level each time. Two come out significant. He writes up one of them as "Students in debate club have higher GPAs" and presents it as the question the study set out to answer.
(a) Suppose none of the 30 clubs is really related to GPA. About how many of the 30 comparisons would you expect to come out significant anyway? (b) Which requirement of an investigative question did he break? (c) His defense is that the finding is real because the data are real and the arithmetic is correct. Answer him. (d) Even if debate club membership and GPA really are associated in this file, name the second thing his sentence claims that the data cannot support. (e) Say what he should do with the debate club finding.
Problem 7
A high school's records already contain, for every one of this year's 1,140 students: class year, whether the student rides a school bus, days absent, and final GPA. Of the 1,140, 486 ride a bus. For each proposed investigative question, say whether the existing records answer it, and if not, say exactly what would have to be collected.
(a) What is the mean number of days absent for this year's students at this school? (b) Are days absent and GPA associated at this school? (c) Does riding the bus cause students to miss more school? (d) How many hours of sleep do this school's students get on a school night? (e) Do students at this school miss more school than students statewide?
(f) Find the proportion of the school that rides a bus, and say whether that proportion is a parameter or a statistic.
Problem 8
A city transportation office says, "We need to figure out whether the bike share is working." Two facts are on hand: the system has 12,000 registered riders on a list the office maintains, and 240 docking stations that are comparable in size and location type. The office can afford to contact 400 riders, and it can afford to build covered shelters at 45 stations.
(a) Turn the wondering into an answerable descriptive question, giving all three components, and name the way of collecting data it calls for. (b) Turn the same wondering into an answerable causal question about shelters, giving all three components, and name the way of collecting data it calls for. (c) For (b), describe the randomization and the group sizes, using 90 of the 240 stations. (d) State what each of the two studies licenses the office to conclude, and about whom. (e) Why can neither study, on its own, tell the office whether the bike share is working?
From statslearn.org/worksheets/investigative-questions-worksheet. Free to print and copy for classroom use.
Investigative question: answer key
(a) It names no measurable variable, since "enough sleep" is a judgment rather than something you can record; the population, the 1,480 students at this school, is already well defined. (b) "On a typical school night, how many hours of sleep do students at this school get?" (c) "What is the mean number of hours of sleep on a school night for the 1,480 students at this school, and what range of values is plausible for it?" The parameter is , the population mean. (d) "Is the mean number of hours of sleep on a school night for students at this school less than 8 hours?" Parameter , direction less than. (e) The conclusion describes the mean sleep of all 1,480 students with no causal claim, and it reaches them because the 120 were randomly selected; , about 8.1% of the school.
(a) Census: the 62 buses are the whole population and the district needs the exact count. (b) Experiment: the manager imposes the condition by deciding which days the second line opens, and the units are days; to earn a causal claim she has to assign those days at random rather than by hand, since otherwise busy days can be the days she chooses. (c) Survey: the data are what people report, and only about 200 of 900 can be reached, , about 22.2%; whether it reaches all 900 depends on how those 200 are chosen, which the stem does not say. (d) Observational study: night-shift history cannot be assigned. (e) Observational study: the drivers chose the course themselves. (f) (d) is retrospective, reading work history that already happened; (e) is prospective, following the 900 drivers forward for three years.
(a) , 63.4% of respondents, and , about 2.3% of the daily readership. (b) Only (i), and exactly, because the 4,812 are all of the respondents rather than a sample of them. (ii) fails because they selected themselves and are not connected to the city, and (iii) fails because nothing was assigned and nothing was measured beforehand. (c) A voluntary response sample, which biases the estimate in a consistent direction because people with strong feelings click more often. (d) No. A bigger voluntary response poll estimates the same wrong quantity more precisely; size reduces variability, not bias. (e) "What proportion of the city's registered voters support the transit levy, and what range of values is plausible for it?" Collect it by contacting a random sample drawn from the voter registration list with identical wording, and stop the conclusion at registered voters.
