Framing a geographical question and designing the enquiry

Overview

Null hypothesis, sequence of stages in a geographical enquiry and difference between a geographical topic and a geographical issue.

What you need to know

Answer each one before you reveal it.

Questions from this subtopic

Each one with the model answer underneath.

1. 1

   What makes a good geographical question for investigation?

   Model answer

   It has to be genuinely geographical, so it concerns place, space, environment or the relationship between people and their surroundings, and it must be answerable with data you can actually collect in the time and area available. It should be narrow enough to be tested - about one process, in one defined place, over one defined period - and it should connect to established geographical ideas so the findings can be discussed rather than merely reported. A question with no possible answer other than yes, and one so broad that no realistic data set could settle it, are both unusable.

2. 2

   What is a hypothesis, and how does it differ from a question?

   Model answer

   A hypothesis is a statement, not a question: it predicts a relationship or a difference that the data can then support or fail to support - channel width increases with distance downstream, or pedestrian density decreases with distance from the central business district. A question asks what is happening; a hypothesis commits to an expected answer in advance and can be tested statistically because it can be paired with a null hypothesis. Both are legitimate starting points, but a hypothesis makes the design of the data collection more obvious.

3. 3

   Give an example of a testable hypothesis for a river investigation and explain what makes it testable.

   Model answer

   Channel cross-sectional area increases with distance downstream from the source. It is testable because both variables are measurable with standard equipment, both produce numbers, the sites can be sampled systematically along the channel, and the resulting paired data can be plotted on a scatter graph and tested with Spearman's rank correlation. It also predicts a direction of change, so the data can contradict it. A hypothesis such as the river is interesting downstream fails all of those tests.

4. 4

   What is a null hypothesis, and why does statistical testing need one?

   Model answer

   The null hypothesis is the statement that there is no significant relationship, difference or pattern, and that anything appearing in the data is the product of chance. Statistical testing needs it because a test cannot prove a positive directly; what it can do is calculate how likely the observed result would be if nothing were going on, and reject the null hypothesis if that likelihood is small enough. Rejecting the null hypothesis is what allows you to accept the alternative hypothesis that a real relationship exists.

5. 5

   Why must a hypothesis be phrased in measurable terms?

   Model answer

   Because every word of it has to become a decision about what to record. A hypothesis stating that the environment improves towards the edge of town cannot be tested until improves is defined as something countable, such as an environmental quality score, and edge of town is defined as a measured distance. Vague wording produces data that do not match the claim, arguments about whether the hypothesis was supported, and no way of choosing a statistical technique.

6. 6

   What is the sequence of stages in a geographical enquiry?

   Model answer

   Identify the issue and frame a question or hypothesis; decide what data are needed and select the sources and methods, including the sampling strategy, and assess the risks; collect the primary and secondary data in the field and from published sources; organise and present the evidence as maps, diagrams and graphs; process and analyse it with appropriate statistical techniques; interpret the results and draw conclusions that answer the original question; and finally evaluate the methods, the reliability of the evidence and the limitations of the conclusions. The stages inform each other - what you can collect shapes what you can ask.

Other ways to work this

The same content, four other ways.

What is inside Levo

Six ways to work the same material.