Turning Field Data Into an Argument
Use field data to support a geographical argument. Choose a graph or map that fits the question, report the sample and method, and explain the uncertainty and limits of the conclusion.
What a learner can do afterwards
- Choose a graph or map form that fits the data type and defend the choice
- State a conclusion with the uncertainty attached, rather than as a flat fact
- Say what a small or unrepresentative sample lets you claim and what it does not
1 · Read
Use the planned question and sample to choose a display. A histogram shows counts in numerical intervals; use equal-width intervals here. A time series puts time on the horizontal axis and measured values on the vertical axis. Bars compare categories. A scatter plot pairs two measurements from each observation. A map can show observation locations. Explain how the display helps answer the question.
On a scatter plot, a rising pattern shows a positive relationship and a falling pattern a negative one. Points close to a sloping straight line show a strong linear relationship. More spread around that pattern means a weaker linear relationship. A curve can also show a strong relationship, but a straight line may fit it poorly. Do not add a rising trend line when no clear rising pattern appears.
An association, where measurements vary together, does not prove cause. A third factor could affect both. For example, weather could affect ice cream sales and swimming activity. Report the observed group and conditions: if more than half the respondents chose shade, say 'Most respondents chose shade that afternoon.' The replies alone do not show what caused their choices or what people elsewhere prefer.
Match conclusions to the sample's coverage. Counts at one shop on one afternoon describe that entrance during the counting period, not the whole town or season. Questionnaire replies describe the respondents; they may not represent every visitor. To support a wider claim, plan observations across relevant places, times and people. More records from the same narrow group do not remove its selection bias.
Report the number of observations or replies, where and when they were collected, and how they were selected. Missing sample size makes a result harder to judge; it does not by itself prove the conclusion false. Volunteer replies on a campaign forum may overrepresent supporters, even if many people reply. State what the records show and which wider claims the method cannot support.
Choose a display that answers the question. Report the sample and method, describe the observed pattern, and keep the conclusion within the data's limits.
Field-data interpretation teaching, questions and flashcards written for this dataset. Numerical teaching examples are illustrative.
The readings and flashcards were written for this dataset. The linked video, Statistical questions, is by Khan Academy.
More to read
Start with what the display must answer
One notebook might contain stone lengths, counts of visitors at different times and the types of litter found. A useful display depends on which question you are answering.
To show the distribution of measured lengths, a histogram counts values within numerical intervals. With equal-width intervals, bar heights show the counts. State the units and boundaries. To show change through the day, a time series keeps observations in time order. Compare counts collected over equal durations.
Separate category bars can compare litter types. A scatter plot needs paired measurements from the same observations, such as each stone's length and width. A location map answers a different question: where were the observations made? Explain the purpose of the chosen display.
Read direction, shape and spread
A scatter plot places each pair of measurements at one point. First look for a direction: do vertical values tend to increase or decrease as horizontal values increase? Next inspect the shape. A curved pattern may describe a relationship that a straight line misses.
For a sloping straight-line pattern, closer clustering means a stronger linear relationship. Greater scatter means a weaker linear relationship, not automatically no relationship. A rising line should not be added when the observations show no clear rising pattern. Describe the pattern visible in the measured range and avoid presenting the line as proof of cause.
Separate the observation from its explanation
Imagine that a survey finds more litter on busier streets. That association is an observation. Saying that pedestrians caused the extra litter is an explanation that needs further evidence. Food outlets, cleaning times and other differences are possible factors to investigate, not causes already established by the pattern.
A conclusion should state the observed result and the records behind it. Then distinguish the possible explanations from what has been tested. If most people in a set of replies preferred shade, report that preference for those respondents and conditions. The replies alone do not identify why they preferred it. A cautious causal claim needs reasons and evidence, not just cautious wording.
Check who could enter the records
A survey at one market entrance on Saturday morning reaches a particular group. Replies come only from people who were there and agreed to answer. The result directly describes their replies. It may not represent other entrances, later visitors or residents who do not use the market.
Report the reply count, place, time and selection method. A missing count makes the result harder to judge, but selection matters as well as size. More volunteers from the same forum or entrance do not automatically restore missing groups. Broader conclusions need a collection plan that reaches the relevant people, places and times. Even collecting from the entire target population would not prevent recording mistakes or misleading questions.
Explore more
2 · Watch
به انگلیسیTake it off screen
Where it sits
Learn first
Optional connections
This opens up
Nothing builds on it yet.
Where this leads
Jobs that lean on this skill. Follow one to see everything it is built on.
8 questions wait behind this lesson, each with its answer explained. Every answer feeds the sky: stars light as they are learned, and dim when it is time to come back.