(a) Confidence interval; parameter , the mean bag weight in grams; no direction. (b) Hypothesis test; parameter , the proportion of packages that arrive late; direction greater than, with the alternative . (c) Hypothesis test; parameter , the difference between the two counties' proportions; direction not equal. (d) Hypothesis test; the claim is that favorite sport and class year are associated, or not independent, and there is no single numeric parameter. (e) (d), because 3 sports crossed with 4 class years leaves no single number to put an interval around, while (c)'s difference between two proportions is a number and could be estimated instead of tested.
(a) Any two measurable responses, for example 20-meter sprint time in seconds and hamstring injuries per season. Both cannot stay open because the question must be fixed before the data arrive, and they need different studies: a sprint time is one afternoon's measurement on every player, while injuries need a season and far more players. (b) Explanatory variable: warm-up routine, with two levels, the new routine and last season's, which are the treatments. (c) An experiment, available because the coach can decide which players do which warm-up. (d) The 48 players are the units, per routine, assigned by random number generator rather than by preference. (e) The 48 are one squad and were not randomly selected, so the causal finding covers them and players like them; random selection of players or squads from across the 14 schools is what would extend it to the district.
(a) comparisons, so about 1 or 2, which is what he got. (b) The investigative question must be fixed before the data are analyzed, and he had none: the data chose the question, so the result that chose it cannot also be evidence for it. (c) Real data and correct arithmetic are not the issue; the issue is that he ran 30 comparisons and reported the ones that came out well, and about 1.5 of 30 come out well by chance alone. (d) The population. The file is one high school's 1,200 students with no random selection, so the claim reaches those students and students like them, not students in general. (e) Write it down as an investigative question in advance, naming the variable, parameter, and direction, then collect fresh data to answer that one question; and note that the answer will still be an association, since debate club membership is never assigned.
(a) Yes, exactly: days absent is recorded for all 1,140, so the mean is a parameter with nothing to infer. (b) Yes, as an association, since both variables are in the records; the records cannot say which one moves the other. (c) No. Bus riding was never assigned, so only an experiment could support cause, and the records support an association at most. (d) No. Sleep is not a recorded variable, so it has to be collected, by surveying a random sample of the school's students about a stated school night. (e) No. The records hold no student from any other school, so a statewide comparison needs absence data from a random sample of the state's students or a published statewide figure. (f) , about 42.6%, and it is a parameter, because the records cover every student at the school.
(a) Collection: how many bike share trips in the past 30 days a registered rider took with a car available to them at the time, which is something a rider knows, unlike a guess about what they would otherwise have done. Analysis: the mean number of such trips for the 12,000 registered riders, estimated within a range, so the parameter is . Conclusion: a description of all 12,000, no cause, valid only if the 400 are randomly selected. Method: a survey of , about 3.3% of the list. (b) Collection: trips started per station over 8 weeks, and whether the station has a shelter. Analysis: is the mean trips per station greater for sheltered stations, so the parameter is the difference in treatment means and the direction is greater than. Conclusion: causal, and reaching the city's 240 comparable stations. Method: an experiment. (c) Draw 90 of the 240 at random ignoring repeats, , then draw 45 of those 90 at random for shelters, leaving unchanged; compare mean trips started per station over the same 8 weeks, with stations as the units. (d) The survey generalizes a description to all 12,000 registered riders; the experiment supports a causal claim and generalizes it to the 240 comparable stations, because it has both randomizations. (e) Working is not a variable, so the office has to name the measure it will use before collecting, and that choice is what selects the method.
Every step is worked at statslearn.org/practice/investigative-questions-practice. Free to print and copy for classroom use.
The key gives final answers only. The full worked solution for each problem is on the matching practice set, and every other sheet is on the worksheets page.
More worksheets from this unit
- Identifying bias worksheet with answer key
- Boxplot and five-number summary worksheet with answer key
- Categorical data worksheet with answer key
- Comparing two distributions worksheet with answer key
- Describing and comparing distributions worksheet
- Experimental design worksheet with answer key
- Dotplot, stemplot, and histogram worksheet with answer key
- Quartiles, IQR, and outliers worksheet with answer key
- Sampling methods worksheet with answer key
- Scope of inference worksheet with answer key
- Standard deviation and variance worksheet with answer key
- Choosing mean vs median and SD vs IQR worksheet with answers
- Classifying variables worksheet with answer key